Home |
Search |
Today's Posts |
#1
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
The 845 amps I have been strugling to find time to build
are approaching a stage where I must decide on the actual working points that would be most suitable for both 211 and 845 including class A2 operation, by means of cathode follower buffer choke loaded to the fixed bias supply. So far having Ea at 1,000V and Ea at 80mA with RL = 12.5k could seem OK but it means Pda = 80 watts, and only KR845 or KR211 look able to take this idling condition, one reason being that the KR heaters are 10V at 1A instead of the RCA GE use of 10V at 3.3A. RCA data gives Pda = 67 watts, and if Pd-filament = 33 watts, the total tube Pd = 100 watts Anyone here with actual experience of using either tube in an SE amp??? ALSO, anyone here have longterm experience of use of Shuguang 845? Patrick Turner. |
#2
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
On Aug 8, 7:32 am, Patrick Turner wrote:
The 845 amps I have been strugling to find time to build are approaching a stage where I must decide on the actual working points that would be most suitable for both 211 and 845 including class A2 operation, by means of cathode follower buffer choke loaded to the fixed bias supply. So far having Ea at 1,000V and Ea at 80mA with RL = 12.5k could seem OK but it means Pda = 80 watts, and only KR845 or KR211 look able to take this idling condition, one reason being that the KR heaters are 10V at 1A instead of the RCA GE use of 10V at 3.3A. RCA data gives Pda = 67 watts, and if Pd-filament = 33 watts, the total tube Pd = 100 watts Anyone here with actual experience of using either tube in an SE amp??? ALSO, anyone here have longterm experience of use of Shuguang 845? Patrick Turner. Yoo-wee Patrick Thats gonna be one hot momma, should be plenty a power. What you gonna drive with that? Have fun! Willie |
#3
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
http://www.pentalaboratories.com/pdfs/845.pdf
I wonder how the 750V operating point sounds? It almost looks, um, practical! "Patrick Turner" wrote in message ... The 845 amps I have been strugling to find time to build are approaching a stage where I must decide on the actual working points that would be most suitable for both 211 and 845 including class A2 operation, by means of cathode follower buffer choke loaded to the fixed bias supply. So far having Ea at 1,000V and Ea at 80mA with RL = 12.5k could seem OK but it means Pda = 80 watts, and only KR845 or KR211 look able to take this idling condition, one reason being that the KR heaters are 10V at 1A instead of the RCA GE use of 10V at 3.3A. RCA data gives Pda = 67 watts, and if Pd-filament = 33 watts, the total tube Pd = 100 watts Anyone here with actual experience of using either tube in an SE amp??? ALSO, anyone here have longterm experience of use of Shuguang 845? Patrick Turner. |
#4
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
drummerwill wrote: On Aug 8, 7:32 am, Patrick Turner wrote: The 845 amps I have been strugling to find time to build are approaching a stage where I must decide on the actual working points that would be most suitable for both 211 and 845 including class A2 operation, by means of cathode follower buffer choke loaded to the fixed bias supply. So far having Ea at 1,000V and Ea at 80mA with RL = 12.5k could seem OK but it means Pda = 80 watts, and only KR845 or KR211 look able to take this idling condition, one reason being that the KR heaters are 10V at 1A instead of the RCA GE use of 10V at 3.3A. RCA data gives Pda = 67 watts, and if Pd-filament = 33 watts, the total tube Pd = 100 watts Anyone here with actual experience of using either tube in an SE amp??? ALSO, anyone here have longterm experience of use of Shuguang 845? Patrick Turner. Yoo-wee Patrick Thats gonna be one hot momma, should be plenty a power. What you gonna drive with that? Have fun! Willie Having given it some more thought, and of course while realizing the "hot momma" factor, specs for "normal 211 say 67 watts of Pda, which means op point of say 1,000V x 67 mA, and that means a higher anode load of 14k for class A2. For just class A, and no grid current, the Ea must be shifted up to about 1,250V and the swing is a little limited by the Ra curve where Eg=0V, so the load can actually be lower at 12, and you get about 21 watts with 211, and with 58mA. Trying to screw every bit out of such tubes required careful setting up. I am using TWO tubes in parallel, so the A2 operation could get me 60 watts with over 1/2 that in normal class A, so its only ever going to be the extreme levels on transients which will force the driving amp to work into some grid current. But the A2 means a direct coupled CF for each output tube and some clever biasing method I have yet to fully finalise. Fixed bias isn't too nice because if the mains rises 5%, you get an extra 10% of Pda, 73 watts instead of 67 watts, not good. With CCS cathode supply, one gets 5% increase in Pda with 5% rise in Ea if the mains rises 5%. The amps are to be used with normal modern speakers of about 88dB/W/M, so that the performance is as good as with a sixpack of KT88 or EL34 set up for pure class A. I'm sorely tempted to convert to using 2 x 13E1 in PP, because I have an OPT with 6k : 6 ohms.... Patrick Turner |
#5
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
maxhifi wrote: http://www.pentalaboratories.com/pdfs/845.pdf I wonder how the 750V operating point sounds? It almost looks, um, practical! Yes, 750V at about 89mA is OK but the 12k to 14k load is way too high. I defy anyone to tell the difference in sound quality between 1,000V x 67mA and 750V x 89mA at ordinary levels providing the load is right for the tube. The higher Ea and lower Ia give less THD/IMD, but for the first two watts with a pair of these tubes in parallel, there is only a tiny "academic" difference in distortions. Rout is slightly lower at the higher idling Ia because the load line is higher on the Ra curves where they are less sloped than at the lower Ia. So any differences heard would be hard to pin down onto technical operational set up. A fair old difference of op points can be used, like with trioded KT88 which can be used with Ea at 350V to 500V at least. I figure the 67 watt Pda limit should not provoke un-reliablity with Chinese Shuguang 845/211 tubes. Patrick Turner. "Patrick Turner" wrote in message ... The 845 amps I have been strugling to find time to build are approaching a stage where I must decide on the actual working points that would be most suitable for both 211 and 845 including class A2 operation, by means of cathode follower buffer choke loaded to the fixed bias supply. So far having Ea at 1,000V and Ea at 80mA with RL = 12.5k could seem OK but it means Pda = 80 watts, and only KR845 or KR211 look able to take this idling condition, one reason being that the KR heaters are 10V at 1A instead of the RCA GE use of 10V at 3.3A. RCA data gives Pda = 67 watts, and if Pd-filament = 33 watts, the total tube Pd = 100 watts Anyone here with actual experience of using either tube in an SE amp??? ALSO, anyone here have longterm experience of use of Shuguang 845? Patrick Turner. |
#6
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
"Patrick Turner" wrote in message ... The 845 amps I have been strugling to find time to build are approaching a stage where I must decide on the actual working points that would be most suitable for both 211 and 845 including class A2 operation, by means of cathode follower buffer choke loaded to the fixed bias supply. So far having Ea at 1,000V and Ea at 80mA with RL = 12.5k could seem OK but it means Pda = 80 watts, and only KR845 or KR211 look able to take this idling condition, one reason being that the KR heaters are 10V at 1A instead of the RCA GE use of 10V at 3.3A. RCA data gives Pda = 67 watts, and if Pd-filament = 33 watts, the total tube Pd = 100 watts Anyone here with actual experience of using either tube in an SE amp??? ALSO, anyone here have longterm experience of use of Shuguang 845? Patrick Turner. I remember reading, not long, ago that the Boss (AJ) found 2 sweet spots with his 845's, 970v & 1046v. The voltages are rough approximates that are the best my memory can do at the moment. Good luck and keep us posted on this venture and why not ping him? west |
#7
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
West wrote: "Patrick Turner" wrote in message ... The 845 amps I have been strugling to find time to build are approaching a stage where I must decide on the actual working points that would be most suitable for both 211 and 845 including class A2 operation, by means of cathode follower buffer choke loaded to the fixed bias supply. So far having Ea at 1,000V and Ea at 80mA with RL = 12.5k could seem OK but it means Pda = 80 watts, and only KR845 or KR211 look able to take this idling condition, one reason being that the KR heaters are 10V at 1A instead of the RCA GE use of 10V at 3.3A. RCA data gives Pda = 67 watts, and if Pd-filament = 33 watts, the total tube Pd = 100 watts Anyone here with actual experience of using either tube in an SE amp??? ALSO, anyone here have longterm experience of use of Shuguang 845? Patrick Turner. I remember reading, not long, ago that the Boss (AJ) found 2 sweet spots with his 845's, 970v & 1046v. The voltages are rough approximates that are the best my memory can do at the moment. Good luck and keep us posted on this venture and why not ping him? Well yes, between 970V and 1,050V seems about right... 211 has Ra = about 3.3k and 845 has Ra about 1.9k. So the damping factor of the 211 is poorer and NFB would seem to me to be almost unavoidable. The A2 operation means THD tends to rise when grid current happens... A SRPP with a pair of EL84 in triode looks promising to make the needed 150Vrms drive at low impedance for the 845, and with a choke to bias the 845 grids so the bias can't change easily when in A2. AJ may not be reading the group at the moment. Patrick Turner. west |
#8
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
"Patrick Turner" schreef in bericht ... West wrote: "Patrick Turner" wrote in message ... The 845 amps I have been strugling to find time to build are approaching a stage where I must decide on the actual working points that would be most suitable for both 211 and 845 including class A2 operation, by means of cathode follower buffer choke loaded to the fixed bias supply. So far having Ea at 1,000V and Ea at 80mA with RL = 12.5k could seem OK but it means Pda = 80 watts, and only KR845 or KR211 look able to take this idling condition, one reason being that the KR heaters are 10V at 1A instead of the RCA GE use of 10V at 3.3A. RCA data gives Pda = 67 watts, and if Pd-filament = 33 watts, the total tube Pd = 100 watts Anyone here with actual experience of using either tube in an SE amp??? ALSO, anyone here have longterm experience of use of Shuguang 845? Patrick Turner. I remember reading, not long, ago that the Boss (AJ) found 2 sweet spots with his 845's, 970v & 1046v. The voltages are rough approximates that are the best my memory can do at the moment. Good luck and keep us posted on this venture and why not ping him? Well yes, between 970V and 1,050V seems about right... 211 has Ra = about 3.3k and 845 has Ra about 1.9k. So the damping factor of the 211 is poorer and NFB would seem to me to be almost unavoidable. The A2 operation means THD tends to rise when grid current happens... A SRPP with a pair of EL84 in triode looks promising to make the needed 150Vrms drive at low impedance for the 845, and with a choke to bias the 845 grids so the bias can't change easily when in A2. AJ may not be reading the group at the moment. Patrick Turner. west Has anyone checked this? http://www.tubelab.com/powerdrive.htm It seems very promising, being able to go into A2 without the grid current problems. I know the powerdrive involves mosfets but, since I really like the sound of a transistorized CCS in the tail of my LTP phase splitter, I am very tempted to try a 845 amp with this powerdrive circuit. Cheers! Reyer |
#9
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
"Patrick Turner" wrote in message ... maxhifi wrote: http://www.pentalaboratories.com/pdfs/845.pdf I wonder how the 750V operating point sounds? It almost looks, um, practical! Yes, 750V at about 89mA is OK but the 12k to 14k load is way too high. I defy anyone to tell the difference in sound quality between 1,000V x 67mA and 750V x 89mA at ordinary levels providing the load is right for the tube. The higher Ea and lower Ia give less THD/IMD, but for the first two watts with a pair of these tubes in parallel, there is only a tiny "academic" difference in distortions. Rout is slightly lower at the higher idling Ia because the load line is higher on the Ra curves where they are less sloped than at the lower Ia. So any differences heard would be hard to pin down onto technical operational set up. A fair old difference of op points can be used, like with trioded KT88 which can be used with Ea at 350V to 500V at least. I figure the 67 watt Pda limit should not provoke un-reliablity with Chinese Shuguang 845/211 tubes. Patrick Turner. Are these tubes really worth the trouble? They appear to have some really special (i.e expensive) requirements in terms of transformers, circuit, etc. |
#10
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
Patrick Turner wrote:
West wrote: "Patrick Turner" wrote in message ... The 845 amps I have been strugling to find time to build are approaching a stage where I must decide on the actual working points that would be most suitable for both 211 and 845 including class A2 operation, by means of cathode follower buffer choke loaded to the fixed bias supply. So far having Ea at 1,000V and Ea at 80mA with RL = 12.5k could seem OK but it means Pda = 80 watts, and only KR845 or KR211 look able to take this idling condition, one reason being that the KR heaters are 10V at 1A instead of the RCA GE use of 10V at 3.3A. RCA data gives Pda = 67 watts, and if Pd-filament = 33 watts, the total tube Pd = 100 watts Anyone here with actual experience of using either tube in an SE amp??? ALSO, anyone here have longterm experience of use of Shuguang 845? Patrick Turner. I remember reading, not long, ago that the Boss (AJ) found 2 sweet spots with his 845's, 970v & 1046v. The voltages are rough approximates that are the best my memory can do at the moment. Good luck and keep us posted on this venture and why not ping him? 960 and 1040V. And I was talking strictly about the 845. I never got on with the brittle sound of the 211 but it is well known that the 211 starts to sound better nearer its 1250max. Kondo got around the 211s problems at less elevated voltages (he ran his Ongaku at 960V at the plate, I seem to remember) by loading it with a 20K (claimed, probably 16K in real life) transformer. Well yes, between 970V and 1,050V seems about right... Hey, Max, I'm not impressed with that 760V 845 you found. It may sound fine to the owners if they haven't heard an 845 as it should be built, in fact better than any other amp they ever heard, but it ain't the real thing, which is worth a lot of trouble, as any 845 owner will tell you. Back in the days of the Joenet still, there was a craze among the Japanese for 500V 845. This was one of the dumbest ideas I ever heard. You don't need to build it (though I did) to know it will sound like ****. Just check the Ia-Ea-Eg curves, and ask yourself what sort of a primary impedance you will have to put on the tube to get any power whatsoever, and I mean less than 2W. Suddenly it becomes clear that the entire negative-going half of the signal, to the right of the quiescent operating point, stands chin-deep in sewerage, and if Pavarotti breathes heavily the entire cast drowns in slurry. I also tried about 800V and while it sounded infinitely better than 500V, and much more like the real thing, it still wasn't the best the 845 is capable of. 211 has Ra = about 3.3k and 845 has Ra about 1.9k. So the damping factor of the 211 is poorer and NFB would seem to me to be almost unavoidable. It seems to me perverse to consider the 211 and the 845 as interchangeable in any way whatsoever. All they share is the glass envelope and a requirement for very high voltages to give their best. The A2 operation means THD tends to rise when grid current happens... A SRPP with a pair of EL84 in triode looks promising to make the needed 150Vrms drive at low impedance for the 845, and with a choke to bias the 845 grids so the bias can't change easily when in A2. While I understand the need to drive the 211 into Class A2 to get any real power, it is too high a price to pay for a little extra power to drive the 845 into Class A2. The 845 is a tube that naturally and natively lives in Class A1, where at an entirely feasible 80mA and around 1000V it delivers near as damnit 20 single-ended watts of unequalled glory. To drive it into Class A2 for some "engineering" wet dream of extracting maximum possible power will unbalance its sound, in particular by giving it a bloated bass. I like the idea of SRPP EL84 -- one of my favourite tubes. But with what you're going to have to charge for this amp, you may as well go the whole hog and drive the 845 with another DHT. I've always had good luck gilding the lily by driving kilovolt transmitters with 300B idling along very, very, very quietly in the driver spot. And while I'm giving advice that you probably don't need, don't even think of loading less than 10K on the plate of the 845. (And higher on the plate of the 211 -- the need for impossible impedances is what makes these tubes difficult, not the high drive requirements; to my mind the impossibility of finding the correct output transformer for a 211 rules it out for anyone who isn't himself a winder.) AJ may not be reading the group at the moment. I was happily designing my new bike when some well-meaning lurkers wrote to say that I am wanted. So, like the Good Scout, here I am to help you across the street. Dib-dib. Patrick Turner. west On longevity: the only 845 I ever accidentally popped dropped off the table onto its head. Mine were all Chinese, some cheap rubbish bought for less than ten bucks each in a wholesale box, some obsessively selected. I ran them up to 84mA for several hours in tests. I detuned the ones that left here to 72mA and sometimes up to 75mA. I ran one for a couple of months at about 80mA and 960V, which sent Bill up the wall (add the filament to the idling power and compare with spec). Mine are all gone now to new deeply caring homes, but all the same I haven't heard of one of those tubes that died a premature death in the hands of new owners. Perhaps I have a thousand hours or two on 845, perhaps even up to 5k hours, but that was across quite a few tubes, maybe as many as twenty, and years ago (maybe current production is even better than then), which doesn't really make my anecdotal evidence the sort of hard-use information you want. But 845 are cheap; so what if you have to replace them every couple of thousand hours? If it bothers you, perhaps from a commercial fear for your reputation, make the customer pay for Krone's fab tubes, which come with a guarantee; I have KR 300B equivalents at least a decade old and certainly with more than 10K hours on them. Why don't you ever have easy problems, Patrick, say like how long KR 300B types will last, or whether gennie WE is worth the money? Every audiophile should hear a good 845 at least once before he dies. Otherwise he will have no reference to judge the music of the spheres in Heaven. Andre Jute Visit Jute on Amps at http://members.lycos.co.uk/fiultra/ "wonderfully well written and reasoned information for the tube audio constructor" John Broskie TubeCAD & GlassWare "an unbelievably comprehensive web site containing vital gems of wisdom" Stuart Perry Hi-Fi News & Record Review |
#11
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
You don't need to build it (though I did) to know it will sound like
****. Just check the Ia-Ea-Eg curves, and ask yourself what sort of a primary impedance you will have to put on the tube to get any power whatsoever, and I mean less than 2W. They say to use 3400 ohms, I did print the curve out and draw the load line - and really, it looks decent - amazing in fact compared to say a triode connected KT88. I believe their 15W at 5% THD spec, but I see what you mean, it hardly uses the tube to its full potential. I was just imagining new ways of using 'normal' parts I have laying around (i.e. output transformers), rather than seriously contemplating building a new amp around 845's. |
#12
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
You don't need to build it (though I did) to know it will sound like
****. Just check the Ia-Ea-Eg curves, and ask yourself what sort of a primary impedance you will have to put on the tube to get any power whatsoever, and I mean less than 2W. They say to use 3400 ohms, I did print the curve out and draw the load line - and really, it looks decent - amazing in fact compared to say a triode connected KT88. I believe their 15W at 5% THD spec, but I see what you mean, it hardly uses the tube to its full potential. I was just imagining new ways of using 'normal' parts I have laying around (i.e. output transformers), rather than seriously contemplating building a new amp around 845's. |
#13
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
On Thu, 09 Aug 2007 19:03:08 -0700, Andre Jute
wrote: Lotsa great stuff. All done better than I coulda, and true-er than almost anything to be found locally. (Truth is local, and r.a.t is pretty provincial. Our truth is local but has a global potential.) It seems to me perverse to consider the 211 and the 845 as interchangeable in any way whatsoever. All they share is the glass envelope and a requirement for very high voltages to give their best. I know where you're coming from, but from a transformer perspective, three type 211's are about the same as two type 845's. But, yeah. I have a pair of Chinese 845's and a pair of Chinese 211's, all purchased more than a decade ago, so not interesting to Patrick's question. They work; there's some disturbing (?) tinkling during warm-up, but both pairs bias out fine; neither pair has an hour's total use, so I have no data. The great thing about type 211's *used to* be that they were everywhere for next to nothing. Today, with the antiques given their proper respect, these old beezers are too expensive for experimenters to drop large for six per stereo set, and Patrick's question is appropo. And OTOH, these are the easiest valves to make in a third-world environment. Is this even a correct description of China today? I suspect not, making the question even more interesting. Thanks, as always, Chris Hornbeck "It's just this little Chromium Switch. You people are SO superstitious." |
#14
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
reyer wrote: "Patrick Turner" schreef in bericht ... West wrote: "Patrick Turner" wrote in message ... The 845 amps I have been strugling to find time to build are approaching a stage where I must decide on the actual working points that would be most suitable for both 211 and 845 including class A2 operation, by means of cathode follower buffer choke loaded to the fixed bias supply. So far having Ea at 1,000V and Ea at 80mA with RL = 12.5k could seem OK but it means Pda = 80 watts, and only KR845 or KR211 look able to take this idling condition, one reason being that the KR heaters are 10V at 1A instead of the RCA GE use of 10V at 3.3A. RCA data gives Pda = 67 watts, and if Pd-filament = 33 watts, the total tube Pd = 100 watts Anyone here with actual experience of using either tube in an SE amp??? ALSO, anyone here have longterm experience of use of Shuguang 845? Patrick Turner. I remember reading, not long, ago that the Boss (AJ) found 2 sweet spots with his 845's, 970v & 1046v. The voltages are rough approximates that are the best my memory can do at the moment. Good luck and keep us posted on this venture and why not ping him? Well yes, between 970V and 1,050V seems about right... 211 has Ra = about 3.3k and 845 has Ra about 1.9k. So the damping factor of the 211 is poorer and NFB would seem to me to be almost unavoidable. The A2 operation means THD tends to rise when grid current happens... A SRPP with a pair of EL84 in triode looks promising to make the needed 150Vrms drive at low impedance for the 845, and with a choke to bias the 845 grids so the bias can't change easily when in A2. AJ may not be reading the group at the moment. Patrick Turner. west Has anyone checked this? http://www.tubelab.com/powerdrive.htm It seems very promising, being able to go into A2 without the grid current problems. I know the powerdrive involves mosfets but, since I really like the sound of a transistorized CCS in the tail of my LTP phase splitter, I am very tempted to try a 845 amp with this powerdrive circuit. Using some kind of voltage follower directly connected to the output 845 or 211 grid is the normal way to over come the sudden low impedance input when the grid of the output suddenly begins to draw grid current when it is forced above 0V potential The A2 distortion in the output signal is due to the sag in the input signal drive, because the output grid appears as a low resistance in series with a diode to 0V, so it conducts above 0V. The mosfet follower isn't too bad a choice because of its very low source resistance of 1/gm and for the mosfet concerned, its probably less than 10 ohms. So there is almost no distortion created at the power grids due to normal plate drive stalling when the load changed to low resistance with grid current, maybe a couple of thousand ohms. I am not sure how much grid current is drawn by either 845 or 211 when the grid is at +50V and the anode has swung down to a low voltage. So I don't know what the equivalent diode + R circuit is that any follower has to deal whith when grid current is drawn. My plan so far for this amp is as follows :- Anode supply, B+, +800V approx, CLC filtering. Cathode supply, B, -400V approx, CLC filtering. 