Home |
Search |
Today's Posts |
#1
|
|||
|
|||
AM Tuner -- Build/Find a high performance modern tube AM Tuner ?
John Byrns wrote: In article , Fred Nachbaur wrote: Patrick Turner wrote: John Byrns wrote: In article , "John Walton" wrote: If you don't go this route, the most difficult item to find for a super-het will be the ganged 365uF (they called 'em 365mmF) capacitor and the IF transformers. Doesn't a TRF receiver also require a "ganged 365uF capacitor"? For TRF, 4 single tuned circuits are required, so that means two double gang tuning caps, and then you have to get them all to track properly. The Q of each tuning circuit dosn't want to be too high, or else the final BW will be so narrow that so will the AF BW. The bigger problem is that the Q changes across the band, tending to be higher at the top end because of the higher reactance (and therefore higher ratio of reactance to resistance). So there'a compromise between good sound but lousy selectivity at the low end, and good selectivity but muddy sound at the high end. The AM broadcast band is bad that way, since it's about a 3:1 frequency range. When you speak of the "low end", and the "high end", are you talking about wavelength, or frequency? Actually you want the Q to vary across the band, you want the Q to be proportional to frequency, so that you can maintain a constant bandwidth as you tune from low frequencies to high frequencies. This effect can be approximated by making as many as possible of the lossy elements in the tuned circuits, series elements. This means the tuned circuit should be driven from a pentode with a high output resistance, and care should be taken in the design and layout of the amplifiers to make the grid conductance as low as possible. series resistors can then be added to the tuned circuits to set the desired circuit Q. Double tuned circuits, using two tuning condenser gangs for each stage, can also be used to improve the filter shape factor and achieve steeper skirts without excessive narrowing of the nose response. Complex coupling impedances can also be used to maintain the desired coupling as you tune across the band, and also to provide increased coupling at the low frequency end of the band to help offset the effects of the remaining parallel losses in the circuit. These techniques can greatly reduce the compromises necessary in TRF AM broadcast receivers. The famous Western Electric No. 10A Radio Receiver is an example of this type of design. J.W Miller also offers a series of radios and components based on this design technique, and the J.W. Miller design was used in Altec's early AM tuners. They invented the superhet to overcome the TRF troubles. Indeed. Only one frequency to worry about, as far as the selectivity vs. bandwidth situation is concerned. Also makes it practical to do stagger-tuning to give a more square-shouldered bandpass characteristic. Why even bother to stagger tune? Double tuned IF transformers, and even higher order filters, like the fourth order filters used in some of the H.H. Scott AM tuners, can accomplish much the same thing by judicious choice of Qs and coupling coefficients. Regards, John Byrns Interesting. I have two tuned LC input circuits in the front end of my superhet reciever. they are loosely coupled to the antenna. I found that the BW was smallest at the lowest RF and there was significant side band cutting and would have been reduction of audio bandwidth, had I not used two LC coils, tuned slightly apart, to give a wide enough BW. It is necessary to have at least 60 dB attenuation when 40 kHz off the wanted F, and 80 to 100 dB is better, and a couple of tuned circuit don't give enough attenuation away from the wanted F, even if the LC circuit has a high Q. If a single 455 kHz IF LC circuit has a BW of 20 kHz, then the Q is only about 22.7. When 4 such circuits are arrayed, the resultant Q is a lot more, and BW is a lot less. Such LC circuits have only say 15 dB attenuation at say 40 kHz off the wanted F, so you need a few LC circuits to get stations 40 kHz away from the one you want to be inaudible when there is silence in the programme. We have a sports station of 300 watts at 1008 kHz, and there is an entertainment station of 10,000 watts KHz less than a Km away, at 1,053 kHz, and I know a radio is OK if I can tune to 1017, and not hear the 1053 in the bacground of 1008 which will be noisy, and distorted. Patrick Turner. Surf my web pages at, http://users.rcn.com/jbyrns/ |
#2
|
|||
|
|||
"Patrick Turner" wrote in message ... Yves Monmagnon wrote: I also tried an all tube direct conversion synchrodyne, but that was a flop for many reasons, and I see why it never became popular. Hi Pat ! Could you tell a bit more about that ? Yves. Crikey, it'd take a bloody book. The first real proper synchrodyne afaik was D.G.Tucker's circuit, in 1947, with an RF amp using two pentodes with NFB and inptut tuning circuit, then a locked oscillator, third pentode, then a balanced detector using a toroidal transformer and 4 diodes. I never got the detector coils to work, and the RF amp teneded to be be very unstable, and the locked oscillator tended to be modulated by the synchronising injection from the input amp, and all in all, it seemed like a real PITA. Then when I eventually got a sychrodyne to work with a self oscillating 6BE6, the output was tiny, it was unstable, and the supposed selectivity due to AF filtering still left plenty of weird tones and crap in the AF output. Commercially viable synchronous detection had to wait until the days of chips and PLL, and then it works OK. And tuning a synchro means lots of whistles and howls until the set locks on a station. Was just to satisfy my curiosity since have never seen such designs with tubes, done ! I expected the synchrodyne was in place of the IF, that means a superhet with synchrodyne detector. Have done that (with chps, not tubes !) in a VHF traffic receiver to detect SSB, AM and FM. In FM, audio in the PLL error voltage. We have chance that regulation authority left room for good (for me) radios in FM band along all those commercials playing heavy compressed beats ! so more tubes are needed for a complex muting circuit. In this country, there were never any built commercially, afaik. Ppl stuck to plain old 5 valve superhets. Then eventually, after an absurdly long wait due to government bungles and vested interests in radio, we got FM in 1972, and the tube era was over then. The RDH4 has some FM circuits, which very few ppl could understand, let alone build. And very few know how multplex decoders work. As an Ham, I explain multiplex to myself as a supressed carrier dual side band 38khz, with a 19Khz pilot to rebuid the carrier. Then there was tubed TV, and that was really baffling. But many radio amateurs at least made their own receivers, and there were kits you could buy, but none of the parts used in 1960 are still available except as pulls from other gear. There are mountains of old but not very collectable valve radios about. something made about 1960 will have fairly miniturised parts, and have at least a two gang tuning cap. At the end of the day, what use is an AM radio? All the commercial stations here are talkback amounst blue collar uneducated folks, so its the blind leading the blind, or it pop music, usually from a small selection of records. So basically, commercial radio is complete crap. But we at least have the print handicapped radio station, where they read the papers each day. Then we have the ABC, the govt rum station, and the afternoon talkback is amoungst those with degrees fro a uni, and its actually interesting to listen in. The music shows on the ONE station and public interest interviews demand a decent AM reciever, so I built one from scratch, because the usual 5 valve superhet has so many inherent problems of small AF bandwidth, and high thd/imd. These problems can easily be addressed, and AM can nearly be as good as FM to listen to. The transmitted signal is usually subject to a filter at 9 kHz, and since OZ stations are all 9 kHz apart, with DX, one gets 9 kHz whistles, with chatter each side due to the beating of the station you are on with sidebands of a nearby station. A 9 kHz notch filter only removes the 9 kHz whistle, and not the other whistly chatter. To avoid this interference AF bandwidth has to be limited to about 4 kHz, and this is the usual AF BW of most tube radios. Many only have 2 kHz. 2 Khz is typical of many SS radios, because the IFT are not transformers but single winding coils, so the shape of the selectivity curve is lousy. at leat 5 tuned circuits are needed for a radio, and 4 of the 5 shoulod be those in an IF amp, where the critical coupling of the IFTs result in a flat topped bandpass filter. Three IFTs are good, but each IFT further reduces AF bandwidth, so to get more AF BW with steep sided bandpass filter action, a higher frequency has to be used for the IF, hence the idea of 2.5 MHz. So that if the staion recieved is at 500 kHz, the oscillator is at 2 Mhz, just above the top station frequency. at a staion of 1,600 Khz, the oscillator is at 3.6 Mhz. The oscillator F range is easily tuned. The coils for IFT at 2.5 MHz might be easier to wind than 455 kHz, and they could be wound on existing 455 kHz formers. They need less turns, and less capacitance. With an IF of 2.5 MHz, the filtering of RF out of the recovered audio would be very easy. The pentodes used for IFamps will work just as well. With two IF amps, only the first one really needs to be subject to AVC for MW locals. But for short wave, obviously, the 2.5 mHz IF would restrict you to stations above 3 Mhz. Once one considers SW, up to say 30 Mhz, one has to think about quitet inputs, quiet mixers, maybe doble conversion perhaps mechanical or crystal filters, then there is the BFO, for sideband reception, and so on. A good SW radio will pick up a few radio hams on the bands. But for SW, antennas become important, and for the amateur bands, rotating beam antennas are the go, so that someone using a single 807 in a transmitter in japan, can be heard OK in Canberra. Or even here ! The Racal communications radio was one of the best ever made, with a cast and machined aluminium chassis, to seperate all the stages, and this had 22 valves. Digital techniques dominate HF radio now, and receivers just digitise the antenna signal, whatever it is, and get a computer to count what is in there, and just extract certain numbers, and there is your siganl, at any requency selected, its more like a noise analyser. I know hams who built everything in the shed, all the test gear, and all the transmitters and modulators, etc. They are mostly retired, and getting old, and they craft will die with them MMMH ! I don't feel such old !! and not ready to die soon !! because HF radio is becoming an old fashioned way to communicate long distance. And that communication is subject to ionsphere condtions in 11 year cycles, and lots and lots of bloody noise and interference. Then there are all the problems of restrictions on radio gear in built up areas. I doubt I will ever become a ham radio guy. I find too much activity at AF frequencies, which prevents me going higher. Yes, but we need DG to talk about what we're doing. The main specificity of amateur radio is to use the rig to speech about the rig ! And I will certainly never had the opportunity to listen your amps except if you tie one as a modulator for an AM transmitter ! With those 12x6550, you could easily modulate a full RF KW . Then I'll need a good variable selectivity receiver Patrick Turner. P.S. Oh ! what a beautifull red tie and assorted handkerchief !! Sure, with radios it's not so easy to have pictures from antipods ! Cheers, Yves. |
#3
|
|||
|
|||
Yves Monmagnon wrote: "Patrick Turner" wrote in message ... Yves Monmagnon wrote: snip for brevity.... Yes, but we need DG to talk about what we're doing. The main specificity of amateur radio is to use the rig to speech about the rig ! And I will certainly never had the opportunity to listen your amps except if you tie one as a modulator for an AM transmitter ! With those 12x6550, you could easily modulate a full RF KW . Then I'll need a good variable selectivity receiver Well maybe you will have to get busy to build the Miller TRF reciever detailed at http://users.rcn.com/jbyrns/Miller570.html To make such a wonderful radio, REAL craftmanship is required. Patrick Turner. P.S. Oh ! what a beautifull red tie and assorted handkerchief !! Sure, with radios it's not so easy to have pictures from antipods ! Cheers, Yves. It was the only time I wore that suit last year...... Anyway, if I used the 300 watt amps to modulate a nice linear RF amp, the atmospheric conditions between here and France would ruin the fidelity. But if I can borrow a few tubes from Radio America, and do a deal with the power company, and borrow one of Radio National's antennas, maybe you get me pretty good. BTW, watch out for our Oz stars in Le Tour, I hear they are giving some guys a hard time. But maybe Lance will win again, who knows? But that Miller radio is not a bad idea, and I tried a similar front end, but its a bit hard to get the flat topped response. The synchrodyne simply is supposed to add a constant level RF tone of the same F as the recieved station, to that station signal. Then all that is passed through a low pass filter, and hence audio is made. if there was a station 9 kHz away, then you get a 9 kHz AF tone detected in the audio from the wanted station. Then if you have a very steep LP AF filter with a cut off of 8.5 kHz, the adjoining carrier is kept out of the wanted station signal But if the station at 9 kHz away from the wanted one has AF modulation, then it is in the form of a 9 kHz carrier modulated by its AF content, and the sharp cut off AF filter cannot keep out the sidebands either sider of 9 kHz, and it sounds bloody awful. So synchronous, or direct conversion is good where there are no other strong signals, but needs to have a small AF BW where stations of similar strength are close to each other. I think a superhet is MUCH easier to build than that Miller radio. And I still think 2.5 MHz IF would be practical, and have 3 IFTs. This would allow a high Q for the total IF coils system, and for 25 kHz AF BW, the Q could be 100. And with a total of six tuned IF circuits, the slope away from the cut off F would indeed be nice and steep. There would not need to be much gain in the tubes used for IF. The secert to getting clean performance from a frequency converter is to have a fair degree of selectivity prior to the converter, so the signal you want isn't modulated by strong stations nearby, also entering the converter. so the double LC tuned stage with mutual L or mutual C coupling in the bottom of the L of each LC circuit will make a nice front end system. Patrick Turner. |
