Home |
Search |
Today's Posts |
#1
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
I have been doing some tests on a mu follower based on an ECC88 (Morgan
Jones circuit) and I get the expected gain close to mu but the distortion at 1KHz is 0.5% at +5dBV and 5% at +25dBV. Jones claims the mu follower is low distortion because it minimises the effects of changing ra. Are these numbers about right? Cheers Ian |
#2
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
"Ian Bell" wrote in message ... I have been doing some tests on a mu follower based on an ECC88 (Morgan Jones circuit) and I get the expected gain close to mu but the distortion at 1KHz is 0.5% at +5dBV and 5% at +25dBV. Jones claims the mu follower is low distortion because it minimises the effects of changing ra. Are these numbers about right? Hello Ian, I have used the 6CG7 for my mu-follower projects.- It is said to be the 9 pin equivalent of the 6SN7, a tube which I like very much. At 1kHz, I have measured 0.025% to 0.03% at +6dBV Cheers Iain |
#3
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Iain Churches wrote:
"Ian Bell" wrote in message ... I have been doing some tests on a mu follower based on an ECC88 (Morgan Jones circuit) and I get the expected gain close to mu but the distortion at 1KHz is 0.5% at +5dBV and 5% at +25dBV. Jones claims the mu follower is low distortion because it minimises the effects of changing ra. Are these numbers about right? Hello Ian, I have used the 6CG7 for my mu-follower projects.- It is said to be the 9 pin equivalent of the 6SN7, a tube which I like very much. At 1kHz, I have measured 0.025% to 0.03% at +6dBV Interesting and more like what I would have expected. Can you post a schematic?? Cheers Ian Cheers Iain |
#4
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Ian Bell wrote: I have been doing some tests on a mu follower based on an ECC88 (Morgan Jones circuit) and I get the expected gain close to mu but the distortion at 1KHz is 0.5% at +5dBV and 5% at +25dBV. Jones claims the mu follower is low distortion because it minimises the effects of changing ra. Are these numbers about right? Cheers Ian The distortion you get with triodes becomes lower as the load approaches a CCS. For lowest possible THD with ECC88 in µ-follower mode, try applying fixed bias to the top triode follower via a 1M to say +150V. Then have 5k R between top cathode and bottom triode anode which you couple to the top grid via say 0.22uF. Ia could be 7mA. B+ needs to be about +270V. Suppose the load to the outside world is 50k from the top tube cathode, then the top triode has internal gain = 28 approx so the 5k between the triodes is seen by the bottom tube as about 28 x 5k = 140k in parallel with 1M bias R. So the gain of the bottom triode appraoches the triode µ of about 33, and for 10Vrms output expect less than 0.3% THD. You should be able to plug a 6CG7 in to replace the ECC88 and see the THD drop even further. Of course at 1Vo, expect 0.03%, and at 0.1Vo, expect 0.003%. Blimey, low enough ain't it? And all without any loop NFB! The µ-follower is a nice circuit to use; high Z input, low Z output and much lower THD than if you have a resistance load on the bottom gain triode. If the bottom triode has Ea = +120V, then there would be 150V from anode to B+ = 270V. The load for 7mA would be 22k approx, which is only about 7 x Ra of the triode at 7mA. When the load becomes 140k as with the µ-follower, the load = 46 x Ra, and THD has become lower. Patrick Turner. |
#5
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Ian Bell wrote: Iain Churches wrote: "Ian Bell" wrote in message ... I have been doing some tests on a mu follower based on an ECC88 (Morgan Jones circuit) and I get the expected gain close to mu but the distortion at 1KHz is 0.5% at +5dBV and 5% at +25dBV. Jones claims the mu follower is low distortion because it minimises the effects of changing ra. Are these numbers about right? Hello Ian, I have used the 6CG7 for my mu-follower projects.- It is said to be the 9 pin equivalent of the 6SN7, a tube which I like very much. At 1kHz, I have measured 0.025% to 0.03% at +6dBV Interesting and more like what I would have expected. Can you post a schematic?? Cheers Try a read of the preamp pages at my website. For CCS loaded triode cap coupled to 100k volume control, and CF buffer after volume pot see http://www.turneraudio.com.au/preamp...+psu-2005.html For µ-follower topologies that work see http://www.turneraudio.com.au/preamp...ated-2006.html Go to about 1/2 way down the page. Patrick Turner. Ian Cheers Iain |
#6
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Patrick Turner wrote:
Ian Bell wrote: I have been doing some tests on a mu follower based on an ECC88 (Morgan Jones circuit) and I get the expected gain close to mu but the distortion at 1KHz is 0.5% at +5dBV and 5% at +25dBV. Jones claims the mu follower is low distortion because it minimises the effects of changing ra. Are these numbers about right? Cheers Ian The distortion you get with triodes becomes lower as the load approaches a CCS. For lowest possible THD with ECC88 in µ-follower mode, try applying fixed bias to the top triode follower via a 1M to say +150V. Then have 5k R between top cathode and bottom triode anode which you couple to the top grid via say 0.22uF. Ia could be 7mA. B+ needs to be about +270V. Interesting, I'll try that. Does the fixed bias have any direct effect on distortion i.e is it any better than self biassing the top triode? Cheers Ian Suppose the load to the outside world is 50k from the top tube cathode, then the top triode has internal gain = 28 approx so the 5k between the triodes is seen by the bottom tube as about 28 x 5k = 140k in parallel with 1M bias R. So the gain of the bottom triode appraoches the triode µ of about 33, and for 10Vrms output expect less than 0.3% THD. You should be able to plug a 6CG7 in to replace the ECC88 and see the THD drop even further. Of course at 1Vo, expect 0.03%, and at 0.1Vo, expect 0.003%. Blimey, low enough ain't it? And all without any loop NFB! The µ-follower is a nice circuit to use; high Z input, low Z output and much lower THD than if you have a resistance load on the bottom gain triode. If the bottom triode has Ea = +120V, then there would be 150V from anode to B+ = 270V. The load for 7mA would be 22k approx, which is only about 7 x Ra of the triode at 7mA. When the load becomes 140k as with the µ-follower, the load = 46 x Ra, and THD has become lower. Patrick Turner. |
#7
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Ian Bell wrote
I have been doing some tests on a mu follower based on an ECC88 (Morgan Jones circuit) and I get the expected gain close to mu but the distortion at 1KHz is 0.5% at +5dBV and 5% at +25dBV. Jones claims the mu follower is low distortion because it minimises the effects of changing ra. Are these numbers about right? You measured. Presumably considerably lower than one triode at the same operating point and with the same HT supply but a resistor load to suit. Using the second triode on top increases the effective anode load, but is not particularly close to a CCS, so mu is not followed perfectly. That, plus the extra distortion arising from the top valve operating as a cathode follower (you don't say what load it was driving when you measured it), more or less, should account for what you have measured. For a version that measures better, try using a MOSFET on top. Alan Kimmel seems to have taken down his site but his stuff is still around, e.g. http://www7.taosnet.com/f10/mustage.html Ian |
#8
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Ian Iveson wrote:
Ian Bell wrote I have been doing some tests on a mu follower based on an ECC88 (Morgan Jones circuit) and I get the expected gain close to mu but the distortion at 1KHz is 0.5% at +5dBV and 5% at +25dBV. Jones claims the mu follower is low distortion because it minimises the effects of changing ra. Are these numbers about right? You measured. I did. Presumably considerably lower than one triode at the same operating point and with the same HT supply but a resistor load to suit. Using the second triode on top increases the effective anode load, but is not particularly close to a CCS, so mu is not followed perfectly. Doesn't matter does it. Provided the dynamic plate load of the bottom tube is ~ 50*ra you have pretty much reached the distortion minima. That, plus the extra distortion arising from the top valve operating as a cathode follower (you don't say what load it was driving when you measured it), 100K more or less, should account for what you have measured. CF distortion should be negligible compared to the figures I am getting. So are you saying a mu follower is not low distortion? For a version that measures better, try using a MOSFET on top. Alan Kimmel seems to have taken down his site but his stuff is still around, e.g. Seems to me the improvement would be marginal. Cheers Ian http://www7.taosnet.com/f10/mustage.html Ian |
#9
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Ian Bell wrote:
I have been doing some tests on a mu follower based on an ECC88 (Morgan Jones circuit) and I get the expected gain close to mu but the distortion at 1KHz is 0.5% at +5dBV and 5% at +25dBV. Jones claims the mu follower is low distortion because it minimises the effects of changing ra. Are these numbers about right? You measured. I did. Presumably considerably lower than one triode at the same operating point and with the same HT supply but a resistor load to suit. Using the second triode on top increases the effective anode load, but is not particularly close to a CCS, so mu is not followed perfectly. Doesn't matter does it. Provided the dynamic plate load of the bottom tube is ~ 50*ra you have pretty much reached the distortion minima. Eh? What doesn't matter? To get the same operating point with the same HT but a resistive anode load would mean using a much lower load resistance than that effectively presented by your mu-stage...perhaps 2-3 times ra. Where does your 50 come from? To return to your original question, you could obtain the data sheet for an ECC88 and draw a horizontal load line through your operating point on the anode characteristics chart. That would allow a good estimate of what distortion to expect. If your actual distortion is much different, then you've made a mistake somewhere, or your ECC88 is misbehaving. That, plus the extra distortion arising from the top valve operating as a cathode follower (you don't say what load it was driving when you measured it), 100K more or less, should account for what you have measured. CF distortion should be negligible compared to the figures I am getting. Assuming a sensible operating point, and now knowing the driven load, then at 1kHz CF distortion would be numerically small, yes. So are you saying a mu follower is not low distortion It's considerably lower than when the same valve is used at the same operating point, with the same HT, and with a resistive anode load, particularly when driving a low impedance. If it follows mu, then it is as linear as mu is constant. You might give some thought to how to pick a valve and operating point that is likely to have the least mu variation. For a version that measures better, try using a MOSFET on top. Alan Kimmel seems to have taken down his site but his stuff is still around, e.g. http://www7.taosnet.com/f10/mustage.html Seems to me the improvement would be marginal. Depends on what you consider to be marginal. Perhaps it would be worth listening. Ian |
