Home |
Search |
Today's Posts |
#1
Posted to rec.audio.tubes
|
|||
|
|||
Phase Locked Loop with vacuum tubes.
I searched in Google for the existance of a schematic for a phase
locked loop using vaccum tubes, and hundreds of images came up but NOT ONE with a phase locked loop that could be used for better AM detection. I must away to my shed to do some work rather than chase fancy fairy circuits on the Internet. Patrick Turner. |
#2
Posted to rec.audio.tubes
|
|||
|
|||
Phase Locked Loop with vacuum tubes.
"Patrick Turner" wrote in message ... I searched in Google for the existance of a schematic for a phase locked loop using vaccum tubes, and hundreds of images came up but NOT ONE with a phase locked loop that could be used for better AM detection. I must away to my shed to do some work rather than chase fancy fairy circuits on the Internet. If they'd ever declassify some of the radar sets I worked on in the late 1960s... 400+ tubes each, and a number of PLL-stabilized thises and thats. |
#3
Posted to rec.audio.tubes
|
|||
|
|||
Phase Locked Loop with vacuum tubes.
On Jul 20, 4:06*am, flipper wrote:
On Tue, 19 Jul 2011 11:40:47 -0400, "Arny Krueger" wrote: "Patrick Turner" wrote in message ... I searched in Google for the existance of a schematic for a phase locked loop using vaccum tubes, and hundreds of images came up but NOT ONE with a phase locked loop that could be used for better AM detection. I must away to my shed to do some work rather than chase fancy fairy circuits on the Internet. If they'd ever declassify some of the radar sets I worked on in the late 1960s... 400+ tubes each, and a number of PLL-stabilized thises and thats. I'm not planning to build AM radios for ppl with 400+ tubes. 4 tubes is about enough, 1 x RF amp, 1 x F converter, 1 x IF amp, and 1 x twin triode for the CF detector. Then 3 tubes for a nice 9W SE amp for the speaker, which will be a modern type, not the old field coil type made in 1935 and crumbling to bits. I like to use 1 paralleled twin triode for tone control, same for SE amp input tube, and an EL34 for output tube in 50% SEUL mode. 9W is doable with B+ at OPT = +390V and 70mA for audio amp and another 15mA for everything else. Some old sets using Si diodes can provide this power easily, but if not then usually its possible to use an EL34 in triode with the existing OPT if its good enough, and get 4.5Watts of SET power which is much better than 6F6, 6BQ5, 6V6, 6M5, 6AR5, 6BW6, etc. The EL34 is placed where the 5Y3, 80, or other rectifier was. A mains fuse and HT fuse is added, so the ****in' old set don't burn down anyone's house. The nice roomy chassis made in the years just before and after WW2 are very easily rewired with stuff that works better. I'm having to build a dual concentric speaker for a set I'm doing now - 10" driver with 2" tweeter slung in the middle so it fits the cabinet made in 1940 without any changes, and all that's needed are slightly heavier guage wood screws. But If I wanted to make a PLL for AM detection, I guess theory tells me I'd have to have circuit to detect the phase difference of the IF signal with a 455kHz crystal oscillator, so something like an FM discriminator which produces a good swinging DC voltage, which will swing enough adjust the variable oscilator in the F converter section of the AM radio, and then hold the oscillator F so darn accurate the IF and crystal oscillator are dead accurately aligned, without being even one wave out. Trouble is trying to vary oscillator F in tube gear and at such LF as 455kHz takes large variations of C in tank circuits, and all I could find to do this was a bunch of paralleld 68V zener diodes where their C varies like a varicap before you reach the zener voltage. Varicaps diodes require very low circuit signal voltages. But higher voltage rated zeners allow the kind of voltages seen in tube gear - always much higher than in SS junk. But it seems no ******* has ever bothered to spend a month or three to get a PLL working for AM reception with tubes. Its like Multiplex Stereo decoding, no ******* has ever produced a fully tubed version of a Gilbert Cell to get stereo signals from an FM composite signal - they all fall back onto a diode matrix circuit. So they say they like to use tubes, but they ****in' won't, and don't have all the answers. At least 5 bjts are needed for a Gilbert Cell, so 5 triodes are needed for a tubes version, and who wants to use more tubes just for detection than the number used for the rest of the set? So the CF detector of mine is lookin' good, so simple and effective. I've even used my detector within my home brew variable RF oscillator so that the modulated RF output is detected, and fed into a diff amp with incoming audio F and the signal sent to modulate the RF has an error signal to linearize the envelope. This reduced the envelope shape THD from about 5% at 95% mod to about 1%, - a very welcome improvement so that detector distortions in receivers being tested would not be confused with distortion in the incoming test signal. Patrick Turner. Every vacuum tube 'electronic' (meaning CRT vs mechanical) television receiver had at least one: the horizontal oscillator/sync. In color sets you also have the chroma osc. |
#4
Posted to rec.audio.tubes
|
|||
|
|||
Phase Locked Loop with vacuum tubes.
I'm not planning to build AM radios for ppl with 400+ tubes That's the whole damn radar set, silly, not 'a' PLL. Of course. I was yus Yoking. . 4 tubes is about enough, 1 x RF amp, 1 x F converter, 1 x IF amp, and 1 x twin triode for the CF detector. Bean counter. But all those manufacturers before me used far fewer tubes..... Then 3 tubes for a nice 9W SE amp for the speaker, which will be a modern type, not the old field coil type made in 1935 and crumbling to bits. I like to use 1 paralleled twin triode for tone control, same for SE amp input tube, and an EL34 for output tube in 50% SEUL mode. 9W is doable with B+ at OPT = +390V and 70mA for audio amp and another 15mA for everything else. Some old sets using Si diodes can provide this power easily, but if not then usually its possible to use an EL34 in triode with the existing OPT if its good enough, and get 4.5Watts of SET power which is much better than 6F6, 6BQ5, 6V6, 6M5, 6AR5, 6BW6, etc. The EL34 is placed where the 5Y3, 80, or other rectifier was. A mains fuse and HT fuse is added, so the ****in' old set don't burn down anyone's house. The nice roomy chassis made in the years just before and after WW2 are very easily rewired with stuff that works better. I'm having to build a dual concentric speaker for a set I'm doing now - 10" driver with 2" tweeter slung in the middle so it fits the cabinet made in 1940 without any changes, and all that's needed are slightly heavier guage wood screws. Yeah. I cut costs by using old chassis/enclosures to stick something else in too. I'm doing this for part of my living. The labour time I spend on a radio rebuild can be hundreds of hours. I enjoy it, but the pay is maybe $2 an hour. I often must wind a new OPT about 4 times bigger than the Walnut Acountant model of OPTs which were included in the scam behind the brand name. But If I wanted to make a PLL for AM detection, I guess theory tells me I'd have to have circuit to detect the phase difference of the IF signal with a 455kHz crystal oscillator, Only if you kept IF conversion. Even if I didn't, one still has to have an oscillator track the RF, and be accurate. DG Tucker's Synchrodyne of 1947 is a difficult thing for anyone to build at home, but at least its tubed, and perhaps could be improved a lot if a couple of modern methods were applied. But I also like the idea of a 2Mhz IF so that the side band cutting is minimised but it probably requires 6 tuned ciricuits in 3 IFTs to get sufficient skirt selectivity. Just means two IF amps. Same CF detector could be used, but 6DJ8 would be a better tube for the higher F. so something like an FM discriminator which produces a good swinging DC voltage, which will swing enough adjust the variable oscilator in the F converter section of the AM radio, and then hold the oscillator F so darn accurate the IF and crystal oscillator are dead accurately aligned, without being even one wave out. It's called a PLL in modern parlance. It isn't quite the 'miracle' you imply with "so darn accurate," although it is. In a tube FM set there is often 4 triodes for the RF amp, mixer, oscilator and rmaining triode is for AFC, where the dc from discriminator is applied to the grid of 1/2 a 6AQ8 or 12AT7 to change gm of the triode which changes the C looking into the anode. The slight change of C is enough to swinf F a long way at 100MHz, but not at 455kHz. Its not the same4 as PLL though. You're thinking like proportional error, which would have to be incredibly small requiring huge gain, but a PLL isn't proportional "automatic frequency control." *A PLL, or "automatic phase control," integrates the difference by comparing phase and because it behaves as a frequency error integrator the frequency error 'naturally' goes to 0. Well if you know so much about PLL, I guess we can look forawrd to you providing us with a schematic next week. Trouble is trying to vary oscillator F in tube gear and at such LF as 455kHz takes large variations of C in tank circuits, and all I could find to do this was a bunch of paralleld 68V zener diodes where their C varies like a varicap before you reach the zener voltage. Varicaps diodes require very low circuit signal voltages. But higher voltage rated zeners allow the kind of voltages seen in tube gear - always much higher than in SS junk. Reactance tube. Does not do enough at 455kHz. But it seems no ******* has ever bothered to spend a month or three to get a PLL working for AM reception with tubes. Of course they did. Try looking up synchrodyne. http://www.thevalvepage.com/radtech/...5/section5.htm The schematic there has NO PLL and **IS D.G. Tucker's 1947 Synchrodyne** circuit. The main difficulty with the synchrodyne is how to make the ring modulator transformer, amoung other problems. By the time someone has ****ed around with Tucker's circuit one would have better spent the time on a superhet. The terminology "Phase lock loop" didn't come about till the 1960s when integrated circuit 'building blocks' prompted the creation of more generic 'building block' names. Indeed. In the tube era circuits tend to take on names related to what's being made. In TV circuits you have horizontal sync detectors, which are phase locked onto the horizontal sync pulses, and "automatic phase control" for the chroma demod. Look at a 'modern' schematic of the same thing and you'll see them called PLLs. So a locked oscilator is a form of PLL? It don't seem like that to me. In an FM set there is a 19kHz pilot tone with constant amplitude to lock an oscilator to so getting a 38kHz carrier which can be adjusted with LC circuits to be in phase with supressed subcarrier L-R signal. But the 38kHz generated can drift a little, because there is no PLL, and then the stereo separation suffers at higher AF. The modern version of the synchrodyne direct conversion is PLL direct conversion. Those things are literally everywhere but the basic principle goes back to the synchrodyne which, in turn, evolved from the homodyne. Its like Multiplex Stereo decoding, no ******* has ever produced a fully tubed version of a Gilbert Cell to get stereo signals from an FM composite signal - they all fall back onto a diode matrix circuit. It's cheaper and sufficient. As I've told you a thousand times, cost is *always* a concern and it had nothing to do with bean counters who don't know a synchrodyne from a dyno-soar. But I don't have to care about costs where the profits are not having to be shared by an accountant, marketing&sales team, rent of factory, CEO supply of Glen Fidich Wiskey, CEO Cadilac, CEO wages, shareholders and all the others who have to get paid out of sales. My activities are completely hand crafty one of a kind specials for mainly poor ppl, like me. Now one might think that in a world of 7 billion ppl there'd be one or two others who'd be able to tell us how they tried to use 4 x 6DJ8 tubes to make a GIlbert Cell, or a PLL, and get a result as good as any chip. Understand my DINO SAW appoach? So they say they like to use tubes, but they ****in' won't, and don't have all the answers. Just who is the 'they' you are babbling about? Excellent question. I have no idea who has built a PLL using tubes. Who ever they are, they are keeping themselves out of the limelight, and certainly away from giving you any opportunty to tip a bucket of **** on them. At least 5 bjts are needed for a Gilbert Cell, so 5 triodes are needed for a tubes version, No, because you use 'special' tubes instead of triodes, like the beam deflection tubes I told you about way back when you were futzing with your transistor Gilbert cell. But where you can use a bjt, you should be able to use a simple triode. and who wants to use more tubes just for detection than the number used for the rest of the set? Bean counter. Depends how many and why and what sort of job is done. I repeat, the CF detector of mine is lookin' good, so simple and effective. And it could improved. The load of the CF could be an RFC and the diode could just drive a resistance so the output looks like one half of the modulated envelope shape, and there is no C used. Then the second CF is used to buffer this to feed to a low pass LC or RC filter which removes the 455kHz signal entirely and you are lefet with just the audio content. The conversion efficiency is less, as it is with the "infinite impedance detector" but linearity of detection should be excellent. Patrick Turner. |
#5
Posted to rec.audio.tubes
|
|||
|
|||
Phase Locked Loop with vacuum tubes.
On Jul 20, 8:23*pm, flipper wrote:
On Wed, 20 Jul 2011 00:10:06 -0700 (PDT), Patrick Turner wrote: I'm not planning to build AM radios for ppl with 400+ tubes That's the whole damn radar set, silly, not 'a' PLL. Of course. I was yus Yoking. . 4 tubes is about enough, 1 x RF amp, 1 x F converter, 1 x IF amp, and 1 x twin triode for the CF detector. Bean counter. But all those manufacturers before me used far fewer tubes..... Doesn't alter the fact you're bean counting. OK, I count tubes, and sure, there's a limit. I offer more than what companies used to offer. They employed accountants, I don't. I don't need an accountant to tell me how I waste my talents or time, all of which are worth money. Just not very much. Not enough business is done here to warrant the employment of an accountant to advise me about where all the money goes, and how to limit the flow, thus controlling design. McIntosh did 'more and bigger' than most others too but I assure you they also took cost into account. They just had a different market strategy. When McIntosh began, like so many ther companies, they went in with bucket fulls of investment capital and armed with a reasonably good idea of an amp. They had good accountants and managers and engineers and quality control so the success happened. But while McIntosh beavered away making stuff that cost an arm and leg for most ordinary low paid folks, a veritable army of others worked away to make junk for the low income earners, and here the margins were much lower, and thus saving the cost of one tube with a production run of 100,000 units meant a large different to a company bottom line. snip, I'm doing this for part of my living. The labour time I spend on a radio rebuild can be hundreds of hours. I enjoy it, but the pay is maybe $2 an hour. I was making the point that after gutting and replacing the insides it's no longer what they brought to you. In a 1935 radio I am now re-engineering, the 3 tubes made in Europe in the 1930s had to be replaced. None are easily available. Two already had been replaced with Oz made octals. The 1935 speaker IS CRAP. The lady will like the tube sound using a better circuit and tube types made between 1945 and 1960. These remain fairly abundant. Doesn't bother me, much, but I'm sure there are collectors out there fuming and ripping out hair at the 'tragedy of it all. I give my customers what suits their listenig needs, and what suits their wallets, and what suits their preferences for preservation of the old circuits. A few just want their gear kept original such as one guy who wanted his Leak and Quad tube systems kept original. When you walked into his lounge, an early vinyl would be playing, and deacor was all exactly like 1955. On the 1955 coffee table, a newspaper headline screamed "SUEZ BOMBED". One had to pinch oneself to remind you were really in 2010. But some have a nice of floor standing cabinet with very attractive dial which IS SACRED as far as I am concerned. The electronics which was fitted into such grandiose bits of furniture was often pretty ordinary, and sub-standard in performance. All I do is make it all work as IT SHOULD HAVE when new, and I'll include to put terminals for a CD player. I have ppl wanting me to fix some very junky un- collectable non valuable old radiograms with very poor TTs. I refuse to work on some of the worst. Some don't give a stuff about the cabinet and just want me to rescue the 4W stereo SE amps and re-house them in a painted plywood box with a tube phono amp for MM TT. So there is a whole range of things I am asked to do. I have a pair of Quad-II-Forty which I'll drastically change to a far better circuit in coming weeks. So the only tragedy is in your short sighted understanding of what occupies my life. I often must wind a new OPT about 4 times bigger than the Walnut Acountant model of OPTs which were included in the scam behind the brand name. Definitely need a 20Hz OPT feeding a 5 inch accordion cone speaker. What were they thinking? But one radio-gram was a Phillips POS with stereo with 2 x 6M5, 2.5W per channel, and 8" speakers which were broken and in a large cabinet for good bass. Existing OPTs were fried. The pair of OPTs I wound went down to 40Hz, but still were much larger than the very high winding loss types made by ****ing Phillips. For 9W, and 20Hz, Afe must be at least 36mm x 36mm. I don't go that far, and in a small mantle radio, 80Hz is OK and for 2 watts, along with 25% winding losses with hair thin primary wire. I have enough spares to last me for those types of radios, but not for the big floor standers which are capable of excellent sound providing OPTs are bettered along with the circuit function. In Oz, digital radio which receives signal at over 250Mhz is the new radio media, and most sets are horrid little boxes, some of which are retro styles and only capable of terrible sound despite the beautiful signal quality that is received so I guess in future there's the trade of making sure old AM sets may be fed with a signal from a digital radio. But If I wanted to make a PLL for AM detection, I guess theory tells me I'd have to have circuit to detect the phase difference of the IF signal with a 455kHz crystal oscillator, Only if you kept IF conversion. Even if I didn't, one still has to have an oscillator track the RF, Yes, but you don't need to do a conversion. and be accurate. DG Tucker's Synchrodyne of 1947 is a difficult thing for anyone to build at home, but at least its tubed, and perhaps could be improved a lot if a couple of modern methods were applied. Whether it's 'difficult' or not isn't the point. The point was you claiming no one did it. You claimed that, I didn't. I just said a google search on tubed circuits labelled as PLL brought up **** all. Nobody gives a rat's arse about developing tubed circuits it seems. There is nothing for me or anyone else to build on. Go do your own R&D, that's what the world tells me. But I also like the idea of a 2Mhz IF so that the side band cutting is minimised but it probably requires 6 tuned ciricuits in 3 IFTs to get sufficient skirt selectivity. Just means two IF amps. Same CF detector could be used, but 6DJ8 would be a better tube for the higher F. The point would be synchronous detection, not to go through that rigmarole just to stick the same ole diode on the end of it. I'd like to think a tubed synchrodne set with PLL could easily be done, but I'm a realist; no such thing is easy. so something like an FM discriminator which produces a good swinging DC voltage, which will swing enough adjust the variable oscilator in the F converter section of the AM radio, and then hold the oscillator F so darn accurate the IF and crystal oscillator are dead accurately aligned, without being even one wave out. It's called a PLL in modern parlance. It isn't quite the 'miracle' you imply with "so darn accurate," although it is. In a tube FM set there is often 4 triodes for the RF amp, mixer, oscilator and rmaining triode is for AFC, where the dc from discriminator is applied to the grid of 1/2 a 6AQ8 or 12AT7 to change gm of the triode which changes the C looking into the anode. The slight change of C is enough to swinf F a long way at 100MHz, but not at 455kHz. Its not the same4 as PLL though. You don't need to 'swing' the PLL far. You *tune* it near the frequency and then it pulls lock. One would think someone has done all with tubes and has it online but no such thing I could find. You're thinking like proportional error, which would have to be incredibly small requiring huge gain, but a PLL isn't proportional "automatic frequency control." �A PLL, or "automatic phase control," integrates the difference by comparing phase and because it behaves as a frequency error integrator the frequency error 'naturally' goes to 0. Well if you know so much about PLL, I guess we can look forawrd to you providing us with a schematic next week. I've used PLLs before, and even designed them from scratch when the 'building blocks' didn't fit my application, but