#4
|
|||
|
|||
JC wrote: I saw a couple of people suggest an old Eico tuner. There is also a Heathkit AM tuner from the 50's that was called hifi back then. I have never used it so I can't comment on its sensitivity or selectivity. I have its companion FM tuner the FM-3 and I used it use it several years ago it was a good performer if you didn't need stereo. They are on ebay constantly here is one http://cgi.ebay.com/ws/eBayISAPI.dll...category=29 6 Heathkit also marketed another AM RX kit, their BC-1. Has most of the features of the Eico HFT-94 with the exception of the switchable IF BW. Not as well built either. I have one here built by someone else. Bought it from one of the guys who worked in the same lab as I about 40 years ago. I used it here in the barn for several years. Worth a look, anyway. Cheers, John Stewart |
#5
|
|||
|
|||
In article , "JC"
wrote: I saw a couple of people suggest an old Eico tuner. There is also a Heathkit AM tuner from the 50's that was called hifi back then. I have never used it so I can't comment on its sensitivity or selectivity. I have its companion FM tuner the FM-3 and I used it use it several years ago it was a good performer if you didn't need stereo. They are on ebay constantly here is one http://cgi.ebay.com/ws/eBayISAPI.dll...category=29 6 Yeah, the heathkits are better receivers than the EICO HFT-94, although they both suffer somewhat from the edgy sound of semiconductor diode detectors. I had an HFT-94 in the early 1960's, and it just didn't do as good a job as the heathkit does. There are at least 4 different models, using similar circuitry, in the heathkit lineup of tube based AM tuners. The famed BC-1A was the progenitor of the family, which also included the AJ-20, AJ-21, and AJ-53. The BC-1A was built in the gold anodized aluminum cabinet, which wasn't the greatest. The BC-1A had only a wide band IF for High Fidelity, and didn't include the switch that was added in the later models for switching to narrow band operation. The antenna circuit also was changed in the later models, along with a change to solid state rectification in the power supply, but otherwise the original BC-1A circuit was carried through all four models. The AJ-20, which is the model I have was built in the black and gold flecked cabinet, and has improved chassis construction. The AJ-20 was re styled for the AJ-21 and AJ-53 which used heathkit's "automotive dashboard" styling motif. The AJ-21 and AJ-53 differed in the color scheme used. It would be hard to beat one of these heatkit's, unless you can find a Fisher AM-80 AM tuner. Although the original BC-1A is best avoided if you are interested in DX reception. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#6
|
|||
|
|||
John Byrns wrote:
Yeah, the heathkits are better receivers than the EICO HFT-94, although they both suffer somewhat from the edgy sound of semiconductor diode detectors. I have a BC-1A myself, and it uses semiconductor detector diodes. I suppose I could remove them and sub a tube like a 6110 or similar sub mini tube with wire leads under the chassis. It's sitting in storage right now, so it will be some time before I can do that.... |
#7
|
|||
|
|||
Sargent-Rayment built what was considered the best commercial AM tuner
ever made in the early-mid 50s using a unique circuit that may be found in the RDH3, I think...they put a huge engineering effort into what was unfortunatley a dying art, they'd been building TRF and SH radios for the military & commercial markets since the 20s...Audio said it was the only AM made that approached FM quality - see the upcoming article in Vacuum Tube Valley #20, coming out soon... Roger in NY " Uncle Peter" wrote in message news:UwmPa.53$zd4.5@lakeread02... "Robert Casey" wrote in message ... John Byrns wrote: Yeah, the heathkits are better receivers than the EICO HFT-94, although they both suffer somewhat from the edgy sound of semiconductor diode detectors. I have a BC-1A myself, and it uses semiconductor detector diodes. I suppose I could remove them and sub a tube like a 6110 or similar sub mini tube with wire leads under the chassis. It's sitting in storage right now, so it will be some time before I can do that.... Stupid question, but would there be any advantage to using a cathode follower to drive a diode detector? Pete |
#9
|
|||
|
|||
In article , Patrick Turner
wrote: I have not tried the use of a balanced IF amp, to eliminate the 2H distortion so prevalent in a typical SE pentode stage which is a variable mu type. Then one could have a balanced CF detector arrangement, and this would reduce thd to even lower. If the IF envelope is amplified by an amp with a distorted transfer characteristic, then the audio detected from it will contain the the distortions. Another method is to detect the +ve and -ve halves of the AM envelope, and sum them to cancel thd. I'm not sure if I am following this last correctly, I may have to think about it some more to be sure, but it sounds like you suggesting this as a way to cancel the some of the distortion when a variable mu pentode stage is used to drive the diode? Of course some AM tuners eliminate this problem at the source by using an ordinary pentode, no variable mu, to drive the diode. Would the full wave voltage doubler detector used in the Heathkit tuners accomplish this same goal, it's a possible advantage for this circuit I hadn't thought of before? No wonder the Heathkits are so great! Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#10
|
|||
|
|||
John Byrns wrote: In article , Patrick Turner wrote: I have not tried the use of a balanced IF amp, to eliminate the 2H distortion so prevalent in a typical SE pentode stage which is a variable mu type. Then one could have a balanced CF detector arrangement, and this would reduce thd to even lower. If the IF envelope is amplified by an amp with a distorted transfer characteristic, then the audio detected from it will contain the the distortions. Another method is to detect the +ve and -ve halves of the AM envelope, and sum them to cancel thd. I'm not sure if I am following this last correctly, I may have to think about it some more to be sure, but it sounds like you suggesting this as a way to cancel the some of the distortion when a variable mu pentode stage is used to drive the diode? Of course some AM tuners eliminate this problem at the source by using an ordinary pentode, no variable mu, to drive the diode. In MY home designed and built from scratch superhet radio, I have a 6BX6 as the IF tube, with an unbypassed Rk to provide some current FB. The linearity at 10vrms of IF signal is fairly good for a sharp cut off pentode in this application, and its better than say a 6BA6. The detector I have produces its AF from the +ve side of the IF envelope, but also another cap and diode is used to produce an AVC -voltage from the -ve side of the envelope, and this AVC is only applied to an RF input stage using a variable mu twin triode in a grounded grid stage, driven by a CF. The signal is so small in the triode stage that its linearity is fine, and so is that of the 6AN7 converter stage, which like the 6BX6, runs with fixed bias. Would the full wave voltage doubler detector used in the Heathkit tuners accomplish this same goal, it's a possible advantage for this circuit I hadn't thought of before? No wonder the Heathkits are so great! I don't know. I haven't seen the schematic and only careful analysis, or trial and measurement would tell you. But if you have the last IFT winding set up with 1 megohm from each end to some fixed positive bias of say +100v, then the grids of two CFs can be direct connected to each side of the winding. Then there are two seperate diode detectors, and filters, and another pair of CF, with an AF 1:1 tranny attatched, and from its secondary there will be little thd when the balanced output is coverted to single phase for use with a following amplifier, if of course it needs an unbalanced input signal. Balanced all the way, including a balanced RF input stage and converter stage has some merits. You just use twice the tube count, but plenty of tubes are there to use. The 1M resistors won't load down the last IF tube significantly, or over damp the Q of the last tuned circuit. In FM IF circuits, no need for balanced amps, because the linearity don't matter much, since the limiter output is just a 10.7 MHz signal which is devoid of amplitude changes, and distortion products of 10.7 MHz envelope can't appear in the audio, since the detector only detects variations in the distances between the up/downs of the 10.7 MHz waves. Patrick Turner. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#11
|
|||
|
|||
In article , Patrick Turner
wrote: John Byrns wrote: In article , Patrick Turner wrote: I have not tried the use of a balanced IF amp, to eliminate the 2H distortion so prevalent in a typical SE pentode stage which is a variable mu type. Then one could have a balanced CF detector arrangement, and this would reduce thd to even lower. If the IF envelope is amplified by an amp with a distorted transfer characteristic, then the audio detected from it will contain the the distortions. Another method is to detect the +ve and -ve halves of the AM envelope, and sum them to cancel thd. I'm not sure if I am following this last correctly, I may have to think about it some more to be sure, but it sounds like you suggesting this as a way to cancel the some of the distortion when a variable mu pentode stage is used to drive the diode? Of course some AM tuners eliminate this problem at the source by using an ordinary pentode, no variable mu, to drive the diode. I have had a chance to think about this a little more, and pending doing the hard math, I don't believe that using a push pull IF stage will help with "the 2H distortion so prevalent in a typical SE pentode stage which is a variable mu type." Push pull stages only cancel even order distortion products, and the following IF transformer also removes the even order distortion products, including "2H", so a push pull IF stage will only help in that it will provide more power to drive the detector diode, which with your cathode follower scheme is pretty much irrelevant anyway. In MY home designed and built from scratch superhet radio, I have a 6BX6 as the IF tube, with an unbypassed Rk to provide some current FB. The linearity at 10vrms of IF signal is fairly good for a sharp cut off pentode in this application, and its better than say a 6BA6. The detector I have produces its AF from the +ve side of the IF envelope, but also another cap and diode is used to produce an AVC -voltage from the -ve side of the envelope, and this AVC is only applied to an RF input stage using a variable mu twin triode in a grounded grid stage, driven by a CF. The signal is so small in the triode stage that its linearity is fine, and so is that of the 6AN7 converter stage, which like the 6BX6, runs with fixed bias. Would the full wave voltage doubler detector used in the Heathkit tuners accomplish this same goal, it's a possible advantage for this circuit I hadn't thought of before? No wonder the Heathkits are so great! I don't know. I haven't seen the schematic and only careful analysis, or trial and measurement would tell you. As above, I have decided it wouldn't help because the IFT has already filtered out the 2H distortion. But if you have the last IFT winding set up with 1 megohm from each end to some fixed positive bias of say +100v, then the grids of two CFs can be direct connected to each side of the winding. Then there are two seperate diode detectors, and filters, and another pair of CF, with an AF 1:1 tranny attatched, and from its secondary there will be little thd when the balanced output is coverted to single phase for use with a following amplifier, if of course it needs an unbalanced input signal. This might help with distortion generated in your first CF, but why not just build a more robust CF in the first place? On the other hand, it isn't clear to me that you really want to drive the diode with a CF in the first place, what is wrong with hooking it directly to the IFT in the traditional fashion? A CF following the diode is definitely a good thing though. Balanced all the way, including a balanced RF input stage and converter stage has some merits. You just use twice the tube count, but plenty of tubes are there to use. The 1M resistors won't load down the last IF tube significantly, or over damp the Q of the last tuned circuit. I don't see the advantage of "balanced all the way" as far as audio distortion goes? In FM IF circuits, no need for balanced amps, because the linearity don't matter much, since the limiter output is just a 10.7 MHz signal which is devoid of amplitude changes, and distortion products of 10.7 MHz envelope can't appear in the audio, since the detector only detects variations in the distances between the up/downs of the 10.7 MHz waves. I'm not an expert in this, but I think this is a fallacy, and the linearity of FM IF stages is actually quite important. Nonlinear IF stages in an FM system that sees only one signal may very well not effect the detected audio, but I believe that nonlinear IF stages can degrade the capture ratio, and affect the performance of the FM receiver when it confronted with multiple signals, such as under multipath conditions, or conditions of co channel interference, but there are many complexities that enter into this. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#12
|
|||
|
|||