211 or 845 is set up with OPT and cathode bias resistor between +800 and -400V and grid is biased via a 100H choke from the -400. The cathode with 845 will be at about 160V above the -400V at -240V. Rk will carry the tube Ia of approx 70mA so Rk = 1k2 approx. The cathode has 10Vdc applied, CLC filtered, and with negative end as the connection for the Rk. Also from this point 1,000uF is used to bypass the cathode to the 0V point in the power supply and rest of the input circuitry. The input tube V1 is one x paralleled 6CG7 with CCS load to +350V to produce a low distortion signal up to about 10Vrms for the driver stage. Driver gain stage stage is EL84/6BQ5 in triode with Ea at 350V approx. Its B+ of +750V is derived from output B+. Ia is about 12mA, RL about 40k, so 200Vrms is easy from this set up and distortion is mainly 2H and has a phase which cancels the 2H of the output tubes. Between this voltage gain stage and the 845/211 grids, there is a buffer cathode follower stage. It does the job of the mosfet used in the "power drive" scheme mentioned by our learned experimenter at the website quoted above. I see little reaon at all to use a solid state follower in series in the signal path. The buffer cathode follower is set up with its anode grounded to 0V. Its cathode is connected to the output tube grids directly. Then there is a small R of perhaps 560 ohms towards -400V, and then choke to -400V. So there is a very low impedance biasing the output grids to -400V, while the cathodes of output tubes adjust themselves ordinarily as with cathode bias to a point of eqilibrium. There is no fixed bias to adjust, and no need. The EL84 cathode follower grid is biased via 220k also from the -400V, and the cathode voltage will rise with the Ia flow of about 15mA to about 15V, depending on the choke winding resistance and the small series resistor. The grid of the CF is cap coupled with two 0.47uF to the voltage amp EL84 whose anode is at +350V. The centre join of the two caps is biased via 1M to 0V. Thus 630V rated coupling caps can be used. This is a simple arrangement, and for changing from 845 to 211, some reduced Rk is needed, and some added R in the cathode power supply so the headroom taken up with cathode bias voltage is used better. Cathode biasing in class AB and B PP amps isn't used because the Ia is non-linear, and positive going currents are much higher than negative going currents, so Ck charge up and biase drifts. This does not occur in class A2 amps. The +Ik and -Ik swings are very similar, changeing only 5% if that is what the THD is at clipping even when into A2 operation. So all we need to guard against is the distortion to the drive signal caused by the grid current stalling the voltage amp. In my scheme, the output tube grid bias voltage is kept constant because of the low DC R of the choke and trimming R for the buffer CF. The output resistance of the EL84 in triode in CF = 1 / gm = approx 100 ohms. AFAIK, this low resiostance is very much lower than the load offered by the output tubes when drawing grid current. The EL84 will itself saturate at about 120mA and I think that current ability would be sufficient when 845 or 211 have their grids forced to being about +50v at the Ea minimum value on negative going anode swings. A 120mA change in Ig for a 50V swing indicates Rg input resistance when positive to be 50 / 0.12 = 416 ohms. So the question is, what is the gid current when output grid is at say +55V above the cathode voltage??? Does one need to use a mosfet?? In the website above or at http://www.tubelab.com/powerdrive.htm the guy has a photo of a 211 which has turned bright blue while in the process of dying, and with 300mA of anode current, and if he had a simple active protection scheem which reacted to a voltage across a cathode resistance, there would be no panic about turning off the amp when ****e like this happens. If you build tube amps, especially high voltage types, you should always take the time to add in the protection and auto shut down circuits. Patrick Turner. Cheers! Reyer |
#15
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
maxhifi wrote: "Patrick Turner" wrote in message ... maxhifi wrote: http://www.pentalaboratories.com/pdfs/845.pdf I wonder how the 750V operating point sounds? It almost looks, um, practical! Yes, 750V at about 89mA is OK but the 12k to 14k load is way too high. I defy anyone to tell the difference in sound quality between 1,000V x 67mA and 750V x 89mA at ordinary levels providing the load is right for the tube. The higher Ea and lower Ia give less THD/IMD, but for the first two watts with a pair of these tubes in parallel, there is only a tiny "academic" difference in distortions. Rout is slightly lower at the higher idling Ia because the load line is higher on the Ra curves where they are less sloped than at the lower Ia. So any differences heard would be hard to pin down onto technical operational set up. A fair old difference of op points can be used, like with trioded KT88 which can be used with Ea at 350V to 500V at least. I figure the 67 watt Pda limit should not provoke un-reliablity with Chinese Shuguang 845/211 tubes. Patrick Turner. Are these tubes really worth the trouble? They appear to have some really special (i.e expensive) requirements in terms of transformers, circuit, etc. If you want 50 watts of pure class A SE power, then you'd have to use about 8 x 300B in parallel. A couple of 211/845 are a viable alternative. I've used 1 x 13E1 to make 25 watts, not bad, and two would make 50W with only a 450v supply. Sadly 13E1 are not being made any more. I've done 35 watts with 4 x 6CA7/EL34, The challenge of the 845 beckons. |
#16
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
Andre Jute wrote: Patrick Turner wrote: West wrote: "Patrick Turner" wrote in message ... The 845 amps I have been strugling to find time to build are approaching a stage where I must decide on the actual working points that would be most suitable for both 211 and 845 including class A2 operation, by means of cathode follower buffer choke loaded to the fixed bias supply. So far having Ea at 1,000V and Ea at 80mA with RL = 12.5k could seem OK but it means Pda = 80 watts, and only KR845 or KR211 look able to take this idling condition, one reason being that the KR heaters are 10V at 1A instead of the RCA GE use of 10V at 3.3A. RCA data gives Pda = 67 watts, and if Pd-filament = 33 watts, the total tube Pd = 100 watts Anyone here with actual experience of using either tube in an SE amp??? ALSO, anyone here have longterm experience of use of Shuguang 845? Patrick Turner. I remember reading, not long, ago that the Boss (AJ) found 2 sweet spots with his 845's, 970v & 1046v. The voltages are rough approximates that are the best my memory can do at the moment. Good luck and keep us posted on this venture and why not ping him? 960 and 1040V. And I was talking strictly about the 845. I never got on with the brittle sound of the 211 but it is well known that the 211 starts to sound better nearer its 1250max. Kondo got around the 211s problems at less elevated voltages (he ran his Ongaku at 960V at the plate, I seem to remember) by loading it with a 20K (claimed, probably 16K in real life) transformer. Kondo has his schematic shown at http://www.drtube.com/schematics/an/ongaku.gif The output 211 doesn't have a cathode shown, but let's assume it is taken to 0V. Ea sure is 960V, and bias at about -47V, and the CF drives the 211 directly to get the +53V max applied to the grid to give a total drive voltage of about +/- 100V pk or 70Vrms. So Kondo knew how to get lots of power but not have to use 1,250V to avoid the grid current. he knew the first 10 watts class A1 up to onset of A2 was going to be utterly blameless. Then he said its due to silver wire and special cores and a lot of BS followed about it all and oh how the $$ rolled in. What a masterful stroke of luck he engineered! Hats off for Kondo. I have heard 211 and find 'em real goooood. It depends who made the darn thing, and all the other issues in the amp, and boy do thay matter because there isn't any NFB loop to correct sonic or technical problems. ( No NFB to further muddle a sound either, but its all not as simple as some would asert..) Well yes, between 970V and 1,050V seems about right... Hey, Max, I'm not impressed with that 760V 845 you found. It may sound fine to the owners if they haven't heard an 845 as it should be built, in fact better than any other amp they ever heard, but it ain't the real thing, which is worth a lot of trouble, as any 845 owner will tell you. Back in the days of the Joenet still, there was a craze among the Japanese for 500V 845. This was one of the dumbest ideas I ever heard. But why so dumb? if the Ea = only 760V, and Pda = 75 watts, then Ia would be 100mA, and load whatever to suit, and you get a reasonable 15+ watts and listen with 2 watts. I bet a bunch of golden eared auio buffs would not hear any difference beteween an 845 and 3 paralled KT88 in triode, or 3 x 300B. Perhaps its good wine for all three options. The 500V option isn't ideal, but if Ia = 100mA, Pda = 50W and if RL = 5k, you get +/- 160V peak across a 5k load for 2.6 watts. But many japanese are happy sipping their drinks very slowly rather than guzzling it down like a round eye westener. And volume is low in their small roomed houses with paper walls. so they marvel at the 845, and its sweetness even at low power. But whatever thei pin onto the 845 isn't all that is tickiling their ears; its all the other tubes in the line up as well.... The last option is to whack up the Ea, and whack up the RL, to have the triode act more like a voltage device than a current one, thus provoking its internal NFB loop to create more heaven in sound by coping with the devils of distortion better. You don't need to build it (though I did) to know it will sound like ****. Just check the Ia-Ea-Eg curves, and ask yourself what sort of a primary impedance you will have to put on the tube to get any power whatsoever, and I mean less than 2W. 5k and 500V at 100mA is a good way to get 2.6 watts of nice power. Like a 2A3 on steroids. A big bottle full in fact. Its like using an elephant to shift a matchstick, and a limited elephant at this Ea and Ia, without being allowed to flex muscle with A2. If the circuit is made to work A2 you get 15 watts. The tube will last forever.... Suddenly it becomes clear that the entire negative-going half of the signal, to the right of the quiescent operating point, stands chin-deep in sewerage, and if Pavarotti breathes heavily the entire cast drowns in slurry. Not if the Ia is made high enough to stop the Ea +ve swinging into Ra curvature when Ia becomes low. I also tried about 800V and while it sounded infinitely better than 500V, and much more like the real thing, it still wasn't the best the 845 is capable of. I agree. My customers all prefer a much higher power ceiling than japanese who don't mind low levels. 211 has Ra = about 3.3k and 845 has Ra about 1.9k. So the damping factor of the 211 is poorer and NFB would seem to me to be almost unavoidable. It seems to me perverse to consider the 211 and the 845 as interchangeable in any way whatsoever. All they share is the glass envelope and a requirement for very high voltages to give their best. Well, the fact is that the same load at the same Ea and Ia can be used to get a maximum of around 31 watts where Pda quiescent = 67 watts and operation is A2. If you dissallow A2, and toy stipulate you want 20 watts, then the op points will have to change for the same Pda; the 211 will have higher Ea and RL and lower Ia than 845. Because the the anode load will about 33% higher than for 845, then the damping factor difference is reduced, and both become usable without any NFB. If the Ra = 3.6k at 55mA with 211, and load = 20k, then DF = 5.5, quite good. 845 would have Ra = 1.8k and RL = 14k and DF = 7.7, only marginally better. It means that if the speaker coils suits 6 ohms, and 6 ohms is the load, then allowing 0.6 ohms total winding R, Rout varies between 1.8 ohms, 211, and 1.5 ohms. So with either tube used wisely, you get DF 4 with either triode without any loop NFB. I can say that 10dB of NFB will not destroy the sound, and probaably make bass sound tighter etc, but NFB is optional when Rout gives a DF 4. (( IMHO )) The A2 operation means THD tends to rise when grid current happens... A SRPP with a pair of EL84 in triode looks promising to make the needed 150Vrms drive at low impedance for the 845, and with a choke to bias the 845 grids so the bias can't change easily when in A2. While I understand the need to drive the 211 into Class A2 to get any real power, it is too high a price to pay for a little extra power to drive the 845 into Class A2. To get the 211 to make power in A1, you simply have to use a high Ea, hence 1,250V is about right, maybe a little more even, as long as the required OPT is available.... OPT with 20k primary loads for operation with HV are thin on the ground. The A2 allows operation of both 211 and 845 at a lower Ea than one might use. But if one insists on A2 AND having Ea high, then RL must be higher to allow the wider swing without cut off. meanwhile if this becomes the case for eithewr tube then at the op point with high ea, and high RL the tube has very low THD/IMD for the first A1 watts before it moves to A2. Then of course the driver tube surely must have a large effect because its 2H cancels the output tube 2H whether you like that or not. To what extent the cancellation occurs depends on a lot of things, and the actual load in use. There may be almost perfect cancelation with 8 ohms at the speaker, but at 4 ohms or 12 ohms the cancelation makes hardly any difference. But my customers say they like clean sound of my SE amps. I suppose I could say why; less than 0.1% THD at any load from 4 ohms to 12 ohms for the first 10 watts out a total of 35 At 2 watts and 5 ohms the CFB 35 had less THD than most PP amps of the same max power and even with 6 dB more global NFB.......... Because the 845 and 211 present pretty low THD at listening levels, AND because the driver stage has to work a bit even to produce these levels, there is a fortitious amount of non contrived natural and unavoidable distortion cancelling going on. Usually the output stage of any amp makes more THD than any of its driver stages, but where output tube gain is low, then the driver stage has to work hard and thus it produces considerable distortions, and in SE design it cancels where all the tubes involved are triodes. The 845 is a tube that naturally and natively lives in Class A1, where at an entirely feasible 80mA and around 1000V it delivers near as damnit 20 single-ended watts of unequalled glory. To drive it into Class A2 for some "engineering" wet dream of extracting maximum possible power will unbalance its sound, in particular by giving it a bloated bass. Well, I fail to see the slightest reason why adding a CF to simply extend the A1 range a bit into A2 would have any sound effect. Remember, its the overall thinge here, and who has made the amp, and how far they went, and at what sort of trouble...... I like the idea of SRPP EL84 -- one of my favourite tubes. But with what you're going to have to charge for this amp, you may as well go the whole hog and drive the 845 with another DHT. I've always had good luck gilding the lily by driving kilovolt transmitters with 300B idling along very, very, very quietly in the driver spot. SRPP EL84 in triode has faded in my mind in favour of a simple anode gain stage driving a direct coupled CF using a secong EL84 triode. The reason I favour EL84, it have it capable of 180Vrms for 845 drive, and only need 10Vrms input, so a nice drive tube like 6SN7, 12AU7, 6CG7 can be used for an input voltage not more than 1Vrms, without global NFB. Using as pair of 845 in parallel improves the sensitivity. 