I'm not interested in designing an AM radio. Same goes for an internal combustion engine. I can tell you how they work too but am not interested in building one of those from scratch either. At least not today. Ah, you don't really give a **** about the subject anyway. Trouble is trying to vary oscillator F in tube gear and at such LF as 455kHz takes large variations of C in tank circuits, and all I could find to do this was a bunch of paralleld 68V zener diodes where their C varies like a varicap before you reach the zener voltage. Varicaps diodes require very low circuit signal voltages. But higher voltage rated zeners allow the kind of voltages seen in tube gear - always much higher than in SS junk. Reactance tube. Does not do enough at 455kHz. Prove it. I did prove it to myself once when I built a 455khz oscillator which has a wobbulator function to create FM with +/- 40kHz of F deviation using 10 x parallel 68V zener diodes. Its possible that Idea could be applied with a circuit which compares RF input phase and oscilator phase and then the Vdc could swing the F enough. In old days, a reactance tube was handy, and then the F was multiplied quite a few times along with the deviation. FM transmitters using tubes were hellishly complex. Now all one needs is a BA1404 chip and a 1.5V battery. Something you clip to your lapel, with two mics for stereo. But it seems no ******* has ever bothered to spend a month or three to get a PLL working for AM reception with tubes. Of course they did. Try looking up synchrodyne. http://www.thevalvepage.com/radtech/...5/section5.htm The schematic there has NO PLL and **IS D.G. Tucker's 1947 Synchrodyne** circuit. Yes, it is Tucker's Synchrodyne and yes it's a PLL. Go back to RDH4 and read p1227. They describe the classic PLL action, albeit in different words. In particular "One advantage of this receiver is it is either correctly tuned or not tuned at all." I.E. It is 'tuned' when phase locked and not tuned at all if the PLL breaks lock. There is no 'in between': classic PLL snap action behavior. RDH4 has only such a tiny bit of info about synchronous detection, and no mention of PLL as it is understood in the more modern post tube world. There is a pile of references listed for the sychrodyne including several to Mr D. G. Tucker, but Tucker's schematic does not appear to be a PLL circuit, although may have a means by whitch AFC control is good enough to get clean synchronous detection and to stop whistles during tuning, which RDH4 later refers to below what you quoted. In Tuckers circuit, V3 is a pentode set up to accept RF from the broadband front end amps and this RF is applied by V3 cathode to a winding on the VFO, and so when a strong enough RF signal appeared, there would be tuning whistles from carrier beating with oscilator until the RF in became strong enough to synchronise the oscillator as its F was tuned close enough to the incoming RF. So one could tune a bit each side of the RF and the oscillator F would stay fixed, allthogh phase would shift a bit. I guess one tuned it best by sound quality. There is no DC amp or very sensitve detection of phase differences with interfering RF square waves and oscilator square waves. The Tucker signal would seem to depend on a threshold of RF signal needed to lock the oscillator, and in 1947, AM modulation rarely went near 100% so there would be enough carrier always present for locking. But now I see 100% on the CRO all the time from AM stations, and even if one tries to amplifiy an incoming RF signal with a couple of pentodes set up like IF limiter tubes in an FM set to ensure no AM exists, there are times when mod = 100% and the RF limited signal just dissaprears. So one must have an oscillator which will "run on" at the locked F for some time without drift or cessation and controlled by a DC voltage feeding a an RC circuit with just the right time constant. Tucker's circuit never caught on. Some say it was before its time. But when did better AM circuits ever catch on? I see and hear no evidence anything better ever occured. The vast majority of SS AM/FM tuners contained AM tuners with 3 bjts, and high Q single tapped coil IFTs about the size of a pea. The AF BW = 1.5kHz max, andf boosting treble made no difference because no treble was there to boost. Later they used even cheaper nastier horribgle 3 pin TRF chips. Audio was frightful, and on a weak community station of 300W at 1008kHz, some 5km away, I could hear a booming 5kW station also nearby and at 1053kHz. Sometimes I'd get a station from 2,000km away as well. None such crap happens with a tubed set with well done IFTs. Note that you don't need 'conversion' for an IF, to get a 'single frequency' for filters, because simply LP filtering the phase discriminator does it. That controls the PLL pull in and lock range so if it gets 'close enough' to a received carrier it pulls in lock and demods. The theory is great but, as with all things, there are practical considerations with the antenna potentially picking up the local osc, which is at the same frequency you're trying to receive, being one of them. Obviously not a 'good thing' to have happen. The main difficulty with the synchrodyne is how to make the ring modulator transformer, amoung other problems. Tucker managed and so did others. There is never ever enough info on such trannys now. No How-To-Do. A radio hame said use a toroid core with proper permeability. I got nowhere with the ideas. Knowing exactly which toroid RF core to use is essential. I gave up. By the time someone has ****ed around with Tucker's circuit one would have better spent the time on a superhet. Depends on what the bean counter says, eh? Bean counters sure made sure the Tucker would never be put onto shop shelving. A superhet is sufficient but the synchrodyne is more accurate. Says he who doesn't really care enough to build anything much. The terminology "Phase lock loop" didn't come about till the 1960s when integrated circuit 'building blocks' prompted the creation of more generic 'building block' names. Indeed. In the tube era circuits tend to take on names related to what's being made. In TV circuits you have horizontal sync detectors, which are phase locked onto the horizontal sync pulses, and "automatic phase control" for the chroma demod. Look at a 'modern' schematic of the same thing and you'll see them called PLLs. So a locked oscilator is a form of PLL? A phase locked oscillator is, which is why it's call a "Phase Locked Loop." AFC (if properly identified) is not a PLL since it's oscillator is driven by a proportional frequency error signal. It get's arbitrarily 'close' (depending on gain), but is not phase locked, and there is always some frequency 'error' that's needed to push the oscillator in the desired direction. Indeed, in FM sets the tendency for RF and oscillator tuning to drift with changes in temperature is well controlled by AFC but still there is *some drift* but its not enough to matter. In a phase locked loop the error signal comes from the phase difference and 'in lock' the local oscillator frequency is exactly the same as the reference with a phase lead/lag error depending on which direction the local osc needs to be 'pushed' to stay on frequency. The difference is in how the error signal is derived. AFC is frequency deviation and PLL is phase deviation. It don't seem like that to me. In an FM set there is a 19kHz pilot tone with constant amplitude to lock an oscilator to so getting a 38kHz carrier which can be adjusted with LC circuits to be in phase with supressed subcarrier L-R signal. But the 38kHz generated can drift a little, because there is no PLL, and then the stereo separation suffers at higher AF. Why in the world do you take a TV horizontal sync circuit and talk about FM demod? If you take an old TV set with a 'horizontal hold' knob you can see the PLL 'phase lock' action. When in lock the picture shifts a bit either left or right when you turn the knob as the phase of the lock shifts with respect to the incoming video (phase error having to 'push' the local osc harder, because you're changing the 'idle' frequency, so the phase shifts) but it remains perfectly aligned until you turn it far enough and... *zap, brrr... the slanty lines suddenly appear as the local osc breaks lock and is then free running at a significantly 'not right' frequency. Turn the knob back to where the PLL can pull ... Google stops me from replying to more. There are simpler synchrodynes which use one RF amp with RF tuning followed by a 6BE6 pentagrid. MUCH simpler than Tucker's set, but I found that a nightmare to build with very feeble performance. The tubed Superhet is King, still. Patrick Turner. |
#6
Posted to rec.audio.tubes
|
|||
|
|||
Phase Locked Loop with vacuum tubes.
"flipper" wrote in message ... On Tue, 19 Jul 2011 11:40:47 -0400, "Arny Krueger" wrote: "Patrick Turner" wrote in message ... I searched in Google for the existance of a schematic for a phase locked loop using vaccum tubes, and hundreds of images came up but NOT ONE with a phase locked loop that could be used for better AM detection. I must away to my shed to do some work rather than chase fancy fairy circuits on the Internet. If they'd ever declassify some of the radar sets I worked on in the late 1960s... 400+ tubes each, and a number of PLL-stabilized thises and thats. Every vacuum tube 'electronic' (meaning CRT vs mechanical) television receiver had at least one: the horizontal oscillator/sync. In color sets you also have the chroma osc. trip down memory lane Yes, those had to be in phase with the transmitter's clocks. I remember some phase detectors in the old NTSC sets - an instant tip-off that there was some kind of phase locked oscillator in play. |
#7
Posted to rec.audio.tubes
|
|||
|
|||
Phase Locked Loop with vacuum tubes.
On Wed, 20 Jul 2011 12:31:17 -0400, "Arny Krueger"
wrote: "flipper" wrote in message .. . On Tue, 19 Jul 2011 11:40:47 -0400, "Arny Krueger" wrote: "Patrick Turner" wrote in message ... I searched in Google for the existance of a schematic for a phase locked loop using vaccum tubes, and hundreds of images came up but NOT ONE with a phase locked loop that could be used for better AM detection. I must away to my shed to do some work rather than chase fancy fairy circuits on the Internet. If they'd ever declassify some of the radar sets I worked on in the late 1960s... 400+ tubes each, and a number of PLL-stabilized thises and thats. Every vacuum tube 'electronic' (meaning CRT vs mechanical) television receiver had at least one: the horizontal oscillator/sync. In color sets you also have the chroma osc. trip down memory lane Yes, those had to be in phase with the transmitter's clocks. I remember some phase detectors in the old NTSC sets - an instant tip-off that there was some kind of phase locked oscillator in play. This is where PAL really won out over NTSC. Each alternate line the chroma phase leads or lags by 45 degrees, and the PLL locks to the average - which is true phase. The result is that when there is an overall phase error in the chroma signal, alternate lines lead and lag the proper phase. Add them together and the error vanishes. With NTSC a phase error means a wrong colour. d |
#8
Posted to rec.audio.tubes
|
|||
|
|||
Some thoughts on PLLs, Synchrodynes, and AM modulation depth
In article ,
Patrick Turner wrote: Google stops me from replying to more. There are simpler synchrodynes which use one RF amp with RF tuning followed by a 6BE6 pentagrid. MUCH simpler than Tucker's set, but I found that a nightmare to build with very feeble performance. The tubed Superhet is King, still. Patrick, No reason we can't combine the tubed superheterodyne with a tubed synchronous detector. I have a few thoughts on PLL detectors, and AM broadcasting modulation depth. Take my comments with a few grains of salt as I have little expertise in either of these fields, although one of my PLL designs did make it into a commercial product, possibly sold in the millions, some in this group, at least when it was larger may even have used my PLL. However just because I got a PLL into production doesn't mean that I had a clue what I was doing. First some thoughts on PLLs. Injection locked oscillators, as used in Tucker's Synchrodyne Receiver, are not PLLs in my book, how could they be as there is no loop involved. The system used in many old FM tuners to stabilize the local oscillator, using the output of a frequency discriminator to control a VCO, often called "AFC" is not even a FLL, or Frequency Locked Loop, it is however a loop which reduces frequency error. Next in the pecking order is the FLL, or Frequency Locked Loop. This loop replaces the frequency discriminator with a phase detector which compares the VCO frequency with a reference such as a crystal oscillator or some other signal, such as the IF singal, to be locked to. This loop locks the frequency of the VCO to the frequency of the "reference" but there will be phase error, which is necessary to create the error signal to control the VCO. While the frequency is locked, the phase may vary and is not locked, so calling this common loop a PLL is somewhat of a misnomer, as the phase is variable and is not locked. This may be what the experts call a first order PLL. Next in the pecking order is what I will call the true PLL, or Phase Locked Loop. This is similar to the FLL circuit described above, with the addition of an integrator circuit between the output of the phase detector and the VCO control input. The phase error drives the integrator to the voltage level necessary to put the VCO on frequency and with zero phase error. When the phase error reaches zero the integrator stops integrating at the voltage level required by the VCO, if a phase error again develops for any reason, the integrator level changes to correct it. This loop may be what the experts call a second order PLL. You said "the main difficulty with the synchrodyne is how to make the ring modulator transformer." This doesn't seem like it should be a major problem, especially for a transformer expert like you. I would think an ordinary IFT could be pressed into service here, deleting the original secondary and winding a new low impedance secondary directly over the primary, the outside of which becomes the cold end of the primary. My concept of the synchronous detector in this discussion, was using a phase detector to compare a local VCO with the incoming IF frequency in a PLL circuit. A second "phase detector" with the local oscillator signal delayed 90 degrees would be used to recover the audio. In this scheme the local oscillator "VCO" would always operate at approximately 455 kHz, so that the reactance change required of the reactance tube is minimal, as there is no need to provide the 10:1 reactance ratio that would be required if the synchronous detector operated at the received frequency. For the less ambitious Tucker's "synchronous" detector could be operated at the IF frequency, eliminating the need for PLLs and multiple quadrature phase demodulators as used in my scheme. Modulation depth of AM broadcast transmitters is an interesting subject. Typically the modulation monitor is connected to the output of the transmitter, in some transmitters a sample of the output voltage is provided for the monitor, while in other transmitters a sample of the transmitters output current is provided. Since the impedance of the transmission line/antenna system is not constant across the broadcast channel the actual carrier and sideband power supplied to the transmission line will vary with frequency. Assume that the transmitter provides a current sample to the modulation monitor and that the resistance seen by the transmitter at the input to the transmission line increases at the sideband frequencies. Assuming that the modulation monitor is reading 100% the actual sideband power radiated by the antenna will be greater than what is required for 100% modulation, and a radio receiving the signal radiated by the broadcast antenna will see greater than 100% modulation, even in the negative direction, causing problems for envelope detectors including the pseudo synchronous detectors. A true synchronous detector will be immune to this type of over modulation. If the transmission line resistance varies in the opposite way the situation will be reversed and the radio will see lower modulation than the station thinks it is using. Voltage sensing for the modulation monitor will reverse all the above conditions, as will an extra quarter wave length of transmission line between the transmitter and antenna will invert the impedance, again reversing all the conditions accordingly. Also the impedance seen by the transmitter may be asymmetrical, causing the amplitude and phase of the two sidebands to be unequal, again resulting in distortion with envelope detectors, which true synchronous detectors will ignore. Finally, perhaps the biggest distortion problem is caused by directional antennas which are nearly universally used in the US, but may not be as widely used in OZ. The radiation pattern of directional broadcast antennas may be considerably different at the sideband frequencies than at the carrier frequency, again resulting in considerable over modulation in some areas, particularly in the areas around the direction of the antennas nulls. This effect can often be observed by driving around or past a directional broadcast antenna while listening to the automobile radio. This is just a simple minded overview, by someone who doesn't know what he is talking about, to give some flavor to the many factors involved in determining the actual modulation seen by the radio receiver in your lounge. Many of these factors can cause the modulation to exceed 100% negative at the antenna of your radio, or can cause an asymmetrical sideband situation, either of which will cause distortion in a simple envelope detector, or even a pseudo synchronous detector. A true synchronous detector will be immune to the distortion from these causes, resulting in a more pleasing listening experience. Many of the AM stereo demodulator chips used in the days of AM stereo used a pseudo synchronous envelope detector during tuning to eliminate all the squawks and whistles, and then switched over to a true synchronous detector once the signal was correctly tuned. So it would seem there is room for both a Turner CF detector and a true synchronous detector in our tubed radios. -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
#9
Posted to rec.audio.tubes
|
|||
|
|||
Some thoughts on PLLs, Synchrodynes, and AM modulation depth
On Jul 21, 10:01*am, John Byrns wrote:
In article , *Patrick Turner wrote: Google stops me from replying to more. There are simpler synchrodynes which use one RF amp with RF tuning followed by a 6BE6 pentagrid. MUCH simpler than Tucker's set, but I found that a nightmare to build with very feeble performance. The tubed Superhet is King, still. Patrick, No reason we can't combine the tubed superheterodyne with a tubed synchronous detector. Except that it is difficult to do as easily as a superhet, even with a locked oscillator operating at 455kHz. One would still get lots whistles while tuning until synchronisation occurs. VFO F - RF is an IF frequency away from 455kHz and is only on 455kHz or close to it when maximum IF gain is achieved at the tops of selectivity curves. I've had nuerous hetero sets which self oscilate at 455kHz or therabouts ths causing all sorts of tuning bothers and noise. I have a few thoughts on PLL detectors, and AM broadcasting modulation depth. * Take my comments with a few grains of salt as I have little expertise in either of these fields, although one of my PLL designs did make it into a commercial product, possibly sold in the millions, some in this group, at least when it was larger may even have used my PLL. *However just because I got a PLL into production doesn't mean that I had a clue what I was doing. First some thoughts on PLLs. Injection locked oscillators, as used in Tucker's Synchrodyne Receiver, are not PLLs in my book, how could they be as there is no loop involved. That was my point to Flipper. RDH4 does not talk of PLL. The system used in many old FM tuners to stabilize the local oscillator, using the output of a frequency discriminator to control a VCO, often called "AFC" is not even a FLL, or Frequency Locked Loop, it is however a loop which reduces frequency error. I agree. The AFC with one triode does make for much less drift after turning on a set, tuning to a station while the components are still cold. So much so that the set stays tuned for days with little drift even after turning off, allowing everything to cool, then turning back on. Next in the pecking order is the FLL, or Frequency Locked Loop. *This loop replaces the frequency discriminator with a phase detector which compares the VCO frequency with a reference such as a crystal oscillator or some other signal, such as the IF singal, to be locked to. *This loop locks the frequency of the VCO to the frequency of the "reference" but there will be phase error, which is necessary to create the error signal to control the VCO. *While the frequency is locked, the phase may vary and is not locked, so calling this common loop a PLL is somewhat of a misnomer, as the phase is variable and is not locked. *This may be what the experts call a first order PLL. Next in the pecking order is what I will call the true PLL, or Phase Locked Loop. *This is similar to the FLL circuit described above, with the addition of an integrator circuit between the output of the phase detector and the VCO control input. *The phase error drives the integrator to the voltage level necessary to put the VCO on frequency and with zero phase error. *When the phase error reaches zero the integrator stops integrating at the voltage level required by the VCO, if a phase error again develops for any reason, the integrator level changes to correct it. *This loop may be what the experts call a second order PLL. That is my understanding of the modern PLL, and a number of chips are supposed to do that. You said "the main difficulty with the synchrodyne is how to make the ring modulator transformer." *This doesn't seem like it should be a major problem, especially for a transformer expert like you. * I'm like the guy who is a brain neurosurgeon. Get me to work on an arsole, and I'm dangerous. Just because I work on AF trannys don't mean I have any experience at RF. So many ppl want me for my expertise at AF I have not had enough time to move up the band. I would think an ordinary IFT could be pressed into service here, deleting the original secondary and winding a new low impedance secondary directly over the primary, the outside of which becomes the cold end of the primary. If only it were that simple. It seems to me the untuned coil inductive reactance would need to be say 10k at the lowest F and with very low self capacitance 10pF to be able to be driven with a CF tube so if lowest F = 455kHz, then L = 3.4mH, which is a lot more than say 200uH needed for an AM input coil to resonate with a 36-400PF tuning gain. Perhaps I could use a solenoid coil on a 25mm former with ferrite rods placed inside the former, and using some solid core telephone hook up wire which has 0.1mm plastic insulation which spaces the wires to keep self C low. I just used such a coil for an AM RF input coil which gave excellent results, and maybe I needed 45t for 200uH, but for 3,400uH I'd need propably maybe at least 5 times the turns, but working stuff out for RF is not my forte. Working with litz wire is a complete PITA. its mainly used to reduce RF resistance to get the tuned Q of a tank high for F between say 50kHz and 2MHz. Above that F the turns needed for a coil are less, so solid wire is OK. My concept of the synchronous detector in this discussion, was using a phase detector to compare a local VCO with the incoming IF frequency in a PLL circuit. * A second "phase detector" with the local oscillator signal delayed 90 degrees would be used to recover the audio. *In this scheme the local oscillator "VCO" would always operate at approximately 455 kHz, so that the reactance change required of the reactance tube is minimal, as there is no need to provide the 10:1 reactance ratio that would be required if the synchronous detector operated at the received frequency. And in other words so everyone can understand, you meant to say............................................... ....., and the schematic may be found at www....................................here. Please fill in the blanks for everyon'es benefit. For the less ambitious Tucker's "synchronous" detector could be operated at the IF frequency, eliminating the need for PLLs and multiple quadrature phase demodulators as used in my scheme. I recall I tried trying to use a locked oscilator at 455kHz, but had so many bothers I gave that idea up too. There is such a thing as "elevated carrier detection" where a square wave at the carrier F or IF is added to the incoming RF or IF and the modulation % is effectively converted to a much smaller %, and the wamted signal is also "raised up out of the noise" and a detector such as mine is then used to get almost distortionless AF recovery because the RF/IF ripple voltage amplitude remains at a constant level during each AF wave without any flattening or slight cutting off near 100% mod. The flats created by mod over 100% mod occur with the same shape as seen on incoming RF/IF which has more than 100%. Modulation depth of AM broadcast transmitters is an interesting subject. * Typically the modulation monitor is connected to the output of the transmitter, in some transmitters a sample of the output voltage is provided for the monitor, while in other transmitters a sample of the transmitters output current is provided. *Since the impedance of the transmission line/antenna system is not constant across the broadcast channel the actual carrier and sideband power supplied to the transmission line will vary with frequency. *Assume that the transmitter provides a current sample to the modulation monitor and that the resistance seen by the transmitter at the input to the transmission line increases at the sideband frequencies. *Assuming that the modulation monitor is reading 100% the actual sideband power radiated by the antenna will be greater than what is required for 100% modulation, and a radio receiving the signal radiated by the broadcast antenna will see greater than 100% modulation, even in the negative direction, causing problems for envelope detectors including the pseudo synchronous detectors. *A true synchronous detector will be immune to this type of over modulation. *If the transmission line resistance varies in the opposite way the situation will be reversed and the radio will see lower modulation than the station thinks it is using. *Voltage sensing for the modulation monitor will reverse all the above conditions, as will an extra quarter wave length of transmission line between the transmitter and antenna will invert the impedance, again reversing all the conditions accordingly.. *Also the impedance seen by the transmitter may be asymmetrical, causing the amplitude and phase of the two sidebands to be unequal, again resulting in distortion with envelope detectors, which true synchronous detectors will ignore. *Finally, perhaps the biggest distortion problem is caused by directional antennas which are nearly universally used in the US, but may not be as widely used in OZ. *The radiation pattern of directional broadcast antennas may be considerably different at the sideband frequencies than at the carrier frequency, again resulting in considerable over modulation in some areas, particularly in the areas around the direction of the antennas nulls. *This effect can often be observed by driving around or past a directional broadcast antenna while listening to the automobile radio. *This is just a simple minded overview, by someone who doesn't know what he is talking about, to give some flavor to the many factors involved in determining the actual modulation seen by the radio receiver in your lounge. AFAIK, some directional AM antennas are used here and I have no idea what AF distortion they cause. But most AM stations are close to town and give plenty of level and low THD with 9kHz BW -- If the set is capable of low THD wide BW, and 95% just ain't. Digital radio will spread to regional areas as time goes by. Time in regional areas is much slower than time in big cities, as you should know, and the further you get from big cities, the less time seems to happen and its not uncommon to see a small town clock stuck on 12:30pm, and its been like that for years. I now such a clock and nothing happens there. People are virtually un-alive. But they serve a very nice coffee and sweetcake. They stopped letting time pass, and it's done 'em the world of good. Stuff that happens in big cities often never happens in small towns. The only time when time will passes in Twelve Thirty is if an ageing pidgeon crashes into the minute hand on the town clock. Many of these factors can cause the modulation to exceed 100% negative at the antenna of your radio, or can cause an asymmetrical sideband situation, either of which will cause distortion in a simple envelope detector, or even a pseudo synchronous detector. *A true synchronous detector will be immune to the distortion from these causes, resulting in a more pleasing listening experience. That's rather difficult to verify. Many of the AM stereo demodulator chips used in the days of AM stereo used a pseudo synchronous envelope detector during tuning to eliminate all the squawks and whistles, and then switched over to a true synchronous detector once the signal was correctly tuned. *So it would seem there is room for both a Turner CF detector and a true synchronous detector in our tubed radios. In 11months, 2 weeks, 15 hours and 14 minutes time, I turn 65 and can retire on the old age pension. It will be a great relief not to have to work for the Greatly Ungrateful ****in Masses Who Pay ****in Peanuts for me to repair their bloody awful generic crap they bring me. I might then have time for spending more days and daze and daiz and deighs working out what might have been done back in about 1950 for the 2% of the General Public willing to pay more for better performance. Patrick Turner. -- Regards, John Byrns Surf my web pages at, *http://fmamradios.com/ |
#10
Posted to rec.audio.tubes
|
|||
|
|||
Some thoughts on PLLs, Synchrodynes, and AM modulation depth
In article ,
flipper wrote: On Wed, 20 Jul 2011 19:01:30 -0500, John Byrns wrote: First some thoughts on PLLs. Injection locked oscillators, as used in Tucker's Synchrodyne Receiver, are not PLLs in my book, how could they be as there is no loop involved. By golly I think you're right. I first thought he was using phase from the mixer but upon closer inspection I see I got turned around in the wiring loops. Amusingly enough, though, injection locking can be modeled as a first order PLL so the net result is conceptually the same. Sounds right, I'll buy that. Not being an RF engineer I could be wrong but for AM demod we don't really give a hoot about phase, do we? I believe we do care about phase, and I don't think we need look any further than our old High School Trigonometry books to find out why, no RF engineering expertise required. IIRC the AM equation goes something like this, sin(wct) * (1 + sin(Wmt)). If we multiply this by a locally generated copy of sin(Wct) in the receiver we get ((sin(wct))^2) * (1 + sin(Wmt)). Without getting out my aforementioned High School Trigonometry book, I think ((sin(wct))^2) = 0.5 + 0.5 sin(2Wct) or something like that. The 0.5 term multiplied by sin(Wmt) gives us back our original modulation. If we try to use a quadrateure carrier in the receiver, i.e. cos(Wct) so that we have cos(Wct) * sin(wct) * (1 + sin(Wmt)), and IIRC, without getting out the book, cos(Wct) * sin(wct) doesn't include a DC term, only a sin(2Wct) term, without the DC term we don't get demodulation of the desired signal, just quadrature noise. Local carrier phase angles between sin(Wct) and cos(Wct) will produce some DC, although less than an in phase carrier, so the demodulated signal will be attenuated relative to an in phase carrier, and it will be full of quadrature noise that is demodulated by the cos(Wct) component of the locally generated carrier. -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
#11
Posted to rec.audio.tubes
|
|||
|
|||
Some thoughts on PLLs, Synchrodynes, and AM modulation depth
On Jul 22, 12:38*pm, John Byrns wrote:
In article , *flipper wrote: On Wed, 20 Jul 2011 19:01:30 -0500, John Byrns wrote: First some thoughts on PLLs. Injection locked oscillators, as used in Tucker's Synchrodyne Receiver, are not PLLs in my book, how could they be as there is no loop involved. By golly I think you're right. I first thought he was using phase from the mixer but upon closer inspection I see I got turned around in the wiring loops. Amusingly enough, though, injection locking can be modeled as a first order PLL so the net result is conceptually the same. Sounds right, I'll buy that. Not being an RF engineer I could be wrong but for AM demod we don't really give a hoot about phase, do we? I believe we do care about phase, and I don't think we need look any further than our old High School Trigonometry books to find out why, no RF engineering expertise required. *IIRC the AM equation goes something like this, sin(wct) * (1 + sin(Wmt)). *If we multiply this by a locally generated copy of sin(Wct) in the receiver we get ((sin(wct))^2) * (1 + sin(Wmt)). *Without getting out my aforementioned High School Trigonometry book, I think ((sin(wct))^2) = 0.5 + 0.5 sin(2Wct) or something like that. *The 0.5 term multiplied by sin(Wmt) gives us back our original modulation. If we try to use a quadrateure carrier in the receiver, i.e. cos(Wct) so that we have cos(Wct) * sin(wct) * (1 + sin(Wmt)), and IIRC, without getting out the book, cos(Wct) * sin(wct) doesn't include a DC term, only a sin(2Wct) term, without the DC term we don't get demodulation of the desired signal, just quadrature noise. *Local carrier phase angles between sin(Wct) and cos(Wct) will produce some DC, although less than an in phase carrier, so the demodulated signal will be attenuated relative to an in phase carrier, and it will be full of quadrature noise that is demodulated by the cos(Wct) component of the locally generated carrier. -- Regards, John Byrns I can announce quite safely that John's expert opinions and explanation has delivered a perfect understanding of how to write two paragraphs which are totally incomprehensible, and thus offer Sweet Phark All information about how to build a better AM detector. Whatever John was trying to say remains pure and detailed, and meaningful to someone@somewhere while remaining 100% fully obscure. There was a man with a radio, who used just one cat's whisker, He prodded the crystal, the cat went beserk, So that idea couldn't be sillier. Patrick Turner. |
#12
Posted to rec.audio.tubes
|
|||
|
|||
Some thoughts on PLLs, Synchrodynes, and AM modulation depth
On Jul 22, 3:23*pm, flipper wrote:
On Thu, 21 Jul 2011 21:38:45 -0500, John Byrns wrote: In article , flipper wrote: On Wed, 20 Jul 2011 19:01:30 -0500, John Byrns wrote: First some thoughts on PLLs. Injection locked oscillators, as used in Tucker's Synchrodyne Receiver, are not PLLs in my book, how could they be as there is no loop involved. By golly I think you're right. I first thought he was using phase from the mixer but upon closer inspection I see I got turned around in the wiring loops. Amusingly enough, though, injection locking can be modeled as a first order PLL so the net result is conceptually the same. Sounds right, I'll buy that. After posting I went back and looked again at Adler's analysis of injection locked LC oscillators to see 'why'. Maybe no one cares but I found it interesting and can conceptually see that the 'phase comparator' and VCO 'loop' are subsumed, or 'hidden', in the process. A 'PLL expert' would probably instantly see it in the math. The basic 'frequency changing' mechanism is similar to how a reactance tube works. I.E The 'injected' signal forces a shift in the tank (current/voltage) phase, which forces the tank to change frequency in order to reestablish proper phase. Our 'VCO'. In the case of the typical reactance tube application the tube is run at a fixed phase offset and the amount of tank 'upset' is due to how much of it is injected, by modulating the grid. The VCO 'control input'. In this case, however, the amount of tank 'upset' will be proportional to the LO-RF phase difference, so it acts a phase comparator. I.E. If they are in phase then there is no (current/voltage) tank upset, and no change to tank frequency, but there must be 'some change' when the LO is not at the right frequency, so there must be a phase difference to 'hold' the tank on frequency. That's a first order PLL. If we built it the conventional 'separate piece for each function' way we'd have a separate phase detector, filter, and reactance tube VCO. Those functions are still there, albeit hidden, in the injection lock but we have much less control over the parameters. Like the reactance tube circuit it also depends on injection amplitude so AM should cause phase jitter we have no convenient way to filter or control and, of course, there's no 'free wheeling' across 0 carrier, not to mention we can't add an integrator. I mean, since it's all subsumed in the process itself we don't have a lot of filtering choices like with separate components.. Does imply the thing might be better with a separate RF limiter before the injection lock, though. Or do a proper 'separate pieces' PLL, like I mistakenly thought he had. Speaking of which, I consider Tucker's injection lock to be 'proof' that a reactance tube VCO would have enough range since they are reducible to the same thing. You only need enough lock range for the channel width with 'getting close enough' being what the station tuning knob is for (antenna tank and VCO Fo). As a side note, for further entertainment, while searching for the injection lock papers I also ran across one on an "injection locked PLL" where the phase detector and 'VCO' functions for what is explicitly called a "PLL" are also subsumed inside the oscillator so you have the rather amusing case where a 'hidden' PLL is called injection locked but the other 'hidden' PLL is called a PLL. Not being an RF engineer I could be wrong but for AM demod we don't really give a hoot about phase, do we? I believe we do care about phase, and I don't think we need look any further than our old High School Trigonometry books to find out why, no RF engineering expertise required. *IIRC the AM equation Looks to me like you've already called on some "RF engineering" goes something like this, sin(wct) * (1 + sin(Wmt)). *If we multiply this by a locally generated copy of sin(Wct) in the receiver we get ((sin(wct))^2) * (1 + sin(Wmt)). *Without getting out my aforementioned High School Trigonometry book, I think ((sin(wct))^2) = 0.5 + 0.5 sin(2Wct) or something like that. *The 0.5 term multiplied by sin(Wmt) gives us back our original modulation. If we try to use a quadrateure carrier in the receiver, i.e. cos(Wct) so that we have cos(Wct) * sin(wct) * (1 + sin(Wmt)), and IIRC, without getting out the book, cos(Wct) * sin(wct) doesn't include a DC term, only a sin(2Wct) term, without the DC term we don't get demodulation of the desired signal, just quadrature noise. *Local carrier phase angles between sin(Wct) and cos(Wct) will produce some DC, although less than an in phase carrier, so the demodulated signal will be attenuated relative to an in phase carrier, and it will be full of quadrature noise that is demodulated by the cos(Wct) component of the locally generated carrier. Again I think your right. I was forgetting we've got a diode ring mixer in there.- Hide quoted text - - Show quoted text -- Hide quoted text - - Show quoted text - Perhaps you are seeing more about how a locked oscillator works. In practice, a locked oscillator produces a constant amplitude of identical frequency to the RF input which needs to be over a threshold to ensure locking, ie, synchronisation. But if the incoming RF is modulated then the oscillator signal phase relative to RF signal which we wish to demodulate becomes phase modulated which gives unwanted IMD effects. So the RF signal fed to the locked oscillator is ideally a highly limited sample of the RF signal derived from the receiver front end. But when 100% modulation occurs, the RF signal reduces to zero, and no synchronization is possible, and the oscillator drifts straight off the RF carrier F. But while the oscillator is locked, slight tuning each side of the "center frequency" alters the RF carrier to oscillator phase and it can be aatered to virtually zero which we might say is where the sound output will be greatest and best sounding. Then there are other bothers with synchrodyne selectivity. Usually most RF front ends on radios have at least one tuned LC tank which would track the tuning of the VFO. But Tucker's set just has a pass band and everything comes in, and so just what sort of selectivity and freedom from monkey chatter and other effects one gets is not clear. The superhet overcomes most problems one has when trying to build a Tucker circuit. Just why did nobody bother to commercially produce a Tucker Synchrodyne radio in the 1950s? I'm sure a lotta experimenty ppl must have tried, and I'm sure they mostly failed for a variety of reasons, and they'd have wasted days if not weeks trying to get the damn thing to stop making whistles and noises and then give sound as good as a well done superhet. All the synchrodyne radio circuits I have on paper files have RF tuning on their front end RF amps. Here is a brief list of what I have, which by now may have been put online by someone somewhere, or by nobody anywhere :- Ideas for the inventor ETI Feb 1980, a synchrodyne with ferrite rod antenna, one 415pF tuning cap, one mosfet in cascode with bjt for RF gain, then a standard FM detector and limiter chip such as LM2111, MC1351, LM1841, and several others similar. Works off +/-9Vdc. Radio, Television & Hobbies, Sep 1963, by Ian Pogson, Synchrodyne AM tuner. Uses 6EH7 RF amp with only input tank, tuned by one gang of 2 gang cap.Has 6BE6 VFO with LC tuned by the second gang. Has LCLC low pass filter for audio output from 6BE6 anode. Very simple. AF is -3dB at 10kHz! Synchrodyne AM Receiver by Linsley Hood, Part 2, Electronics and Wireless World, Feb 1986. Part 3 in March 1986. Has fiendishly complex circuit needing very cafeful PCB board desugn using GENUINE PLL with MC1496 phase detector, TLO71 dc amp, varicap diode to control oscillator F. Enormous expertise needed to reproduce such circuits. Synchronous Detection in Radio Reception - 1, by Pat Hawker, Wireless World, Sep 1972, part 2 Nov 1972. Lots of explanations about Tucker's efforts, then lots of discrete SS devices used in schematics shown. Extremely complex schematics shown alongside some simple types, all SS. A Homodyne Receiver by J W Herbert, Wireless World, Sep 1973, Not very complex, uses CA3028A plus MC1330P, dedicated for broadcast band, but has tuning by a pair of ganged variable inductors. Patrick Turner. |
#13
Posted to rec.audio.tubes
|
|||
|
|||
Phase Locked Loop with vacuum tubes.