#10
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Ian Iveson wrote:
Ian Bell wrote: I have been doing some tests on a mu follower based on an ECC88 (Morgan Jones circuit) and I get the expected gain close to mu but the distortion at 1KHz is 0.5% at +5dBV and 5% at +25dBV. Jones claims the mu follower is low distortion because it minimises the effects of changing ra. Are these numbers about right? You measured. I did. Presumably considerably lower than one triode at the same operating point and with the same HT supply but a resistor load to suit. Using the second triode on top increases the effective anode load, but is not particularly close to a CCS, so mu is not followed perfectly. Doesn't matter does it. Provided the dynamic plate load of the bottom tube is ~ 50*ra you have pretty much reached the distortion minima. Eh? What doesn't matter? Doesn't matter that it is not particularly close to a CCS. Provided the mu stage achieves an effect Rl 50*ra. To get the same operating point with the same HT but a resistive anode load would mean using a much lower load resistance than that effectively presented by your mu-stage...perhaps 2-3 times ra. Correct, I think we are just talking at cross purposes. I was just saying that if the mu stage achieves and effective Rl 50*ra then the distortion due to ra changes should be insignificant. I agree it should be much better than a pure resistive load. Where does your 50 come from? To return to your original question, you could obtain the data sheet for an ECC88 and draw a horizontal load line through your operating point on the anode characteristics chart. That would allow a good estimate of what distortion to expect. If your actual distortion is much different, then you've made a mistake somewhere, or your ECC88 is misbehaving. That, plus the extra distortion arising from the top valve operating as a cathode follower (you don't say what load it was driving when you measured it), 100K more or less, should account for what you have measured. CF distortion should be negligible compared to the figures I am getting. Assuming a sensible operating point, and now knowing the driven load, then at 1kHz CF distortion would be numerically small, yes. So are you saying a mu follower is not low distortion It's considerably lower than when the same valve is used at the same operating point, with the same HT, and with a resistive anode load, particularly when driving a low impedance. If it follows mu, then it is as linear as mu is constant. You might give some thought to how to pick a valve and operating point that is likely to have the least mu variation. So, from your experience, are the distortion figures I am seeing unusually high for a mu stage? For a version that measures better, try using a MOSFET on top. Alan Kimmel seems to have taken down his site but his stuff is still around, e.g. http://www7.taosnet.com/f10/mustage.html Seems to me the improvement would be marginal. Depends on what you consider to be marginal. Perhaps it would be worth listening. Indeed, but at the moment I have 5% distortion at +25dBV! |Cheers Ian Ian |
#11
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Ian Bell wrote: Patrick Turner wrote: Ian Bell wrote: I have been doing some tests on a mu follower based on an ECC88 (Morgan Jones circuit) and I get the expected gain close to mu but the distortion at 1KHz is 0.5% at +5dBV and 5% at +25dBV. Jones claims the mu follower is low distortion because it minimises the effects of changing ra. Are these numbers about right? Cheers Ian The distortion you get with triodes becomes lower as the load approaches a CCS. For lowest possible THD with ECC88 in µ-follower mode, try applying fixed bias to the top triode follower via a 1M to say +150V. Then have 5k R between top cathode and bottom triode anode which you couple to the top grid via say 0.22uF. Ia could be 7mA. B+ needs to be about +270V. Interesting, I'll try that. Does the fixed bias have any direct effect on distortion i.e is it any better than self biassing the top triode? The fixed bias R of 1M has negligible loading effect on the bottom gain triode. The larger the R between top cathode and bottom anode then the less distortion you get because the load on the bottom triode becomes closer to a CCS when THD is then a minimum. You can have a self biasing Rk for the top tube same as the Rk for the bottom tube. Between the top tube's rk and the bottom tube anode, you can have the 5k, and bias the top grid with 1M from the junction of top Rk and top of the 5k. Then you have to couple the bottom anode to top grid with a cap. But I prefer the fixed bias arrangement. With a regulated or well filtered bias voltage, the output of the top cathode stays more stable at dc, and you don't get a wandery output signal. Patrick Turner. Cheers Ian Suppose the load to the outside world is 50k from the top tube cathode, then the top triode has internal gain = 28 approx so the 5k between the triodes is seen by the bottom tube as about 28 x 5k = 140k in parallel with 1M bias R. So the gain of the bottom triode appraoches the triode µ of about 33, and for 10Vrms output expect less than 0.3% THD. You should be able to plug a 6CG7 in to replace the ECC88 and see the THD drop even further. Of course at 1Vo, expect 0.03%, and at 0.1Vo, expect 0.003%. Blimey, low enough ain't it? And all without any loop NFB! The µ-follower is a nice circuit to use; high Z input, low Z output and much lower THD than if you have a resistance load on the bottom gain triode. If the bottom triode has Ea = +120V, then there would be 150V from anode to B+ = 270V. The load for 7mA would be 22k approx, which is only about 7 x Ra of the triode at 7mA. When the load becomes 140k as with the µ-follower, the load = 46 x Ra, and THD has become lower. Patrick Turner. |
#12
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Ian Iveson wrote: Ian Bell wrote: I have been doing some tests on a mu follower based on an ECC88 (Morgan Jones circuit) and I get the expected gain close to mu but the distortion at 1KHz is 0.5% at +5dBV and 5% at +25dBV. Jones claims the mu follower is low distortion because it minimises the effects of changing ra. Are these numbers about right? You measured. I did. Presumably considerably lower than one triode at the same operating point and with the same HT supply but a resistor load to suit. Using the second triode on top increases the effective anode load, but is not particularly close to a CCS, so mu is not followed perfectly. Doesn't matter does it. Provided the dynamic plate load of the bottom tube is ~ 50*ra you have pretty much reached the distortion minima. Eh? What doesn't matter? To get the same operating point with the same HT but a resistive anode load would mean using a much lower load resistance than that effectively presented by your mu-stage...perhaps 2-3 times ra. Where does your 50 come from? To return to your original question, you could obtain the data sheet for an ECC88 and draw a horizontal load line through your operating point on the anode characteristics chart. That would allow a good estimate of what distortion to expect. If your actual distortion is much different, then you've made a mistake somewhere, or your ECC88 is misbehaving. Hmm, usually when you measure a tube's distortion with a CCS load which might be 500 times greater than Ra, the 2H is much less than what you'd see with a line across a set of ancient plotted data curves which are not very accurate. Think of the rudimentary gear used in 1955. All tubed, and riddled with more distortion than what they may have been measuring. The curves are beautifully important guide, but that's about all, and to really know you must get away from books and PC, build a circuit and measure and tweak it until you get the lowest THD. That, plus the extra distortion arising from the top valve operating as a cathode follower (you don't say what load it was driving when you measured it), 100K more or less, should account for what you have measured. CF distortion should be negligible compared to the figures I am getting. Assuming a sensible operating point, and now knowing the driven load, then at 1kHz CF distortion would be numerically small, yes. So are you saying a mu follower is not low distortion It's considerably lower than when the same valve is used at the same operating point, with the same HT, and with a resistive anode load, particularly when driving a low impedance. If it follows mu, then it is as linear as mu is constant. You might give some thought to how to pick a valve and operating point that is likely to have the least mu variation. Just don't confuse the µ-follower with a SRPP where the 2H is supposed to become lowest when equal changes in Ia for top and bottom tubes occur for a particular load value and Rk used. The µ-follower goes beyond the SRPP and uses a cathode follower to provide a very high load to the gain tube and approaching CCS when THD becomes minimal. Patrick Turner. For a version that measures better, try using a MOSFET on top. Alan Kimmel seems to have taken down his site but his stuff is still around, e.g. http://www7.taosnet.com/f10/mustage.html Seems to me the improvement would be marginal. Depends on what you consider to be marginal. Perhaps it would be worth listening. Ian |
#13
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Patrick Turner wrote