211 may require a little bit of GNFB, but because its gain is twice 845, GNFB can be used, and sensitivity is still sensible, under 1.5vrms input. And while I'm giving advice that you probably don't need, don't even think of loading less than 10K on the plate of the 845. (And higher on the plate of the 211 -- the need for impossible impedances is what makes these tubes difficult, not the high drive requirements; to my mind the impossibility of finding the correct output transformer for a 211 rules it out for anyone who isn't himself a winder.) I agree entirely. I want to have about 6k minimum for the two paralleled tubes, so 12k each. SE loads refelected from speakers vary. Many speaker makers lie all their way to the bank about Z and sensitivity. Nearly all speakers have a low Z between bass and midranges where most power is needed because crossovers shunt a portion of the current if they are second order and damped properly. an Xo set for typical 6 ohm driver has Zin = 4.5 ohms at Fo and if the Fo of bass and mids is the same, then Zin = 2.3 ohms at Fo. I try to avoid this horror by overlapping the bass-midrange, so that Zmin is 4 ohms perhaps, over a wider F band. I arrange my PP amps to make max PO at 3 ohms, and SE amps to peak at 4 ohms. But in the real world a load mismatch often cannot be avoided, and still we get blameless sound from SET class A where the clipping ceiling is way above the peak levels used, even with the wrong load. But a couple of my customers cheat, and are sneaky to a point. Thy have bought ZeroImpedance load matching toroidal transformers, and both say they like having a 2:1 Z ratio change so that an 8 ohm speaker looks like a 16 ohm, and a 3 ohm load at some point along the AF band becomes a more reasonable 6 ohms. They don't like the reduced sensitivity and lower ceiling, but they say it sounds better. AJ may not be reading the group at the moment. I was happily designing my new bike when some well-meaning lurkers wrote to say that I am wanted. So, like the Good Scout, here I am to help you across the street. Dib-dib. Blink and you'll miss me for I do try to ride fast these days..... Patrick Turner. west On longevity: the only 845 I ever accidentally popped dropped off the table onto its head. Mine were all Chinese, some cheap rubbish bought for less than ten bucks each in a wholesale box, some obsessively selected. I ran them up to 84mA for several hours in tests. I detuned the ones that left here to 72mA and sometimes up to 75mA. I ran one for a couple of months at about 80mA and 960V, which sent Bill up the wall (add the filament to the idling power and compare with spec). Mine are all gone now to new deeply caring homes, but all the same I haven't heard of one of those tubes that died a premature death in the hands of new owners. Perhaps I have a thousand hours or two on 845, perhaps even up to 5k hours, but that was across quite a few tubes, maybe as many as twenty, and years ago (maybe current production is even better than then), which doesn't really make my anecdotal evidence the sort of hard-use information you want. But 845 are cheap; so what if you have to replace them every couple of thousand hours? If it bothers you, perhaps from a commercial fear for your reputation, make the customer pay for Krone's fab tubes, which come with a guarantee; I have KR 300B equivalents at least a decade old and certainly with more than 10K hours on them. Well, people say don't drop DH tubes because the cathodes are brittle. KR have oxide coated DH cathodes, and probably having the tubes turn on for a coupla minutes without B+ so that gass molecules are absorbed by the getter BEFORE becoming stripped of electrons and bombarding the cathode is a good idea. Why don't you ever have easy problems, Patrick, say like how long KR 300B types will last, or whether gennie WE is worth the money? I don't have a clue how long some tubes might last. But a customer here has been running his 13E1 NOS made probably in 1960 for 5 years nearly every day for sometimes all day, and there has never been smoke, always been real music, and technically, they still measure almost as new' last service I did 2 years ago all was as new except some grid current had started to make grids slightly +ve, so I reduced the Rg from 220k to 47k, and changed the ECC32 paralleled driver to 6V6 in triode. FB was effectively slightly reduced, and the client said the sound was slightly better. I will soon be giving the amps a check up again; maybe swap the plain UL config to CFB if the OPT allows it. Its 11 years since I made those amps, and because I thought I could teach the local audio club that one didn't have to spend $140,000 on an Ongaku, a humble single big fat beam tetrode could do the business, and without silver wire, and fancy-shmancy caps, resistors, chokes and contacts et all. I have a pair of KR300B I was given to demo to customers. So I did, and nobody can pick them from any other 300B, and they say well I have a spare dozen Sveltlanas, and don't need anything else. The KR can be run continuously at higher Pda than most other 300b... Same goes for KR845 and 211, both of which have 10V x 1A cathodes, yea, 1/3 of the heat in the heaters!!! Hence along with a bit of a fiddle hear and there the Pda of 100 watts is OK. But I wouldn't. Sooner or later someone will try to use an old RCA etc, and there would be smoke. Every audiophile should hear a good 845 at least once before he dies. Otherwise he will have no reference to judge the music of the spheres in Heaven. Hmm, Angels have orbs they say, glorious wobbalititybitty spheres they say. If ever i go Upstairs, it'll be a case of Pardon me God, lemme try to get a root here because they didna want root much where I jus came from. Sigh, but the rich guys will have all the shielas Up there I guess, because in the latter day churches they now preach that Prosperity Is Good. Go to Jesus with a millyun. If ya poor, youse live in a drain pipe of Heaven. But turds of gold don't smell as they float past though. When ya see a discarded 845 float past, ya grab it. I'm told the hills are slight and winds are favourable when cycling around heaven and you never get a puncture. Car drivers stop to salute, and wait until youv'e passed. And there ain't no drug cheats in La Tour De Heaven. Patrick Turner. Andre Jute Visit Jute on Amps at http://members.lycos.co.uk/fiultra/ "wonderfully well written and reasoned information for the tube audio constructor" John Broskie TubeCAD & GlassWare "an unbelievably comprehensive web site containing vital gems of wisdom" Stuart Perry Hi-Fi News & Record Review |
#17
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
On Aug 9, 9:06 pm, "maxhifi" wrote:
You don't need to build it (though I did) to know it will sound like ****. Just check the Ia-Ea-Eg curves, and ask yourself what sort of a primary impedance you will have to put on the tube to get any power whatsoever, and I mean less than 2W. They say to use 3400 ohms, I did print the curve out and draw the load line - and really, it looks decent - amazing in fact compared to say a triode connected KT88. I believe their 15W at 5% THD spec, but I see what you mean, it hardly uses the tube to its full potential. I was just imagining new ways of using 'normal' parts I have laying around (i.e. output transformers), rather than seriously contemplating building a new amp around 845's. maxhifi wrote: You don't need to build it (though I did) to know it will sound like ****. Just check the Ia-Ea-Eg curves, and ask yourself what sort of a primary impedance you will have to put on the tube to get any power whatsoever, and I mean less than 2W. They say to use 3400 ohms, I did print the curve out and draw the load line - and really, it looks decent - amazing in fact compared to say a triode connected KT88. I believe their 15W at 5% THD spec, but I see what you mean, it hardly uses the tube to its full potential. I was just imagining new ways of using 'normal' parts I have laying around (i.e. output transformers), rather than seriously contemplating building a new amp around 845's. I've drawn the 500V 5K (warning magenta lines), 750V 3.4K (seasick cyan lines) and 1000V 10K (reliable deep blue lines) operating points on the curves to aid discussion: http://members.lycos.co.uk/fiultra/D...20Circuits.htm The key thing isn't the voltage but the plate load necessitated at that voltage by (often unrealistic) power expectations. I haven't drawn the 500V 3K combination that was suggested at one stage but you can easily see that it will be down in the curved grunge at the bottom of the transfer function. Even with 5K this isn't such a hot amp: it will operate at current cutoff rather often, right after it makes an unplanned discursion into grid current, with the uncontrolled consequences Patrick has already alluded to. The 750V 3.4K suggestion is one made by engineers who think THD of 5 per cent acceptable... It might have been made acceptable by accepting a lower power output consequent on a more conservative load choice. It uses huge current swings for absolutely no good purpose. It isn't brilliant even by engineering standards, never mind audio standards. Notice also that my minimum acceptable combination of 1000V 10K has *uses more current over half its output*, that is up to nearly 10W! Notice that the output of the 1000V 10K combination is *visibly* less distorted over its entire output: you don't have to measure, it hits you in the face when you eyeball it. It is in fact virtually undistorted to more power than the total output of the 750V suggestion. Notice "minimum acceptable combination": Ideally the loadline should be tilted to cross the negative bias lines at right angles, but I have never had big transformers of such elevated impedance, and Patrick, who is presumably capable of winding something good enough for 100W at 12-14K, doesn't want to pay the price in lost power. I do however know that a major improvement lies beyond 10K if you don't need all the power. I discovered this when I seriesed two high quality 5K6 Lundahl transformers for 11K2 altogether and loaded up an 845 with their primaries. Of course, the 845 is such a fabulously linear tube, what Patrick and I and others are shouting about like high tragedy are in practice minor pecuniations that in a less superior tube will not even be noticed in its general raucousness. We can get altogether too bloody refeened for our own good. Andre Jute Visit Jute on Amps at http://members.lycos.co.uk/fiultra/ "wonderfully well written and reasoned information for the tube audio constructor" John Broskie TubeCAD & GlassWare "an unbelievably comprehensive web site containing vital gems of wisdom" Stuart Perry Hi-Fi News & Record Review |
#18
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
On Aug 9, 10:49 pm, Chris Hornbeck
wrote: On Thu, 09 Aug 2007 19:03:08 -0700, Andre Jute wrote: I have a pair of Chinese 845's and a pair of Chinese 211's, all purchased more than a decade ago, so not interesting to Patrick's question. They work; there's some disturbing (?) tinkling during warm-up, but both pairs bias out fine; neither pair has an hour's total use, so I have no data. It occurred to me last night after I sent my post that you have 845 in use, that it might be smart to refer Patrick to you as the longevity expert. Then I though, Nah, Chris is so rich, he probably burns NOS GE 845 and has a couple for nightlights on his bedside tables too. Just goes to prove, I'm a scary judge of character! Andre Jute |
#19
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
On Aug 10, 6:47 am, Patrick Turner wrote:
Andre Jute wrote: [[[[big snip]]]] The last option is to whack up the Ea, and whack up the RL, to have the triode act more like a voltage device than a current one, thus provoking its internal NFB loop to create more heaven in sound by coping with the devils of distortion better. Shh, man, we're trying to keep our magic a secret from the credit cards of the hoi-polloi, not to mention pretending to the ultrafidelista that we absolutely abhor NFB, the devil's own wickedness. Andre Jute Descended from three Popes |
#20
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
Hi Andre,
All is clear now, thanks for the diagram - I drew the load line incorrectly, for some reason (I'd blame it on being tired, but more likely lack of any recent practice with tube design!). The slope I drew is actually about 9300 ohms, which at 750V looks pretty nice - hence my confusion. |
#21
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
On Fri, 10 Aug 2007 14:01:01 -0700, Andre Jute
wrote: It occurred to me last night after I sent my post that you have 845 in use, that it might be smart to refer Patrick to you as the longevity expert. Then I thought, Nah, Chris is so rich, he probably burns NOS GE 845 and has a couple for nightlights on his bedside tables too. I'm strictly blue collar working class, but I'm rich in tubes (bought, or sometimes given to me, back before they became Unobtainium. Shoot, I have more than a dozen GE 211/VT4C's; many are the 1942 production still in boxes with papers, etc. that I don't even use. Also four RCA's from the War era and a variety of others. I have Mullard GZ34's new in their boxes and dozens of Telefunken 12AX7's, even a pair of tall plate Bugle Boys. In 845's I have RCA's of various vintages, but I use my favorite, United Electronics, given to me by a friend who was scrapping out a lightning-struck carrilon. Power and output transformer were melted, and some of the six or eight original 845's got hurt, but some survived (!) to play music here since 1994. This isn't to brag or anything, just to marvel at my good fortune. I'd rather be lucky than smart any day. Just goes to prove, I'm a scary judge of character! Arf! Thanks, as always, Chris Hornbeck "It's just this little Chromium Switch. You people are SO superstitious." |