"Don Pearce" wrote in message ... Yes, those had to be in phase with the transmitter's clocks. I remember some phase detectors in the old NTSC sets - an instant tip-off that there was some kind of phase locked oscillator in play. This is where PAL really won out over NTSC. Each alternate line the chroma phase leads or lags by 45 degrees, and the PLL locks to the average - which is true phase. The result is that when there is an overall phase error in the chroma signal, alternate lines lead and lag the proper phase. Add them together and the error vanishes. With NTSC a phase error means a wrong colour. NTSC = Never The Smae Color Twice. ;-) West German color TV was PAL, right? When I was there in the late 60s, the color was IMO better and more consistent than NTSC was at that time. With *only* about another 20 years of development, NTSC became generally quite stable and generally looked good by the 80s. ;-) |
#14
Posted to rec.audio.tubes
|
|||
|
|||
Phase Locked Loop with vacuum tubes.
On Fri, 22 Jul 2011 10:08:46 -0400, "Arny Krueger"
wrote: "Don Pearce" wrote in message ... Yes, those had to be in phase with the transmitter's clocks. I remember some phase detectors in the old NTSC sets - an instant tip-off that there was some kind of phase locked oscillator in play. This is where PAL really won out over NTSC. Each alternate line the chroma phase leads or lags by 45 degrees, and the PLL locks to the average - which is true phase. The result is that when there is an overall phase error in the chroma signal, alternate lines lead and lag the proper phase. Add them together and the error vanishes. With NTSC a phase error means a wrong colour. NTSC = Never The Smae Color Twice. ;-) West German color TV was PAL, right? When I was there in the late 60s, the color was IMO better and more consistent than NTSC was at that time. With *only* about another 20 years of development, NTSC became generally quite stable and generally looked good by the 80s. ;-) And now it is all digital, so no longer matters. Actually Germany was PAL B/G, while UK was PAL I. The only real difference was in the luminance bandwidth, which was a fair bit wider in PAL I. Better horizontal resolution was the result. I should have added that the two adjacent lines' chroma signals were added together by means of a delay line. This happened in good quality TVs. In cheap ones, almost the same effect was obtained visually by the lines being close enough together to effectively merge on screen. Poor beam focus helped enormously. d |
#15
Posted to rec.audio.tubes
|
|||
|
|||
Some thoughts on PLLs, Synchrodynes, and AM modulation depth
http://www.thevalvepage.com/radtech/...5/section5.htm However, from what the reprint says, Tucker was not trying to make a 'radio'. But it was a radio, damn it. and so just what sort of selectivity and freedom from monkey chatter and other effects one gets is not clear. The superhet overcomes most problems one has when trying to build a Tucker circuit. And the Tucker circuit overcomes problems with the superhet. I am not so sure it does. The superhet offered very good selectivity at the cost of audio bandwidth, but you could make a superhet with wide AF BW. Makers just didn't, because they'd successfully got people addicted to low AF BW whichsuited bean counters. AFTER you've tried to make a synchrodyne successfully, then see what you think. Neither are 'perfect' and I didn't present it as 'perfect'. It was intended to be an example and 'aid' to your quest, not a hobby construction kit. Indeed that's true. Just why did nobody bother to commercially produce a Tucker Synchrodyne radio in the 1950s? For one, because it isn't a 'radio'. Second it has, at least in raw form, some consumer irritating aspects. And, probably most important, it costs more than a superhet which has flawed but, nevertheless, satisfactory performance. But it IS a radio. I'm sure a lotta experimenty ppl must have tried, and I'm sure they mostly failed for a variety of reasons, and they'd have wasted days if not weeks trying to get the damn thing to stop making whistles and noises and then give sound as good as a well done superhet. So does a SSB receiver but people discovered how to make squelch circuits. Sure, but communications radio techniques are all mainly utterly unacceptable for Joe Punter, who wants excellent broadcast sound at the flick of a switch, and different stations by turning one knob, with no noises. So all the fancy techniches used to comm radio are irrelevant, because they focus on using small level signals to be able to be recieved in a crowded band fulla noise. Tucker's original already has the beginnings of one possible solution with the 'tuning indicator' (which, btw, is the beginnings of a second order PLL). I.E. MUTE the darn thing when off tune. That, of course, adds complexity so see third reason above. Indeed the mute was needed to block loud whistles when tuning. Today we'd use touch tuning in 10kHz steps in US or 9kHz here. The oscillator F is selected, not tuned with LC. If one wanted to, one probably could use an SS step tuned oscilator for use in a tubed superhet or synchrodyne. AM Stations here are all expected to have accurate station frequencies at 9kHz along the AM band and in fact there's no need for mechanical tuning. The type of oscillator won't affect the sound quality. But I thought you were all for the 'absolute best' instead of being a 'bean counter'. The absolute best cannot contain more tubes than it has to. And no less than it has to. I don't need to employ a bean counter. All the synchrodyne radio circuits I have on paper files have RF tuning on their front end RF amps. Here is a brief list of what I have, which by now may have been put online by someone somewhere, or by nobody anywhere :- Ideas for the inventor ETI Feb 1980, a synchrodyne with ferrite rod antenna, one 415pF tuning cap, one mosfet in cascode with bjt for RF gain, then a standard FM detector and limiter chip such as LM2111, MC1351, LM1841, and several others similar. Works off +/-9Vdc. Radio, Television & Hobbies, Sep 1963, by Ian Pogson, Synchrodyne AM tuner. Uses 6EH7 RF amp with only input tank, tuned by one gang of 2 gang cap.Has 6BE6 VFO with LC tuned by the second gang. Has LCLC low pass filter for audio output from 6BE6 anode. Very simple. AF is -3dB at 10kHz! Synchrodyne AM Receiver by Linsley Hood, Part 2, Electronics and Wireless World, Feb 1986. Part 3 in March 1986. Has fiendishly complex circuit needing very cafeful PCB board desugn using GENUINE PLL with MC1496 phase detector, TLO71 dc amp, varicap diode to control oscillator F. Enormous expertise needed to reproduce such circuits. Synchronous Detection in Radio Reception - 1, by Pat Hawker, Wireless World, Sep 1972, part 2 Nov 1972. Lots of explanations about Tucker's efforts, then lots of discrete SS devices used in schematics shown. Extremely complex schematics shown alongside some simple types, all SS. A Homodyne Receiver by J W Herbert, Wireless World, Sep 1973, Not very complex, uses CA3028A plus MC1330P, dedicated for broadcast band, but has tuning by a pair of ganged variable inductors. I'm sure they're interesting but without the articles they're not much else. But these things were leading edge thoughts on radio design in their day. You have not told me how your construction progress is proceeding on anything better than in all those articles by greater minds than yours. You seem to think that PLLs, double balanced mixers, and direct conversion are all 'failed' concepts but they're all over the place, like every cell phone to name but one incredibly common application. They became 'uber cheap' with integrated circuits but there's no reason variations of the same techniques can't be applied with tubes. I didn't say the ideas failed; I just said Synchrodyne radios failed to be manufactured because superhets were easier and cheaper. Trying transfer ideas used in mobile telephone chips to tubes is often virtually impossible, just making a Gilbert Cell needs about 4 twin triodes so just to do one function tales as many tubes as have been used in a complete tube radio. Mobile phones sound dreadful. Amusingly enough, the injection locked oscillator you seem to disdain also seems to be all the rage these days. If I were so inclined I'd probably look for some pre-made diode ring mixers to avoid the misery of trying recreate that wheel. Now, here's one that's interesting, the Injection locked PLL I mentioned earlier http://my.ece.ucsb.edu/yorklab/publi...20-%20mtt%20au... It uses a FET but that should be tube transferable. The schematic is figurative, and can't be built from info offered. It is als designed for 10GHz. But probably varying a triode's bias current would change F a bit. There are a number of reactance tube tuners around. I'd still tune with an LC, for the same kind of 'Tucker' injection loc and 'wide tuning range' (with the 'big mechanical cap'), and see if I could somehow duplicate the second order PLL function with the varactor(s), like they did, since that has to pull only enough to bring phase in lock. Or I'd just use a reactance tube and make a 'proper' PLL loop with it but that might take a second limited mixer to get the AM off it. That's what makes the ILPLL interesting since it uses the gain element as a mixer so you'd just need to use an RF limiter for the injection. I recall making a limiter like that used in a tubed FM radio 10.7MHz IF strip. You set up a 6AU6 so it spends its life grossly overloaded, and anode current is just basically square waves, so very little input signal is need for that, especially if you have two such simple tube stages cascaded. But still you get drop out on 100% mod peaks. The changes in carrier F are faithfully preserved during such techniques. The better way to make a syhchronous AM radio with only as many tubes as needed for signal handling is to use a digital controller for oscillator F. These have been around for 30+ years. You get push button tuning. One could have push button oscillator tuning and then have input RF tuning with conventional tuning gang. One could even have a switchtable set of say 12 preset frequencies with 12 variable preset caps. One could copy part of what has been done already in many push button type AM/FM tuners. I'd prefer to used ganged tuning caps and have a muting device which turns off when lock is achieved. That would be easy with a some discrete bjts. Although I like green eyed tube tuning indicators, a meter or row of LEDS is just as good. Patrick Turner. |
#16
Posted to rec.audio.tubes
|
|||
|
|||
Phase Locked Loop with vacuum tubes.
Don't forget SECAM, System European, and Contrary to Any other Method
|
#17
Posted to rec.audio.tubes
|
|||
|
|||
Phase Locked Loop with vacuum tubes.
On Fri, 22 Jul 2011 23:41:18 -0400, Don
wrote: Don't forget SECAM, System European, and Contrary to Any other Method The French. Says it all, really. d |
#18
Posted to rec.audio.tubes
|
|||
|
|||
Some thoughts on PLLs, Synchrodynes, and AM modulation depth
In article ,
Patrick Turner wrote: http://www.thevalvepage.com/radtech/...5/section5.htm Today we'd use touch tuning in 10kHz steps in US or 9kHz here. The oscillator F is selected, not tuned with LC. If one wanted to, one probably could use an SS step tuned oscilator for use in a tubed superhet or synchrodyne. AM Stations here are all expected to have accurate station frequencies at 9kHz along the AM band and in fact there's no need for mechanical tuning. The type of oscillator won't affect the sound quality. Perhaps not for you, however different oscillators have different amounts of sideband noise that can degrade RF performance of the radio, and some people with good hearing may be sensitive to the artifacts introduced by different types of oscillators. But I thought you were all for the 'absolute best' instead of being a 'bean counter'. The absolute best cannot contain more tubes than it has to. And no less than it has to. That's a little vague; it sounds like the number of tubes needed is subjective? You seem to think that PLLs, double balanced mixers, and direct conversion are all 'failed' concepts but they're all over the place, like every cell phone to name but one incredibly common application. They became 'uber cheap' with integrated circuits but there's no reason variations of the same techniques can't be applied with tubes. I didn't say the ideas failed; I just said Synchrodyne radios failed to be manufactured because superhets were easier and cheaper. Trying transfer ideas used in mobile telephone chips to tubes is often virtually impossible, just making a Gilbert Cell needs about 4 twin triodes so just to do one function tales as many tubes as have been used in a complete tube radio. Mobile phones sound dreadful. I think the point was that many cell phones use direct conversion receivers, which is essentially what we are talking about here. In the case of cell phones the radio handles digital data, not analog signals, and the low rate digital speech coders used are the reason for the dreadful sound. There are apparently ³HD² cell phones available now in some countries that use higher bit rates and ³better² speech coders. What we are talking about is using direct conversion receivers to demodulate good old analog AM, not digital data, so the dreadful mobile phone sound is irrelevant. Direct conversion receiver concepts are easily transferable to tubes. The better way to make a syhchronous AM radio with only as many tubes as needed for signal handling is to use a digital controller for oscillator F. These have been around for 30+ years. You get push button tuning. One could have push button oscillator tuning and then have input RF tuning with conventional tuning gang. Using ³a digital controller for oscillator F² doesn't give you a synchronous AM radio, you still need a synchronous detector with some kind of PLL or injection locked oscillator, whether it operates at the RF or IF frequency. -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
#19
Posted to rec.audio.tubes
|
|||
|
|||
Phase Locked Loop with vacuum tubes.
"Don Pearce" wrote in message ... On Fri, 22 Jul 2011 23:41:18 -0400, Don wrote: Don't forget SECAM, System European, and Contrary to Any other Method The French. Says it all, really. No only the French, but the Russians originally chose SECAM too. I do not know why the French preferred SECAM to PAL, but for the Russians the answere was obvious. The thing is, colours in SECAM are *least* affected by nonlinearity of amplitude/phase response of the signal chain (around the colour subcarrier(s), which are two in SECAM). For example, in the direct vicinity of a TV station a good NTSC TV set will potentially give the best picture quality, because colour demodulation is linear, and each line is independent. PAL will be the second because of averaging of two lines, and SECAM the worst because FM colour subcarrier modulation inherently is limited by slew rate, and colour transitions are not that sharp. And now suppose you pass your TV signal through a couple of analogue radio-relay links. Each would introduce a significant phase shift (up to +/-45deg), especially the Russian ones, where everything electronic was of inferior quality. NTSC will lose its colour completely. PAL will preserve colour, but will lose saturation or the colours will be "pulsating" because of phase jitter. SECAM will still show correct well saturated colours. Not surprising, because though phase might be unstable intermittently, the frequency on average does not change. The colours will be more noisy (phase jitter--frequency jitter), transitions will not be as crisp (like the picture will be rather coloured than colour), but on large areas the colour reproduction and saturation will be the same as in the original studio material. No wonder the Russians in those days chose SECAM because vast size of the country assumed lots of radio-relay links. Now when everything is digital and satellite, PAL is the most common in Europe and better than SECAM, and Russia switched to PAL. |
#20
Posted to rec.audio.tubes
|
|||
|
|||
Some thoughts on PLLs, Synchrodynes, and AM modulation depth
On Jul 23, 3:36*pm, flipper wrote:
On Fri, 22 Jul 2011 18:27:12 -0700 (PDT), Patrick Turner wrote: http://www.thevalvepage.com/radtech/...5/section5.htm However, from what the reprint says, Tucker was not trying to make a 'radio'. But it was a radio, damn it. The man who built it says otherwise and his opinion on what he himself intended trumps yours. Well I don't mind much about being trumped by a bloke who knew +20dB more than I do about radio sets. But for me and thousands of others, a Tucker Synchrodyne = radio, right? and so just what sort of selectivity and freedom from monkey chatter and other effects one gets is not clear. The superhet overcomes most problems one has when trying to build a Tucker circuit. And the Tucker circuit overcomes problems with the superhet. I am not so sure it does. I trust the opinion of the man who built it more than the one who hasn't. I suggest you BUILD a Tuckerdyne radio, before you place any trust in anyone. Having a permantly skeptical attitude is part of the method for developing ideas. The superhet offered very good selectivity at the cost of audio bandwidth, but you could make a superhet with wide AF BW. Makers just didn't, because they'd successfully got people addicted to low AF BW whichsuited bean counters. You need to decide whether your going to keep claiming it's wonderful or crap. AFTER you've *tried to make a synchrodyne successfully, then see what you think. I already told you I have no interest in building one and neither do I intend to build an F-22 to see what it's like. Then you'll end up more ignorant about matters raised in this discussion than I am. I found Tucker's idea was difficult to make work. Superhets are easier to make work. Sure, but communications radio techniques are all mainly utterly unacceptable for Joe Punter, who wants excellent broadcast sound at the flick of a switch, and different stations by turning one knob, with no noises. Irrelevant. Of course. YOU rabbit on with irrelvant BS then claim my comments are irrelevant. So all the fancy techniches used to comm radio are irrelevant, because they focus on using small level signals to be able to be recieved in a crowded band fulla noise. Nonsense. Communications radio sets share many features with entertainment mass media tube radios, but the comm sets use many techniques specific to their use which are not wanted in the 6 tube set on the mantle piece used for local station AM reception. Indeed the mute was needed to block loud whistles when tuning. Today we'd use touch tuning in 10kHz steps in US or 9kHz here. The oscillator F is selected, not tuned with LC. If one wanted to, one probably could use an SS step tuned oscilator for use in a tubed superhet or synchrodyne. AM Stations here are all expected to have accurate station frequencies at 9kHz along the AM band and in fact there's no need for mechanical tuning. The type of oscillator won't affect the sound quality. See? Try actually solving problems instead of perpetually bitching and you might get somewhere. And just how many good ideas about PLL have you offered to the group so far which can be implemented NOW, using tubes, and perhaps improved upon? I'm sure they're interesting but without the articles they're not much else. But these things were leading edge thoughts on radio design in their day. Doesn't matter if they were reverse engineered from crashed space aliens at Area 51. Titles don't tell me anything other than someone wrote something. That's it, be determined to stay ignorant. But the author list I mention do have lots to say that s relevant. You have not told me how your construction progress is proceeding on anything better than in all those articles by greater minds than yours. Yes I have. Precisely 0% seeing as how I still have no interest in building one, just like I've told you every time you babble that crap. Ah, just an arm chair solderer & general bull**** artist. Go make something or do something useful. Now, put a cleverly sized ... Indeed, but I don't have time to R&D a better AM radio right now. Most of my work with radios is re-engineering existing superhet AM radios mainly wanted for strong local stations, so even considering changing such things to use synchronous AM detection is never my aim, mainly because I get excellent results with a very simple use of CF plus diode +R+C AM detector. It seems nobody has ever bothered to build a real PLL with tubes for BC band radios. Its all become old hat in the world of chips. But I'm into tubes, and to do better than Tucker did I'd have to work longer and harder at it. Patrick Turner. |
#21
|
|||
|
|||
Quoted by Teflon John B
"Just because I'm not in the mood to build your toy for you." Don't depend on Teflon John B to build, provide schematics or anything else for us. Tefln John B likes others to do the sweat while he sits back & comments. But in the real world action talks & sheisse walks. I would not want Teflon John B on my team. What a drag!! Cheers to all, John S |
#22
Posted to rec.audio.tubes
|
|||
|
|||
Phase Locked Loop with vacuum tubes.