To return to your original question, you could obtain the data sheet for an ECC88 and draw a horizontal load line through your operating point on the anode characteristics chart. That would allow a good estimate of what distortion to expect. If your actual distortion is much different, then you've made a mistake somewhere, or your ECC88 is misbehaving. Hmm, usually when you measure a tube's distortion with a CCS load which might be 500 times greater than Ra, the 2H is much less than what you'd see with a line across a set of ancient plotted data curves which are not very accurate. Think of the rudimentary gear used in 1955. All tubed, and riddled with more distortion than what they may have been measuring. The curves are beautifully important guide, but that's about all, and to really know you must get away from books and PC, build a circuit and measure and tweak it until you get the lowest THD. Fair enough, Patrick, but the question was about whether the measured distortion of an actual circuit is "about right". It seems reasonable therefore to compare the actual results with what theory might expect. An examination of published valve characteristics should offer a workable approximation. A glance at a 6DJ8 datasheet on screen suggests to me that 0.5% distortion is a pretty good result. It may possibly be reduced for a particular signal level by messing with the operating point of the bottom valve, but then it would require frequent readjustment to follow the sweet spot as the valve ages. I don't measure distortion. I haven't got good enough instruments and can't be bothered anyway. Seems too late by then. I just check that my products are working properly as designed, and listen. I don't make stuff to sell. That, plus the extra distortion arising from the top valve operating as a cathode follower (you don't say what load it was driving when you measured it), 100K more or less, should account for what you have measured. CF distortion should be negligible compared to the figures I am getting. Assuming a sensible operating point, and now knowing the driven load, then at 1kHz CF distortion would be numerically small, yes. So are you saying a mu follower is not low distortion It's considerably lower than when the same valve is used at the same operating point, with the same HT, and with a resistive anode load, particularly when driving a low impedance. If it follows mu, then it is as linear as mu is constant. You might give some thought to how to pick a valve and operating point that is likely to have the least mu variation. Just don't confuse the µ-follower with a SRPP where the 2H is supposed to become lowest when equal changes in Ia for top and bottom tubes occur for a particular load value and Rk used. I wasn't confused. My mu-stages use MOSFETs on top, and I hope they aren't into any of that push-pull hanky-panky. I always forget what SRPP stands for...there's a little cluster of acronyms applicable to totem-pole configurations that are easily confused. I tend to remember the circuits but not their names. But I'm OK with mu-stage. The µ-follower goes beyond the SRPP and uses a cathode follower to provide a very high load to the gain tube and approaching CCS when THD becomes minimal. When THD due to the bottom valve is minimised, yes. However, it is possible that circuit THD of an SRPP stage could be less, because of the active cancelation of some distortion products. It could also be that the spectrum produced by an SRPP stage would be more euphonic in the context of particular preamp circuit. AFAIK ppl don't generally build these variations in a search for minimum THD. Ian |
#14
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Ian Iveson wrote:
Patrick Turner wrote To return to your original question, you could obtain the data sheet for an ECC88 and draw a horizontal load line through your operating point on the anode characteristics chart. That would allow a good estimate of what distortion to expect. If your actual distortion is much different, then you've made a mistake somewhere, or your ECC88 is misbehaving. Hmm, usually when you measure a tube's distortion with a CCS load which might be 500 times greater than Ra, the 2H is much less than what you'd see with a line across a set of ancient plotted data curves which are not very accurate. Think of the rudimentary gear used in 1955. All tubed, and riddled with more distortion than what they may have been measuring. The curves are beautifully important guide, but that's about all, and to really know you must get away from books and PC, build a circuit and measure and tweak it until you get the lowest THD. Fair enough, Patrick, but the question was about whether the measured distortion of an actual circuit is "about right". It seems reasonable therefore to compare the actual results with what theory might expect. An examination of published valve characteristics should offer a workable approximation. A glance at a 6DJ8 datasheet on screen suggests to me that 0.5% distortion is a pretty good result. It may possibly be reduced for a particular signal level by messing with the operating point of the bottom valve, but then it would require frequent readjustment to follow the sweet spot as the valve ages. I have made a much more radical change. I redesigned the ECC88 mu follower with an increased standing current of 5mA (up from Morgan Jones 1.34mA). The result is 1KHz distortion at 20V rms is 0.64% and at 2V rms it is down to 0.054%. I also built the Morgan Jones ECC88 cascode stage and also a version with 5mA standing current and again the higher current version had lower distortion. Cheers Ian |
#15
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Ian Bell wrote
To return to your original question, you could obtain the data sheet for an ECC88 and draw a horizontal load line through your operating point on the anode characteristics chart. That would allow a good estimate of what distortion to expect. If your actual distortion is much different, then you've made a mistake somewhere, or your ECC88 is misbehaving. Hmm, usually when you measure a tube's distortion with a CCS load which might be 500 times greater than Ra, the 2H is much less than what you'd see with a line across a set of ancient plotted data curves which are not very accurate. Think of the rudimentary gear used in 1955. All tubed, and riddled with more distortion than what they may have been measuring. The curves are beautifully important guide, but that's about all, and to really know you must get away from books and PC, build a circuit and measure and tweak it until you get the lowest THD. Fair enough, Patrick, but the question was about whether the measured distortion of an actual circuit is "about right". It seems reasonable therefore to compare the actual results with what theory might expect. An examination of published valve characteristics should offer a workable approximation. A glance at a 6DJ8 datasheet on screen suggests to me that 0.5% distortion is a pretty good result. It may possibly be reduced for a particular signal level by messing with the operating point of the bottom valve, but then it would require frequent readjustment to follow the sweet spot as the valve ages. I have made a much more radical change. I redesigned the ECC88 mu follower with an increased standing current of 5mA (up from Morgan Jones 1.34mA). The result is 1KHz distortion at 20V rms is 0.64% and at 2V rms it is down to 0.054%. Just wondering if it follows that the lower harmonic distortion will contain a higher proportion of higher-order harmonics. Some say it's the spectrum that counts. Also to what extent we might be concerned about distortion modulation. I know that artifacts are less noticeable if they are of constant amplitude with respect to signal ground, such as hum or hiss might be. Are they also less noticeable if they are constant with respect to signal amplitude and/or frequency? If one amplifier distorts at a constant x% of 2H, y% of 3H, etc. and another distorts at various proportions depending on amplitude and frequency of input, then how would the difference sound with real music, given the same aggregate THD and IMD, etc? Not a new question, I know, and I think it just ends up underlining the importance of listening...so complicated it's best to trust your ears. You also need to trust your speakers and your room. At last I've managed to quell my screaming Slovak neighbours. I also built the Morgan Jones ECC88 cascode stage and also a version with 5mA standing current and again the higher current version had lower distortion. Morgan is a bit quirky, and occasionally irrational. He seems to believe that an ECC88 has only one sensible operating point. All things considered, including his also questionable belief in the convenience of 330V HT, he might be right. In general it may be worth making the point that increased current can impinge on distortion in several ways. It may be that the gain valve is more linear in that region of higher current, but it could also be that it is less prone to distortion induced by the following input. If the grid of the follower is being overdriven, for example, then the higher current driver, having a lower output impedance especially at higher signal levels, will distort less in response. Ian |
#16
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Ian Iveson wrote: Patrick Turner wrote To return to your original question, you could obtain the data sheet for an ECC88 and draw a horizontal load line through your operating point on the anode characteristics chart. That would allow a good estimate of what distortion to expect. If your actual distortion is much different, then you've made a mistake somewhere, or your ECC88 is misbehaving. Hmm, usually when you measure a tube's distortion with a CCS load which might be 500 times greater than Ra, the 2H is much less than what you'd see with a line across a set of ancient plotted data curves which are not very accurate. Think of the rudimentary gear used in 1955. All tubed, and riddled with more distortion than what they may have been measuring. The curves are beautifully important guide, but that's about all, and to really know you must get away from books and PC, build a circuit and measure and tweak it until you get the lowest THD. Fair enough, Patrick, but the question was about whether the measured distortion of an actual circuit is "about right". From what you said initially, the THD seemed rather high. It seems reasonable therefore to compare the actual results with what theory might expect. I faced the same questions when I started seriously mucking around with tubes some 15 years ago after a temporary 30 year long break. An examination of published valve characteristics should offer a workable approximation. Indeed, but don't expect accuracy with tiny little RCA tube manual drawings. A glance at a 6DJ8 datasheet on screen suggests to me that 0.5% distortion is a pretty good result. It may possibly be reduced for a particular signal level by messing with the operating point of the bottom valve, but then it would require frequent readjustment to follow the sweet spot as the valve ages. I doubt the linearity changes much with ageing. I've measured new tubes, and then measured them 5,000 hours later, and they are basically the same if the emission is healthy and there isn't any grid current. I don't measure distortion. I haven't got good enough instruments and can't be bothered anyway. Seems too late by then. I just check that my products are working properly as designed, and listen. I don't make stuff to sell. After a life of building house extensions and then suffering some knee troubles I had a vision to help people by providing some sound gear worth owning and paying for. Building amps was easier than wrestling with barrows of concrete and 2X4 timbers. I wondered a lot about THD and the resulting IMD, and so I ended all doubts in my mind by designing and building my own low THD oscillator with less than 0.002% THD, and then building a distortion measuring unit which could reliably display 0.002% THD in a 10Vrms signal on a CRO. I became pretty design savvy before I had anyone to talk to online in 2000. When you re-educate yourself, you have to make sacrifices. No personal relationships, no family life, no TV watching ( because why watch other people doing their lives when you could be doing yur own? ) no boozing at the pub, no sport, no holidays, luxuries, no stupid vain expenses, and so on and on, all the while working to learn, and to be intellectually independant. If someone does not get into Groove Frugal, then they just don't learn much very fast. Then you have to learn to doubt all you observe including your own understanding unless you can proove you are right. 