John Byrns wrote: In article , Patrick Turner wrote: John Byrns wrote: In article , Patrick Turner wrote: I have not tried the use of a balanced IF amp, to eliminate the 2H distortion so prevalent in a typical SE pentode stage which is a variable mu type. Then one could have a balanced CF detector arrangement, and this would reduce thd to even lower. If the IF envelope is amplified by an amp with a distorted transfer characteristic, then the audio detected from it will contain the the distortions. Another method is to detect the +ve and -ve halves of the AM envelope, and sum them to cancel thd. I'm not sure if I am following this last correctly, I may have to think about it some more to be sure, but it sounds like you suggesting this as a way to cancel the some of the distortion when a variable mu pentode stage is used to drive the diode? Of course some AM tuners eliminate this problem at the source by using an ordinary pentode, no variable mu, to drive the diode. I have had a chance to think about this a little more, and pending doing the hard math, I don't believe that using a push pull IF stage will help with "the 2H distortion so prevalent in a typical SE pentode stage which is a variable mu type." Push pull stages only cancel even order distortion products, and the following IF transformer also removes the even order distortion products, including "2H", so a push pull IF stage will only help in that it will provide more power to drive the detector diode, which with your cathode follower scheme is pretty much irrelevant anyway. I assure you there are references to IMD caused by IF amp nonlinearities. Maybe you should inspect the output of a typical pentode IF amp. You may not easily see less than 2% thd on the modulation waveform on a CRO. Perahps you ought to measure a typical IFamp, by using a linear detector, as described. But I am sure that a balanced amp with pentodes, with a common cathode. each with a lot of 2H when used alone as an SE amp, would display less thd than the SE case. They use PP balanced oscillators when a signal output is required to be cleaner than an SE type. In MY home designed and built from scratch superhet radio, I have a 6BX6 as the IF tube, with an unbypassed Rk to provide some current FB. The linearity at 10vrms of IF signal is fairly good for a sharp cut off pentode in this application, and its better than say a 6BA6. The detector I have produces its AF from the +ve side of the IF envelope, but also another cap and diode is used to produce an AVC -voltage from the -ve side of the envelope, and this AVC is only applied to an RF input stage using a variable mu twin triode in a grounded grid stage, driven by a CF. The signal is so small in the triode stage that its linearity is fine, and so is that of the 6AN7 converter stage, which like the 6BX6, runs with fixed bias. Would the full wave voltage doubler detector used in the Heathkit tuners accomplish this same goal, it's a possible advantage for this circuit I hadn't thought of before? No wonder the Heathkits are so great! I don't know. I haven't seen the schematic and only careful analysis, or trial and measurement would tell you. As above, I have decided it wouldn't help because the IFT has already filtered out the 2H distortion. If you examine the plate output of a pentode IF amp, with a AM envelope, the audio signal outlines the extent of RF voltage swing on the top of the wave form, and on the bottom, but the two AF modulation waves are out of phase. The upper half of the envelope is amplified more than the bottom half, since the Gm of the tube is higher when more instantaneous current flows. The AF envelope is thus endowed with the the tube transfer characteristics and the AF recovered by the following detector will contain such distortions, and with an SE stage, there is a lot of 2H introduced to the recovered AF, and as well some other H because a pentode generates a lot of thd at high outputs. The impedance of the IFT input might be 50k at Fo, but at AF, its maybe only 100 ohms, so any even order assymetry in the top and bottom halves of the envelope wave form tends to balance out. But the damage to the fidelity has already been done by generation of imds. I can't see why ppl should try a balanced RF and IF and oscillator, just to see how clean they could get a tube stage to be. I don't believe they would be dissappointed. Just as we tryto get thd of an audio amp down to a respectable 0.1%, then why not aim in the same direction for all our tuners, where it is applicable? But if you have the last IFT winding set up with 1 megohm from each end to some fixed positive bias of say +100v, then the grids of two CFs can be direct connected to each side of the winding. Then there are two seperate diode detectors, and filters, and another pair of CF, with an AF 1:1 tranny attatched, and from its secondary there will be little thd when the balanced output is coverted to single phase for use with a following amplifier, if of course it needs an unbalanced input signal. This might help with distortion generated in your first CF, but why not just build a more robust CF in the first place? Mine is as robust as I think I need it to be. All I am saying, it could all be better, in terms of measured imd, and that can't be a bad thing On the other hand, it isn't clear to me that you really want to drive the diode with a CF in the first place, what is wrong with hooking it directly to the IFT in the traditional fashion? The load seen by the IF amp is the impedance of the LC circuit of the IFT primary, plus the reflected load of the secondary multiplied by the mutual coupling factor, usually a lot less than 1.0. Nevertheless, we realise the best IFT/tube performance if the load driven by the IFT is as high as possible a value. The CF input grid is high enough. So the IFT and IF amp tube are buffered from the clipping of the peaks of the envelope due to the diode charging. The CF itself sees a benign load, and the diiode is biased on, and I assure you its a more precise way to demodulate an AM envelope. There is less latch up of the CR network being charged by the CF, because the CF has plenty of current, and is a low Ro sigbnal source. A CF following the diode is definitely a good thing though. Balanced all the way, including a balanced RF input stage and converter stage has some merits. You just use twice the tube count, but plenty of tubes are there to use. The 1M resistors won't load down the last IF tube significantly, or over damp the Q of the last tuned circuit. I don't see the advantage of "balanced all the way" as far as audio distortion goes? Maybe you should try building one to find out. But I have schematics using twin gated fets for RF input, and for a balanced converter, and there is no reason why it couldn't be balanced all the way through, and benefit from the greater dynamic range of tubes. In FM IF circuits, no need for balanced amps, because the linearity don't matter much, since the limiter output is just a 10.7 MHz signal which is devoid of amplitude changes, and distortion products of 10.7 MHz envelope can't appear in the audio, since the detector only detects variations in the distances between the up/downs of the 10.7 MHz waves. I'm not an expert in this, but I think this is a fallacy, and the linearity of FM IF stages is actually quite important. Nope, and a good FM IF stage starts to limit whenh a low threshold of input signal is applied, and the output of the limiter, often a 6AU6 is that of a tube enduring 30 dB of gross overload, ie, the tube is being driven well into clipping, and it it wasn't for the fact of the discriminator LC input circuit, the wave form at the 6AU6 would be a 10.7 MHz square wave. The discriminator would still only produce the same audio, even with a perfectly square IF wave, since the frequency changes of +/- 75 kHz carry the audio info and that's all that the discriminator detects, unless some unwanted AM is present, such as noise. FM is inherently immune to noise, such as lightning bolts, which rarely manifests itself as FM, but usually only as a burst of AM. Nonlinear IF stages in an FM system that sees only one signal may very well not effect the detected audio, but I believe that nonlinear IF stages can degrade the capture ratio, and affect the performance of the FM receiver when it confronted with multiple signals, such as under multipath conditions, or conditions of co channel interference, but there are many complexities that enter into this. Perhaps so, but I doubt that matters if the IF is receieving enough signal to drive it well into limiting. If the limiting was marginal, and some other received out of phase signal could beat with the wanted signal, causing the IF to drop below limiting, and then subject the discriminator to AM, and intermittent AM, and the sound goes scratchy, and distorted. Patrick Turner. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#13
|
|||
|
|||
John Byrns wrote: Exactly what is the advantage of this "honeycomb" construction method over other IF transformer construction techniques, say for example the cup core style of construction? Why was McIntosh so interested in the "honeycomb" design? I would guess that the primary advantage of the "honeycomb" construction in a simple double tuned IF transformer is not technical, but is in the marketing area, giving the manufacturer another unusual point to advertise. As you suspect - the primary "advantage" was in print. Having said that - I'll noted that there was a school of thought that since the wires in a honeycomb coil weren't "wound tightly together side-by side and in tight layers"- the capacitance between layers / adjacent windings was less - therefore providing a more "pure" inductance. best regards... -- randy guttery A Tender Tale - a page dedicated to those Ships and Crews so vital to the United States Silent Service: http://tendertale.com |
#14
|
|||
|
|||
In article ,
(firedome) wrote: As for the circuit's provenance, I never said S-R invented it...it was, as I said, an adaptation of the Selden -Smith 2 tube decector circuit (did any one else use 2 tube detectors?) found in the RDH3. Will Rayment never claimed to have invented it, but was the only one to use it in a commercial product AFAIK. I tried to email you about some information that I received from Walter Selsted on the "two tube" detector, but your email address bounces. Perhaps you already talked to Walter, did you interview him for your article on the S-R tuner? Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#15
|
|||
|
|||
In article , Patrick Turner
wrote: John Byrns wrote: In article , Patrick Turner wrote: I have had a chance to think about this a little more, and pending doing the hard math, I don't believe that using a push pull IF stage will help with "the 2H distortion so prevalent in a typical SE pentode stage which is a variable mu type." Push pull stages only cancel even order distortion products, and the following IF transformer also removes the even order distortion products, including "2H", so a push pull IF stage will only help in that it will provide more power to drive the detector diode, which with your cathode follower scheme is pretty much irrelevant anyway. I assure you there are references to IMD caused by IF amp nonlinearities. Maybe you should inspect the output of a typical pentode IF amp. You may not easily see less than 2% thd on the modulation waveform on a CRO. Perahps you ought to measure a typical IFamp, by using a linear detector, as described. Read what I said again, I didn't say that there is no "IMD" caused by the IF amp nonlinearities, what I said was that the even order distortion products, including "2H", are filtered out by the IF transformer before reaching the detector. What mechanism are you suggesting that allows a significant amount of the even order distortion products make it through the IF transformer? As far as measuring the distortion at the detector output, I don't have an AM generator with sufficiently low distortion to do this, I wish that I did have a generator good enough to measure the effects we are talking about. But I am sure that a balanced amp with pentodes, with a common cathode. each with a lot of 2H when used alone as an SE amp, would display less thd than the SE case. This is not a "SE" audio amplifier, this is an IF amplifier, which in a tube radio is typically followed by a double tuned IF transformer. The double tuned IF transformer filters out the even order distortion products, allowing only the odd order distortion to pass, just like your push pull stage. The only thing the push pull IF stage is going to do is increase the power available to drive the detector. They use PP balanced oscillators when a signal output is required to be cleaner than an SE type. An oscillator is a whole different matter, we don't want oscillator harmonics, or RF harmonics, getting into the mixer where they can create spurious responses. In FM IF circuits, no need for balanced amps, because the linearity don't matter much, since the limiter output is just a 10.7 MHz signal which is devoid of amplitude changes, and distortion products of 10.7 MHz envelope can't appear in the audio, since the detector only detects variations in the distances between the up/downs of the 10.7 MHz waves. I'm not an expert in this, but I think this is a fallacy, and the linearity of FM IF stages is actually quite important. Nope, I suggest you read up on your FM theory, there were a lot of good papers published on this subject in the 1950s. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#16
|
|||
|
|||
"Randy and/or Sherry" wrote in message ... As you suspect - the primary "advantage" was in print. Having said that - I'll noted that there was a school of thought that since the wires in a honeycomb coil weren't "wound tightly together side-by side and in tight layers"- the capacitance between layers / adjacent windings was less - therefore providing a more "pure" inductance. best regards... -- randy guttery A Tender Tale - a page dedicated to those Ships and Crews so vital to the United States Silent Service: http://tendertale.com Randy I remember reading somewhere that one of the advantages to that sort of winding was to reduce the proximity effect of having wires laying to close to each other, which would cause a loss of Q? Pete |
#17
|
|||
|
|||
OK, going back to the original premise of this thread.