On Jul 23, 8:59*pm, John L Stewart John.L.Stewart.
wrote: Quoted by Teflon John B "Just because I'm not in the mood to build your toy for you." Don't depend on Teflon John B to build, provide schematics or anything else for us. Tefln John B likes others to do the sweat while he sits back & comments. But in the real world action talks & sheisse walks. I would not want Teflon John B on my team. What a drag!! Cheers to all, John S -- John L Stewart I've always found "Teflon John B" a tolerable, and even likable fellow and I think he'd like us to know he's here for the intellectual discussion and not to show off by his prolific efforts with a soldering iron. He seems to have a respectable dislike for "ego trips". Like myself and some others here, he struggles with ageing. There are others here that "jarg the jargon" about electronics with or without using tubes but whose lengthy posts don't really contribute anything to making a better tube amp, and better AM radio, or a better tubed anything, such as a more linear AM detector. Unfortunately, the nursing home where Mr Huelett and Mr Packard reside at the ages of 112 and 113 does not have an internet facility, and of course the zeal of these gentlemen now only extends to scrabble and chess. I heard Peter Walker and Peter Baxandal have developed a mildly allergic reaction to when anyone says the word "valve" or "transistor", and the nurses in the UK establishments have to warn visitors about their langauge, should they be granted an audience. What seemed to start all this crap about PLL in AM radios was the idea synchronous detection might work better than a diode + R + C detector. People said there was never any need for a cathode follower to be used to which I basically said "******** ! , of course there is a damn need", and I recommend a pair of CF, not just one, to avoid the dreadful cut off distortion at high % mod. Anyway, I can think of two more ways to use a CF to drive a detector circuit. First is to have the usual CF, but its cathode is loaded with RF choke to 0V. This RFC should be pie wound to avoid any C 10pF and to ensure CF and RFC do not act like a C+R detector. The CF triode may need an RC circuit between cathode and RFC to give cathode biasing. The IFT2 sec is at 0V potential, and direct couples to CF grid. Then one has a Ge diode working from cathode into 22k resistance with one end to 0V, and thus you get a waveform which is 1/2 the envelope which is then directly applied to to the second grid of a CF. No capacitors are used yet, and circuit Z is low enough to allow use of Ge diodes or 6H6, 6AL5 etc. The second CF cathode has 47k taken to -150Vdc rail, and the output from cathode is taken to an LC filter with just enough R across the C to critically damp the Fo of the LC, which should be at about 32kHz. The XL and XC should be about 33k at 32kHz, ie, use 160mH plus 150pF, with R damping of 47k. The choke should be pie wound to reduce self capacitance. It could be a toroidal cored type with just the right core permeability. Such a filter works like a choke input power supply which detects the average peak voltage of the applied ac waveform. Output is low, but free of most unwanted characteristics of most "conventional" or "booooring" detectors to which Alex has referred. Methinks the technical term for this detector is "infinite impedance detector" becase the load on the IFT2 sec is indeed a virtual very high Z of the CF. To get Q to where its wanted, R loading on the LC between 100k and 220k should be used. Such R loading won't adversely load the IF tube. The second method is to use CF1 set up as above, but the RFC between cathode and 0V is in fact a tapped autotransformer which raises the 455kHz IF waves to 3 times their level at cathode. The autotransformer might also be tuned with a cap across the whole coil for 455kHz. This increased IF voltage level is then applied to the diode and say a 47k to 0V and thus the following levels of detected AF will be 3 times higher. The CF need to be set up to cope with the dynamic range of expected signals which should not be a problem in a set with AGC voltages applied. Because the detected waves are taken from a grounded winding it is possible to use two diodes so that a positive going and negative going output voltage is detected and thus TWO CF would be needed to buffer the outputs from diode + R. Nice if one wants to drive a PP amp in the radio. Meanwhile, many retired bean counters in nursing homes have been denied access to rec.audio.tubes. The nurses decided that bean counters endured severe depression and trauma after reading my posts which recommend nothing they can ever agree with. Meanwhile, Cadel Evans finally seems to have won the Tour De France which goes to show that a short stature squeaky voiced humble pie Ozzie could beat the pair tall good looking Schlek brothers from Luxenburg. And Cadel is 8 years older than Andy Schlek, and didn't have so much team support. When you go for a ride today, watch you don't go to hard, and ride to enjoy the afterglow of doing something, and not to spend 4 hours so exhausted you can't do other sunday jollies. It is forecast to be only 8C here today, maybe some rain, but I think I'll do 70km, weather permitting - Too much yellow glow about to stay indoors. Patrick Turner. |
#23
Posted to rec.audio.tubes
|
|||
|
|||
Some thoughts on PLLs, Synchrodynes, and AM modulation depth
In article ,
Patrick Turner wrote: On Jul 23, 3:36*pm, flipper wrote: On Fri, 22 Jul 2011 18:27:12 -0700 (PDT), Patrick Turner wrote: http://www.thevalvepage.com/radtech/...5/section5.htm and so just what sort of selectivity and freedom from monkey chatter and other effects one gets is not clear. The superhet overcomes most problems one has when trying to build a Tucker circuit. Can you give a brief overview of the type of problems you encountered when you tried to build a "Tucker circuit"? And the Tucker circuit overcomes problems with the superhet. I am not so sure it does. I trust the opinion of the man who built it more than the one who hasn't. I suggest you BUILD a Tuckerdyne radio, before you place any trust in anyone. Having a permantly skeptical attitude is part of the method for developing ideas. This comment suggests that you encountered substantial problems implementing the "Tucker circuit"? The superhet offered very good selectivity at the cost of audio bandwidth, but you could make a superhet with wide AF BW. Makers just didn't, because they'd successfully got people addicted to low AF BW whichsuited bean counters. You need to decide whether your going to keep claiming it's wonderful or crap. AFTER you've *tried to make a synchrodyne successfully, then see what you think. I already told you I have no interest in building one and neither do I intend to build an F-22 to see what it's like. Then you'll end up more ignorant about matters raised in this discussion than I am. I found Tucker's idea was difficult to make work. Did you try to build "tucker's circuit" from an existing schematic design, or did you do your own design from the ground up? If you used an existing schematic could you indicate which one it was? I take it that what you built operated the detector at the actual frequency of the broadcast station being received? Did your design have only a tuned oscillator, or did it also include a tuned RF or antenna stage? What circuit did you use for the synchronous detector itself? This link to the "Tucker circuit" was given at the top of this post. http://www.thevalvepage.com/radtech/...5/section5.htm The "Tucker circuit" in Figure 8 looks virtually identical to the tube circuits used to demodulate the L-R stereo sub carrier in the FM stereo system. The only reason it should be anymore difficult to get operating than an FM stereo decoder, and it is a big problem, is the fact that it is designed to be tuned over the three to one frequency range of the MW broadcast band, rather than operating at a single frequency, like a stereo decoder. That is why I earlier suggested using the circuit as a detector in a superheterodyne, where it would only have to operate at a single frequency, and wouldn't have to accommodate variable tuning. Reviewing this page I just discovered the "Curtis Homodyn Receiver" in figure 6 on the previous web page, here. http://www.thevalvepage.com/radtech/...4/section4.htm This circuit is essentially identical to what I proposed in an earlier post, and might be easier to get working than the "Tucker circuit". The downside is that the schematic is something of an artist's misconception, while the schematic of the "tucker circuit" is more like a real schematic that you could directly build from. One thing I don't understand about the "Curtis Homodyn Receiver" is that in addition to the Reactance tube that controls the VCO frequency in the PLL, there is also a "Synchronization By Injection" circuit feeding the 4th grid of the oscillator tube with some signal from the RF amplifier. At first glance this seems redundant, however from my experience with "old" FM transmitters employing PLLs, I know that these circuits don't always lock at turn on from a cold start without some assistance from a startup circuit. Does anyone know if that is the function served by the "Synchronization By Injection" circuit in the "Curtis Homodyn Receiver"? -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
#24
Posted to rec.audio.tubes
|
|||
|
|||
Phase Locked Loop with vacuum tubes.
On Jul 22, 10:08*am, "Arny Krueger" wrote:
"Don Pearce" wrote in message ... Yes, those had to be in phase with the transmitter's clocks. I remember some phase detectors in the old NTSC sets - an instant tip-off that there was some kind of phase locked oscillator in play. This is where PAL really won out over NTSC. Each alternate line the chroma phase leads or lags by 45 degrees, and the PLL locks to the average - which is true phase. The result is that when there is an overall phase error in the chroma signal, alternate lines lead and lag the proper phase. Add them together and the error vanishes. With NTSC a phase error means a wrong colour. NTSC = Never The Smae Color Twice. ;-) West German color TV was PAL, right? When I was there in the late 60s, the color was IMO better and more consistent than NTSC was at that time. With *only* about another 20 years of development, NTSC became generally quite stable and generally looked good by the 80s. ;-) NTSC NEVER looked good . It was the oldest color TV standard in the world . Nobody will miss it . |
#25
Posted to rec.audio.tubes
|
|||
|
|||
Phase Locked Loop with vacuum tubes.
On Jul 23, 11:38*pm, flipper wrote:
On Sat, 23 Jul 2011 20:24:18 -0700 (PDT), wrote: On Jul 22, 10:08*am, "Arny Krueger" wrote: "Don Pearce" wrote in message ... Yes, those had to be in phase with the transmitter's clocks. I remember some phase detectors in the old NTSC sets - an instant tip-off that there was some kind of phase locked oscillator in play. This is where PAL really won out over NTSC. Each alternate line the chroma phase leads or lags by 45 degrees, and the PLL locks to the average - which is true phase. The result is that when there is an overall phase error in the chroma signal, alternate lines lead and lag the proper phase. Add them together and the error vanishes. With NTSC a phase error means a wrong colour. NTSC = Never The Smae Color Twice. ;-) West German color TV was PAL, right? When I was there in the late 60s, the color was IMO better and more consistent than NTSC was at that time. With *only* about another 20 years of development, NTSC became generally quite stable and generally looked good by the 80s. ;-) NTSC NEVER looked good . It was the oldest color TV standard in the world . Nobody will miss it . Neither did PAL or SECAM. The 50Hz flicker drove me near insane, negating any other alleged 'features'.- Hide quoted text - - Show quoted text - How is that ? Even the motion picture standard was adopted as a mere 24 frames per second . Do you possess an unusual biological-optical ability to see thing moving that fast ? |
#26
Posted to rec.audio.tubes
|
|||
|
|||
Phase Locked Loop with vacuum tubes.
SNIP,
NTSC NEVER looked good . It was the oldest color TV standard in the world . Nobody will miss it . Neither did PAL or SECAM. The 50Hz flicker drove me near insane, negating any other alleged 'features'.- Hide quoted text - - Show quoted text - How is that ? Even the motion picture standard was adopted as a mere 24 frames per second . Do you possess an unusual biological-optical ability to see thing moving that fast ?- Hide quoted text - - Show quoted text - Flipper evolved by means of a spurious genetic transmogrification leading to him having the extra sensitive optical capabilities of a fish, which needs to be able to see shark approaching with mouth wide open, lest he be gobbled up. Works two ways though. If I see Angelina Jollie coming at me with mouth wide open I let her gobble me up, and I don't mind slower frame speeds at all, to be sure now. Patrick Turner. |
#27
Posted to rec.audio.tubes
|
|||
|
|||
Phase Locked Loop with vacuum tubes.
On Jul 24, 1:31*am, Patrick Turner wrote:
SNIP, NTSC NEVER looked good . It was the oldest color TV standard in the world . Nobody will miss it . Neither did PAL or SECAM. The 50Hz flicker drove me near insane, negating any other alleged 'features'.- Hide quoted text - - Show quoted text - How is that ? Even the motion picture standard was adopted as a mere 24 frames per second . Do you possess an unusual biological-optical ability to see thing moving that fast ?- Hide quoted text - - Show quoted text - Flipper evolved by means of a spurious genetic transmogrification leading to him having the extra sensitive optical capabilities of a fish, which needs to be able to see shark approaching with mouth wide open, lest he be gobbled up. Works two ways though. If I see Angelina Jollie coming at me with mouth wide open I let her gobble me up, and I don't mind slower frame speeds at all, to be sure now. Patrick Turner. Amen . |
#28
Posted to rec.audio.tubes
|
|||
|
|||
Some thoughts on PLLs, Synchrodynes, and AM modulation depth
In article ,
flipper wrote: On Sat, 23 Jul 2011 21:32:25 -0500, John Byrns wrote: This link to the "Tucker circuit" was given at the top of this post. http://www.thevalvepage.com/radtech/...5/section5.htm The "Tucker circuit" in Figure 8 looks virtually identical to the tube circuits used to demodulate the L-R stereo sub carrier in the FM stereo system. The only reason it should be anymore difficult to get operating than an FM stereo decoder, and it is a big problem, is the fact that it is designed to be tuned over the three to one frequency range of the MW broadcast band, rather than operating at a single frequency, like a stereo decoder. That is why I earlier suggested using the circuit as a detector in a superheterodyne, where it would only have to operate at a single frequency, and wouldn't have to accommodate variable tuning. I still don't understand why you and Patrick see the 3:1 ratio as a PLL problem because the PLL doesn't have to pull and lock over that range, just whatever max station deviation from channel frequency is. You turn the tuning knob to get things 'in range'. I missed the post where Patrick mentioned the 3:1 ratio as a problem? My only point is that I can't see how having to deal with the 3:1 oscillator tuning ratio, including tracking the Antenna and RF tuning sections, could possibly make it easier to get this circuit working properly in your workshop in the back garden shed. Operating the synchrodyne detector circuit at IF should simplify getting it working properly, once one's confidence has been bolstered by conquering this first step, then one could move on to the version operating directly on frequency. My thought was simply to learning to crawl before trying to learn to walk. I would want my receiver to receive any station whose "primary service area", as defined by the FCC in the US, I was located within. This probably would call for a tuned antenna stage to minimize 3rd order IM. Also the RF amplifier output should probably also be tuned to maximize signal to noise ratio, especially for those stations where I am located on the fringe of the "primary service area. This would call for a three gang tuning capacitor, or three varactors. If the PLL pulled and locked over the entire MW band how the heck would you select a station? I'm not sure I understand the question?! The oscillator has some means of being manually tuned, a variable capacitor or other tuning device, stations are tuned by setting the oscillator frequency to the frequency of the station it is desired to receive, then the loop is made to lock on that frequency. An IF might be a good idea from the aspect of eliminating Lo feedback into the antenna and RF amp, though. And of course as we would expect it also brings some new problems of its own. Actually, I've seen some 'synchronous detector' kits intended to be tacked onto an existing radio's 455 kHz IF. I have three AM broadcast receivers here in the house that use a superheterodyne circuit and a synchronous detector. Reviewing this page I just discovered the "Curtis Homodyn Receiver" in figure 6 on the previous web page, here. http://www.thevalvepage.com/radtech/...4/section4.htm I was just about to suggest Patrick read the whole Tucker article. The earlier Bellescize circuit 'looks' more like a PLL but I suspect it suffers from DC offset and phase modulation. [Snip] You have to 'translate' their 'radio speak' into 'PLL speak'. Like, when Tucker explains "it is evidently necessary to prevent the natural frequency of the oscillator from drifting very far from the frequency to which it is made to synchronize" he's simply saying we need phase lock and 'the problem' is caused by injection lock being a first order PLL, with the attendant phase shift depending on how far the phase difference has to 'push' on the Lo to get it on frequency. And then there's AM phase distortion which I'm not sure either addressed. What is the nature of the "AM phase distortion" you are speaking of? One of the advantages of the synchronous detector over Patrick's "C+D" envelope detector is that it is relatively immune to phase and amplitude distortion of the modulated carrier being received. -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
#29
Posted to rec.audio.tubes
|
|||
|
|||
Phase Locked Loop with vacuum tubes.