0.002% in a 10Vrms signal = 100uV, or 1/10 of 1mV, a signal at -94dBV. Its quite inaudible, BTW. But with 50 frequencies present in an audio signal at the same instant, the THD must measure low to get low IMD. The THD itself isn't important in music because when one F is present, so is its many harmonics, and when the amounts of harmonics are altered quite a lot you won't change the sound much; a violin will still sound just like a violin. But the IMD products generated are often not musically related to all the harmonics in the music, and thus are like small black spots on a white clean page; you only need a low level of IMD and the music is stuffed. It is my belief that the IMD created by class A SE triode stages in preamps or in power amps is less objectionable than that produced in solid state gear, even if the SS measures 40dB better. It may seem like stoopid belief, but my customers' ears tend to support my belief. And nobody has yet explained why a 6CG7 made by Seimans in 1965 will sound "better" than one made by Sovtek last year. In a preamp both could easily measure less than 0.015% THD at the maximum peak level of signal. Is it subtle microphony causing more "artifacts" than THD&IMD? perhaps. Is it some kind of dynamic IMD distortion than we are not fully aware of? damned if I know. But one preamp triode will do sonically better than an SS opamp with lots of devices and a loop of NFB. To win prizes for the best wine at a wine show it takes more than employing a good chemist. One should have a virgin of 10 **** into the vat of crushed grapes at midnight during a full moon. Every vigneron knows this. Can't talk about it though. That, plus the extra distortion arising from the top valve operating as a cathode follower (you don't say what load it was driving when you measured it), 100K more or less, should account for what you have measured. CF distortion should be negligible compared to the figures I am getting. Assuming a sensible operating point, and now knowing the driven load, then at 1kHz CF distortion would be numerically small, yes. So are you saying a mu follower is not low distortion It's considerably lower than when the same valve is used at the same operating point, with the same HT, and with a resistive anode load, particularly when driving a low impedance. If it follows mu, then it is as linear as mu is constant. You might give some thought to how to pick a valve and operating point that is likely to have the least mu variation. Just don't confuse the µ-follower with a SRPP where the 2H is supposed to become lowest when equal changes in Ia for top and bottom tubes occur for a particular load value and Rk used. I wasn't confused. My mu-stages use MOSFETs on top, and I hope they aren't into any of that push-pull hanky-panky. With a mosfet "on top", it sounds like having a female body builder performing difficult tasks according to the wishes and whims of a skinny male "rooting her from below" :-) They look like they are pushing and pulling, but they ain't. The wimp does the message, and womp does the grunt. The triode wimp makes the perfect undistorted message, and the source output of the mosfet just slavishly follows. BUT, and there is always a but, the mosfet usually works best with far more dc than the triode, and the gm of the mosfet is low and noise high when is Id is low. I've never seen the attraction of having mosfets in µ-follower circuits. Usually, having say 5mA through 2 seriesed ttriodes in a µ-folower is just fine. I always forget what SRPP stands for...there's a little cluster of acronyms applicable to totem-pole configurations that are easily confused. I tend to remember the circuits but not their names. But I'm OK with mu-stage. µ-follower circuits were once always called bootstrapped followers. It was a far more accurate name. The µ-follower goes beyond the SRPP and uses a cathode follower to provide a very high load to the gain tube and approaching CCS when THD becomes minimal. When THD due to the bottom valve is minimised, yes. However, it is possible that circuit THD of an SRPP stage could be less, because of the active cancelation of some distortion products. True, but you can't get full 2H cancelation in a SRPP. And when you have conditions for best 2H cancelation, usually you may have considerable 3H. So although I beagn with SRPP, I soon realised µ-follower was "better" imho. Both sonically, and with measurements. With 6SN7, well set up for µ-follower, expect 0.1% at 10Vo. This is much better than with the usual 33k load with ia = 4mA. It could also be that the spectrum produced by an SRPP stage would be more euphonic in the context of particular preamp circuit. AFAIK ppl don't generally build these variations in a search for minimum THD. I'm into music, not euphony. Patrick Turner. Ian |
#17
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Ian Bell wrote: Ian Iveson wrote: Patrick Turner wrote To return to your original question, you could obtain the data sheet for an ECC88 and draw a horizontal load line through your operating point on the anode characteristics chart. That would allow a good estimate of what distortion to expect. If your actual distortion is much different, then you've made a mistake somewhere, or your ECC88 is misbehaving. Hmm, usually when you measure a tube's distortion with a CCS load which might be 500 times greater than Ra, the 2H is much less than what you'd see with a line across a set of ancient plotted data curves which are not very accurate. Think of the rudimentary gear used in 1955. All tubed, and riddled with more distortion than what they may have been measuring. The curves are beautifully important guide, but that's about all, and to really know you must get away from books and PC, build a circuit and measure and tweak it until you get the lowest THD. Fair enough, Patrick, but the question was about whether the measured distortion of an actual circuit is "about right". It seems reasonable therefore to compare the actual results with what theory might expect. An examination of published valve characteristics should offer a workable approximation. A glance at a 6DJ8 datasheet on screen suggests to me that 0.5% distortion is a pretty good result. It may possibly be reduced for a particular signal level by messing with the operating point of the bottom valve, but then it would require frequent readjustment to follow the sweet spot as the valve ages. I have made a much more radical change. I redesigned the ECC88 mu follower with an increased standing current of 5mA (up from Morgan Jones 1.34mA). The result is 1KHz distortion at 20V rms is 0.64% and at 2V rms it is down to 0.054%. I also built the Morgan Jones ECC88 cascode stage and also a version with 5mA standing current and again the higher current version had lower distortion. Now ya talkin'!! Patrick Turner. Cheers Ian |
#18
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Ian Iveson wrote:
Ian Bell wrote To return to your original question, you could obtain the data sheet for an ECC88 and draw a horizontal load line through your operating point on the anode characteristics chart. That would allow a good estimate of what distortion to expect. If your actual distortion is much different, then you've made a mistake somewhere, or your ECC88 is misbehaving. Hmm, usually when you measure a tube's distortion with a CCS load which might be 500 times greater than Ra, the 2H is much less than what you'd see with a line across a set of ancient plotted data curves which are not very accurate. Think of the rudimentary gear used in 1955. All tubed, and riddled with more distortion than what they may have been measuring. The curves are beautifully important guide, but that's about all, and to really know you must get away from books and PC, build a circuit and measure and tweak it until you get the lowest THD. Fair enough, Patrick, but the question was about whether the measured distortion of an actual circuit is "about right". It seems reasonable therefore to compare the actual results with what theory might expect. An examination of published valve characteristics should offer a workable approximation. A glance at a 6DJ8 datasheet on screen suggests to me that 0.5% distortion is a pretty good result. It may possibly be reduced for a particular signal level by messing with the operating point of the bottom valve, but then it would require frequent readjustment to follow the sweet spot as the valve ages. I have made a much more radical change. I redesigned the ECC88 mu follower with an increased standing current of 5mA (up from Morgan Jones 1.34mA). The result is 1KHz distortion at 20V rms is 0.64% and at 2V rms it is down to 0.054%. Just wondering if it follows that the lower harmonic distortion will contain a higher proportion of higher-order harmonics. Some say it's the spectrum that counts. I think there is quite a lot of evidence that that is true. I have seen a fairly recent paper that used mathcad to emulate various transfer functions through which audio was processed to produce output with differing spectral content. These were then played to a panel of listeners who were asked to comment on the 'quality' of the sound. The results supported the view that the relative level of harmonics is important and that higher order harmonics were deemed more objectionable. This led to a means