Parts are still readily available. I remember the original poster being concerned about the tuning cap. Antique Electronic Supply has a beautiful 3 gang 500 pf/gang unit for $24.95. Coils and IF transformers are all easily available, and for not much cash, either. The Miller coils are available as NOS from a number of places. AES has coils and IF transformers. Tubes are easy. Using the BA6/BE6 lineup would be a snap. The tubes are about $3 or $4 each. What I'm saying is that the electronic part of the project is easy, and not all that much more expensive than 25 years ago. Everything is PC mount now. No socket hole punches needed. It seems like actually deciding how to build the thing is the biggest problem! Hardware (chassis, dial, etc.) will be much harder than building the electronics. I'm a big fan of the modular technique. You work with smaller boards, and build in steps. Each module can be tested separately. It was always my dream from a very young age to build my own receiver, which is why I've been following this. |
#18
|
|||
|
|||
In article , donut wrote:
Coils and IF transformers are all easily available, and for not much cash, either. The Miller coils are available as NOS from a number of places. AES has coils and IF transformers. Where are some places I can get NOS Miller coils and transformers? Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#19
|
|||
|
|||
donut wrote: What I'm saying is that the electronic part of the project is easy, and not all that much more expensive than 25 years ago. Everything is PC mount now. No socket hole punches needed. As I noted before - my senior class project was building an AM radio. And the "idea" was to build it in sections... power supply / audio then IF and converter. I just built mine in a straight line layout - used the brake in the sheet metal shop to form the chassis - chassis punches for the tube sockets and a bit of careful drilling / filing for the IF can mounts. It was a "simple" AA5 - but it worked fine (for an AA5+1 - I added a 6E5 to the lineup - just "because"). It wasn't fancy - but looked as good as any "factory" chassis - I guess I could have chrome plated it --- since I was getting some stuff for my car chromed about the same time - probably wouldn't have been that big a deal just never crossed my mind. Point is - with a little effort - standard chassis work is certainly within the reach of anyone willing to do just a bit of work - either starting with a "bud box" or forming one from scratch. Unless someone decides to get a bit funky - you only need one size of chassis punch (either for miniature or octal tubes)... usually for less than $30.00 - or "borrowed" from a friendly electrician. just my .02... -- randy guttery A Tender Tale - a page dedicated to those Ships and Crews so vital to the United States Silent Service: http://tendertale.com |
#20
|
|||
|
|||
John Byrns wrote: In article , Patrick Turner wrote: John Byrns wrote: In article , Patrick Turner wrote: I have had a chance to think about this a little more, and pending doing the hard math, I don't believe that using a push pull IF stage will help with "the 2H distortion so prevalent in a typical SE pentode stage which is a variable mu type." Push pull stages only cancel even order distortion products, and the following IF transformer also removes the even order distortion products, including "2H", so a push pull IF stage will only help in that it will provide more power to drive the detector diode, which with your cathode follower scheme is pretty much irrelevant anyway. I assure you there are references to IMD caused by IF amp nonlinearities. Maybe you should inspect the output of a typical pentode IF amp. You may not easily see less than 2% thd on the modulation waveform on a CRO. Perahps you ought to measure a typical IFamp, by using a linear detector, as described. Read what I said again, I didn't say that there is no "IMD" caused by the IF amp nonlinearities, what I said was that the even order distortion products, including "2H", are filtered out by the IF transformer before reaching the detector. Indeed, if the top half of the IF envelope is amplified more than the bottom half, then the waveform is as if somebody added an AF component to the 455 kHz wave form, and this is shunted by the low impedance of the IFTs at AF. But the IMDs remain, and if the envelope is subject to limiting or clipping then this will distort the detected AF. If the peaks of an AM envelope are compressed on top and bottom, then that's odd order distortion of the whole wave form, but when one side of the compressed IF envelope is detected, the AF thd is mainly even order! In most sets the output voltage from the detector is around an average of 1 to 2 vrms, from a 30% modulated carrier. Working back from that, the peak to peak envelope of the IF amp has to be about 8 to 16 v. The maximum before IF amp clipping occurs could be perhaps 100v, at which voltage the thd is severe. Luckily, tube thd at lower voltage outputs is not severe, but nevertheless, the more linear the IF stage, the less thd there is in the AF. What mechanism are you suggesting that allows a significant amount of the even order distortion products make it through the IF transformer? I agree, the even order harmonic mauling of the modulation don't make it through. But the IMDs must surely get through, as would any odd order harmonic mauling of the wave form. As far as measuring the distortion at the detector output, I don't have an AM generator with sufficiently low distortion to do this, I wish that I did have a generator good enough to measure the effects we are talking about. I do have enough gear make some measurements. My RF gene has an SE pentode with modulated AM output, with an RF output transformer, and there is some even order generated on the shape of the modulation envelope. But with a dual trace CRO, I can monitor the input AM wave form, and overlay the detected AF on top and adjust until they line up very closely at some low value of AM input to the set under test, and with the AVC shunted to a fixed negative voltage which is usual for most local recieved stations. Then when the input AM is increased, there should be an eqaul increase in the detected AF, free of much difference between the AF and the modulation envelope shape. If about 15vrms at 1 kHz can be detected before any serious distortions occur, you have a good RF input, mixer, IF amp, and detector. I don't have a really pure sig gene, with a certified less than 0.01% thd. So just measuring the thd in the AF output from a detector doesn't tell you much. But if you do have less than 1% at say 10v output, which ius impossible to see on a CRO, then both the sig gene, and radio are doing OK. I am simply trying to suggest that there may be some benefits, even though small, to using balanced circuitry right through an AM radio reciever, even up to the audio amp speaker. But I am sure that a balanced amp with pentodes, with a common cathode. each with a lot of 2H when used alone as an SE amp, would display less thd than the SE case. This is not a "SE" audio amplifier, I didn't say it was. this is an IF amplifier, which in a tube radio is typically followed by a double tuned IF transformer. The double tuned IF transformer filters out the even order distortion products, allowing only the odd order distortion to pass, just like your push pull stage. The only thing the push pull IF stage is going to do is increase the power available to drive the detector. But what of the imd generated by the non linear operation of the SE tube? RDH4 does mention the problems I have mentioned on thr bottom of page 1,067. Using a PP balanced amp for at least the final IF amp would have nothing but a beneficial effect on the sound, and I suggest ppl try it. The extra 6 dB of voltage ability would be welcome. But of course unless the stage is done with an IFT with centre tapped windings, then parafeed might have to be used, with a centre tapped RF choke used to feed the DC to each half of the PP IF amp. The CF buffers following the IF amp can be quite linear, and if 12AT7 is used, the thd at 10v output would be indeed low, even if there is a simple SE detector. The gain reduction in a a 12AT7 CF is about 50 times, so the thd introduced is 0.2% maximum. The diode distortion following is lowest when Vo is high. They use PP balanced oscillators when a signal output is required to be cleaner than an SE type. An oscillator is a whole different matter, we don't want oscillator harmonics, or RF harmonics, getting into the mixer where they can create spurious responses. Agreed. In FM IF circuits, no need for balanced amps, because the linearity don't matter much, since the limiter output is just a 10.7 MHz signal which is devoid of amplitude changes, and distortion products of 10.7 MHz envelope can't appear in the audio, since the detector only detects variations in the distances between the up/downs of the 10.7 MHz waves. I'm not an expert in this, but I think this is a fallacy, and the linearity of FM IF stages is actually quite important. Nope, I suggest you read up on your FM theory, there were a lot of good papers published on this subject in the 1950s. But what about when the amp is under serious limiting? The IF amp is then saturated, and the current distortion is huge. The whole stage simply acts to convert the incomming IF wave a simple clipped sine wave, and the IF can be quite non linear in this condition, when AF fidelity is then highest. FM radios represented nice big move away from the constrictions we need to carefully follow in AM recievers, where any FM would be deadly. Patrick Turner. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#21
|
|||
|
|||
"John Byrns" wrote in message ... In article , donut wrote: Where are some places I can get NOS Miller coils and transformers? Regards, John Byrns A dwindling supply was offered here... Not much left.. http://www.metcoelectronics.com/coils.htm#antenna Peter |
#22
|
|||
|
|||
In article , Patrick Turner
wrote: John Byrns wrote: In article , Patrick Turner wrote: John Byrns wrote: In article , Patrick Turner wrote: I have had a chance to think about this a little more, and pending doing the hard math, I don't believe that using a push pull IF stage will help with "the 2H distortion so prevalent in a typical SE pentode stage which is a variable mu type." Push pull stages only cancel even order distortion products, and the following IF transformer also removes the even order distortion products, including "2H", so a push pull IF stage will only help in that it will provide more power to drive the detector diode, which with your cathode follower scheme is pretty much irrelevant anyway. I assure you there are references to IMD caused by IF amp nonlinearities. Maybe you should inspect the output of a typical pentode IF amp. You may not easily see less than 2% thd on the modulation waveform on a CRO. Perahps you ought to measure a typical IFamp, by using a linear detector, as described. Read what I said again, I didn't say that there is no "IMD" caused by the IF amp nonlinearities, what I said was that the even order distortion products, including "2H", are filtered out by the IF transformer before reaching the detector. Indeed, if the top half of the IF envelope is amplified more than the bottom half, then the waveform is as if somebody added an AF component to the 455 kHz wave form, and this is shunted by the low impedance of the IFTs at AF. But the IMDs remain, and if the envelope is subject to limiting or clipping then this will distort the detected AF. "2H" distortion is just a special case of 2nd order "IMD". In looking up your reference to the RDH4th, I found that the page you referenced doesn't really say much, but it does refer in turn to pages 944 and 945, which briefly mentions that the transfer characteristic of the vari-mu tube can be modeled by a power series, and that the even order products don't make it through the IFT. It also says vari-mu tubes are especially designed to minimize 3rd order distortions, which do pass through the IFT to the detector. If the peaks of an AM envelope are compressed on top and bottom, then that's odd order distortion of the whole wave form, but when one side of the compressed IF envelope is detected, the AF thd is mainly even order! That sounds like a problem in the detector, not the IF amplifier. This raises an interesting point that I will have to think about. The third order distortions, which do make it through the IFT, may unbalance the amplitudes of the upper and lower sidebands around the carrier, resulting in the type of distortion you are talking about when an envelope detector is used, but if this is the case, I don't think your idea of detecting both the "ve- & ve+" sides of the wave form would fix the problem. A synchronous detector would eliminate distortion due to unbalanced sidebands, if indeed 3rd order distortion products can cause this. When I get a chance in the next few days I will try to run through the math for this and see if 3rd order distortion does unbalance the sidebands, resulting in distortion when envelope detectors are used. In FM IF circuits, no need for balanced amps, because the linearity don't matter much, since the limiter output is just a 10.7 MHz signal which is devoid of amplitude changes, and distortion products of 10.7 MHz envelope can't appear in the audio, since the detector only detects variations in the distances between the up/downs of the 10.7 MHz waves. I'm not an expert in this, but I think this is a fallacy, and the linearity of FM IF stages is actually quite important. Nope, I suggest you read up on your FM theory, there were a lot of good papers published on this subject in the 1950s. But what about when the amp is under serious limiting? The IF amp is then saturated, and the current distortion is huge. The whole stage simply acts to convert the incomming IF wave a simple clipped sine wave, and the IF can be quite non linear in this condition, when AF fidelity is then highest. FM radios represented nice big move away from the constrictions we need to carefully follow in AM recievers, where any FM would be deadly. I put together a bibliography on this subject a while back, which I can email to you if you are interested. I don't pretend to understand it all, in fact I am far from completely understanding it. I think the problem comes about because co channel signals, either other stations on the same channel, or multipath from the desired station result in amplitude modulation of the IF signal, which the limiters convert to phase modulation. I'm not sure if I have the right grip on this, but one way to deal with the problem seems to be to use a very wide band FM detector, with no filtering between the limiter and detector. The other approach, which I don't pretend to even begin to understand involves multiple limiters with filters between which have specially chosen bandwidths, which somehow allows the use of a narrower bandwidth FM detector, but the filter bandwidths must be carefully chosen. The bottom line seems to be if these factors aren't taken into consideration, then the suppression of multipath, and co channel interference, will suffer, and the full advantage of FM will not be realized. This just my possibly wrong, foggy understanding of what some of the papers were trying to say. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#23