In article ,
John L Stewart wrote: Quoted by Teflon John B "Just because I'm not in the mood to build your toy for you." Don't depend on Teflon John B to build, provide schematics or anything else for us. Tefln John B likes others to do the sweat while he sits back & comments. But in the real world action talks & sheisse walks. I would not want Teflon John B on my team. What a drag!! Cheers to all, John S Hi John, Is the ³Teflon John B² you are referring to me, or is it just my ego that makes me think that you are talking about me? If ³Teflon John B² is not me, who is it? Or are you saying that you believe ³Flipper² is one of my aliases? If ³Teflon John B² is me, then I would suggest that you are slightly confused, or maybe more than just slightly. I did not make the original statement "Just because I'm not in the mood to build your toy for you" that you quote. It was Flipper that made the original comment, and I don't think I even repeated the quote as part of a reply to Flipper's post. There are two possible causes for your confusion. One, it may be an artifact of ³AudioBanter², the site you use as an interface to the Usenet, or two, it may be something more serious which I will leave for you to contemplate privately. -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
#30
Posted to rec.audio.tubes
|
|||
|
|||
Phase Locked Loop with vacuum tubes.
On Jul 25, 9:15*am, John Byrns wrote:
In article , *John L Stewart wrote: Quoted by Teflon John B "Just because I'm not in the mood to build your toy for you." Don't depend on Teflon John B to build, provide schematics or anything else for us. Tefln John B likes others to do the sweat while he sits back & comments. But in the real world action talks & sheisse walks. I would not want Teflon John B on my team. What a drag!! Cheers to all, John S Hi John, Is the ³Teflon John B² you are referring to me, or is it just my ego that makes me think that you are talking about me? If ³Teflon John B² is not me, who is it? Or are you saying that you believe ³Flipper² is one of my aliases? If ³Teflon John B² is me, then I would suggest that you are slightly confused, or maybe more than just slightly. I did not make the original statement "Just because I'm not in the mood to build your toy for you" that you quote. *It was Flipper that made the original comment, and I don't think I even repeated the quote as part of a reply to Flipper's post. There are two possible causes for your confusion. *One, it may be an artifact of ³AudioBanter², the site you use as an interface to the Usenet, or two, it may be something more serious which I will leave for you to contemplate privately. -- Regards, John Byrns Surf my web pages at, *http://fmamradios.com/ Just who did say "Just because I'm not in the mood to build your toy for you." ?? It is a Flippereque statement as there is a style that is different to "Teflon John". But if Flipper said it, or John Byrns said it, probably it matters not, because both John Byrns and Flipper are loathe to build anything at any time and both wanna just sit around all day chatting while doing nothing. OK for them. I have work to do. Soldering up tube circuits that work well. Patrick Turner. |
#31
Posted to rec.audio.tubes
|
|||
|
|||
Turner's Synchrodyne (was: Some thoughts on PLLs, Synchrodynes)
In article ,
flipper wrote: On Sun, 24 Jul 2011 17:36:31 -0500, John Byrns wrote: In article , flipper wrote: On Sun, 24 Jul 2011 12:03:43 -0500, John Byrns wrote: In article , flipper wrote: I still don't understand why you and Patrick see the 3:1 ratio as a PLL problem because the PLL doesn't have to pull and lock over that range, just whatever max station deviation from channel frequency is. You turn the tuning knob to get things 'in range'. I missed the post where Patrick mentioned the 3:1 ratio as a problem? He claimed, first a varactor and then, a reactance tube couldn't handle the range. Yes, I do remember those complaintsof Patrick's, I didn't associate them with the problem of integrating the oscillator and antenna tuning though. Well, not being able to tune the range at all sort is sort of 'integral' to everything It isn't clear what kind of schematic he was trying to build from, I can see that he may have had trouble getting a reactance tube working properly although I would think it would be child's play to get a varactor working for the VCO/PLL function. I don't think he was working from anything. Just in a "won't work" mood. Patrick has been unusually closed mouthed about what it was that he actually tried building. [Snip] I think the reason for the plethora of approaches is there isn't really a 'project'. I got involved when Patrick said he couldn't find any 'tube' PLLs and I was just explaining that he wouldn't find them described as such because the term didn't exist until later, and then provided some examples. Trying to explain they existed sort of drove the conversation in all manner of directions, like Patrick trying to 'coerce' me into building a 'Tucker'. They may even have existed in the tube days, in exactly the form Patrick is looking for, during the first run at AM stereo in the US in the late 1950s. There were half a dozen or so systems proposed by various companies, most of which used FM or PM to convey the stereo difference signal, but there may have been a QAM proposal among them that would have included the tube PLL circuit Patrick is looking for. I have descriptions of all the systems buried in my library, but due to extreme domestic entropy, which is this summers project to make a dent in, I don't have the data to hand to verify if one of the systems was a QAM system with a PLL that Patrick could use. Since Patrick has been loath to rise to the bait I have dangled and give us an overview of what sort of Synchrodyne he has attempted to solder together, with unsatisfactory results, I decided to root through the archives to see what he said on the subject in the past. I found the following quotes in the rec.audio.tubes archives, except for the first which was part of a personal email I received from Patrick. I hope he will forgive my lapse of Netiquette in quoting it here in this public group. From a private email: Date: Sat, 18 Aug 2001 11:35:27 +1000 "I guess true synchronous detection could be used for the 38 Khz signal, just like a Synchrodyne reciever. The one I have in mind was By D.G. Tucker from 1947, but it never caught on, as it really did not offer any better overall performance, and this method of direct detection did not catch on until chips came along. There was a Synchrodyne circuit for AM which came out in Electronics Australia in 1962, and it used a 6BE6 as the self oscillating mixer with the oscillations locked to the incoming carrier, and a simple R for a load, and a C to remove RF ripple." From the usenet group rec.audio.tubes: Date: Thu, 17 Jan 2002 15:32:00 +1100 "Less is more sometimes with radio,and two detectors in series seem unecessary." "I tried building a Synchrodyne reciever, with tubes, and it quite stumped me, for a variety of reasons, so I returned to the idea of a decent superhet, and have not looked back. D.G.Tucker brought out a tubed pure Synchrodyne in 1947, and I started with a similar idea to his, and found it difficult to stop it oscillating where I didn't want it too." Date: Sat, 12 Jun 2004 01:02:30 +1000 "My paper files have around 20 different circuits for AM tuners including a fairly simple synchrodyne ( or direct conversion ) two tube sets which use a 6EJ7 for an RF amp, and followed by a 6BE6 synchronous detector." "I tried building one, but audio output was low, stability was difficult, and and a superhet proved far better." "Then there were several if not many synchrodyne and some homodyne designs in Wireless World over the years, but these were all chip based, except the early D.G.Tucker circuit of 1947." Date: Wed, 23 Jun 2004 02:10:33 +1000 "I did play around with the Tucker tubed synchrodyne but gave up when I concluded that my superhet idea was easier to build." Date: Fri, 12 Nov 2010 02:11:07 -0800 (PST) "I guess a PLL oscillator which would "run on" during short time absences of a synchronized and limited carrier would be better. The last time I tried was 1999 when I spent a week trying t get a tubed "Synchrodyne" first designed by a Mr Tucker in 1947, and published by Wireless World that year. It never was a commercial success and could not compete with good superhet performence, at least for broadcast AM bands. Besides, stations today routinely use right up to 100% modulation." "Anyway, my synchrodyne was very hard to get working anywhere near as well as a good superhet even when I resorted to using a much simpller circuit with 6EJ7 input RF gain amp and a 6BE6 self oscillating AM detector; plus there's the howls and whistles when tuning and the the monkey chatter from other stations." Date: Mon, 11 Jul 2011 03:13:34 -0700 (PDT) Subject: VLF stability in Williamson-type amplifiers "I tried to build a version of the DG Tucker Synchrodyne circuit published in Wireless World in about 1947. Maybe its now online some place, and what prevented progess was the making of a suitable balanced synchronous demodulator. Probably the best way is to use a toroidal transformer with the core ? just right for range of RF frequencies. But I never had time to explore how to make such a tranny. The tucker circuit uses lots more tubes than a superhet and needs much more practical expertise to get running without loud whistles while tuning. It was a potentially excellent thing but it was never to become popular because the superhet was king. Then tried making a synchrodyne based on a self oscillating 6BE6, and there are some simple circuits of those around, just RF input tube and 6BE6 needed only. That sort of worked a bit but monkey chatter and whistles and controlling oscillations just right were easier said than done so I abandoned the idea and proceeded with a good superheat." That brings us to the current thread which Patrick started on 18 Jul 2011 under the subject "Phase Locked Loop with vacuum tubes." I later changed the subject to "Some thoughts on PLLs, Synchrodynes, and AM modulation depth", and with this post am changing again to " Turner's Synchrodyne" The above quotes make it fairly clear that what Patrick built was a project described in a 1962 issue of "Electronics Australia". Patrick says this design used two tubes in a direct conversion circuit, with a 6EJ7 as an RF amplifier stage, and a 6BE6 as a combined oscillator and synchronous detector. Presumably the oscillator section of the 6BE6 was "injection locked". I suspect that an easier to get working beginner's circuit, requiring no special parts, could be created by adding only one tube to the Electronics Australia project. First in keeping with my notion of crawling before walking, I would change the design from a direct conversion receiver to a superheterodyne receiver with a synchronous detector. I would replace the 6EJ7 RF stage with a second 6BE6 operating as a converter from the broadcast frequency to a 455 kHz IF. The synchronous detector and oscillator would remain essentially the same as in the "Electronics Australia" design using a 6BE6. The big change to the circuit would be the means of synchronizing the oscillator portion of the synchronous detector. Instead of injection locking the oscillator in the second 6BE6, a PLL would be used. A third tube, a 6AU6 would be added, with its grid feed from the same IF signal feeding the 6BE6 synchronous detector. A slightly modified 455 kHz discriminator transformer, an easily available part, is used along with two GE diodes as a phase detector with the 6AU6 feeding one input and the 455 kHz oscillator section of the second 6BE6 feeding the phase detector's second input. The output of the phase detector, through a loop filter, would then control a varactor diode connected in the 455 kHz oscillator circuit, "phase locking" it to the IF signal. Achieving the initial "phase lock" with this simple circuit would probably make tuning in a station rather finicky. Some additional circuitry to facilitate the initial lock would eliminate this problem, at the cost of some added complexity. This might be as simple as a tune/listen switch to change the bandwidth of the loop filter. A slight variation of this circuit would be to fix the frequency of the 455 kHz oscillator, and connect the varactor into the superheterodyne oscillator circuit. Each variation has pluses and minuses. Once one has gained some confidence with this beginner's IF Synchrodyne, a direct conversion variant, replacing the 6BE6 converter with a 6EJ7 RF amplifier as in the original "Electronics Australia" design, but retaining the PLL concept, could be tried with its added problems. One problem with this PLL direct conversion approach is that there isn't an obvious off the shelf phase detector transformer available. -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
#32
Posted to rec.audio.tubes
|
|||
|
|||
Turner's Synchrodyne (was: Some thoughts on PLLs, Synchrodynes)
On Jul 26, 10:37*am, John Byrns
snip Trying to explain they existed sort of drove the conversation in all manner of directions, like Patrick trying to 'coerce' me into building a 'Tucker'. ** I strongly recommended you try my cathode follower detector, and I don't coerce anyone into building anything. Of course when I do recommend something, usually it is a very simple easy thing which might occupy a wet sunday arvo which I forsee might be a pleasure besides being educational. I've never coerced anyone into using more than one tube at a time. A whole synchrodyne radio is several tubes at least, and probably 3 month's worth of sundays to begin to overcome numerous design problems. Its no use me even cordially inviting anyone to build anything because the mindset of everyone here is so stultifyingly moribund. Cold soldering irons are normal, and people dare to never walk outside their front gate to explore anything any more. They may even have existed in the tube days, in exactly the form Patrick is looking for, during the first run at AM stereo in the US in the late 1950s. ** Er, just what did you ever really think I was looking for? Your powers of imagination have run away like an over excited dog off the lead in a nice big parkland. There were half a dozen or so systems proposed by various companies, most of which used FM or PM to convey the stereo difference signal, but there may have been a QAM proposal among them that would have included the tube PLL circuit Patrick is looking for. ** So where are the full schematics with explanations and waveforms? ** Here in Oz, stereo AM just never ever took off. Stereo FM sure did, followed by surround sound. But with AM, nobody used much more than an SE output tube for 3 watts, one channel only, until the 1970's, when Japanese junk imports began to trickle and then flood into Oz after trade tarrifs were reduced because there were more australians willing to vote for cheaper prices than there were voters who'd vote to preserve their mate's jobs. I have descriptions of all the systems buried in my library, but due to extreme domestic entropy, which is this summers project to make a dent in, I don't have the data to hand to verify if one of the systems was a QAM system with a PLL that Patrick could use. ** It sounds like you must be about 94 John, and doing just about anything takes +20dB longer than it did when you were 30, and getting it right might take +20dB more capability. My mother of 94 was an excellent pianist and secretary at 26 before she ruined her development by bending to the urge towards motherhood, after which all her talents fell by the wayside in favour of becoming an excellent cook and house keeper. I now understand her incapacity, and her extreme limited ability to do anything much at all. I will be like her one day if I live that long. SO, I have to wonder where people are here, where they are at, and I smell no fumes from soldering iron just like I smell no wondrous aromas from my mother's kitchen when she used to cook. Since Patrick has been loath to rise to the bait I have dangled and give us an overview of what sort of Synchrodyne he has attempted to solder together, with unsatisfactory results, I decided to root through the archives to see what he said on the subject in the past. ** What bait had you ever dangled John? ** I simply stated I tried Tucker's idea but was stumped on the transformers for the balanced mixer with 4 diodes. The R&D would have taken too long, and nobody I contacted knew anything about anything much, including a bunch of local radio amateurs, most of whom are dead now. I found the following quotes in the rec.audio.tubes archives, except for the first which was part of a personal email I received from Patrick. I hope he will forgive my lapse of Netiquette in quoting it here in this public group. ** I doubt anyone would need to apologise for re-broadcasting my private emails. Wikileaks couldn't offend me. I have nothing evil to hide. From a private email: Date: Sat, 18 Aug 2001 11:35:27 +1000 "I guess true synchronous detection could be used for the 38 Khz signal, just like a Synchrodyne reciever. The one I have in mind was By D.G. Tucker from 1947, but it never caught on, as it really did not offer any better overall performance, and this method of direct detection did not catch on until chips came along. There was a Synchrodyne circuit for AM which came out in Electronics Australia in 1962, and it used a 6BE6 as the self oscillating mixer with the oscillations locked to the incoming carrier, and a simple R for a load, and a C to remove RF ripple." ** The 2 tube circuit was by Ian Pogson, published in Radio TV & Hobbies. It had a 6EH7 tuned RF amp and self oscilating 6BE6 into which was fed the tuned RF signal. AF came out of the anode circuit. So very very simple, except it wasn't as good a superhet. From the usenet group rec.audio.tubes: Date: Thu, 17 Jan 2002 15:32:00 +1100 "Less is more sometimes with radio,and two detectors in series seem unecessary." "I tried building a Synchrodyne reciever, with tubes, and it quite stumped me, for a variety of reasons, so I returned to the idea of a decent superhet, and have not looked back. D.G.Tucker brought out a tubed pure Synchrodyne in 1947, and I started with a similar idea to his, and found it difficult to stop it oscillating where I didn't want it too." **Indeed. Since that time, I've developed the CF detector in various forms to render further investigation of synchronous tube radios to be pointless. I can get excellent AF BW, low noise, excellent skirt selectivity, freedom from whistles or monkey chatter from a superhet. Perhaps someone at somewhere might prove they can build a better AM radio with tubes or chips but NOBODY has even come close to proving it with a practical example. I apologise in advance for my arrogance, but it must be remembered by all that I have been repairing or rebuilding several AM radios each year at least for the last 15 years and NOT ONE set in its original condition sounds as well as the radio in my kitchen which I designed & cobbled together in 1999. I'll eat that radio when someone fronts up with a better sounding radio. Date: Sat, 12 Jun 2004 01:02:30 +1000 "My paper files have around 20 different circuits for AM tuners including a fairly simple synchrodyne ( or direct conversion ) two tube sets which use a 6EJ7 for an RF amp, and followed by a 6BE6 synchronous detector." "I tried building one, but audio output was low, stability was difficult, and and a superhet proved far better." ** Indeed. Your point in quoting me????? "Then there were several if not many synchrodyne and some homodyne designs in Wireless World over the years, but these were all chip based, except the early D.G.Tucker circuit of 1947." Date: Wed, 23 Jun 2004 02:10:33 +1000 "I did play around with the Tucker tubed synchrodyne but gave up when I concluded that my superhet idea was easier to build." Date: Fri, 12 Nov 2010 02:11:07 -0800 (PST) "I guess a PLL oscillator which would "run on" during short time absences of a synchronized and limited carrier would be better. The last time I tried was 1999 when I spent a week trying t get a tubed "Synchrodyne" first designed by a Mr Tucker in 1947, and published by Wireless World that year. It never was a commercial success and could not compete with good superhet performence, at least for broadcast AM bands. Besides, stations today routinely use right up to 100% modulation." "Anyway, my synchrodyne was very hard to get working anywhere near as well as a good superhet even when I resorted to using a much simpller circuit with 6EJ7 input RF gain amp and a 6BE6 self oscillating AM detector; plus there's the howls and whistles when tuning and the the monkey chatter from other stations." ** yeah yeah, .... So??? Date: Mon, 11 Jul 2011 03:13:34 -0700 (PDT) Subject: VLF stability in Williamson-type amplifiers "I tried to build a version of the DG Tucker Synchrodyne circuit published in Wireless World in about 1947. Maybe its now online some place, and what prevented progess was the making of a suitable balanced synchronous demodulator. Probably the best way is to use a toroidal transformer with the core ? just right for range of RF frequencies. But I never had time to explore how to make such a tranny. The tucker circuit uses lots more tubes than a superhet and needs much more practical expertise to get running without loud whistles while tuning. It was a potentially excellent thing but it was never to become popular because the superhet was king. Then tried making a synchrodyne based on a self oscillating 6BE6, and there are some simple circuits of those around, just RF input tube and 6BE6 needed only. That sort of worked a bit but monkey chatter and whistles and controlling oscillations just right were easier said than done so I abandoned the idea and proceeded with a good superheat." That brings us to the current thread which Patrick started on 18 Jul 2011 under the subject "Phase Locked Loop with vacuum tubes." I later changed the subject to "Some thoughts on PLLs, Synchrodynes, and AM modulation depth", and with this post am changing again to " Turner's Synchrodyne" ** I think you have let intellectual licence run away with the imagination dog in the park. ** What exact arrangement of tubes and coils etc could ever be worthy to be named after me? The above quotes make it fairly clear that what Patrick built was a project described in a 1962 issue of "Electronics Australia". ** I tried to build both Tucker's Synchrodyne published in WW, and Pogson's Synchrodyne published in R,TV&H which became Electronics Australia. Patrick says this design used two tubes in a direct conversion circuit, with a 6EJ7 as an RF amplifier stage, and a 6BE6 as a combined oscillator and synchronous detector. Presumably the oscillator section of the 6BE6 was "injection locked". ** With both RF input and oscillator operating at the same F and with same L and C tank values, and applied into the same single 6BE6 frequency converter tube, the oscillations tended to lock easily while modulation % was low. I suspect that an easier to get working beginner's circuit, requiring