of weighting higher order harmonics tp better reflect their effect on perceived quality which pretty much supported similar work that I think was conducted back in the 50s. Also to what extent we might be concerned about distortion modulation. I know that artifacts are less noticeable if they are of constant amplitude with respect to signal ground, such as hum or hiss might be. Are they also less noticeable if they are constant with respect to signal amplitude and/or frequency? If one amplifier distorts at a constant x% of 2H, y% of 3H, etc. and another distorts at various proportions depending on amplitude and frequency of input, then how would the difference sound with real music, given the same aggregate THD and IMD, etc? Good question but not one I think that has been tested. Not a new question, I know, and I think it just ends up underlining the importance of listening...so complicated it's best to trust your ears. Indeed. You also need to trust your speakers and your room. A factor too often overlooked. At last I've managed to quell my screaming Slovak neighbours. I also built the Morgan Jones ECC88 cascode stage and also a version with 5mA standing current and again the higher current version had lower distortion. Morgan is a bit quirky, and occasionally irrational. He seems to believe that an ECC88 has only one sensible operating point. All things considered, including his also questionable belief in the convenience of 330V HT, he might be right. He does seem to be stuck at one operating point and keeps using the phrase 'this puts it in a very linear region' or words to that effect without further explanation. In general it may be worth making the point that increased current can impinge on distortion in several ways. It may be that the gain valve is more linear in that region of higher current, but it could also be that it is less prone to distortion induced by the following input. If the grid of the follower is being overdriven, for example, then the higher current driver, having a lower output impedance especially at higher signal levels, will distort less in response. Well, it is certainly true that at higher currents both mu and gm are higher and ra is lower. If we agree with Jones that ra variations are a/the principal cause of distortion in triodes then higher currents ought to produce less distortion. Ian Cheers Ian |
#19
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Patrick Turner wrote:
Ian Bell wrote: Ian Iveson wrote: Patrick Turner wrote To return to your original question, you could obtain the data sheet for an ECC88 and draw a horizontal load line through your operating point on the anode characteristics chart. That would allow a good estimate of what distortion to expect. If your actual distortion is much different, then you've made a mistake somewhere, or your ECC88 is misbehaving. Hmm, usually when you measure a tube's distortion with a CCS load which might be 500 times greater than Ra, the 2H is much less than what you'd see with a line across a set of ancient plotted data curves which are not very accurate. Think of the rudimentary gear used in 1955. All tubed, and riddled with more distortion than what they may have been measuring. The curves are beautifully important guide, but that's about all, and to really know you must get away from books and PC, build a circuit and measure and tweak it until you get the lowest THD. Fair enough, Patrick, but the question was about whether the measured distortion of an actual circuit is "about right". It seems reasonable therefore to compare the actual results with what theory might expect. An examination of published valve characteristics should offer a workable approximation. A glance at a 6DJ8 datasheet on screen suggests to me that 0.5% distortion is a pretty good result. It may possibly be reduced for a particular signal level by messing with the operating point of the bottom valve, but then it would require frequent readjustment to follow the sweet spot as the valve ages. I have made a much more radical change. I redesigned the ECC88 mu follower with an increased standing current of 5mA (up from Morgan Jones 1.34mA). The result is 1KHz distortion at 20V rms is 0.64% and at 2V rms it is down to 0.054%. I also built the Morgan Jones ECC88 cascode stage and also a version with 5mA standing current and again the higher current version had lower distortion. Now ya talkin'!! Patrick Turner. Cheers Ian One thing I forgot to add was that I tried the mu follower with both an ECC88 and a 6N1P. Distortion was pretty much the same in both cases. One interesting thing I did notice was that I was using ac heaters elevated to 75V. Tweaking the hum bucking pot made sod all difference to the hum at the output of the ECC88 version but the 6N1P version needed the pot tweaked to one extreme to reduce the hum - don't think I has actually reached the minima even then. I wonder why the ECC88 is so little prone to heater hum? Could it be some sort of cancellation effect - hum injected into the lower cathode appears at the anode then gets injected into the top tube grid and appears in ant-phase with the hum induced in the top tube cathode??? I took care to use the recommended triode half for the top triode. Any ideas? Cheers Ian |
#20
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
In article ,
Patrick Turner wrote: To win prizes for the best wine at a wine show it takes more than employing a good chemist. One should have a virgin of 10 **** into the vat of crushed grapes at midnight during a full moon. Every vigneron knows this. Can't talk about it though. So what is the equivalent procedure in tube audio design? -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
#21
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Ian Bell wrote
To return to your original question, you could obtain the data sheet for an ECC88 and draw a horizontal load line through your operating point on the anode characteristics chart. That would allow a good estimate of what distortion to expect. If your actual distortion is much different, then you've made a mistake somewhere, or your ECC88 is misbehaving. Hmm, usually when you measure a tube's distortion with a CCS load which might be 500 times greater than Ra, the 2H is much less than what you'd see with a line across a set of ancient plotted data curves which are not very accurate. Think of the rudimentary gear used in 1955. All tubed, and riddled with more distortion than what they may have been measuring. The curves are beautifully important guide, but that's about all, and to really know you must get away from books and PC, build a circuit and measure and tweak it until you get the lowest THD. Fair enough, Patrick, but the question was about whether the measured distortion of an actual circuit is "about right". It seems reasonable therefore to compare the actual results with what theory might expect. An examination of published valve characteristics should offer a workable approximation. A glance at a 6DJ8 datasheet on screen suggests to me that 0.5% distortion is a pretty good result. It may possibly be reduced for a particular signal level by messing with the operating point of the bottom valve, but then it would require frequent readjustment to follow the sweet spot as the valve ages. I have made a much more radical change. I redesigned the ECC88 mu follower with an increased standing current of 5mA (up from Morgan Jones 1.34mA). The result is 1KHz distortion at 20V rms is 0.64% and at 2V rms it is down to 0.054%. Just wondering if it follows that the lower harmonic distortion will contain a higher proportion of higher-order harmonics. Some say it's the spectrum that counts. I think there is quite a lot of evidence that that is true. I have seen a fairly recent paper that used mathcad to emulate various transfer functions through which audio was processed to produce output with differing spectral content. These were then played to a panel of listeners who were asked to comment on the 'quality' of the sound. The results supported the view that the relative level of harmonics is important and that higher order harmonics were deemed more objectionable. This led to a means of weighting higher order harmonics tp better reflect their effect on perceived quality which pretty much supported similar work that I think was conducted back in the 50s. Also to what extent we might be concerned about distortion modulation. I know that artifacts are less noticeable if they are of constant amplitude with respect to signal ground, such as hum or hiss might be. Are they also less noticeable if they are constant with respect to signal amplitude and/or frequency? If one amplifier distorts at a constant x% of 2H, y% of 3H, etc. and another distorts at various proportions depending on amplitude and frequency of input, then how would the difference sound with real music, given the same aggregate THD and IMD, etc? Good question but not one I think that has been tested. Not a new question, I know, and I think it just ends up underlining the importance of listening...so complicated it's best to trust your ears. Indeed. You also need to trust your speakers and your room. A factor too often overlooked. At last I've managed to quell my screaming Slovak neighbours. I also built the Morgan Jones ECC88 cascode stage and also a version with 5mA standing current and again the higher current version had lower distortion. Morgan is a bit quirky, and occasionally irrational. He seems to believe that an ECC88 has only one sensible operating point. All things considered, including his also questionable belief in the convenience of 330V HT, he might be right. He does seem to be stuck at one operating point and keeps using the phrase 'this puts it in a very linear region' or words to that effect without further explanation. I think it may follow from his assertion that the ECC88 was designed for cascode operation. That would be in an early stage of equipment powered by UK mains, perhaps. Once you've dropped that once or twice to isolate stages and reduce hum, you're left with 200V or so, at low current, for the cascode stage. If the valve was made for such a circumstance, then it is possible that it was optimised at Morgan's operating point of around 90V, 1.3mA, and perhaps there should be a particularly linear region round there somewhere. Your observation is not obvious from a casual examination of the anode characteristics. It is hard to judge by placing a ruler across and moving it up and down, because of the optical illusion caused by the changing gradients of the grid lines. If it's a general truth, it presumably follows from a simple application of Child's Law? It should therefore show up clearly in simulation using the simplest of models. Ian In general it may be worth making the point that increased current can impinge on distortion in several ways. It may be that the gain valve is more linear in that region of higher current, but it could also be that it is less prone to distortion induced by the following input. If the grid of the follower is being overdriven, for example, then the higher current driver, having a lower output impedance especially at higher signal levels, will distort less in response. Well, it is certainly true that at higher currents both mu and gm are higher and ra is lower. If we agree with Jones that ra variations are a/the principal cause of distortion in triodes then higher currents ought to produce less distortion. Ian |