#22
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
Has anyone checked this? http://www.tubelab.com/powerdrive.htm It seems very promising, being able to go into A2 without the grid current problems. I know the powerdrive involves mosfets but, since I really like the sound of a transistorized CCS in the tail of my LTP phase splitter, I am very tempted to try a 845 amp with this powerdrive circuit. Interesting. Because a MOSFET follower with grid bias set by gate V is pretty much my 'standard' fixed bias, positive grid drive, setup. Patrick may remember me mentioning it way back when I did my first 6V6 SE design where it was my solution to triode mode A2. The problem with MOSFETs is the high gate capacitance with the one I like using in my comparatively pipsqueak power designs being the STP2NK60Z, because it's cheap, has built in gate protection, and a gate capacitance of 170pF. By comparison, the ubiquitous IRF740 has 1400pF gate capacitance. Yes but its in follower mode, so because the source chases after the gate, the effective capacitance is low. But should ever the mosfet short circuit, its B+ supply will pass a heck of a lot of current into the 845 grid. Hence active shut down protection is a real must have. Within in a particular design family the more current, and lower RdsOn, the higher the gate capacitance, because of the increased gate and channel width to get it, so 'over sizing' is not necessarily the way to go, if you're concerned with gate capacitance. MOSFETs also have a rather pronounced gate threshold temperature dependence (along with Ids) but if grid bias -V is large then MOSFET gate drift may not be significant enough to worry much about, but it's something to keep in mind. Keeping the mosfet cool is easy in this app. Patrick Turner. Cheers! Reyer |
#23
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
flipper wrote: On Fri, 10 Aug 2007 10:15:51 GMT, Patrick Turner wrote: reyer wrote: trim Has anyone checked this? http://www.tubelab.com/powerdrive.htm It seems very promising, being able to go into A2 without the grid current problems. I know the powerdrive involves mosfets but, since I really like the sound of a transistorized CCS in the tail of my LTP phase splitter, I am very tempted to try a 845 amp with this powerdrive circuit. trim Cathode biasing in class AB and B PP amps isn't used because the Ia is non-linear, and positive going currents are much higher than negative going currents, so Ck charge up and biase drifts. This does not occur in class A2 amps. The +Ik and -Ik swings are very similar, changeing only 5% if that is what the THD is at clipping even when into A2 operation. So all we need to guard against is the distortion to the drive signal caused by the grid current stalling the voltage amp. In my scheme, the output tube grid bias voltage is kept constant because of the low DC R of the choke and trimming R for the buffer CF. What about positive grid current charging Ck? In class A2, there isn't any non linear increase in Ik to charge Ck as in the case of one tube in an AB PP amp. Grid current occurs over a threshold, that's all, and you just need a direct coupled voltage follower to provide it. Patrick Turner. trim Patrick Turner. Cheers! Reyer |
#24
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
Andre Jute wrote: On Aug 9, 9:06 pm, "maxhifi" wrote: You don't need to build it (though I did) to know it will sound like ****. Just check the Ia-Ea-Eg curves, and ask yourself what sort of a primary impedance you will have to put on the tube to get any power whatsoever, and I mean less than 2W. They say to use 3400 ohms, I did print the curve out and draw the load line - and really, it looks decent - amazing in fact compared to say a triode connected KT88. I believe their 15W at 5% THD spec, but I see what you mean, it hardly uses the tube to its full potential. I was just imagining new ways of using 'normal' parts I have laying around (i.e. output transformers), rather than seriously contemplating building a new amp around 845's. maxhifi wrote: You don't need to build it (though I did) to know it will sound like ****. Just check the Ia-Ea-Eg curves, and ask yourself what sort of a primary impedance you will have to put on the tube to get any power whatsoever, and I mean less than 2W. They say to use 3400 ohms, I did print the curve out and draw the load line - and really, it looks decent - amazing in fact compared to say a triode connected KT88. I believe their 15W at 5% THD spec, but I see what you mean, it hardly uses the tube to its full potential. I was just imagining new ways of using 'normal' parts I have laying around (i.e. output transformers), rather than seriously contemplating building a new amp around 845's. I've drawn the 500V 5K (warning magenta lines), 750V 3.4K (seasick cyan lines) and 1000V 10K (reliable deep blue lines) operating points on the curves to aid discussion: http://members.lycos.co.uk/fiultra/D...20Circuits.htm That 3k4 load looks the wrongest of them all IMHO. I'd reckon nothing less than 5k should wever be used. Ra = 1.8k, and RL should never be less than 2Ra, preferably 3Ra of above. The key thing isn't the voltage but the plate load necessitated at that voltage by (often unrealistic) power expectations. I haven't drawn the 500V 3K combination that was suggested at one stage but you can easily see that it will be down in the curved grunge at the bottom of the transfer function. Even with 5K this isn't such a hot amp: it will operate at current cutoff rather often, right after it makes an unplanned discursion into grid current, with the uncontrolled consequences Patrick has already alluded to. 3k is a horrible load, and in the real world of speaker having some Z half the speaker nominal Z means you could easily have RL = Ra Then with say 3.6k:6 ohms, the OPT Z ratio = 600:1, so the Ra of 1.8k becomes 3 ohms at the output coil. With Rw = say 0.7 ohms, Rout = 3.7 ohms and too high; Its perfectly lousy design to use around 3k with such a tube. The 750V 3.4K suggestion is one made by engineers who think THD of 5 per cent acceptable... It might have been made acceptable by accepting a lower power output consequent on a more conservative load choice. It uses huge current swings for absolutely no good purpose. It isn't brilliant even by engineering standards, never mind audio standards. Agreed. Notice also that my minimum acceptable combination of 1000V 10K has *uses more current over half its output*, that is up to nearly 10W! Notice that the output of the 1000V 10K combination is *visibly* less distorted over its entire output: you don't have to measure, it hits you in the face when you eyeball it. It is in fact virtually undistorted to more power than the total output of the 750V suggestion. Now we are getting somewhere. I thought Ea at 1,100V, Ia at 70mA with 12k with A1 was just fine, at least with KR845. But its Pda = 77 watts, and I'd be happier with 65mA. Now if the load line is 14k, the excursion into the curves on Ea +ve swings is less. AND we could afford to shift the Q point left by about 150V and run A2. I intend constructing the amp to be able to run A2 because I simply want to be able to use any known 845/211 ever made without worry, and with only cathode biasing the change. Anyway, I worked out you could get 31 watts A2 with 12 to 14k with 68Watts input, or about 45% efficiency which is superb for any class A amp. Notice "minimum acceptable combination": Ideally the loadline should be tilted to cross the negative bias lines at right angles, but I have never had big transformers of such elevated impedance, and Patrick, who is presumably capable of winding something good enough for 100W at 12-14K, doesn't want to pay the price in lost power. I do however know that a major improvement lies beyond 10K if you don't need all the power. I discovered this when I seriesed two high quality 5K6 Lundahl transformers for 11K2 altogether and loaded up an 845 with their primaries. All SET triode amps and even SEUL or CFB amps tend to sound better the more you flatten the load line; ie, make it a higher value so there is a wide voltage swing for a small current swing. Of course, the 845 is such a fabulously linear tube, what Patrick and I and others are shouting about like high tragedy are in practice minor pecuniations that in a less superior tube will not even be noticed in its general raucousness. We can get altogether too bloody refeened for our own good. Indeed. But the 300B, or a trioded KT88/6550/KT90 or 2A3, 45, et all will benefit from a high Ea and low Ia and wide voltage swing and low current swing. None are better with low Ea and high Ia and low load. Its silly to use say a KT88 at 250V, 140mA, and RL about 2k. Rout and THD/IMD are all much higher than they need to be. Tubes like voltage, not current. Patrick Turner. Andre Jute Visit Jute on Amps at http://members.lycos.co.uk/fiultra/ "wonderfully well written and reasoned information for the tube audio constructor" John Broskie TubeCAD & GlassWare "an unbelievably comprehensive web site containing vital gems of wisdom" Stuart Perry Hi-Fi News & Record Review |
#25
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
Andre Jute wrote: On Aug 10, 6:47 am, Patrick Turner wrote: Andre Jute wrote: [[[[big snip]]]] The last option is to whack up the Ea, and whack up the RL, to have the triode act more like a voltage device than a current one, thus provoking its internal NFB loop to create more heaven in sound by coping with the devils of distortion better. Shh, man, we're trying to keep our magic a secret from the credit cards of the hoi-polloi, not to mention pretending to the ultrafidelista that we absolutely abhor NFB, the devil's own wickedness. Well I won't even mention that the internal NFB in the triode was put there by the God Of Triodes. See, I haven't said that have i? Let them all think we are crazy and mad, I for one don't mind if that's the reason they don't get it. Patrick Turner. Andre Jute Descended from three Popes |
#26
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
On Aug 11, 12:52 am, Patrick Turner wrote:
flipper wrote: In class A2, there isn't any non linear increase in Ik to charge Ck as in the case of one tube in an AB PP amp. Grid current occurs over a threshold, that's all, and you just need a direct coupled voltage follower to provide it. Patrick Turner. First, the threshold isn't necessarily at zero grid volts. It could, depending on tube type and the parameters the individual tube received as its manufacturing inheritance, be several or maybe a handful of volts either sides of no grid bias. Second, much more amusing: Oh, what fun it is to design someone else's amp! Since the power tranny is the least of your worries, you could feed the transmitter tube plate 120-180V extra and power a direct-coupled driver by a drop over the power tube's cathode, a circuit first suggested to me by that excellent Australian tubie Anthony Mills. My fave 417A/5842 is good up to 25mA and the 7044 has lower plate voltage and is as solid as a rock. Both of these would drive 211 easy-peasy, and to get the signal elevation when you sub 845, use a step-up IT. For parts economy, arrange the switching so that the primary of the IT is the choke constant current load on the 417A or 7044 plate when the 211 is in the socket, the switch simply unconnecting the secondaries and cutting in a direct route for the signal from the 417A or 7044 plate to the 845 grid. This direct coupling via the cathode also avoids the tricky/ expensive problem of finding high-voltage switches which is otherwise inherent in your desire to switch 845 and 211 on the same chassis. This unconsidered detail nearly tripped me on my "Millennium's End" which, far less ambitiously, merely switched kilovolt tubes between PSE and shunting the plate voltage in SE. I nearly had a heart attack when I cruised the milspec catalogues and saw the prices for high- voltage switches; it turned out I had a handful an old ham, chief engineer at the airport since the days of tubes, gave me with a truckload of other stuff. By golly, I coulda been a production "engineer"! Why, that's how many $ $$$ parts just saved? It is one of the curiosities of tube amps that the simple ZNFB SE DHT amp always ends up costing more than a complicated PP amp (and causes more headscratching too), and that actually has very little to do with the price of the DHT, even at WE/KR prices. Andre Jute Visit Jute on Amps at http://members.lycos.co.uk/fiultra/ "wonderfully well written and reasoned information for the tube audio constructor" John Broskie TubeCAD & GlassWare "an unbelievably comprehensive web site containing vital gems of wisdom" Stuart Perry Hi-Fi News & Record Review |
#27
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