no special parts, could be created by adding only one tube to the Electronics Australia project. First in keeping with my notion of crawling before walking, I would change the design from a direct conversion receiver to a superheterodyne receiver with a synchronous detector. I would replace the 6EJ7 RF stage with a second 6BE6 operating as a converter from the broadcast frequency to a 455 kHz IF. The synchronous detector and oscillator would remain essentially the same as in the "Electronics Australia" design using a 6BE6. ** You can sit back in your arm chair and type up your suspicions all day long John but it cuts no ice with me. I'll only start to be more receptive to recomendations AFTER YOU have done the donkey work of making up the chassis and building the coils and ensuring its stable and getting it all to work IN YOUR TIME and NOT MINE. ** Its a tough love approach John. But after the superhet began to become king by about 1933, it has managed to stay king for the last 78 years *if one thinks in terms of using tubes and IFTs* Many other ways to receive normal broadcast AM or short wave have been devised and I've not heard anything better than a good tubed superhet. Remarkable multiband sets were made with SS devices but all concentrated on low AF bandwidth, and none offered hi-fi sound. About the most elaborate tubed set I worked on was a 26 tube Racal radio which offered astounding communications abilities. But for good local BC band reception all the fancy techniques in a Racal or any other damned complex bit of gear are just not needed. ** Synchronous detection along with other techniques became possible and even popular when silicon chips allowed fast enough wave forms with predictable outcomes with perhaps hundreds of bjts within chips. some of the in-chip techniques could be repllcated with tubes, but you'd need box fulls of them, and it just isn't on, right, and nobody has ever bothered. Even to make a Gilbert Cell with tubes takes 4 x 6DJ8, and lord knows what sort of troubles you'd have to get it to works as well as a chip with Gilbert Cell within. ** As tubes faded away from being mainstream in the radio receiver building industry, nobody was able to make any further improvements which would be lauded by industry leaders who became totally devoted herd members in persuit of solid state solutions to every single problem preferably with ICs, because they were CHEAPER than using room fulls of girls all soldering bjts into PCBs. Now all circuits are never touched by interfering human hands and are designed by a computer, and wave soldered together. The AM broadcasts I now hear from the sets which now cost $00.20 cents each to include on a PCB are basically attrocious, except on my old tube set. ** I come along in about 2000, some 40 years after everyone in the radio industry now wonderes what the heck tubes are or were, and tell ppl that a double CF type of AM detector works better than all the other crap that was industry standard in 1960. I am then hounded down and shot at from all directions. ** Well, all the buckshot missed. The big change to the circuit would be the means of synchronizing the oscillator portion of the synchronous detector. Instead of injection locking the oscillator in the second 6BE6, a PLL would be used. A third tube, a 6AU6 would be added, with its grid feed from the same IF signal feeding the 6BE6 synchronous detector. A slightly modified 455 kHz discriminator transformer, an easily available part, is used along with two GE diodes as a phase detector with the 6AU6 feeding one input and the 455 kHz oscillator section of the second 6BE6 feeding the phase detector's second input. There is NO easily available 455kHz discriminator transformer. One must DESIGN and MAKE the unit. The output of the phase detector, through a loop filter, would then control a varactor diode connected in the 455 kHz oscillator circuit, "phase locking" it to the IF signal. ** Someone at Somewhere is very welcome to build such a circuit in THEIR TIME. Achieving the initial "phase lock" with this simple circuit would probably make tuning in a station rather finicky. Some additional circuitry to facilitate the initial lock would eliminate this problem, at the cost of some added complexity. This might be as simple as a tune/listen switch to change the bandwidth of the loop filter. ** Define "finicky". I suggest it is NOT an engineering term, but I might assume that when tuning, lock might occur when you were say 20kHz off the local station, then lock would "hang on" until you were 20kHz away, so whistles while tuning would not be heard. In simple sets made to recieve morse code without modulation, a BFO is used to tune 1 kHz away from the IF frequency, so you hear the morse. Similarly, the BFO is used to put a carrier into a SSB signal thus making SSB comprehensible. But in synchronous, no such beating of close RF signals is permissible, and oscillator needs to be locked to RF or IF signal F. ** But the details of how you get rid of "finicketty-ness" and make the circuit easily usable by an average dumb punter take some effort. A slight variation of this circuit would be to fix the frequency of the 455 kHz oscillator, and connect the varactor into the superheterodyne oscillator circuit. Each variation has pluses and minuses. ** One could easily build a crystal controlled 455kHz oscillator. Crystals are available. Once one has gained some confidence with this beginner's IF Synchrodyne, a direct conversion variant, replacing the 6BE6 converter with a 6EJ7 RF amplifier as in the original "Electronics Australia" design, but retaining the PLL concept, could be tried with its added problems. One problem with this PLL direct conversion approach is that there isn't an obvious off the shelf phase detector transformer available. ** I await you detailed schematic telling us all how-to-do-it but no doubt the date for the release of such schematic info from the John Byrns Establishment is years away, pending the overcoming of a multitude of problems not mentioned right now, but which certainly exist, or will soon, during the process. Patrick Turner. -- Regards, John Byrns |
#33
Posted to rec.audio.tubes
|
|||
|
|||
Turner's Synchrodyne (was: Some thoughts on PLLs, Synchrodynes)
In article ,
Patrick Turner wrote: snip Hi Patrick, I was having trouble understanding why you started this thread, originally titled "Phase Locked Loop with vacuum tubes", especially considering that you have complained over the years about your inability to build a properly working "Tucker Sybchrodyne", and your eventual discovery that a superheterodyne using your "C+D" detector circuit provided unexcelled performance, especially considering its high bean counter appeal. After reviewing this thread, and some of the related currently running AM detector threads, I now see that you seem to want to build a PLL type AM detector for use in a superheterodyne receiver. Why you want to do this is not at all clear since you already have the "C+D" detector which you say is the ultimate. Flipper has already a link to one schematic for a tube based PLL operating at 3.58 MHz. In one of the AM detector threads I pointed out two or three other tube PLL applications for you, although unlike Flipper I didn't provide any links. The schematic for one of the tube based PLLs I mentioned, operating at about 5 MHz, is also available on the web, complete with explanations and waveforms. Either of these PLLs could be adapted for operation at 455 kHz by changing the frequency determining components. Unfortunately you seem to be a lazy guy, unwilling to sharpen your pencil and get out your slide rule, to modify one of the available tube PLL circuits for use as part of an AM detector. Now you are demanding that we do your work for you and provide ³full schematics with explanations and waveforms² that you can use to build your tube PLL AM detector. I suggest that you take your own advice and continue using your ³C+D² detector, at least until next year when you retire on your old age pension. That will eliminate your need to spend so much time repairing old radio and Hi-Fi equipment to earn money, with your time thus freed up you will have time to further develop the old tube PLL circuits into a form suitable for use as an AM detector. -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
#34
Posted to rec.audio.tubes
|
|||
|
|||
Turner's Synchrodyne (was: Some thoughts on PLLs, Synchrodynes)
On Jul 27, 11:02*am, John Byrns wrote:
In article , *Patrick Turner wrote: snip Hi Patrick, I was having trouble understanding why you started this thread, originally titled "Phase Locked Loop with vacuum tubes", especially considering that you have complained over the years about your inability to build a properly working "Tucker Sybchrodyne", and your eventual discovery that a superheterodyne using your "C+D" detector circuit provided unexcelled performance, especially considering its high bean counter appeal. I cannot see what stopped you being able to understand the FACT that virtually no PLL circuits using tubes have ever been built, as equivalents of what may be done using chips, and for application in AM radios. Sure I gave up trying to build a synchrodyne, but it doesn't mean I was unable to, had I been more persistant. There was limited time then, as there is now for the R&D needed to make a decent balanced AM de-modulator one way or another, using tubes. After reviewing this thread, and some of the related currently running AM detector threads, I now see that you seem to want to build a PLL type AM detector for use in a superheterodyne receiver. *Why you want to do this is not at all clear since you already have the "C+D" detector which you say is the ultimate. Tucker's original circuit appeals because of the reputed benefits of direct conversion receivers. Thousands of radio hams with direct conversion receivers can't be all wrong, although I expect hi-fi sound, when they don't. My double CF detector does work better than most if not all other ways to extract AF from AM using just 2 triodes and no more other parts than used in conventional AM detectors. The bean counters of 1950 could never have approved the use of my two added triodes ( 1 x 12AU7 tube) so don't talk to me about my ideas having bean counter appeal. About none of my ideas suit bean counters, ie, anyone seeking to remove function quality to favour shareholder returns, or to make out they are a hero of the revolution.. Flipper has already a link to one schematic for a tube based PLL operating at 3.58 MHz. Sure some examples exist but nothing devoted to the BC band. *In one of the AM detector threads I pointed out two or three other tube PLL applications for you, although unlike Flipper I didn't provide any links. *The schematic for one of the tube based PLLs I mentioned, operating at about 5 MHz, is also available on the web, complete with explanations and waveforms. *Either of these PLLs could be adapted for operation at 455 kHz by changing the frequency determining components. I'll leave you to provide the info, and produce the addapted schematic and get it running as a prototype and post your results. Meanwhile what I read here is just old men's hot air about what might be, might not be, and I can only really relate to REALITY. Unfortunately you seem to be a lazy guy, unwilling to sharpen your pencil and get out your slide rule, to modify one of the available tube PLL circuits for use as part of an AM detector. Call me lazy if you like, but where are your efforts on the subject? So before anyone calls me lazy, they have to show thay are doing more than I am to produce working gear. I'm employed on client projects and although I think always about better ways of doing things I don't have unlimited time to implement them. Even if I did manage to get a synchronous AM radio or homodyne radio to perform better than what I know now, there would be virtually no justification whatsoever to incorporate into anyone's existing radio set, as I would be having to build so much new stuff and remove stuff that is more easily improved. Now you are demanding that we do your work for you and provide full schematics with explanations and waveforms that you can use to build your tube PLL AM detector. * I'll always demand of other what I demand of myself. If people here expect intellectual respect, or respect of their constructional cabilities, then they need to get off their arses to their work sheds and get busy. We could all sit around ALL DAY but I can't afford it. Too many people want me to fix their junk, and I need de munny. I suggest that you take your own advice and continue using your C+D detector, at least until next year when you retire on your old age pension. * That will eliminate your need to spend so much time repairing old radio and Hi-Fi equipment to earn money, with your time thus freed up you will have time to further develop the old tube PLL circuits into a form suitable for use as an AM detector. If I had to try to build a direct conversion again, i'd probably use a 3 gang tuning cap. Two would be for a an RF input amp to allow use of a ferrite rod antenna, and a following tank after a pentode gain tube. Tucker's set dosn't have this feature, but has a two pentode cascaded untuned amp with NFB to raise RF input levels at the output of what is a low impedance output amp. I'd prefer a tunable RF front end and a simble CF at its OP will reduce the Zout which is best to drive a balanced modulator circuit. That could be similar to Tuckers and possibly give real hi-fi, but the reality can only be found by doing, not spending hours bull****ting here about it. I already have a working FM generator which works to produce 455kHz AM with +/- 40kHz each side for examining IFT selectivity curve shapes. Its F is controlled by a tuning gang plus use of about 11 paralleled 68V x 5W zener diodes which won't conduct like ordinary veractor diodes when used in tube circuits which produce quite high oscillator output voltages in tank circuits where one wants to use such purpose built varicap diodes. I did try, but ordinary varicaps were useless. Anyway, zeners make fine varicaps. Now the synchrodyne set needs a VFO so the third of 3 tuning gangs is used to tune it. But its actual C on the LC tank might be controlled by DC working on zeners to change C. But I have to work out details of how to apply the Vdc to cause locking, so a limiter is needed, so a 6AU6 powered off the recovered and amplified RF is needed. Tube count is mounting up. Anyway, I have ideas about how to do a synchrodyne and I have all the hardware, but I shall never say it will work better than what I can now already do with a suprerhet. Sure I could probably just build a synchrodyne module which worked off an existing 455khz IF strip - maybe easier because there is far less F variation to have to worry about and no worry about tracking by RF input stages and VFO tunable oscillators. Its obvious to me neither you or flipper can ever be relied upon to provide the world with alternative working circuits for DIYers which work better than existing. Stay seated and never raise a sweat if you must, but while nothing seems to happen at your place, I have to do things, and because I like doing them. Someon'e's knocking at my door again. Patrick Turner. |
#35
Posted to rec.audio.tubes
|
|||
|
|||
Turner's Synchrodyne (was: Some thoughts on PLLs, Synchrodynes)
In article ,
Patrick Turner wrote: On Jul 27, 11:02*am, John Byrns wrote: In article , *Patrick Turner wrote: snip Hi Patrick, I was having trouble understanding why you started this thread, originally titled "Phase Locked Loop with vacuum tubes", especially considering that you have complained over the years about your inability to build a properly working "Tucker Sybchrodyne", and your eventual discovery that a superheterodyne using your "C+D" detector circuit provided unexcelled performance, especially considering its high bean counter appeal. I cannot see what stopped you being able to understand the FACT that virtually no PLL circuits using tubes have ever been built, as equivalents of what may be done using chips, and for application in AM radios. Because chips came after tube PLLs, why would anyone want to duplicate chip circuitry with tubes, they are different technologies and the circuits that suit them best differ. I am not sure why you obsess over building Gilbert cells with tubes? Relative to schematics, I forgot to mention that quite a few late 1930s high end AM radios with motorized tuning included an AFC system similar to what was used in tube FM tuners. These circuits included a 455 kHz discriminator and reactance tube circuit to control the local oscillator. With a slight change to the wiring of the discriminator transformer, to feed the centertap of the secondary from the local oscillator rather than from the transformer primary, presto you have the schematic of a complete PLL circuit that will lock the oscillator to the IF frequency. Of course the oscillator must also be modified to operate at 455 kHz. -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
#36
Posted to rec.audio.tubes
|
|||
|
|||
Turner's Synchrodyne (was: Some thoughts on PLLs, Synchrodynes)
In article ,
flipper wrote: On Tue, 26 Jul 2011 20:02:45 -0500, John Byrns wrote: In article , Patrick Turner wrote: snip Hi Patrick, I was having trouble understanding why you started this thread, originally titled "Phase Locked Loop with vacuum tubes", especially considering that you have complained over the years about your inability to build a properly working "Tucker Sybchrodyne", and your eventual discovery that a superheterodyne using your "C+D" detector circuit provided unexcelled performance, especially considering its high bean counter appeal. After reviewing this thread, and some of the related currently running AM detector threads, I now see that you seem to want to build a PLL type AM detector for use in a superheterodyne receiver. Why you want to do this is not at all clear since you already have the "C+D" detector which you say is the ultimate. I think you misinterpreted the 'purpose' of that post. Near as I can tell the genesis of this jumble began way back in the "VLF stability in Williamson-type amplifiers" thread because Pogosso mentioned using his shelf in AM radios. That diverged into various sub threads among which detectors came up. Yes, I think the first follow on thread was Pogosso's "AM detector, part 1". Somewhere in the jumble Patrick apparently decided his CF detector was being criticized and went into Last Starfighter Death Blossom mode shooting anything that even smacked of 'AM detector'. I must plead guilty that I have criticized Patrick's "C+D" detector to the extent that I question the need for the first cathode follower. I don't know who first mentioned 'PLL' but his post as a " It was Pogosso in his "AM detector, part 1" thread that first mentioned "PLL" in connection with a solid state PLL detector module he fits to old radios. Patrick was right on it the next day with a request for schematics of vacuum tube phase locked loops. I still don't understand Patrick's interest in vacuum tube PLL circuits given that he already has his "C+D" detector, and PLL detector circuits up the tube count considerably. -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
#37
Posted to rec.audio.tubes
|
|||
|
|||
Turner's Synchrodyne (was: Some thoughts on PLLs, Synchrodynes)
In article ,
flipper wrote: On Tue, 26 Jul 2011 23:18:49 -0500, John Byrns wrote: In article , flipper wrote: On Tue, 26 Jul 2011 20:02:45 -0500, John Byrns wrote: In article , Patrick Turner wrote: snip Hi Patrick, I was having trouble understanding why you started this thread, originally titled "Phase Locked Loop with vacuum tubes", especially considering that you have complained over the years about your inability to build a properly working "Tucker Sybchrodyne", and your eventual discovery that a superheterodyne using your "C+D" detector circuit provided unexcelled performance, especially considering its high bean counter appeal. After reviewing this thread, and some of the related currently running AM detector threads, I now see that you seem to want to build a PLL type AM detector for use in a superheterodyne receiver. Why you want to do this is not at all clear since you already have the "C+D" detector which you say is the ultimate. I think you misinterpreted the 'purpose' of that post. Near as I can tell the genesis of this jumble began way back in the "VLF stability in Williamson-type amplifiers" thread because Pogosso mentioned using his shelf in AM radios. That diverged into various sub threads among which detectors came up. Yes, I think the first follow on thread was Pogosso's "AM detector, part 1". Somewhere in the jumble Patrick apparently decided his CF detector was being criticized and went into Last Starfighter Death Blossom mode shooting anything that even smacked of 'AM detector'. I must plead guilty that I have criticized Patrick's "C+D" detector to the extent that I question the need for the first cathode follower. I don't know who first mentioned 'PLL' but his post as a " It was Pogosso in his "AM detector, part 1" thread that first mentioned "PLL" in connection with a solid state PLL detector module he fits to old radios. Patrick was right on it the next day with a request for schematics of vacuum tube phase locked loops. Oh. Well, I knew there was something somewhere but I couldn't recall. Now that you mention it I went back and it appears the beginning was Pearce mentioning a synchronous detector. Then, in a follow up, you suggested a PLL for over 100% modulation and Alex agreed. Alex also mentioned his SS detector module. I see it now. You mentioned PLLs in TV sets, Patrick said he didn't work on TV sets and hadn't ever seen any tube PLLs but was going to do a google. The post in question was the result. Sounds about right. That is where I came in because I had seen your post, his, and the 'google result' so I was simply trying to explain they wouldn't be called "phase locked loops" because that term didn't not come into being until later. So when did the term or TLA "PLL" come into use. I'm trying to remember what the PLL in old Television sets was called, I'm thinking it might have been called the Horizontal AFC circuit. I still don't understand Patrick's interest in vacuum tube PLL circuits given that he already has his "C+D" detector, and PLL detector circuits up the tube count considerably. Sorry but I'm sure you noticed by now I accidentally hit some 'magic' key that sent the message before it was done. What I had meant to finally explain is his post was a he 'looked', there ain't none, and he was going back to something useful instead of chasing fairy circuits. Well he sort of changed the goal posts at some point, first he asked for vacuum tube PLL circuits that could be used for better AM reception, later he changed it to demand schematics of complete PLL based AM detectors, no engineering required. -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
#38
Posted to rec.audio.tubes
|
|||
|
|||
Turner's Synchrodyne (was: Some thoughts on PLLs, Synchrodynes)
JB said.....