#22
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Ian Iveson wrote:
Ian Bell wrote To return to your original question, you could obtain the data sheet for an ECC88 and draw a horizontal load line through your operating point on the anode characteristics chart. That would allow a good estimate of what distortion to expect. If your actual distortion is much different, then you've made a mistake somewhere, or your ECC88 is misbehaving. Hmm, usually when you measure a tube's distortion with a CCS load which might be 500 times greater than Ra, the 2H is much less than what you'd see with a line across a set of ancient plotted data curves which are not very accurate. Think of the rudimentary gear used in 1955. All tubed, and riddled with more distortion than what they may have been measuring. The curves are beautifully important guide, but that's about all, and to really know you must get away from books and PC, build a circuit and measure and tweak it until you get the lowest THD. Fair enough, Patrick, but the question was about whether the measured distortion of an actual circuit is "about right". It seems reasonable therefore to compare the actual results with what theory might expect. An examination of published valve characteristics should offer a workable approximation. A glance at a 6DJ8 datasheet on screen suggests to me that 0.5% distortion is a pretty good result. It may possibly be reduced for a particular signal level by messing with the operating point of the bottom valve, but then it would require frequent readjustment to follow the sweet spot as the valve ages. I have made a much more radical change. I redesigned the ECC88 mu follower with an increased standing current of 5mA (up from Morgan Jones 1.34mA). The result is 1KHz distortion at 20V rms is 0.64% and at 2V rms it is down to 0.054%. Just wondering if it follows that the lower harmonic distortion will contain a higher proportion of higher-order harmonics. Some say it's the spectrum that counts. I think there is quite a lot of evidence that that is true. I have seen a fairly recent paper that used mathcad to emulate various transfer functions through which audio was processed to produce output with differing spectral content. These were then played to a panel of listeners who were asked to comment on the 'quality' of the sound. The results supported the view that the relative level of harmonics is important and that higher order harmonics were deemed more objectionable. This led to a means of weighting higher order harmonics tp better reflect their effect on perceived quality which pretty much supported similar work that I think was conducted back in the 50s. Also to what extent we might be concerned about distortion modulation. I know that artifacts are less noticeable if they are of constant amplitude with respect to signal ground, such as hum or hiss might be. Are they also less noticeable if they are constant with respect to signal amplitude and/or frequency? If one amplifier distorts at a constant x% of 2H, y% of 3H, etc. and another distorts at various proportions depending on amplitude and frequency of input, then how would the difference sound with real music, given the same aggregate THD and IMD, etc? Good question but not one I think that has been tested. Not a new question, I know, and I think it just ends up underlining the importance of listening...so complicated it's best to trust your ears. Indeed. You also need to trust your speakers and your room. A factor too often overlooked. At last I've managed to quell my screaming Slovak neighbours. I also built the Morgan Jones ECC88 cascode stage and also a version with 5mA standing current and again the higher current version had lower distortion. Morgan is a bit quirky, and occasionally irrational. He seems to believe that an ECC88 has only one sensible operating point. All things considered, including his also questionable belief in the convenience of 330V HT, he might be right. He does seem to be stuck at one operating point and keeps using the phrase 'this puts it in a very linear region' or words to that effect without further explanation. I think it may follow from his assertion that the ECC88 was designed for cascode operation. That would be in an early stage of equipment powered by UK mains, perhaps. Once you've dropped that once or twice to isolate stages and reduce hum, you're left with 200V or so, at low current, for the cascode stage. If the valve was made for such a circumstance, then it is possible that it was optimised at Morgan's operating point of around 90V, 1.3mA, and perhaps there should be a particularly linear region round there somewhere. Possibly. ISTR the ECC88 was originally intended as a VHF cascode front end for UK FM receivers and it may well be that its normal RF operating point is the same as Jones uses. Whether that makes it also a good point from the AF point of view is moot. Cheers Ian Your observation is not obvious from a casual examination of the anode characteristics. It is hard to judge by placing a ruler across and moving it up and down, because of the optical illusion caused by the changing gradients of the grid lines. If it's a general truth, it presumably follows from a simple application of Child's Law? It should therefore show up clearly in simulation using the simplest of models. Ian In general it may be worth making the point that increased current can impinge on distortion in several ways. It may be that the gain valve is more linear in that region of higher current, but it could also be that it is less prone to distortion induced by the following input. If the grid of the follower is being overdriven, for example, then the higher current driver, having a lower output impedance especially at higher signal levels, will distort less in response. Well, it is certainly true that at higher currents both mu and gm are higher and ra is lower. If we agree with Jones that ra variations are a/the principal cause of distortion in triodes then higher currents ought to produce less distortion. Ian |
#23
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Ian Bell wrote: Ian Iveson wrote: Ian Bell wrote To return to your original question, you could obtain the data sheet for an ECC88 and draw a horizontal load line through your operating point on the anode characteristics chart. That would allow a good estimate of what distortion to expect. If your actual distortion is much different, then you've made a mistake somewhere, or your ECC88 is misbehaving. Hmm, usually when you measure a tube's distortion with a CCS load which might be 500 times greater than Ra, the 2H is much less than what you'd see with a line across a set of ancient plotted data curves which are not very accurate. Think of the rudimentary gear used in 1955. All tubed, and riddled with more distortion than what they may have been measuring. The curves are beautifully important guide, but that's about all, and to really know you must get away from books and PC, build a circuit and measure and tweak it until you get the lowest THD. Fair enough, Patrick, but the question was about whether the measured distortion of an actual circuit is "about right". It seems reasonable therefore to compare the actual results with what theory might expect. An examination of published valve characteristics should offer a workable approximation. A glance at a 6DJ8 datasheet on screen suggests to me that 0.5% distortion is a pretty good result. It may possibly be reduced for a particular signal level by messing with the operating point of the bottom valve, but then it would require frequent readjustment to follow the sweet spot as the valve ages. I have made a much more radical change. I redesigned the ECC88 mu follower with an increased standing current of 5mA (up from Morgan Jones 1.34mA). The result is 1KHz distortion at 20V rms is 0.64% and at 2V rms it is down to 0.054%. Just wondering if it follows that the lower harmonic distortion will contain a higher proportion of higher-order harmonics. Some say it's the spectrum that counts. I think there is quite a lot of evidence that that is true. I have seen a fairly recent paper that used mathcad to emulate various transfer functions through which audio was processed to produce output with differing spectral content. These were then played to a panel of listeners who were asked to comment on the 'quality' of the sound. The results supported the view that the relative level of harmonics is important and that higher order harmonics were deemed more objectionable. This led to a means of weighting higher order harmonics tp better reflect their effect on perceived quality which pretty much supported similar work that I think was conducted back in the 50s. Weighting is usally based on Nsquared/4, where N is the harmonic. 