flipper wrote: On Sat, 11 Aug 2007 07:52:37 GMT, Patrick Turner wrote: flipper wrote: On Fri, 10 Aug 2007 10:15:51 GMT, Patrick Turner wrote: reyer wrote: trim Has anyone checked this? http://www.tubelab.com/powerdrive.htm It seems very promising, being able to go into A2 without the grid current problems. I know the powerdrive involves mosfets but, since I really like the sound of a transistorized CCS in the tail of my LTP phase splitter, I am very tempted to try a 845 amp with this powerdrive circuit. trim Cathode biasing in class AB and B PP amps isn't used because the Ia is non-linear, and positive going currents are much higher than negative going currents, so Ck charge up and biase drifts. This does not occur in class A2 amps. The +Ik and -Ik swings are very similar, changeing only 5% if that is what the THD is at clipping even when into A2 operation. So all we need to guard against is the distortion to the drive signal caused by the grid current stalling the voltage amp. In my scheme, the output tube grid bias voltage is kept constant because of the low DC R of the choke and trimming R for the buffer CF. What about positive grid current charging Ck? In class A2, there isn't any non linear increase in Ik to charge Ck as in the case of one tube in an AB PP amp. Grid current occurs over a threshold, that's all, and you just need a direct coupled voltage follower to provide it. +ve grid current dumps into Ik, but only over 0, That's 'linear'? If the grid input resistance is 1k over Eg = 0V, and then Rout of a follower = 100 ohms, distortion is low. If you have a follower with gain into a high Z load = 0.95, ie, 6BQ5/CF, then to change the grid voltage by 50V with no grid current requires a change to CF grid = 52.77V. Internal tube gain is assumed to be about 18, ie, Vk / Vgk = 18. But once you hit grid current the internal gain changes because RL = 1k, or whatever it is. Assume 1k, so internal gain = µ x RL / ( RL + Ra ) = 20 x 1 / ( 1 + 2 ) = 6.7. CF gain = A / ( A + 1 ) = 6.7 / 7.7 = 0.87. So if +52.77V is applied to the CF grid from the point where the output tube grid has reached 0V, instead of getting 50.0V output if the lad was a high load like for the rest of the action, you get 0.87 x 52.77 = 45.9V, and thus ther has been a sag in applied output tube grid signal of 50V - 45.9V = 4.1V, which is 8% distortion. If the bias started at -155, and applied Eg went to 0V, distortion due to grid input resistance is nil, but from 0V to +50V, there is 4.1V error. Total applied +ve goinf grid signal = 155 + 50 = 205V. If the total error is 4.1V, its 2% distortion, and this will be in the output signal. There will also be sag in output signal due to the cut off character of the tube and so tou get sag at tops and bottoms of the waves, so in effect, more 3H and other harmonics are created because of the grid current and finite resistance of the driving CF. The increase in THD due to grid current with 845 is not of much significance because its only in the final 25% of grid voltage swing, and at a level where only drum beats and transients will occur, so the distortion does not matter at all. If the Rout of the CF driver was say 10 ohms, there may only be an added 0.2% added to final THD at clipping. meanwhile at all levels below where grid current begins, the distortion is quite low, and because the driver distortion is made at about the same rate and amplitude there will be quite nice distortion cancelation possible. The CF buffer allows the load seen by the preceding gain tube to be a much higher load than the bias resistance we might otherwise use. Anyway, the CF with Rout of 100ohms would be OK in order to snare a few more watts and have a 31 watt output level rather than say 21 watts. Don't forget the PO quoted is at the anode, and if the OPT has losses of 10% it eats into quoted power levels available. In a Williamson amp with a pair of 6L6 in triode, you get a nice class AB 16 watts, mainly class A, if Pda for both tubes = 50 watts. This can be increased to about 24 watts if A2 operation is used. I tried it once; its what i got, and most people would not want the extra 6SN7 for a CF to achieve this, and they just settle for UL taps, and A1, and presto, 25 watts A1, but Rout of the amp is 3 times higher than triode3. Maybe you mean the bias shift isn't enough to matter. I hope you can see now about what the "bias shift" would be with a CF and a 1k A2 load. The CF needs to be at least a 6BQ5 with gm = around 10mA/V A 1/2 of a 6SN7 has Gm = 2.5mA/V, so CF Rout = 400 ohms, and the distortion powering into 1k load is considerable. The mosfet will have Rout 10 ohms, and be an ideal driver. It would also be quite feasable to use a pair of MJE340 a darlington pair, and the base input resistance is very high, and drivable by most low Ra triodes. The use of a 211 means the extent of grid swing needing input current is a much greater % of total swing. See the Ongaku circuit for what one Japanese man did about it all, and consider my idea of using a higher gm triode with much lower Rout in CF. I'd rather use a tube than some SS device. In my case I want to run two output tubes in parallel, so a pair of EL84 for the CF is called for. Patrick Turner. |
#28
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
Andre Jute wrote: On Aug 11, 12:52 am, Patrick Turner wrote: flipper wrote: In class A2, there isn't any non linear increase in Ik to charge Ck as in the case of one tube in an AB PP amp. Grid current occurs over a threshold, that's all, and you just need a direct coupled voltage follower to provide it. Patrick Turner. First, the threshold isn't necessarily at zero grid volts. It could, depending on tube type and the parameters the individual tube received as its manufacturing inheritance, be several or maybe a handful of volts either sides of no grid bias. Well indeed, grid current can begin when the grid is still at say -5V. The load created by the grid current increase may not be a linear load either. Its basically a dirty business, A2, or AB2, but Kondo didn't think that would spoil the fun in providing the world with Ongaku amps. Second, much more amusing: Oh, what fun it is to design someone else's amp! Since the power tranny is the least of your worries, you could feed the transmitter tube plate 120-180V extra and power a direct-coupled driver by a drop over the power tube's cathode, a circuit first suggested to me by that excellent Australian tubie Anthony Mills. Hmm, I don't recall seeing this man's schematic dealing with this, and so cannot quite see what you mean now. I think you mean driving the 845 grid direct coupled to a small triode anode? yes, could be done if there is sufficient headroom, but making an output tube swing positive, its due to a drive anode being turned off, so the grid current must come from a load resistance, and maybe 50mA is needed. My fave 417A/5842 is good up to 25mA and the 7044 has lower plate voltage and is as solid as a rock. Both of these would drive 211 easy-peasy, and to get the signal elevation when you sub 845, use a step-up IT. Hmm, the IT allows a phase reversal of the drive signal so that the +ve going 845/211 grid drawing current is powered by a triode turning on, rather than turning off. With a step up tranny and two paralleled output tubes I want to use, the driven load seen by a driving small triode anode becomes way too low, and even a 6BQ5 with Ra of 2k will stall when it hits grid current. The follower almost seems like a must. I figured biasing the 211/845 with a choke, and driving the live end from a CF, so instead of IT and all the problems one has with BW and extra iron distortion, the simplest option is the choke with CF drive. For parts economy, arrange the switching so that the primary of the IT is the choke constant current load on the 417A or 7044 plate when the 211 is in the socket, the switch simply unconnecting the secondaries and cutting in a direct route for the signal from the 417A or 7044 plate to the 845 grid. This direct coupling via the cathode also avoids the tricky/ expensive problem of finding high-voltage switches which is otherwise inherent in your desire to switch 845 and 211 on the same chassis. This unconsidered detail nearly tripped me on my "Millennium's End" which, far less ambitiously, merely switched kilovolt tubes between PSE and shunting the plate voltage in SE. I nearly had a heart attack when I cruised the milspec catalogues and saw the prices for high- voltage switches; it turned out I had a handful an old ham, chief engineer at the airport since the days of tubes, gave me with a truckload of other stuff. I was considering a board with solderable links rather than any crazy switch ideas. Maybe you'd like to do a basic schematic of what Mills suggested and of those you just mention. Its possible also to have a small triode drive the 211/845 grid with direct coupling from the small T anode but with a mosfet or second small triode in there as a direct coupled follower. So no IT needed, no choke needed, but to stabilize the whole thing for DC bias is a nightmare. One needs to be able to set the bias by biasing the follower in A2. But I wanted the follower to be choke loaded to the bias voltage... A1 doesn't matter, bias is simple to the output grids and no follower is needed. By golly, I coulda been a production "engineer"! Why, that's how many $ $$$ parts just saved? It is one of the curiosities of tube amps that the simple ZNFB SE DHT amp always ends up costing more than a complicated PP amp (and causes more headscratching too), and that actually has very little to do with the price of the DHT, even at WE/KR prices. To me, various topologies present themselves as viable. I really don't want to have to make an IT in this case. An RC coupled 6BQ5 with 40k load and B+ of say 750V will make a swing of +/- 240V OK. There will be 750V available. Patrick Turner. Andre Jute Visit Jute on Amps at http://members.lycos.co.uk/fiultra/ "wonderfully well written and reasoned information for the tube audio constructor" John Broskie TubeCAD & GlassWare "an unbelievably comprehensive web site containing vital gems of wisdom" Stuart Perry Hi-Fi News & Record Review |
#29
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
Patrick Turner wrote:
Andre Jute wrote: Since the power tranny is the least of your worries, you could feed the transmitter tube plate 120-180V extra and power a direct-coupled driver by a drop over the power tube's cathode, a circuit first suggested to me by that excellent Australian tubie Anthony Mills. Hmm, I don't recall seeing this man's schematic dealing with this, and so cannot quite see what you mean now. Not a published schematic. A personal suggestion in a letter to my mailbox to solve a specific problem I had. I think you mean driving the 845 grid direct coupled to a small triode anode? yes, could be done if there is sufficient headroom, but making an output tube swing positive, its due to a drive anode being turned off, so the grid current must come from a load resistance, and maybe 50mA is needed. Okay, draw it with Anthony and me. First draw the driver, say requiring 150V and 27mA (e.g. a parallel pair of 5687). Now draw a power tube with cathode bias. Connect the top of the cathode resistor of the power tube to the plate of the driver via first a resistor and next a choke. Connect the plate of the driver to the grid of the power tube, no capacitor. From power tube grid drop the grid leak resistor for the power tube to ground. For the sake of somewhere to start, make the grid leak resistor 50K and draw an arrow next to it to show it sinks 3mA. Make the choke 100H or whatever you think necessary. Run a cap to ground from the junction of the resistor on the small tube power line and the plate choke. Scale the resistor between the choke and the cathode of the power tube so that it drops 50V between the cathode and the plate of the driver. So now we know we need 200V and 30mA from the cathode of the power tube to operate the driver circuit. Let us say you want to operate your power tube at 40mA (just for the sake of argument), then the power tube cathode resistor must be 5K to drop the 200V. So the B+ must be the 200V for the driver (including its "load resistor" as you have it), xV for the plate, yV dropped over the trx primary, and it must all add up to 70mA coming and going. Now scale the capacitors to your chosen time constant (the same all over in SE amps, hmm) in relation to the associated resistances and you're done. When you have drawn it you will kick yourself because it is so obvious, the entire driver plate circuit being merely a pair of voltage dividers off the cathode of the power tube. My fave 417A/5842 is good up to 25mA and the 7044 has lower plate voltage and is as solid as a rock. Both of these would drive 211 easy-peasy, and to get the signal elevation when you sub 845, use a step-up IT. Hmm, the IT allows a phase reversal of the drive signal so that the +ve going 845/211 grid drawing current is powered by a triode turning on, rather than turning off. That's digitial thinking. The tube doesn't care where the current comes from, as long as it is there. With a step up tranny and two paralleled output tubes I want to use, the driven load seen by a driving small triode anode becomes way too low, and even a 6BQ5 with Ra of 2k will stall when it hits grid current. The follower almost seems like a must. Maybe. I'm not big on the soft sound of CF. I think, since the choke load above, alias the primary of the IT in at least one of your dual applications on the same chassis, is under your own control, you can make it whatever you want, and the resistor beyond it is to choice too, within reason. I figured biasing the 211/845 with a choke, and driving the live end from a CF, so instead of IT and all the problems one has with BW and extra iron distortion, the simplest option is the choke with CF drive. I'm not arguing, Patrick. I was merely making a semi-lighthearted alternative suggestion, the sort of thing I would consider myself if I could wind my own trannies. For parts economy, arrange the switching so that the primary of the IT is the choke constant current load on the 417A or 7044 plate when the 211 is in the socket, the switch simply unconnecting the secondaries and cutting in a direct route for the signal from the 417A or 7044 plate to the 845 grid. This direct coupling via the cathode also avoids the tricky/ expensive problem of finding high-voltage switches which is otherwise inherent in your desire to switch 845 and 211 on the same chassis. This unconsidered detail nearly tripped me on my "Millennium's End" which, far less ambitiously, merely switched kilovolt tubes between PSE and shunting the plate voltage in SE. I nearly had a heart attack when I cruised the milspec catalogues and saw the prices for high- voltage switches; it turned out I had a handful an old ham, chief engineer at the airport since the days of tubes, gave me with a truckload of other stuff. I was considering a board with solderable links rather than any crazy switch ideas. Wise. Maybe you'd like to do a basic schematic of what Mills suggested and of those you just mention. What I have is a hand-drawn circuit of what I actually did with Anthony's suggestion (different tubes, different values, different execution -- I used silicon pseudo-chokes and the primary of an IT in lieu of a high-Henry audio choke). I can't make head or tail of the scribbles and quite a few milliamp seem to have gone AWOL but Bill managed to build a working amp from it. If you insist, I'll photograph it and e-mail it to you, but you'll spend more time trying to puzzle out my chicken tracks than drawing the description above, which contains all the important principles of the circuit. Andre Jute Visit Jute on Amps at http://members.lycos.co.uk/fiultra/ "wonderfully well written and reasoned information for the tube audio constructor" John Broskie TubeCAD & GlassWare "an unbelievably comprehensive web site containing vital gems of wisdom" Stuart Perry Hi-Fi News & Record Review |