Hi Patrick, I was having trouble understanding why you started this thread, originally titled "Phase Locked Loop with vacuum tubes", especially considering that you have complained over the years about your inability to build a properly working "Tucker Sybchrodyne", and your eventual discovery that a superheterodyne using your "C+D" detector circuit provided unexcelled performance, especially considering its high bean counter appeal. After reviewing this thread, and some of the related currently running AM detector threads, I now see that you seem to want to build a PLL type AM detector for use in a superheterodyne receiver. *Why you want to do this is not at all clear since you already have the "C+D" detector which you say is the ultimate. Flipper then says..... I think you misinterpreted the 'purpose' of that post. Near as I can tell the genesis of this jumble began way back in the "VLF stability in Williamson-type amplifiers" thread because Pogosso mentioned using his shelf in AM radios. That diverged into various sub threads among which detectors came up. You are approximately correct. Pogooooso who invented the word booooring ro describe old radio AM detectors did indeed suggest I use an extra C in a FB loop to supposedly counter problems which I mentioned were basically non existant in anything I built, and most of everything else. Somewhere in the jumble Patrick apparently decided his CF detector was being criticized and went into Last Starfighter Death Blossom mode shooting anything that even smacked of 'AM detector'. JB didn't see the need for the extra use of CF triodes in a an AM detector. I tried to describe the benefits, suggesting he try my idea before condemning it, but of course that was never to happen when there's an egotistical opportunity to keep dismissing a good idea, rather than to move even 1 centimetre towards a soldering iron and CRO. I don't know who first mentioned 'PLL' but his post as a " snip- Hide quoted text - - Show quoted text - And you can see what a mess Goggle does with NG delivery to me. - no quoted text able to be seen. Anyway, I tried to look at the schematic of the PLL mentioned on page 461 within the .pdf for the 37 tube colour TV at the site at http://antiqueradio.org/sitemap.htm But the pdf was largely un-readable, having been scanned by an incompetent. And so I found far too much time would be needed to de- cypher the schematic to see just what flipper was talking about. I ain't no fukkin expert on tubed color TV sets from 1950s, and I'll stay that way with very scant information, and posted so you can't read what you're lookin at. I guess flipper or JB will NEVER provide a copied version of the tubed PLL schematic with full ammended details to allow it to be used in a synchrodyne radio. Its WAAAAAAAAAAAAAAAAY beyond their willingness to give, while they remain extremely willing to shoot everyone else down who challenges them in in any way. Since posting to JB in my last post I had a good look again at my 455kHz FM generator. It uses 5 x 68V x 5W zener diodes in parallel as varicap diodes to alter C across the LC tank in a cathode follower type of triode oscillator using 1/2 a 12AU7. While in FM mode, the RF gene does not use a tuning gang, although some fine tuning of the centre F is possible with a pot which slightly alters Vdc on ZDs. The 5 ZDs have grounded anodes, and the 5 cathodes connect via 2N6 to the grid end of the oscillator coil. This "top" end of the coil is cap coupled to triode grid, while triode cathode is taken to a tap about 25% up from earthy end of coil which is at 0V. The Vo output from the coil goes via 3k9 to the grid of CF buffer stage, so that anything connected to VFO output can't change the tuned F. The voltage range applied to the ZD cathodes may be varied by a pot from +24Vdc to + 40Vdc. The applied Vac to "wobbulate" or FM modulate the 455kHz is about within this range from a saw tooth generator I have built in, but any old AC signal will do to produce FM to trace the selectivity curves for a 455kHz radio IF strip. A switch allows a change of ZDs to just 1, not 5, thus allowing a 930kHz signal to be FM'd, and applied and detected by the AM set. Now a phase detector **could be** be fairly easily be built using a suitable 1:1 RF transformer with secondary having a CT, and two diodes, plus some R&C bits, all very much like an FM radio ratio detector which is arranged to produce a Vdc output which is applicable to a reactance triode, 1/2 12AT7, which then controls the FM set RF oscillator F at over 100MHz. The VDC generated by the phase detector goes +ve for where VFO F goes higher than carrier, -ve if VFO F goes lower, and we might *HOPE* that the Vdc change is enough to lock a VFO and that the time constant is long enough with Vdc LPF network that VFO will run on without significant F change if the carrier signal drops to nothing during 100% modulation. The only reason I see for use of PLL techniques for a synchrodyne are to get over the problem of intermittent locking of a locked oscillator even with a limiter stage present. But Tucker's radio has its VFO tunable for all of the BC band. This requires a large change in C between say 400pF and 30pF. Let us assume we might find that a limiter stage using a pentode would help to give a better signal to use to lock the VFO, and if you have a tuned VFO with a tuning gang and you have a parallel tuning diode, then getting the C change needed by a Vdc applied might be difficult to get working right, so perhaps all the tuning might be done using a pot and varied DC and nothing but veractors. I cannot list all forseable bothers trying to do all this. But if it can be done for where VFO must make F between 530kHz and 1,710kHz, then its better to do it than retain the superhet mixer stage, which we might want to chuck out BECAUSE we wish to use a synchrodyne instead, and gain selectivity offered by the audio LPF rather than by the IFT selectivity. So if people want to know how I might get better hi-fi from a tubed synchrodyne set which I would want to work better than Tucker's 1947 job, then they might see agreement about using a following summary of stages :- TRF input stage, ferrite rod antenna input coil, AVC applied to grid of V1 pentode. One tuning gang, 30pF-400pF. RFT at anode, secondary tuned with second tuning gang, 30pF-400pF. V2 CF buffer stage. Q of LC should not be too high to caused side band cutting and RF BW should be at least 8kHz at the low end of band. V2 buffer drives balanced modulator using Tucker's circuit with transformers using details to be worked out. V2 buffer also drives V3 pentode limiter to make near constant amplitude output signal. V4 CF buffer used after V3 to drive phase detector to produce Vdc to change veractor C of oscillator. LPF for dc path keeps Vdc constant for limiter signal drop outs when AM % goes to 100%. V5 triode CF oscillator with tunable tank using the 3rd available tuning gang. V6 triode oscillator output buffer used to drive VFO input to balanced demodulator. Veractor used in parallel with tuning gang. Application of Vdc from phase detector to ensure VFO stays locked to limiter carrier F. Tuning of oscillator should track RF stage tuning The AF from balanced de-mod stage buffered by V7, and active second order RC + V8 CF LPF filter used before V9, V10 audio output stages. It obviously might be easier to use a PLL and oscillator which worked at 455kHz only, to avoid having to tune the oscillator over a wide range. Then the RF input stage might be simplified. There are probably 101 other ways which might be utilised, but the synchrodyne will probably end up more complex than the standard 2 tube superhet AM tuner used in most AM radios for strong local stations. I do not know if the synchrodyne balanced demodulator can produce a Vdc which may be used for VFO F control in addition to producing audio output. But one thing is for sure, to make a synchrodyne with tubes to meet modern use expectations is a real big ask. Patrick Turner. |
#39
Posted to rec.audio.tubes
|
|||
|
|||
Turner's Synchrodyne (was: Some thoughts on PLLs, Synchrodynes)
In article ,
Patrick Turner wrote: Anyway, I tried to look at the schematic of the PLL mentioned on page 461 within the .pdf for the 37 tube colour TV at the site at http://antiqueradio.org/sitemap.htm But the pdf was largely un-readable, having been scanned by an incompetent. And so I found far too much time would be needed to de- cypher the schematic to see just what flipper was talking about. I ain't no fukkin expert on tubed color TV sets from 1950s, and I'll stay that way with very scant information, and posted so you can't read what you're lookin at. I guess flipper or JB will NEVER provide a copied version of the tubed PLL schematic with full ammended details to allow it to be used in a synchrodyne radio. Its WAAAAAAAAAAAAAAAAY beyond their willingness to give, while they remain extremely willing to shoot everyone else down who challenges them in in any way. How am I trying to shoot your PLL efforts down? I'm trying to encourage you to experiment with a PLL based radio, and your latest post gives me hope that perhaps you are giving the matter some serious consideration. I am trying to give you a leg up by filling in a few details that you seem to be unaware of, you are free to use, or not use, this information in any way you are comfortable with. [Snip] Now a phase detector **could be** be fairly easily be built using a suitable 1:1 RF transformer with secondary having a CT, and two diodes, plus some R&C bits, all very much like an FM radio ratio detector which is arranged to produce a Vdc output which is applicable to a reactance triode, 1/2 12AT7, which then controls the FM set RF oscillator F at over 100MHz. The VDC generated by the phase detector goes +ve for where VFO F goes higher than carrier, -ve if VFO F goes lower, and we might *HOPE* that the Vdc change is enough to lock a VFO and that the time constant is long enough with Vdc LPF network that VFO will run on without significant F change if the carrier signal drops to nothing during 100% modulation. The ³LPF² is called the ³Loop Filter², the design of this filter is something you should read up on. I'm not a ³PLL² expert but I think you want a double time constant filter here, this typically involves two resistors and two capacitors. This consists of a short time constant filter to remove the carrier frequency components and other grunge from the output of the phase detector. This is followed by a long time constant which serves as a modified integrator to allow the VCO to track through the periods of carrier lock. The long time constant part of the filter is modified so that while it integrates the phase error signal over the long term, it allows some higher frequency components to pass through to provide faster locking and better loop stability, remember this is a negative feedback loop so there may be stability issues. Hopefully there is a fully fledged PLL expert reading this who can better explain this. The only reason I see for use of PLL techniques for a synchrodyne are to get over the problem of intermittent locking of a locked oscillator even with a limiter stage present. That sounds about right, we need a second order loop to ride us through the periods of carrier loss, or even effective phase reversal as may happen with heavy modulation during selective fading. But Tucker's radio has its VFO tunable for all of the BC band. This requires a large change in C between say 400pF and 30pF. Let us assume we might find that a limiter stage using a pentode would help to give a better signal to use to lock the VFO, and if you have a tuned VFO with a tuning gang and you have a parallel tuning diode, then getting the C change needed by a Vdc applied might be difficult to get working right, so perhaps all the tuning might be done using a pot and varied DC and nothing but veractors. Yes, there are certainly problems here, it's not immediately obvious how to connect a varactor to the oscillator section of a standard multi gang tuning condenser while providing both the needed capacity change at the low end of the band, and also not unduly increasing the minimum capacity at the high end of the band, that is required to provide the 9:1 capacity change needed to tune the entire MW band. This 9:1 impedance change across the band will also affect the gain of the PLL circuit, influencing its performance including lock range and stability. It's not clear to me how much substituting varactors for a mechanical tuning gang simplifies the problem, it presumably would if we were willing to use opamps and piecewise linear summing circuits in the DC circuits controlling the varactors. Maybe there is a simple solution that I am just not seeing. In any event there is something very satisfying about a mechanical tuning gang in a tube radio. I cannot list all forseable bothers trying to do all this. Yes, while the direct conversion, or ³Zero-IF², approach has a strongly seductive allure, its realization does raise a multitude of bothers. I'm not sure how common direct conversion radios are outside of the ³Ham² radio community, does anyone know? I suspect that most of the so called ³Zero-IF² receivers are actually superhetrodynes, where what at one time would have been double conversion superheterodynes have had their second IFs replaced with a ³Zero-IF² and detector. All these bothers are why I like my crawl before walking superhetrodyne with PLL synchronous detector approach, at least for a first project. But if it can be done for where VFO must make F between 530kHz and 1,710kHz, then its better to do it than retain the superhet mixer stage, which we might want to chuck out BECAUSE we wish to use a synchrodyne instead, and gain selectivity offered by the audio LPF rather than by the IFT selectivity. Yes, the direct conversion approach is very seductive, no question. With the superheterodyne approach, if you make the IF very wide, perhaps using a 2 MHz IF, then you can control the audio bandwidth with the audio LPF just as with the direct conversion approach. So if people want to know how I might get better hi-fi from a tubed synchrodyne set which I would want to work better than Tucker's 1947 job, then they might see agreement about using a following summary of stages :- TRF input stage, ferrite rod antenna input coil, AVC applied to grid of V1 pentode. One tuning gang, 30pF-400pF. RFT at anode, secondary tuned with second tuning gang, 30pF-400pF. V2 CF buffer stage. Q of LC should not be too high to caused side band cutting and RF BW should be at least 8kHz at the low end of band. V2 buffer drives balanced modulator using Tucker's circuit with transformers using details to be worked out. V2 buffer also drives V3 pentode limiter to make near constant amplitude output signal. V4 CF buffer used after V3 to drive phase detector to produce Vdc to change veractor C of oscillator. LPF for dc path keeps Vdc constant for limiter signal drop outs when AM % goes to 100%. V5 triode CF oscillator with tunable tank using the 3rd available tuning gang. V6 triode oscillator output buffer used to drive VFO input to balanced demodulator. Veractor used in parallel with tuning gang. Application of Vdc from phase detector to ensure VFO stays locked to limiter carrier F. Tuning of oscillator should track RF stage tuning The AF from balanced de-mod stage buffered by V7, and active second order RC + V8 CF LPF filter used before V9, V10 audio output stages. It obviously might be easier to use a PLL and oscillator which worked at 455kHz only, to avoid having to tune the oscillator over a wide range. Then the RF input stage might be simplified. There are probably 101 other ways which might be utilised, but the synchrodyne will probably end up more complex than the standard 2 tube superhet AM tuner used in most AM radios for strong local stations. That's a very good start! Using a two tube superhetrodyne as a base line is bean counter thinking. We want the best possible High Fidelity reception under as many conditions as possible, this is only possible using a true synchronous detector, pretenders need not apply I do not know if the synchrodyne balanced demodulator can produce a Vdc which may be used for VFO F control in addition to producing audio output. No, you need separate detectors for the ³PLL² and to recover the audio. The two detectors must be provided with carrier signals from the oscillator 90 degrees out of phase with each other. Maintaining the required 90 degree phase shift over a 3:1 frequency range is another big potential gotcha for a tube based direct conversion receiver. In a solid state design this problem would be simply dealt with by operating the oscillator at 4 or 8 times the received frequency, and using digital dividers to generate the 90 degree phasing between the two local oscillator signals. A large number of tubes would probably be required to do it this way in a tube based circuit. You may also want to add a lock detector circuit, and an envelope detector to your list above. The lock detector could switch the parameters of the ³Loop Filter² when out of lock to facilitate locking, and also switch to an envelope detector during tuning. But one thing is for sure, to make a synchrodyne with tubes to meet modern use expectations is a real big ask. Yes, it's a very big ask! -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
#40
|
|||
|
|||
Did anyone mention IBOC in North America? That will pretty well ZAP any improved AM detector & many ordinary ones as well. Listen in for IBUZ!
Cheers, John |
Reply |
|
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
FS: Electron Tubes/ Vacuum Tubes/ radiotron/ RCA, GE, Tung-Sol , Vintage | Vacuum Tubes | |||
box of vacuum tubes for sale... | Vacuum Tubes | |||
F.S. : RCA 6L6/5881 vacuum tubes | Marketplace | |||
Vacuum tubes in vacuum | Vacuum Tubes |