|
|||
|
|||
In article , donut wrote:
(John Byrns) wrote in news:jbyrns-1207032036480001@216-80- 74-132.d.enteract.com: Where are some places I can get NOS Miller coils and transformers? AES has some. Also I found: http://www.oldradioparts.com/2a6fla.txt Thanks, I will check the "oldradioparts" site, I never noticed that AES had any of the J.W. Miller IF Transformers that I am interested in. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#24
|
|||
|
|||
In article ,
(firedome) wrote: (John Byrns) wrote in message ... In article , (firedome) wrote: Who considered the S-R tuner the best commercial tuner ever made?... Audio Engineering (July 1953)..."there has never before been available a radio receiver which provided AM quality which approached FM" ...and Consumer Research Bulletin (March 1953)..."the first satisfactory detector circuit yet observed for high fidelity use in an AM tuner". Those are certainly some fantastic superlatives! As for the circuit's provenance, I never said S-R invented it... And for my part I never said you said S-R invented it. it was, as I said, an adaptation of the Selden -Smith 2 tube decector circuit (did any one else use 2 tube detectors?) found in the RDH3. I missed the part where you said that. The Ampex AM/FM simulcast stereo tuner also used a similar "two tube" detector circuit. I forget the actual Ampex model number of this tuner, but many people refer to it as the "008" tuner, but that is not the actual model number. I mentioned the circuit in RDH3, not the inventors names... I'm going on what Will Rayment told me when I interviewed him for VTV, and these are his recollections of what happened 50 yrs ago. Yes, I think the problem is that the circuit was invented more than 60 years ago, and nobody remembers the details. I have not been able to find the circuit in the RDH3, and suspect that Will Rayment is misremembering this. I suspect that Will is actually probably thinking of the RDH4, which does show the circuit in an appendix of some printings. It would also be interesting to know how S-R found out about the circuit, as I don't think Water Selsted knows, although he did give S-R permission to use the circuit when they asked him. I have to suspect that the circuit probably became known by word of mouth in the California technical community during the 1940s. McIntosh obviously thought highly of the S-R IF coils as they pursued it with diligence. I myself am not sure of the advantage of honeycomb construction, but McIntosh apparently thought there was one beyond marketing hype. I'll ask Sid Corderman if he knows when I interview him. That would be an interesting question to ask, why do you say "McIntosh apparently thought there was one beyond marketing hype", I would think the opposite? As for SR-51s, there was one on eBay a few months ago, and I believe the SR-68 and later SR-1000 had essentially the same circuit. Was the SR-51 the AM only version? I know they made a bewildering array of versions, AM only, AM/FM, and with and without built in preamps. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#26
|
|||
|
|||
In article , wrote:
John Byrns wrote: A synchronous detector would eliminate distortion due to unbalanced sidebands, if indeed 3rd order distortion products can cause this. Yeah but a really good synch detector can set you back major bucks - Are they really worth it? I guess 30hz - 15Khz +/- 3db on AM is pretty good (Sherwood Eng.'s is rated that) - but I wonder about distortion - there is just something in my gut that says when you sub a local carrier for the original - something isn't going to be as clean as original. But I'll admit - I've always had a time understanding full pll synch systems... --- I'm saving my pennies... but at $550+ --- it isn't a high priority. Come on, today it only costs pennies, we do it all the time for FM stereo, although in that case the carrier isn't transmitted, so your gut doesn't have to worry. C-Quam AM stereo chips do it too, no big deal. Well, I suppose it is not trivial doing it with tubes, and tube type FM stereo used filters and doublers, rather than PLLs, but it would be doable. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#27
|
|||
|
|||
John Byrns wrote: In article , Patrick Turner wrote: John Byrns wrote: In article , Patrick Turner wrote: John Byrns wrote: In article , Patrick Turner wrote: I have had a chance to think about this a little more, and pending doing the hard math, I don't believe that using a push pull IF stage will help with "the 2H distortion so prevalent in a typical SE pentode stage which is a variable mu type." Push pull stages only cancel even order distortion products, and the following IF transformer also removes the even order distortion products, including "2H", so a push pull IF stage will only help in that it will provide more power to drive the detector diode, which with your cathode follower scheme is pretty much irrelevant anyway. I assure you there are references to IMD caused by IF amp nonlinearities. Maybe you should inspect the output of a typical pentode IF amp. You may not easily see less than 2% thd on the modulation waveform on a CRO. Perahps you ought to measure a typical IFamp, by using a linear detector, as described. Read what I said again, I didn't say that there is no "IMD" caused by the IF amp nonlinearities, what I said was that the even order distortion products, including "2H", are filtered out by the IF transformer before reaching the detector. Indeed, if the top half of the IF envelope is amplified more than the bottom half, then the waveform is as if somebody added an AF component to the 455 kHz wave form, and this is shunted by the low impedance of the IFTs at AF. But the IMDs remain, and if the envelope is subject to limiting or clipping then this will distort the detected AF. "2H" distortion is just a special case of 2nd order "IMD". In looking up your reference to the RDH4th, I found that the page you referenced doesn't really say much, but it does refer in turn to pages 944 and 945, which briefly mentions that the transfer characteristic of the vari-mu tube can be modeled by a power series, and that the even order products don't make it through the IFT. It also says vari-mu tubes are especially designed to minimize 3rd order distortions, which do pass through the IFT to the detector. It says enough to confirm that IF amps are one of the worst sources of thd of a radio, because it is a voltage amplifier producing considerable voltages. There is no nfb involved, and the IF tube operates at a region where introduced thd should be a worry for those who like real fidelity. If the peaks of an AM envelope are compressed on top and bottom, then that's odd order distortion of the whole wave form, but when one side of the compressed IF envelope is detected, the AF thd is mainly even order! That sounds like a problem in the detector, not the IF amplifier. No, it isn't. I have just outlined that the IF amp is capable of compessing the waveforms, ie, adding distortions, and these deformations of the AF modulation shapes are detected, and thus the IF tube has added muck to the AF signal. The ideal IF amp should be as linear as one can make it, period!, if one wants maximum fidelity. And a balance has to be struck to keep its voltage output lowish, but not too low to let noise start to rasie its ugly head. This raises an interesting point that I will have to think about. The third order distortions, which do make it through the IFT, may unbalance the amplitudes of the upper and lower sidebands around the carrier, resulting in the type of distortion you are talking about when an envelope detector is used, but if this is the case, I don't think your idea of detecting both the "ve- & ve+" sides of the wave form would fix the problem. I agree, once the odd distortions have mauled the IF envelope, both top and bottom modulation shapes become distorted. The thd cannot be undone, or filtered out. The PP CF detector isn't really needed, because its thd with a tube like 12AT7, or 6EJ7 or 6EJ7 will be very low, But the CF tube thd is cancelled by PP action, so 12AU7 is quite OK to use. A synchronous detector would eliminate distortion due to unbalanced sidebands, if indeed 3rd order distortion products can cause this. When I get a chance in the next few days I will try to run through the math for this and see if 3rd order distortion does unbalance the sidebands, resulting in distortion when envelope detectors are used. Good luck with the maths. I prefer to build and measure things to find out the ultimate truth. I have tried synchronous detection, and got bogged down with spurious oscillations and strange noises. But balanced synchronous detectros have less thd than SE types, and the purist would go for these in preference to an SE type. In FM IF circuits, no need for balanced amps, because the linearity don't matter much, since the limiter output is just a 10.7 MHz signal which is devoid of amplitude changes, and distortion products of 10.7 MHz envelope can't appear in the audio, since the detector only detects variations in the distances between the up/downs of the 10.7 MHz waves. I'm not an expert in this, but I think this is a fallacy, and the linearity of FM IF stages is actually quite important. Nope, I suggest you read up on your FM theory, there were a lot of good papers published on this subject in the 1950s. But what about when the amp is under serious limiting? The IF amp is then saturated, and the current distortion is huge. The whole stage simply acts to convert the incomming IF wave a simple clipped sine wave, and the IF can be quite non linear in this condition, when AF fidelity is then highest. FM radios represented nice big move away from the constrictions we need to carefully follow in AM recievers, where any FM would be deadly. I put together a bibliography on this subject a while back, which I can email to you if you are interested. I don't pretend to understand it all, in fact I am far from completely understanding it. I think the problem comes about because co channel signals, either other stations on the same channel, or multipath from the desired station result in amplitude modulation of the IF signal, which the limiters convert to phase modulation. FM sets are prone to reflection of signals off buildings, thus cancelling the recieved waves due to phase cancelling, reducing input, and the IF signal goes out of the limited state, causing crackles, buzzes, intermittent drop outs. We all remember trying to get a pair of rabbit ear antennas on top of a TV to give us a clear picture. There usually was only one postion which seemed to work best, and that was the one where reflections were minimal, and maximum wanted signal came through. I have had FM sets whose signals died when I've walked across the room. We have a lot of separation between FM stations here. Its impossible for two stations to be recieved at once. I have never heard it, unless tuned well away from both stations, then sometimes you get a noisy version of various stations. But once tuned, with plenty of signal to drive the IF into limiting, no problems. I'm not sure if I have the right grip on this, but one way to deal with the problem seems to be to use a very wide band FM detector, with no filtering between the limiter and detector. The detector is wide band, the IF could be a wide range of F, several MHz, and the FM wouls till be detected. IF BW in FM sets is usually 300 kHz, and stations are not permitted to use more than +/- 75 kHz deviation to express the amplitude changes to their AF signals. The sidebands produced by FM are far more complex than AM, but at the end of the day, FM offers far better fidelity. The other approach, which I don't pretend to even begin to understand involves multiple limiters with filters between which have specially chosen bandwidths, which somehow allows the use of a narrower bandwidth FM detector, but the filter bandwidths must be carefully chosen. Narrow bandwidth fm was used by the military at all sorts of carrier frequencies, and not for hi-fi. I have a 1959 army tranciever set using a single output tube, at 28 Mhz, for local army communications. Beautifully made inside, very heavy, and totally obsolete by today's standards. Quite a few sets with 3 stage IF amps will go into limiting, ie, the last two IF tubes are being driven into grid current, and well into overload, so any input signal amplitude variations do not get through to cause distortions to the audio detected bu the discriminator, which is designed to respond to FM, not AM. The bottom line seems to be if these factors aren't taken into consideration, then the suppression of multipath, and co channel interference, will suffer, and the full advantage of FM will not be realized. This just my possibly wrong, foggy understanding of what some of the papers were trying to say. Maybe you are right, and not everyone can get good FM reception, or TV. I have line of sight to a 10,000 watt transmitter on top of a big tower on a large hill I see clearly about 7 Km away, so I get excellent TV and FM signals. Ppl living in shadows of hills here get poor reception, and then some areas suffer multipath, which interferes with the waves phase, causing unstable on-off interferences. I am not sure that such things can be all fixed by the quality of the receiver; a lot depends on the transmitted signal. Some suburbs have auxilliary transmitters at a different frequency, placed elsewhere locally, to overcome reception problems in shadow areas. Patrick Turner. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#28
|
|||
|
|||
John Byrns wrote: Come on, today it only costs pennies, we do it all the time for FM stereo, although in that case the carrier isn't transmitted, so your gut doesn't have to worry. C-Quam AM stereo chips do it too, no big deal. Well, I suppose it is not trivial doing it with tubes, and tube type FM stereo used filters and doublers, rather than PLLs, but it would be doable. Then how do they get $$$ for these? http://www.sherweng.com/indepth.html what's different here than say the C-Quam chips - or perhaps that is an answer - before these chips $$$$ - if somebody figures how to adapt these chips to accomplish similar performance???? If I could get that kind of performance out of a cheap chip... hmmmmm....! best regards.. -- randy guttery A Tender Tale - a page dedicated to those Ships and Crews so vital to the United States Silent Service: http://tendertale.com |
#29
|
|||
|
|||
John Byrns wrote: In article , wrote: John Byrns wrote: A synchronous detector would eliminate distortion due to unbalanced sidebands, if indeed 3rd order distortion products can cause this. Yeah but a really good synch detector can set you back major bucks - Are they really worth it? I guess 30hz - 15Khz +/- 3db on AM is pretty good (Sherwood Eng.'s is rated that) - but I wonder about distortion - there is just something in my gut that says when you sub a local carrier for the original - something isn't going to be as clean as original. But I'll admit - I've always had a time understanding full pll synch systems... --- I'm saving my pennies... but at $550+ --- it isn't a high priority. Come on, today it only costs pennies, we do it all the time for FM stereo, although in that case the carrier isn't transmitted, so your gut doesn't have to worry. C-Quam AM stereo chips do it too, no big deal. Well, I suppose it is not trivial doing it with tubes, and tube type FM stereo used filters and doublers, rather than PLLs, but it would be doable. One could use a locked oscillator instead of PLL. The early synchrodynes relied on this instead of PLL. Patrick Turner. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#30
|
|||
|
|||
Randy and/or Sherry wrote: Then how do they get $$$ for these? http://www.sherweng.com/indepth.html what's different here than say the C-Quam chips - or perhaps that is an answer - before these chips $$$$ - if somebody figures how to adapt these chips to accomplish similar performance???? If I could get that kind of performance out of a cheap chip... hmmmmm....! best regards.. Audiophiles aren't the only one's to suffer from phoolishness. Besides, it's like the early PC's way OVER priced, because they could. Don't get me wrong, I'm sure the Sherwood stuff works every bit as well as advertized. So does Motorola. But even though the radios now use a computer interface to add and subtract channels, the pricing is still based on cost of channel elements. Jeff -- "They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety." Benjamin Franklin "A life lived in fear is a life half lived." Tara Morice as Fran, from the movie "Strictly Ballroom" |
#31
|
|||
|
|||
(John Byrns) wrote in message ...
In article , (firedome) wrote: (John Byrns) wrote in message ... In article , (firedome) wrote: Who considered the S-R tuner the best commercial tuner ever made?... Audio Engineering (July 1953)..."there has never before been available a radio receiver which provided AM quality which approached FM" ...and Consumer Research Bulletin (March 1953)..."the first satisfactory detector circuit yet observed for high fidelity use in an AM tuner". Those are certainly some fantastic superlatives! As for the circuit's provenance, I never said S-R invented it... And for my part I never said you said S-R invented it. it was, as I said, an adaptation of the Selden -Smith 2 tube decector circuit (did any one else use 2 tube detectors?) found in the RDH3. I missed the part where you said that. The Ampex AM/FM simulcast stereo tuner also used a similar "two tube" detector circuit. I forget the actual Ampex model number of this tuner, but many people refer to it as the "008" tuner, but that is not the actual model number. I mentioned the circuit in RDH3, not the inventors names... I'm going on what Will Rayment told me when I interviewed him for VTV, and these are his recollections of what happened 50 yrs ago. Yes, I think the problem is that the circuit was invented more than 60 years ago, and nobody remembers the details. I have not been able to find the circuit in the RDH3, and suspect that Will Rayment is misremembering this. I suspect that Will is actually probably thinking of the RDH4, which does show the circuit in an appendix of some printings. It would also be interesting to know how S-R found out about the circuit, as I don't think Water Selsted knows, although he did give S-R permission to use the circuit when they asked him. I have to suspect that the circuit probably became known by word of mouth in the California technical community during the 1940s. Ok, went back and looked at my interview notes...in late '51 before Will's father passed away Walter Selsted had apparently developed some kind of tweeter capable of super high frequencies (I'm not aware of any details) in addition to the circuit found in RDH3 (Will may have meant 4, maybe, but I wrote down 3 as that's what he recalled) that was supposed to have a low distortion of around .35%. Selsted contacted L.C. Rayment about SR using the circuit as they had a long reputation for high quality products back to 1927, and arrangement was made for S-R to use it. Two permeability tuned IF 6BA6 stages terminated into the low distortion detector (1N48 & 12AU7). That's all the detail I have on it...it was used in the SR 51 and 68 and 808 AM/FMs and also on the SR-58 AM only, and later, in the late 50s, in the SR1000, which had provision for mpx and modern styling. Also in the Cantilever Control Brentwood tuner (SR100) of the mid 50s... according to my notes, McIntosh was interested in the coils specifically because of their superior specs, and attempted to have J-R winding of LA duplicate them, unsuccessfully. They also copied the SR variable rumble filter that was designed specifically to deal with the rumble from Garrard turntables. All of this info is direct from Will Rayment, but as you said, filtered thru 60 yrs of time, and who knows what all? Anyhow, a bit of contribution to the history of those days. best, Roger in NY McIntosh obviously thought highly of the S-R IF coils as they pursued it with diligence. I myself am not sure of the advantage of honeycomb construction, but McIntosh apparently thought there was one beyond marketing hype. I'll ask Sid Corderman if he knows when I interview him. That would be an interesting question to ask, why do you say "McIntosh apparently thought there was one beyond marketing hype", I would think the opposite? As for SR-51s, there was one on eBay a few months ago, and I believe the SR-68 and later SR-1000 had essentially the same circuit. Was the SR-51 the AM only version? I know they made a bewildering array of versions, AM only, AM/FM, and with and without built in preamps. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#32
|
|||
|
|||
I pulled Sturley down off the bookshelf to check for more detail from the reference given in the RDH4 for "Modulation Envelope Distortion". So far I have only quickly scanned the first two of six pages Sturley devotes to the subject, but it is clear what you say about "Modulation Envelope Distortion" is correct, and second harmonic envelope components are generated that would affect a synchronous detector as well as the simple envelope detector. Sturley of course backs up the notion that the 2nd order distortion products don't make it through the IFT, but the thing I didn't realize was that third order distortion in the IF amplifier not only produces sidebands around the carrier at three times the modulating frequency, but also at twice the modulating frequency, and at the modulating frequency. Sturley doesn't show the derivation of this, but I now realize that it should have been immediately obvious to me simply from the basics of third order distortion, without even going through the equations. Getting back to your original claims that a push pull IF amplifier would reduce this distortion, and that rectifying both "-ve & +ve" sides of the wave form in the detector would cancel second harmonic distortion, I can't see how either of these would really help matters a great deal. The push pull IF would only cancel the 2nd order distortions, which are already not a problem due to the filtering action of the IFT, the only distortion reducing benefit of push pull operation would seem to be that with two tubes sharing the load, each could operate at half the power otherwise required, with a consequent reduction in distortion, but this has nothing to do with the cancellation of 2nd order distortion. Similarly, rectifying both the "-ve" and "+ve" sides of the IF wave form, and taking the difference, won't help because the damage has already been done before the IFT feeding the detector stage, the "-ve" and "+ve" sides of the IF wave form are mirror images of each other at the output of the IFT, and the detector, even a synchronous detector, can't put it right again. So it seems we can't depend on much help from push pull operation of the IF amplifier stage, or the detector, and must simply design the IF amplifier tube, and supporting circuit, to be as free from 3rd, and higher, order distortions as possible. Regards, John Byrns In article , Patrick Turner wrote: John Byrns wrote: In article , Patrick Turner wrote: Indeed, if the top half of the IF envelope is amplified more than the bottom half, then the waveform is as if somebody added an AF component to the 455 kHz wave form, and this is shunted by the low impedance of the IFTs at AF. But the IMDs remain, and if the envelope is subject to limiting or clipping then this will distort the detected AF. "2H" distortion is just a special case of 2nd order "IMD". In looking up your reference to the RDH4th, I found that the page you referenced doesn't really say much, but it does refer in turn to pages 944 and 945, which briefly mentions that the transfer characteristic of the vari-mu tube can be modeled by a power series, and that the even order products don't make it through the IFT. It also says vari-mu tubes are especially designed to minimize 3rd order distortions, which do pass through the IFT to the detector. It says enough to confirm that IF amps are one of the worst sources of thd of a radio, because it is a voltage amplifier producing considerable voltages. There is no nfb involved, and the IF tube operates at a region where introduced thd should be a worry for those who like real fidelity. If the peaks of an AM envelope are compressed on top and bottom, then that's odd order distortion of the whole wave form, but when one side of the compressed IF envelope is detected, the AF thd is mainly even order! That sounds like a problem in the detector, not the IF amplifier. No, it isn't. I have just outlined that the IF amp is capable of compessing the waveforms, ie, adding distortions, and these deformations of the AF modulation shapes are detected, and thus the IF tube has added muck to the AF signal. The ideal IF amp should be as linear as one can make it, period!, if one wants maximum fidelity. And a balance has to be struck to keep its voltage output lowish, but not too low to let noise start to rasie its ugly head. Surf my web pages at, http://users.rcn.com/jbyrns/ |