#30
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
Andre Jute wrote: Patrick Turner wrote: Andre Jute wrote: Since the power tranny is the least of your worries, you could feed the transmitter tube plate 120-180V extra and power a direct-coupled driver by a drop over the power tube's cathode, a circuit first suggested to me by that excellent Australian tubie Anthony Mills. Hmm, I don't recall seeing this man's schematic dealing with this, and so cannot quite see what you mean now. Not a published schematic. A personal suggestion in a letter to my mailbox to solve a specific problem I had. I think you mean driving the 845 grid direct coupled to a small triode anode? yes, could be done if there is sufficient headroom, but making an output tube swing positive, its due to a drive anode being turned off, so the grid current must come from a load resistance, and maybe 50mA is needed. Okay, draw it with Anthony and me. First draw the driver, say requiring 150V and 27mA (e.g. a parallel pair of 5687). Now draw a power tube with cathode bias. Connect the top of the cathode resistor of the power tube to the plate of the driver via first a resistor and next a choke. Connect the plate of the driver to the grid of the power tube, no capacitor. From power tube grid drop the grid leak resistor for the power tube to ground. For the sake of somewhere to start, make the grid leak resistor 50K and draw an arrow next to it to show it sinks 3mA. Make the choke 100H or whatever you think necessary. Run a cap to ground from the junction of the resistor on the small tube power line and the plate choke. Scale the resistor between the choke and the cathode of the power tube so that it drops 50V between the cathode and the plate of the driver. So now we know we need 200V and 30mA from the cathode of the power tube to operate the driver circuit. Let us say you want to operate your power tube at 40mA (just for the sake of argument), then the power tube cathode resistor must be 5K to drop the 200V. So the B+ must be the 200V for the driver (including its "load resistor" as you have it), xV for the plate, yV dropped over the trx primary, and it must all add up to 70mA coming and going. Now scale the capacitors to your chosen time constant (the same all over in SE amps, hmm) in relation to the associated resistances and you're done. When you have drawn it you will kick yourself because it is so obvious, the entire driver plate circuit being merely a pair of voltage dividers off the cathode of the power tube. To see if I have understood the principles you have outlined, and applied to an SET circuit using one EL84 in triode as the driver, and one 845 as the output... B+ for 845, 1,500V Normal OPT in the anode circuit. Rk for 845, 11.4k. Ck bypass 470uF. Ia = 65mA. Ik = 65mA, with 35mA flow in 11.4k, and 30mA out thru 4k7 then 100H to EL84 anode. Junction of L and 4k7 bypassed with 470uF to 0V. EL84 is thus fed with 30mA, and Ea = 250V. Its Ek is about 10V, and bypassed with 470uF to 0V. EL84 anode directly connected to 845 grid. The difference in voltage between EL84 anode 845 cathode = -150V, the wanted bias voltage and the EL84 Ea of 250V allows the swing of 150V to be made. I see no purpose for the 50k grid leak at this time. However, for A2 operation, I am not sure how well the realease of stored energy in the 100H would give a wanted extra +50V at 50mA of current. The other concern is that the 845 grid bias voltage is a large -150V, and by the time one adds all the headrooms, the B+ needs to be +1,500V. I was hoping to confine the B+ with two supplies, +800V and -400V, or pehaps +/- 600V, so that under extreme signal voltage movements the OPT insulation is less likely to puncture, and under ordinary ope the OPT insulation isn't under any greater stress than in many other designs using a 600v anode suply voltage. When we examine the normal function of an SE output stage with a choke or OPT, and if we apply a a non linear load, ie, a resistor and diode in series to mimic the action of a load suddlenly becoming a much lower value at above a threshold, then we will see the reducing load cause a sag in gain. Just becuase we have a magnnetic choke in there does not mean the Rout of the stage suddenly reduces to ensure constant gain with a load that suddenly reduces. From what I can see there is a need for a follower to get A2 drive; nothing else will work unless the anode driver is converted to an anode follower with overal gain about similar to a cathode follower gain, ie, say 1.0 instead of say 18. My fave 417A/5842 is good up to 25mA and the 7044 has lower plate voltage and is as solid as a rock. Both of these would drive 211 easy-peasy, and to get the signal elevation when you sub 845, use a step-up IT. Hmm, the IT allows a phase reversal of the drive signal so that the +ve going 845/211 grid drawing current is powered by a triode turning on, rather than turning off. That's digitial thinking. The tube doesn't care where the current comes from, as long as it is there. A normal common cathode R loaded gain stage which directly drives a following grid which draws Ig will have current limiting according to the supply R, and the voltage across it. If you want say 50mA maximum Ig, and the voltage across the RL is say 100V, RL has to be 2k, and when the driven Eg is 200V lower at the Q point bias level, there would be 300V across the 2k, and 150mA is needed for the Ia of the driver. And the driver will still suffer appalling distortion due to its high Ra with respect to the change of RL due to Ig flow being powered. For A1, your direct drive scheme can work fine, bit not for A2, which is where I want to go, because I want all I can get from both 845, and the 211, this needing more A2 drive than the 845. With a step up tranny and two paralleled output tubes I want to use, the driven load seen by a driving small triode anode becomes way too low, and even a 6BQ5 with Ra of 2k will stall when it hits grid current. The follower almost seems like a must. Maybe. I'm not big on the soft sound of CF. Well I didn't think it'd be long before someone said they didn't like CF somewhere because of their alledged signature colouration, and lord knows if anyone agrees about its nature. I have never heard bad things from CF stages; of all the tube stages one can use, the CF is the least blamable IMHO. I think, since the choke load above, alias the primary of the IT in at least one of your dual applications on the same chassis, is under your own control, you can make it whatever you want, and the resistor beyond it is to choice too, within reason. I figured biasing the 211/845 with a choke, and driving the live end from a CF, so instead of IT and all the problems one has with BW and extra iron distortion, the simplest option is the choke with CF drive. I'm not arguing, Patrick. I was merely making a semi-lighthearted alternative suggestion, the sort of thing I would consider myself if I could wind my own trannies. Indeed. The use of the driver tube used as part of the cathode biasing resistance makes sense and I have not considered it previously. For the A2 drive a completely direct coupled follower could be added in. Then there is the problem of bias stabilty. To over come that one a DC feedback path between output tube cathode voltage and driver input grid voltage needs to be established, so 845 Ek rises, driver anode voltage falls, so 845 Ek is prevented from rising past its equilibrium. Its a truly dangerous business to stray from the simplest biasing one can think of, and with any loop NFB in the presence of DC FB there is always the ugly possibility of LF oscillations being extremely difficult to tame. I don't like bias servo networks. In any case one must allow for mains voltage changes. The good SE amp will try to minimise the Pda changes of the output tubes with changeing mains voltage. With fixed bias a mains increase of 5% can tip an output tube into an unsafe operating area. Plain cathode biasing reduces the tendency a lot, and CCS cathode current does even better, but a system where Ik reduces if B+ rises is the best. Hence fixed bias amps should never have really fixed regulated bias but must allow the bias to increase with B+ if the mains rises. One can make a fixed bias supply so that if B+ rises then the bias rises slightly more in proprtionally to ensure Pda remains constant. This way when the mains fall, and hence the B= falls, the the Ia rises. There is onlt one mains voltage that really suits the RL, But however there would be always sufficient PO to not have to worry. For parts economy, arrange the switching so that the primary of the IT is the choke constant current load on the 417A or 7044 plate when the 211 is in the socket, the switch simply unconnecting the secondaries and cutting in a direct route for the signal from the 417A or 7044 plate to the 845 grid. This direct coupling via the cathode also avoids the tricky/ expensive problem of finding high-voltage switches which is otherwise inherent in your desire to switch 845 and 211 on the same chassis. This unconsidered detail nearly tripped me on my "Millennium's End" which, far less ambitiously, merely switched kilovolt tubes between PSE and shunting the plate voltage in SE. I nearly had a heart attack when I cruised the milspec catalogues and saw the prices for high- voltage switches; it turned out I had a handful an old ham, chief engineer at the airport since the days of tubes, gave me with a truckload of other stuff. I was considering a board with solderable links rather than any crazy switch ideas. Wise. Maybe you'd like to do a basic schematic of what Mills suggested and of those you just mention. I think I will because other readers of the group have now lost track of WTF we are discussing. Time is my enemy though. What I have is a hand-drawn circuit of what I actually did with Anthony's suggestion (different tubes, different values, different execution -- I used silicon pseudo-chokes and the primary of an IT in lieu of a high-Henry audio choke). I can't make head or tail of the scribbles and quite a few milliamp seem to have gone AWOL but Bill managed to build a working amp from it. If you insist, I'll photograph it and e-mail it to you, but you'll spend more time trying to puzzle out my chicken tracks than drawing the description above, which contains all the important principles of the circuit. Lend me the time to do a schematic in MS paint with proposed options, and from there it can be mauled to suit anyone. Patrick Turner. Andre Jute Visit Jute on Amps at http://members.lycos.co.uk/fiultra/ "wonderfully well written and reasoned information for the tube audio constructor" John Broskie TubeCAD & GlassWare "an unbelievably comprehensive web site containing vital gems of wisdom" Stuart Perry Hi-Fi News & Record Review |
#31
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
snip,
Maybe you mean the bias shift isn't enough to matter. I hope you can see now about what the "bias shift" would be with a CF and a 1k A2 load. I wasn't speaking of voltage distortion. I was speaking of the fact that even if you had a zero impedance, infinite power, 'perfect' driver, with no THD at all, you still have positive grid current added to Ia non linearly charging Ck. If it's from small enough transients, in both magnitude and duration, then the upset will be minimal but if it's sustained, or the repetition rate often enough, Vk will march up to where it pushes the tube back out of +ve. In class A2 operation of an SE tube, and a music signal there is no large change in Ek with a bypassed Rk where the extent of A2 drive is a small fraction of the peak grid drive required, as is the case of the 211/845 we are considering. Some bias wobbling does occur but only when things are going flat out. The bias drift at CK is a lesser effect in class A2 than in a PP AB amps where the output tubes are allowed to cut off each 1/2 cycle, and an Ek can double when driven with a sine wave, even with AB1 drive, or triple with AB2 drive. Fixed bias overcomes these problems. Fixed bias don't stay fixed though, and one needs active protections against runaway effects. Owners are notorious for not adjusting bias and ignoring meters. But if there isn't any Ik cut off in SE A2 with cathode bias, the bias drift with a sine wave input at clipping should not drastically change as in an AB PPamp. There is always going to be some bias drift in a cathode biased pure class A amp of A1 or A2, simply because of the 2H distortions. But with music and a huge value for Ck the drift is SFA. Patrick Turner The CF needs to be at least a 6BQ5 with gm = around 10mA/V A 1/2 of a 6SN7 has Gm = 2.5mA/V, so CF Rout = 400 ohms, and the distortion powering into 1k load is considerable. The mosfet will have Rout 10 ohms, and be an ideal driver. It would also be quite feasable to use a pair of MJE340 a darlington pair, and the base input resistance is very high, and drivable by most low Ra triodes. The use of a 211 means the extent of grid swing needing input current is a much greater % of total swing. See the Ongaku circuit for what one Japanese man did about it all, and consider my idea of using a higher gm triode with much lower Rout in CF. I'd rather use a tube than some SS device. In my case I want to run two output tubes in parallel, so a pair of EL84 for the CF is called for. Patrick Turner. |
#32
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
On Sun, 12 Aug 2007 04:04:56 GMT, Patrick Turner
wrote: I see no purpose for the 50k grid leak at this time. It's by definition extraneous. However, for A2 operation, I am not sure how well the realease of stored energy in the 100H would give a wanted extra +50V at 50mA of current. That ain't happenin'. The larger problem with the mystery e-mailer's proposed topo is that output valve can never be pulled up into A2, unless I've missed something... And that could never happen. Arf! Thanks, as always, Chris Hornbeck "It's just this little Chromium Switch. You people are SO superstitious." |
#33
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
Chris Hornbeck wrote: On Sun, 12 Aug 2007 04:04:56 GMT, Patrick Turner wrote: I see no purpose for the 50k grid leak at this time. It's by definition extraneous. However, for A2 operation, I am not sure how well the realease of stored energy in the 100H would give a wanted extra +50V at 50mA of current. That ain't happenin'. The larger problem with the mystery e-mailer's proposed topo is that output valve can never be pulled up into A2, unless I've missed something... And that could never happen. Arf! I think you confused the cornflakes this morning with those dog biscuits, "arf" indeed. :-) But if you have a direct coupled voltage follower of some sort then the topology works just fine for A2. My only objection with the Mills connection with 845 is the higher B+ voltage total required, and I want to run two output tubes, so a larger than wanted voltage is across the 845 Rk. Then there is the problem of bias stability. Patrick Turner. Thanks, as always, Chris Hornbeck "It's just this little Chromium Switch. You people are SO superstitious." |
#34
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
There is always going to be some bias drift in a cathode biased pure class A amp of A1 or A2, simply because of the 2H distortions. But with music and a huge value for Ck the drift is SFA. Patrick Turner Ok, So you're counting on only short term transients pushing it into A2. Exactly. Patrick Turner. |
#35
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
On Aug 11, 9:04 pm, Patrick Turner wrote:
Andre Jute wrote: Patrick Turner wrote: Andre Jute wrote: [[[snip]]] To see if I have understood the principles you have outlined I've drawn the generic circuit to avoid long discussion with many misunderstandings... See http://members.lycos.co.uk/fiultra/D...20Circuits.htm HTH. Andre Jute Visit Jute on Amps at http://members.lycos.co.uk/fiultra/ "wonderfully well written and reasoned information for the tube audio constructor" John Broskie TubeCAD & GlassWare "an unbelievably comprehensive web site containing vital gems of wisdom" Stuart Perry Hi-Fi News & Record Review |
#36
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
flipper wrote:
On Sat, 11 Aug 2007 13:54:40 -0700, Andre Jute wrote: trim. What I have is a hand-drawn circuit of what I actually did with Anthony's suggestion (different tubes, different values, different execution -- I used silicon pseudo-chokes and the primary of an IT in lieu of a high-Henry audio choke). I can't make head or tail of the scribbles and quite a few milliamp seem to have gone AWOL but Bill managed to build a working amp from it. I'd be interested to hear more details on the silicon pseudo choke. Judging from the brief description I guess it bucks into the secondary? I was talking about an audio-band choke as a plate load. A "silicon pseudo-choke" is merely a constant current load (more commonly called a constant current source) made with transistors; it is a surprisingly good substitute for a real but often unobtanium plate choke of the right spec. I'm sure you know how to make a CCS but if you, or anyone else, don't I can draw a schemo and put it on my netsite. For this sort of application I recommend the combo of the 2N2907 and MJE350 (well, there isn't much else capable of working with tubes). HTH. Andre Jute Visit Jute on Amps at http://members.lycos.co.uk/fiultra/ "wonderfully well written and reasoned information for the tube audio constructor" John Broskie TubeCAD & GlassWare "an unbelievably comprehensive web site containing vital gems of wisdom" Stuart Perry Hi-Fi News & Record Review |
#37
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
Andre Jute wrote: On Aug 11, 9:04 pm, Patrick Turner wrote: Andre Jute wrote: Patrick Turner wrote: Andre Jute wrote: [[[snip]]] To see if I have understood the principles you have outlined I've drawn the generic circuit to avoid long discussion with many misunderstandings... See http://members.lycos.co.uk/fiultra/D...20Circuits.htm Thanks for that, I think I got what you described ok. I just sent you a schematic of a full amp using the Mills idea suited for EL84 driver, and 845, and with DC FB to stabilise biasing. Patrick Turner. HTH. Andre Jute Visit Jute on Amps at http://members.lycos.co.uk/fiultra/ "wonderfully well written and reasoned information for the tube audio constructor" John Broskie TubeCAD & GlassWare "an unbelievably comprehensive web site containing vital gems of wisdom" Stuart Perry Hi-Fi News & Record Review |
#38
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
flipper wrote:
On Sun, 12 Aug 2007 17:43:09 -0700, Andre Jute wrote: flipper wrote: On Sat, 11 Aug 2007 13:54:40 -0700, Andre Jute wrote: trim. What I have is a hand-drawn circuit of what I actually did with Anthony's suggestion (different tubes, different values, different execution -- I used silicon pseudo-chokes and the primary of an IT in lieu of a high-Henry audio choke). I can't make head or tail of the scribbles and quite a few milliamp seem to have gone AWOL but Bill managed to build a working amp from it. I'd be interested to hear more details on the silicon pseudo choke. Judging from the brief description I guess it bucks into the secondary? I was talking about an audio-band choke as a plate load. A "silicon pseudo-choke" is merely a constant current load (more commonly called a constant current source) made with transistors; it is a surprisingly good substitute for a real but often unobtanium plate choke of the right spec. I'm sure you know how to make a CCS but if you, or anyone else, don't I can draw a schemo and put it on my netsite. For this sort of application I recommend the combo of the 2N2907 and MJE350 (well, there isn't much else capable of working with tubes). Oh, yeah, sure I know the CCS although, here lately, I've become rather enamored with bootstrapping. AKA SRPP. A great circuit component if you're more interested in music than in the fact that it isn't really a CCS. I misunderstood. I thought you had something working into the mentioned IT, especially in light of the circuit being discussed because a CCS can't swing over it's supply rail. Gee, now you tell me. Patrick might think I'm trying to discourage him from using 845 in A2. (Is there an emoticon for one's tongue making a bulge in one's cheek?) Oh, well, oops. English isn't my first language, not by several. The misunderstanding is definitely my fault. Andre Jute Visit Jute on Amps at http://members.lycos.co.uk/fiultra/ "wonderfully well written and reasoned information for the tube audio constructor" John Broskie TubeCAD & GlassWare "an unbelievably comprehensive web site containing vital gems of wisdom" Stuart Perry Hi-Fi News & Record Review |
#39
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
Andre Jute wrote: flipper wrote: On Sun, 12 Aug 2007 17:43:09 -0700, Andre Jute wrote: flipper wrote: On Sat, 11 Aug 2007 13:54:40 -0700, Andre Jute wrote: trim. What I have is a hand-drawn circuit of what I actually did with Anthony's suggestion (different tubes, different values, different execution -- I used silicon pseudo-chokes and the primary of an IT in lieu of a high-Henry audio choke). I can't make head or tail of the scribbles and quite a few milliamp seem to have gone AWOL but Bill managed to build a working amp from it. I'd be interested to hear more details on the silicon pseudo choke. Judging from the brief description I guess it bucks into the secondary? I was talking about an audio-band choke as a plate load. A "silicon pseudo-choke" is merely a constant current load (more commonly called a constant current source) made with transistors; it is a surprisingly good substitute for a real but often unobtanium plate choke of the right spec. I'm sure you know how to make a CCS but if you, or anyone else, don't I can draw a schemo and put it on my netsite. For this sort of application I recommend the combo of the 2N2907 and MJE350 (well, there isn't much else capable of working with tubes). Oh, yeah, sure I know the CCS although, here lately, I've become rather enamored with bootstrapping. AKA SRPP. A great circuit component if you're more interested in music than in the fact that it isn't really a CCS. I misunderstood. I thought you had something working into the mentioned IT, especially in light of the circuit being discussed because a CCS can't swing over it's supply rail. Gee, now you tell me. Patrick might think I'm trying to discourage him from using 845 in A2. (Is there an emoticon for one's tongue making a bulge in one's cheek?) I've just done the design and all the details for an 845 amp using one tube based on any 845 one might find, and using Ea = 1,015V and Ia = 66mA, for Pda = 67W. I have also done the Load lines for 5 loads between 5k and 24k and for both A1 and A2 operation, and completed the LL graph with a PO vs RL graph to show the differences in PO results with A1 and A2, at clipping, and from this you see where the increase in PO is possible with A2. All certainly isn't as anyone might have assumed it to have been. Bootstrapping is not a pretty picture if its to allow a drove stage to create an unclipped drive voltage such as in a McIntosh with 12BH7 in the OP stage. Its PFB and distortion gets fed back, and increases... But BSing is OK in a signal amp where it is derived from SRPP. SRPP is where the Rk top and bottom is about the same value. But where Rk of the top triode = say 10 times that of the bottom, then its BSing. If a j-fet is used for the top tube Rk, then you get a CCS. The load seen by the bottom triode = the gain of the top tube x its Rk, so if Rk is a very high value, then indeed the bottom triode operates as if with a CCS load apart perhaps from some biasing resistors for the top tube grid. Its a great preamp circuit. But I like just using a plain old single MJE350 for a CCS to feed a twin triode with both halves paralleled, and this can easily drive a volume control of 50k with low THD and great sound. This same CCS plus twin triode makes the best single triode input stage in a power amp. There are few reactive elements to make it difficult to apply NFB if you want to. Tongue is now against cheek. Didja get the Mills app 845 amp schema I sent? Patrick Turner. Oh, well, oops. English isn't my first language, not by several. The misunderstanding is definitely my fault. Andre Jute Visit Jute on Amps at http://members.lycos.co.uk/fiultra/ "wonderfully well written and reasoned information for the tube audio constructor" John Broskie TubeCAD & GlassWare "an unbelievably comprehensive web site containing vital gems of wisdom" Stuart Perry Hi-Fi News & Record Review |
#40
Posted to rec.audio.tubes
|
|||
|
|||
211 and 845 operating Ea/Ia bias points.
On Wed, 08 Aug 2007 12:32:18 GMT, Patrick Turner
wrote: The 845 amps I have been strugling to find time to build are approaching a stage where I must decide on the actual working points that would be most suitable for both 211 and 845 including class A2 operation, by means of cathode follower buffer choke loaded to the fixed bias supply. So far having Ea at 1,000V and Ea at 80mA with RL = 12.5k could seem OK but it means Pda = 80 watts, and only KR845 or KR211 look able to take this idling condition, one reason being that the KR heaters are 10V at 1A instead of the RCA GE use of 10V at 3.3A. RCA data gives Pda = 67 watts, and if Pd-filament = 33 watts, the total tube Pd = 100 watts Oh, I forgot to say that the RCA and GE type 211's and 845's will run forever at 100 watts plate dissipation. It's not a problem. Other than the very first 1920's versions, that's their CCS rating. Thanks, as always, Chris Hornbeck "It's just this little Chromium Switch. You people are SO superstitious." |
Reply |
|
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
Which operating system | Pro Audio | |||
Operating Manual | Car Audio | |||
Cue Points Don't Appear on CD | Pro Audio | |||
Operating temperature of CD players? | General | |||
Operating temperature of CD players? | Tech |