#33
|
|||
|
|||
In article ,
(firedome) wrote: Ok, went back and looked at my interview notes...in late '51 before Will's father passed away Walter Selsted had apparently developed some kind of tweeter capable of super high frequencies (I'm not aware of any details) That was another Selsted & Smith collaboration, they published an article on it in the January 1950 issue of "Audio Engineering" magazine. in addition to the circuit found in RDH3 (Will may have meant 4, maybe, but I wrote down 3 as that's what he recalled) that was supposed to have a low distortion of around .35%. I can't find it in the RDH3, unfortunately it appears that they never published an article on the "low distortion AM detector", the only thing I have ever seen published about it is in the RDH4, and that doesn't delve into its operation. Selsted contacted L.C. Rayment about SR using the circuit as they had a long reputation for high quality products back to 1927, and arrangement was made for S-R to use it. If that is true, it would explain how S-R came to learn of the circuit, I had come to the probably mistaken conclusion that it was the other way around and S-R had contacted Walter about using it. Did you interview Water Selsted about the detector? Two permeability tuned IF 6BA6 stages terminated into the low distortion detector (1N48 & 12AU7). Interesting that S-R used a Germanium diode in some models, the SR-68 used the two halves of a 6SN7, one half connected as a diode, and the other half as a triode. That's all the detail I have on it...it was used in the SR 51 and 68 and 808 AM/FMs and also on the SR-58 AM only, and later, in the late 50s, in the SR1000, which had provision for mpx and modern styling. Also in the Cantilever Control Brentwood tuner (SR100) of the mid 50s... I had what must have been the matching "Cantilever Control" amplifier, used a pair of EL34s in the output, and had a "scratch" filter with a variable cutoff point controlled by a two or three gang variable capacitor like you would find in a radio, I think it may also have had a matching rumble filter, whose cutoff frequency was controlled by a ganged pot. according to my notes, McIntosh was interested in the coils specifically because of their superior specs, and attempted to have J-R winding of LA duplicate them, unsuccessfully. I wonder what characteristic it was about the S-R IF transformers, that McIntosh was after? They also copied the SR variable rumble filter that was designed specifically to deal with the rumble from Garrard turntables. Which McIntosh amplifier used the variable rumble filter, I don't remember one, but then I am not that familiar with old McIntosh equipment? All of this info is direct from Will Rayment, but as you said, filtered thru 60 yrs of time, and who knows what all? Anyhow, a bit of contribution to the history of those days. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#34
|
|||
|
|||
John Byrns wrote: I pulled Sturley down off the bookshelf to check for more detail from the reference given in the RDH4 for "Modulation Envelope Distortion". So far I have only quickly scanned the first two of six pages Sturley devotes to the subject, but it is clear what you say about "Modulation Envelope Distortion" is correct, and second harmonic envelope components are generated that would affect a synchronous detector as well as the simple envelope detector. Sturley of course backs up the notion that the 2nd order distortion products don't make it through the IFT, but the thing I didn't realize was that third order distortion in the IF amplifier not only produces sidebands around the carrier at three times the modulating frequency, but also at twice the modulating frequency, and at the modulating frequency. Sturley doesn't show the derivation of this, but I now realize that it should have been immediately obvious to me simply from the basics of third order distortion, without even going through the equations. OK, say you have an IF of 455 kHz, and a AF tone of 1 kHz for the modulation. The carrier of 455 kHz is present, and sidebands of 444 kHz, and 456 kHz. Then if the envelope amplifier slightly compresses the envlope peaks, as would be the case with added flattening 3rd harmonic distortion, then the envelope then contains all the above fequencies, plus additional sidebands at 452 kHz and 458 kHz, and this extra muck makes it through the IF tranny BW, and its all detected loud and clear and can be measured as distortion to the AF waves from the detector. Getting back to your original claims that a push pull IF amplifier would reduce this distortion, and that rectifying both "-ve & +ve" sides of the wave form in the detector would cancel second harmonic distortion, I can't see how either of these would really help matters a great deal. I agree. But rectifying two phases with two tubes increases the dynamic range of the detector by 6 dB. The detector needs to be in reciept of signals from a low distortion voltage amp if we want the receiver to be low distortion, as a whole. If we want the best AM that tubes can provide, then a balanced IF amp makes sense, and this could produce up to maybe 100 vpeak to peak envelope voltage, so one has to think carefully about the following detector voltage capability. Once such an enormous dynamic range is established, the thd should be low at a few volts of output. The push pull IF would only cancel the 2nd order distortions, which are already not a problem due to the filtering action of the IFT, the only distortion reducing benefit of push pull operation would seem to be that with two tubes sharing the load, each could operate at half the power otherwise required, with a consequent reduction in distortion, but this has nothing to do with the cancellation of 2nd order distortion. But if there is 2H distortion in a tube amplifying an enevelope, does this not create sidebands at 453 kHz, and 457 kHz, in the above case, and would they not proceed throught the IFT, and get detected? I haven't measured this phenomena for awhile, so I just don't know. Similarly, rectifying both the "-ve" and "+ve" sides of the IF wave form, and taking the difference, won't help because the damage has already been done before the IFT feeding the detector stage, the "-ve" and "+ve" sides of the IF wave form are mirror images of each other at the output of the IFT, and the detector, even a synchronous detector, can't put it right again. Exactly, but the PP detector has 6 dB more dynamic headroom, so shouls operate with less thd. In practice a 12AT7 cathode follower with both halves paralleled operating with an Ea of 200v, and 6 mA of current should cope with the unbalanced secondary output from an IFT driven by a balanced PP IF amp. So it seems we can't depend on much help from push pull operation of the IF amplifier stage, or the detector, and must simply design the IF amplifier tube, and supporting circuit, to be as free from 3rd, and higher, order distortions as possible. A PP IF amp will achieve such an outcome, easily, because of the increased dynamic range. The PP thd is all mainly 3H, but it usually isn't any more than twice what it is in a SE amp using the same tube, and pentodes when carefully set up don't have enormous 3H at moderate voltage levels, although they quickly start to show an awful range of distortion products once past about 10v of output. In the radio I have built, I use a 6BX6, with about 6 dB current FB, from an unbypassed Rk, and this amp is adequately linear for me, but a balanced amp doing the same thing probably would have 1/4 the thd at the same voltage output. I control the voltage output with AVC applied only to the twin vari-mu triode RF amp ahead of the mixer. There is very little voltage produced by the triode input stage, so its linearity wouldn't benefit much with a balanced circuit, and the mixer also works with fixed bias, a high oscillator current change, but small RF input signal, so it is linear. With short wave, allowing the final IF amp to have a higher dynamic range should better the SNR, But then the approach is different, with an extra IF stage needed, maybe double conversion, mechanical or crystal filters, depending how far we want to go, but hi-fi can't be our main aim with SW reception. I will build an extra IF tube into my set one day, and measure before and after. I envisage a DC feed to the tubes using an RF choke with CT, then a normal IFT could be used, a sort of parafeed PP IF amp. Patrick Turner. Regards, John Byrns In article , Patrick Turner wrote: John Byrns wrote: In article , Patrick Turner wrote: Indeed, if the top half of the IF envelope is amplified more than the bottom half, then the waveform is as if somebody added an AF component to the 455 kHz wave form, and this is shunted by the low impedance of the IFTs at AF. But the IMDs remain, and if the envelope is subject to limiting or clipping then this will distort the detected AF. "2H" distortion is just a special case of 2nd order "IMD". In looking up your reference to the RDH4th, I found that the page you referenced doesn't really say much, but it does refer in turn to pages 944 and 945, which briefly mentions that the transfer characteristic of the vari-mu tube can be modeled by a power series, and that the even order products don't make it through the IFT. It also says vari-mu tubes are especially designed to minimize 3rd order distortions, which do pass through the IFT to the detector. It says enough to confirm that IF amps are one of the worst sources of thd of a radio, because it is a voltage amplifier producing considerable voltages. There is no nfb involved, and the IF tube operates at a region where introduced thd should be a worry for those who like real fidelity. If the peaks of an AM envelope are compressed on top and bottom, then that's odd order distortion of the whole wave form, but when one side of the compressed IF envelope is detected, the AF thd is mainly even order! That sounds like a problem in the detector, not the IF amplifier. No, it isn't. I have just outlined that the IF amp is capable of compessing the waveforms, ie, adding distortions, and these deformations of the AF modulation shapes are detected, and thus the IF tube has added muck to the AF signal. The ideal IF amp should be as linear as one can make it, period!, if one wants maximum fidelity. And a balance has to be struck to keep its voltage output lowish, but not too low to let noise start to rasie its ugly head. Surf my web pages at, http://users.rcn.com/jbyrns/ |
#35
|
|||
|
|||
I pulled Sturley down off the bookshelf to check for more detail from the
reference given in the RDH4 for "Modulation Envelope Distortion". So far When you get done reading, John, I'd be interested to know some typical distortion figures. I'll bet they're signficant only for very strong signals. I know that some receivers attenuate the AVC to the last IF stage to minimize envelope distortion since that's where the highest signal levels occur. Still, my guess is that this source of distortion pales in comparison to that due to low modulation acceptance. I'm restoring a Hallicrafters S-76 boatanchor. Stock, it was the worst receiver I've ever encountered for modulation acceptance, clipping on the downward side at 53% modulation or more. Brian |
#36
|
|||
|
|||
In article , Patrick Turner
wrote: John Byrns wrote: I pulled Sturley down off the bookshelf to check for more detail from the reference given in the RDH4 for "Modulation Envelope Distortion". So far I have only quickly scanned the first two of six pages Sturley devotes to the subject, but it is clear what you say about "Modulation Envelope Distortion" is correct, and second harmonic envelope components are generated that would affect a synchronous detector as well as the simple envelope detector. Sturley of course backs up the notion that the 2nd order distortion products don't make it through the IFT, but the thing I didn't realize was that third order distortion in the IF amplifier not only produces sidebands around the carrier at three times the modulating frequency, but also at twice the modulating frequency, and at the modulating frequency. Sturley doesn't show the derivation of this, but I now realize that it should have been immediately obvious to me simply from the basics of third order distortion, without even going through the equations. OK, say you have an IF of 455 kHz, and a AF tone of 1 kHz for the modulation. The carrier of 455 kHz is present, and sidebands of 444 kHz, and 456 kHz. Then if the envelope amplifier slightly compresses the envlope peaks, as would be the case with added flattening 3rd harmonic distortion, then the envelope then contains all the above fequencies, plus additional sidebands at 452 kHz and 458 kHz, and this extra muck makes it through the IF tranny BW, and its all detected loud and clear and can be measured as distortion to the AF waves from the detector. The thing that I missed, that Sturley points out, and I would have eventually seen once I got around to the math, but which should have been obvious to me with only a moments thought, is that 3rd order non linearity in the IF amplifier not only produces sidebands at 452 kHz, 454 kHz, 456 kHz, and 458 kHz as you mention, but also at 453 kHz and 457 kHz, creating the "2H audio distortion you have observed. This is why push pull IF amplifiers can't eliminate this "2H" audio distortion, because it is the result of a third order effect in the IF amplifier, which isn't canceled by the push pull action that cancels even order non linearity. To put it in other words, the 3rd order distortions in the IF amplifier symmetrically squash the IF wave form, but the result of that is that the envelope is squashed on only the positive modulation peaks, creating a second order distortion in the recovered audio. The push pull IF would only cancel the 2nd order distortions, which are already not a problem due to the filtering action of the IFT, the only distortion reducing benefit of push pull operation would seem to be that with two tubes sharing the load, each could operate at half the power otherwise required, with a consequent reduction in distortion, but this has nothing to do with the cancellation of 2nd order distortion. But if there is 2H distortion in a tube amplifying an enevelope, does this not create sidebands at 453 kHz, and 457 kHz, in the above case, and would they not proceed throught the IFT, and get detected? I haven't measured this phenomena for awhile, so I just don't know. No, the second order distortion in the tube creates sidebands around 910 kHz, or twice the IF frequency, at 908 kHz, and 912 kHz, and these get blocked by the IFT, just as the DC and audio frequency components produced by 2nd order distortion. This has all been very useful, because I now realize that to minimize audio distortion you want to avoid aperiodic couplings between the last IF amplifier stage, or RF amplifier stage in a TRF receiver, and the detector. This is a rule that is violated by some well known TRF receivers, including the Western Electric No. 10A, the Weeden, and the J.W. Miller 570. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#37
|
|||
|
|||
donut wrote: OK - the IF distortion was one reason that the TRF was favored for really true hi-fidelity AM. How many stages of RF amplification would a TRF need to equal the sensitivity of an AA5? I know there's a problem with selectivity but the more tuned circuits, the better that would get. Would 5 tank circuits and 5 RF amplifier stages, all operating near unity gain (except for the first one) be sufficient? Would each RF stage add some distortion of it's own? I ask this question because I can get a 5 gang variable cap very inexpensively. There may be more knowledgable folk who will differ, but it seems to me that it's immaterial whether you amplify the RF directly, or after conversion to a fixed intermediate frequency. The only real difference is that in a superhet, tracking difficulties are greatly reduced because you're dealing with only one frequency. In principle, at least, a TRF with the same number of amplifier stages and the same number of tuned circuits with the same Q will perform as well as a superhet. But there's that tracking problem again... What's more, as John Byrn pointed out in an earlier post, an IF transformer's coupling can be tailored for a reasonably square-shouldered response curve. However, this only works properly at a single frequency, so your TRF's response will vary across the band. That may not be an issue if you only plan to listen to a single station, in which case you could, in principle, tailor your tuned circuits exactly for that frequency. You'd essentially have a superhet without the "het". ;-) (I.e. just the RF strip, detector, and amplifier, with no heterodyne conversion necessary.) Five tuned stages would be an overkill, IMO, even at low gain. Don't forget that high-Q tuned circuits have an apparent voltage gain of their own, so even low-mu triodes can have gains as TRF amplifiers many times the published amplification factor (mu). You have to be careful to shield each stage, or you'll end up with a string of TPTG (tuned-plate, tuned-grid) oscillators. And yes, each stage can introduce its own distortion, whether it be RF, IF, or audio. John Byrn's and Patrick Turner's posts on the subject speak volumes. Cheers, Fred -- +--------------------------------------------+ | Music: http://www3.telus.net/dogstarmusic/ | | Projects: http://dogstar.dantimax.dk | +--------------------------------------------+ |
#38
|
|||
|
|||
"Fred Nachbaur" wrote in message . .. There may be more knowledgable folk who will differ, but it seems to me that it's immaterial whether you amplify the RF directly, or after conversion to a fixed intermediate frequency. The only real difference is that in a superhet, tracking difficulties are greatly reduced because you're dealing with only one frequency. The biggest problem with TRF isn't the varying bandwidth, or even the tracking.. but the tendency for feedback. The more stages you have at the same frequency, the more likely it is that you will have feedback problems. This is where the characteristic squeal and swishing sounds you hear associated with very old radios come from. TRF's would break into oscillation on a whim, unless each stage was very well shielded from all the others. IIRC, this is how regenerative receivers were discovered/invented. |
#39
|
|||
|
|||
Brenda Ann wrote: "Fred Nachbaur" wrote in message . .. There may be more knowledgable folk who will differ, but it seems to me that it's immaterial whether you amplify the RF directly, or after conversion to a fixed intermediate frequency. The only real difference is that in a superhet, tracking difficulties are greatly reduced because you're dealing with only one frequency. The biggest problem with TRF isn't the varying bandwidth, or even the tracking.. but the tendency for feedback. The more stages you have at the same frequency, the more likely it is that you will have feedback problems. This is where the characteristic squeal and swishing sounds you hear associated with very old radios come from. TRF's would break into oscillation on a whim, unless each stage was very well shielded from all the others. IIRC, this is how regenerative receivers were discovered/invented. Yes, instability is a problem with TRF, and if you had 5 tuned circuits, with 4 tubes between them, then at least you wouldn't need high transconductance tubes. But an IF amp in any radio one might build can be instable, especially a two stage type. The answer is very short connecting wires, and well thought out layout, and screening. Quite a few communication receivers had cast and machined aluminium chassis, with "pockets" for each RF stage, none of this cheap pressed metal garbage seen on consumer grade radios. The 22 tube Racal is such an example. The DIYer can achieve excellent results, but it all takes time, time, and more time. If you had a single LC Q of 100 at 1,000 kHz, then the BW is 10 kc, so with 4 more other circuits after this, the effective Q would be higher, and BW lot lower, so expect only 3kHz RF BW, so 1.5 kHz AF bandwidth. The skirt steepness away from resonance will be 5th order, or 30 dB/octave, which would just give enough rejection of a strong station at 50 kHz above or below the wanted station at 1,000 kHz. The receiver will be selective, but the sound won't be hi-fi. For hi-fi, we want the characteristic of a critically coupled IF transformer, which only has to deal with one F, and the magnetic coupling which is determined by the distance between the coils will determine the shape of the bandpass filter. A single tuned LC gives a pointy to to the nose shape, while the transformer gives either a really sharp nosed response, or flat or double peaked, when the coils are close. The tuned transformer still gives skirt roll off shapes at an ultimate second order slope. I suggest ppl who have doubt about this should go back to the old books, and do some real study, to save us from having to regurgitate what is in those textbooks. So off your bums, away from the PC and start reading!!!! Not in too many books is the idea of using about 2.4 Mhz as the IF frequency. If the Q of the LC was 100, then BW with a tuned circuit could be 24 kHz, so with a lot of tuned circuits, you'd still get a fair RF BW. But IFT with 2.4 Mhz and with spaced coils for a critically coupled flat response, the BW would be quite wide, and 3 such IFTs could be used to make a BW of at least 20 kHz, giving 10 kHz AF bandwidth, and the six tuned IF LC units would give 6th order ultimate slope of the skirt, so selectivity for local stations would be fine. The use of a PP IF amp has been criticised roundly by the proponents of the status quo, for all sorts of reasons, the main one being that some of the distortion don't matter, its obstructed/cancelled by an IFT. I think many of you are all missing my point, which is that AM should be as hi-fi as possible, as much of what is transmitted is now hi-fi at the transmitter, ie, low thd, but slightly restricted in AF BW, but very listenable indeed, if ONLY we had a decent reciever that didn't dump 7% of thd into the signal, and further restrict the BW to less than two bleeding kHz!!!! This is typical of many old radios. It is NORMAL for most solid state AM reciever sections that have been foistered onto the public for the last 40 years. FM is king, and AM is the poor relation. To address this problem I suggest a PP IF amp isn't as crazy /wacky as you think. Its simply an application of the well known fact that a PP amp using two tubes produces a far cleaner result than one SE tube, for the amount of power produced by one tube on its own. The dynamic range is potentially increased 6dB, or doubled, but what I suggest is worth having is the simple absence of distortion. In my home designed superhet, I have a sharp cut off pentode, with 6 dB of current FB, with cathode bias, which stays fixed, and no AVC applied to the IF amp, and I hear the difference between this and most other radios, which are just historical junk by comparison, ie, bean counter designed consumer electronics, lowest common denominator stuff. I do like them as works of art, but good performers they are not. For example, I have had a 1939? STC MW/SW receiver here for a week to repair. I did the usual, replace ALL the caps; even the paper ones rated at 1,500 volts were leaky. Modern poly caps made after 1960 are far better. The tubes all tesed fine, and I finally got all the switches to work, and then I aligned and tuned it, and sure, its a great perfomer, for its time, with the peaking in the speaker AF response at between 1 and 4 kHz partially compensating for the 2 kHz of normal AF response I measured at the detector output. Not all tube superhets have very wide AF bandwidth. But with such a small RF BW, it is not too bad a performer on SW, for a set without an RF stage. With a long wire antenna, it even picks up amateurs OK, except that only a few of them use AM, and most use SSB, and there is no BFO, and the tuning rate is way to fast. I have a Radiola from 1957, with an RF stage, and it was no better, and it to gave that "melow sound" resonating woodwork, combined with no treble, or treble simply replaced by distortion products, when turned up a bit. By means of strapping 100k resistors across each LC circuit in the IF, I was able to lower the Q, and get about 7 kHz max, but its only good for local stations. I revised the AVC to give the IF tube a bit of a break..... I added the CF detector circuit, and with a PP audio amp, with a better circuit than Radiola ever had, it now sounds like a radio should, crisp treble, creamy, rich bass, and a pleasure to listen to. It has about 11 tubes. I have another of these chassis, (which originally were bought by only wealthy folks,) and I eventually will upgrade it to being a lot better for SW, if I can only work out how to align the tracking and coil gain since there are six SW bands. I won't bother with a PP IF, because the signals with SW are usually small, and N&D from atmospheric conditions make it useless to get a better sound. I doubt it is practical to make a simple AM radio which can wear all types of hats, ie, be great at both DX and local MW stations, and also be good at SW. So In the fullness of time, I will have a seperate set for local MW, and another for SW, and I don't care about MW DX, because there is so much programme crap, and then there is the noise and fade outs. Patrick Turner. |
#40
|
|||
|
|||
"Brenda Ann" wrote in message ...
"Fred Nachbaur" wrote in message . .. There may be more knowledgable folk who will differ, but it seems to me that it's immaterial whether you amplify the RF directly, or after conversion to a fixed intermediate frequency. The only real difference is that in a superhet, tracking difficulties are greatly reduced because you're dealing with only one frequency. The biggest problem with TRF isn't the varying bandwidth, or even the tracking.. but the tendency for feedback. The more stages you have at the same frequency, the more likely it is that you will have feedback problems. This is where the characteristic squeal and swishing sounds you hear associated with very old radios come from. TRF's would break into oscillation on a whim, unless each stage was very well shielded from all the others. IIRC, this is how regenerative receivers were discovered/invented. Hey Brenda Ann,Nice to see you hangin around here. :-) |
Reply |
|
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
on topic: we need a rec.audio.pro.ot newsgroup! | Pro Audio | |||
Some Recording Techniques | Pro Audio | |||
common mode rejection vs. crosstalk | Pro Audio |