2H is the basis point. So 3H gives you 9/4, so you only need 0.44% of 3H to cause music to sound as bad as 1% of 2H. 5H gives you 24/4, 6.25, so you only need 0.16% of 5H to be as bad as 1% of 2H and so on. 7H is particularly bad because it does not relate to the musical scale. However. There is plenty of 7H in many tones produced by instruments. But luckily, not much in amplifiers. Also to what extent we might be concerned about distortion modulation. I know that artifacts are less noticeable if they are of constant amplitude with respect to signal ground, such as hum or hiss might be. Are they also less noticeable if they are constant with respect to signal amplitude and/or frequency? If one amplifier distorts at a constant x% of 2H, y% of 3H, etc. and another distorts at various proportions depending on amplitude and frequency of input, then how would the difference sound with real music, given the same aggregate THD and IMD, etc? Good question but not one I think that has been tested. "distortion modulation" ?? To me that's just IMD. If you watch the THD being produced when you measure and amp, you can see the waveform containing typical amounts of 2H and 3H, and sometimes they look like they are being amplitude modulated by the fundemental tone. That simply means IMD products are present. All AM waves are the sum of the fundemental "carrier" and the sum and difference F between the fundemental and modulating wave form. When an amp with only 2H has a mild amount of loop NFB applied the 2H is reduced but sum and difference F are produced, ie, more fundemental and also 3H where none was present when no FB was applied. When the 2H is about 10%, using 10dB of NFB won't make things sound any better. But if you applied 40dB FB instead of 10dB, then you have much lower 2H and the second order IMD products are also very low, but nevertheless numerous in H spectra. It doesn't matter though because when THD without NFB 1% at is is in all good amps in normal listening, the second order IMD products are usually very low indeed when mild NFB is used. Not a new question, I know, and I think it just ends up underlining the importance of listening...so complicated it's best to trust your ears. Indeed. You also need to trust your speakers and your room. A factor too often overlooked. At last I've managed to quell my screaming Slovak neighbours. I also built the Morgan Jones ECC88 cascode stage and also a version with 5mA standing current and again the higher current version had lower distortion. Morgan is a bit quirky, and occasionally irrational. He seems to believe that an ECC88 has only one sensible operating point. All things considered, including his also questionable belief in the convenience of 330V HT, he might be right. He does seem to be stuck at one operating point and keeps using the phrase 'this puts it in a very linear region' or words to that effect without further explanation. In general it may be worth making the point that increased current can impinge on distortion in several ways. It may be that the gain valve is more linear in that region of higher current, but it could also be that it is less prone to distortion induced by the following input. If the grid of the follower is being overdriven, for example, then the higher current driver, having a lower output impedance especially at higher signal levels, will distort less in response. Well, it is certainly true that at higher currents both mu and gm are higher and ra is lower. If we agree with Jones that ra variations are a/the principal cause of distortion in triodes then higher currents ought to produce less distortion. Its the variations in gm and Ra that count; µ is fairly constant. If the load line is situated in the straighter portions of Ra curves, there are only small variations in Ra and gm in the wave cycle and hence distortion is lower than if you locate the load swinging each side of an Ia that is too low. But raising the RL value to say 20Ra at the low level of Ia makes linearity become acceptable anyway. A CCS load shows linearity to be quite good for low Ia in most triodes. Patrick Turner. Ian Cheers Ian |
#24
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Ian Bell wrote: Patrick Turner wrote: Ian Bell wrote: Ian Iveson wrote: Patrick Turner wrote To return to your original question, you could obtain the data sheet for an ECC88 and draw a horizontal load line through your operating point on the anode characteristics chart. That would allow a good estimate of what distortion to expect. If your actual distortion is much different, then you've made a mistake somewhere, or your ECC88 is misbehaving. Hmm, usually when you measure a tube's distortion with a CCS load which might be 500 times greater than Ra, the 2H is much less than what you'd see with a line across a set of ancient plotted data curves which are not very accurate. Think of the rudimentary gear used in 1955. All tubed, and riddled with more distortion than what they may have been measuring. The curves are beautifully important guide, but that's about all, and to really know you must get away from books and PC, build a circuit and measure and tweak it until you get the lowest THD. Fair enough, Patrick, but the question was about whether the measured distortion of an actual circuit is "about right". It seems reasonable therefore to compare the actual results with what theory might expect. An examination of published valve characteristics should offer a workable approximation. A glance at a 6DJ8 datasheet on screen suggests to me that 0.5% distortion is a pretty good result. It may possibly be reduced for a particular signal level by messing with the operating point of the bottom valve, but then it would require frequent readjustment to follow the sweet spot as the valve ages. I have made a much more radical change. I redesigned the ECC88 mu follower with an increased standing current of 5mA (up from Morgan Jones 1.34mA). The result is 1KHz distortion at 20V rms is 0.64% and at 2V rms it is down to 0.054%. I also built the Morgan Jones ECC88 cascode stage and also a version with 5mA standing current and again the higher current version had lower distortion. Now ya talkin'!! Patrick Turner. Cheers Ian One thing I forgot to add was that I tried the mu follower with both an ECC88 and a 6N1P. Distortion was pretty much the same in both cases. One interesting thing I did notice was that I was using ac heaters elevated to 75V. Tweaking the hum bucking pot made sod all difference to the hum at the output of the ECC88 version but the 6N1P version needed the pot tweaked to one extreme to reduce the hum - don't think I has actually reached the minima even then. I wonder why the ECC88 is so little prone to heater hum? Could it be some sort of cancellation effect - hum injected into the lower cathode appears at the anode then gets injected into the top tube grid and appears in ant-phase with the hum induced in the top tube cathode??? I took care to use the recommended triode half for the top triode. Any ideas? Cheers µ-follower is prone to hum, so *always* use dc for heaters. Retain the biasing of the dc supply, and bypass it to 0V via 100uF. Patrick Turner. Ian µ |
#25
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
John Byrns wrote: In article , Patrick Turner wrote: To win prizes for the best wine at a wine show it takes more than employing a good chemist. One should have a virgin of 10 **** into the vat of crushed grapes at midnight during a full moon. Every vigneron knows this. Can't talk about it though. So what is the equivalent procedure in tube audio design? I hoid it on the grape vine that female heat must be applied to tubes prior to obtaining the best music. I leave you to consider how this may be achieved. Patrick Turner. -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
#26
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
On Nov 22, 11:40*am, Patrick Turner wrote:
But raising the RL value to say 20Ra at the low level of Ia makes linearity become acceptable anyway. A CCS load shows linearity to be quite good for low Ia in most triodes. Patrick Turner. Ha! Back in the day when I was interested in pre-amps, I sent one of my mentors a design with some inoffensive pre-amp tube loaded about 100x Rp in response to him wanting me to operate it about 90V because that was what they did back before the War, rather than at at the 300V, its max rating, which was my natural inclination. "You'll get nothing out," he said on the phone, so disturbed that he called rather than writing to me (by post, not e-mail). "Ah, yes," I said, "but it is a wonderfully silent nothing." 20Ra? You're such a piker, Patrick. Let loose the dogs of war and live a little! ***** I suppose someone has already referred Ian to Steve Bench's page of schematics and a table for getting pre-amp tube noise right down. I haven't been there for a few years but I seem to recall that Steve compared the SRPP and various other topologies and outright tricks, using 6SN7 and 6SL7 and 417A, God's own faves. Steve's table was my inspiration until I decided to work solely with the 417A, the best pre- amp tube ever made. Steve also shows how you can use a ground lift to get a tube, or a stack of tubes, up into a linear region out of grid current even with big signal swings, a twist that is so useful and so cheap, it is surprising it isn't seen more often. 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
|
|||
|
|||
mu follower distortion
Your signature block is ancient and ho-hum history, mate.
What have you accomplished during the last 10 years, except for pathological self-promotion? Just curious . . . . 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 |
#28
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
On Nov 23, 4:55*pm, Jon Yaeger wrote:
Your signature block is ancient and ho-hum history, mate. The last refuge of the nobody, calling his betters "hasbeens". We've seen it all before, sonny. The office cleaner Frank Deutschmann loved calling me a hasbeen. You should ask your bumbuddy Michael LeFevre of Magnequest what happened to little Frankie when I got bored with him. What have you accomplished during the last 10 years, except for pathological self-promotion? Good golly, has it been that long since I mastered electronics? High time to get a new hobby, even if only to give slimy vermin like Jon Yaeger a miss. Just curious . . . . All the answers you want are on my site, Mr Curious. Andre Jute Visit Jute on Amps athttp://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
|
|||
|
|||
mu follower distortion
Andre Jute wrote:
On Nov 23, 4:55 pm, Jon Yaeger wrote: Your signature block is ancient and ho-hum history, mate. The last refuge of the nobody, calling his betters "hasbeens". We've seen it all before, sonny. The office cleaner Frank Deutschmann loved calling me a hasbeen. You should ask your bumbuddy Michael LeFevre of Magnequest what happened to little Frankie when I got bored with him. Andre, I am unable to find Steve Bench's web site. All I can find is a message from AOL that it was shut down last month. Cheers Ian |
#31
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
In article ,
Andre Jute wrote: On Nov 23, 4:55*pm, Jon Yaeger wrote: Your signature block is ancient and ho-hum history, mate. The last refuge of the nobody, calling his betters "hasbeens". We've seen it all before, sonny. The office cleaner Frank Deutschmann loved calling me a hasbeen. You should ask your bumbuddy Michael LeFevre of Magnequest what happened to little Frankie when I got bored with him. Whatever did become of "Frankie", where is he today? Unlike most of the rest of the "gang" he seems to have completely disappeared from the face of the earth, I hope you didn't encourage any unsavory types to fashion a set of concrete galoshes for him? -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
#32
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Ian Bell wrote:
snip Andre, I am unable to find Steve Bench's web site. All I can find is a message from AOL that it was shut down last month. Cheers Ian here's the new URL of steve's site: http://greygum.net/sbench/sbench101/ |
#33
Posted to rec.audio.tubes
|
|||
|
|||
Oh the irony
Jutey-kins,
Surely I am not the only one who noticed the irony and hypocrisy of you calling me to task for a bit of "self-promotion", am I? Garage vermin / trader / whatever in article , Jon Yaeger at wrote on 11/23/08 1:05 PM: I do respect someone like yourself . . . with such a glaring personality disorder . . . much as I respect a skunk that crosses my path. I don't know what you did to Frank, but I'm certainly aware of what nasty stuff you tried to pull on Henry, with the Italian sockpuppet. I'm comfortable being a garage vermin, compared to your treasure trove of psychopathology. Jon in article , Andre Jute at wrote on 11/23/08 12:29 PM: On Nov 23, 4:55*pm, Jon Yaeger wrote: Your signature block is ancient and ho-hum history, mate. The last refuge of the nobody, calling his betters "hasbeens". We've seen it all before, sonny. The office cleaner Frank Deutschmann loved calling me a hasbeen. You should ask your bumbuddy Michael LeFevre of Magnequest what happened to little Frankie when I got bored with him. What have you accomplished during the last 10 years, except for pathological self-promotion? Good golly, has it been that long since I mastered electronics? High time to get a new hobby, even if only to give slimy vermin like Jon Yaeger a miss. Just curious . . . . All the answers you want are on my site, Mr Curious. Andre Jute Visit Jute on Amps athttp://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 |
#34
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Paul Babiak wrote:
Ian Bell wrote: snip Andre, I am unable to find Steve Bench's web site. All I can find is a message from AOL that it was shut down last month. Cheers Ian here's the new URL of steve's site: http://greygum.net/sbench/sbench101/ Thanks Paul. I notice it has not been updated since 2004. Is that correct? Cheers Ian |
#35
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
On Nov 23, 9:45*pm, Ian Bell wrote:
Paul Babiak wrote: Ian Bell wrote: snip Andre, I am unable to find Steve Bench's web site. All I can find is a message from AOL that it was shut down last month. Cheers Ian here's the new URL of steve's site: http://greygum.net/sbench/sbench101/ Thanks Paul. I notice it has not been updated since 2004. Is that correct? Eh? What does it matter when Steve last updated his site? He doesn't owe us a thing. Good information is good information forever. The RDH was first conceived about the time of the Great Crash -- no, not last week's. 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
|
|||
|
|||
mu follower distortion
On Nov 23, 7:04*pm, John Byrns wrote:
In article , *Andre Jute wrote: On Nov 23, 4:55*pm, Jon Yaeger wrote: Your signature block is ancient and ho-hum history, mate. The last refuge of the nobody, calling his betters "hasbeens". We've seen it all before, sonny. The office cleaner Frank Deutschmann loved calling me a hasbeen. You should ask your bumbuddy Michael LeFevre of Magnequest what happened to little Frankie when I got bored with him. Whatever did become of "Frankie", where is he today? *Unlike most of the rest of the "gang" he seems to have completely disappeared from the face of the earth, I hope you didn't encourage any unsavory types to fashion a set of concrete galoshes for him? -- Regards, John Byrns Surf my web pages at, *http://fmamradios.com/ Ha! At one of my grad schools I sat next to a fellow with an Italian name who took me home. His father sat me down in his study and said, "Unless your intentions are honourable, if you touch my daughters I'll bury you in an oil drum in the East River." But no, I didn't need their help to take care of Franki -- he was such an arsehole, he did most of the work himself; it took me less than ten minutes at the strategic moment to pull the rug from under Deutschmann's hooves forever, as witness the fact that we haven't heard even a squeak from him in a whole lotta years. Andre Jute Zero tolerance for scum |
#37
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Andre Jute wrote:
On Nov 23, 9:45 pm, Ian Bell wrote: Paul Babiak wrote: Ian Bell wrote: snip Andre, I am unable to find Steve Bench's web site. All I can find is a message from AOL that it was shut down last month. Cheers Ian here's the new URL of steve's site: http://greygum.net/sbench/sbench101/ Thanks Paul. I notice it has not been updated since 2004. Is that correct? Eh? What does it matter when Steve last updated his site? He doesn't owe us a thing. Good information is good information forever. The RDH was first conceived about the time of the Great Crash -- no, not last week's. Of course it does not matter. I merely wanted to be sure I had the right site. Cheers Ian |
#38
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Ian Bell wrote:
Paul Babiak wrote: Ian Bell wrote: snip Andre, I am unable to find Steve Bench's web site. All I can find is a message from AOL that it was shut down last month. Cheers Ian here's the new URL of steve's site: http://greygum.net/sbench/sbench101/ Thanks Paul. I notice it has not been updated since 2004. Is that correct? Cheers Ian Hi Ian I dunno about the latest update - just passing on some info from the joelist. I'm just glad the info hasn't been lost. Paul |
#39
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
Andre Jute wrote: On Nov 22, 11:40 am, Patrick Turner wrote: But raising the RL value to say 20Ra at the low level of Ia makes linearity become acceptable anyway. A CCS load shows linearity to be quite good for low Ia in most triodes. Patrick Turner. Ha! Back in the day when I was interested in pre-amps, I sent one of my mentors a design with some inoffensive pre-amp tube loaded about 100x Rp in response to him wanting me to operate it about 90V because that was what they did back before the War, rather than at at the 300V, its max rating, which was my natural inclination. "You'll get nothing out," he said on the phone, so disturbed that he called rather than writing to me (by post, not e-mail). "Ah, yes," I said, "but it is a wonderfully silent nothing." 20Ra? You're such a piker, Patrick. Let loose the dogs of war and live a little! Well, depends a bit on the triode concerned, and where its used. With 6DJ8, Ea = 90V is about right if you want to run Ia at say 8mA to operate the tube where Ra becomes close to the data specs of 2.6k. Ea = 300V? Gee, that's absurd for nearly all signal preamp and driver triodes in anything, but ideal for a 6BQ5/6AR5 strapped as a triode for driving output tubes. I often run RL = far higher than 20Ra because I sometimes use a choke + R as the dc carrying load, thus leaving the cap coupled load to be driven as the dominant load. Or I'll have a CCS dc supply, same deal..... You don't see the trees where my dogs **** old son. ***** I suppose someone has already referred Ian to Steve Bench's page of schematics and a table for getting pre-amp tube noise right down. I haven't been there for a few years but I seem to recall that Steve compared the SRPP and various other topologies and outright tricks, using 6SN7 and 6SL7 and 417A, God's own faves. Steve's table was my inspiration until I decided to work solely with the 417A, the best pre- amp tube ever made. Steve also shows how you can use a ground lift to get a tube, or a stack of tubes, up into a linear region out of grid current even with big signal swings, a twist that is so useful and so cheap, it is surprising it isn't seen more often. The noise is a bit a matter of luck with many triodes. And supposedly dependant on the gm, so the higher the gm is with the Iadc, the lower the noise. One has to try things, but for MC amp and low input noise, the highish gm of a suitable j-fet costing $1.20 usually beats any triode you care to point a stick at. 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 |
#40
Posted to rec.audio.tubes
|
|||
|
|||
mu follower distortion
On Nov 27, 10:15*am, Patrick Turner wrote:
Andre Jute wrote: On Nov 22, 11:40 am, Patrick Turner wrote: But raising the RL value to say 20Ra at the low level of Ia makes linearity become acceptable anyway. A CCS load shows linearity to be quite good for low Ia in most triodes. Patrick Turner. Ha! Back in the day when I was interested in pre-amps, I sent one of my mentors a design with some inoffensive pre-amp tube loaded about 100x Rp in response to him wanting me to operate it about 90V because that was what they did back before the War, rather than at at the 300V, its max rating, which was my natural inclination. "You'll get nothing out," he said on the phone, so disturbed that he called rather than writing to me (by post, not e-mail). "Ah, yes," I said, "but it is a wonderfully silent nothing." 20Ra? You're such a piker, Patrick. Let loose the dogs of war and live a little! Well, depends a bit on the triode concerned, and where its used. With 6DJ8, Ea = 90V is about right if you want to run Ia at say 8mA to operate the tube where Ra becomes close to the data specs of 2.6k. Ea = 300V? Gee, that's absurd for nearly all signal preamp and driver triodes in anything, It may be absurd for those designers who automatically reach for the little economy 9-pins, but I already has the good sense to reach only for the indestructible octals, 6SN7 and 6SL7. The only ninepin worth having is 417A (though a 12AT7 come close to the precision of a 6SN7. but ideal for a 6BQ5/6AR5 strapped as a triode for driving output tubes. Run 'em high and run 'em hot, and sweetness follows, as everyone knows with the slightest acquaintance with the folk wisdow of separating honey from earwax. These days I have come to understand that anything bigger than a KT66 *needs* a power tube as a driver. I often run RL = far higher than 20Ra because I sometimes use a choke + R as the dc carrying load, thus leaving the cap coupled load to be driven as the dominant load. Or I'll have a CCS dc supply, same deal..... You don't see the trees where my dogs **** old son. You're getting old if your dogs have bigger constant current streams than you. ***** I suppose someone has already referred Ian to Steve Bench's page of schematics and a table for getting pre-amp tube noise right down. I haven't been there for a few years but I seem to recall that Steve compared the SRPP and various other topologies and outright tricks, using 6SN7 and 6SL7 and 417A, God's own faves. Steve's table was my inspiration until I decided to work solely with the 417A, the best pre- amp tube ever made. Steve also shows how you can use a ground lift to get a tube, or a stack of tubes, up into a linear region out of grid current even with big signal swings, a twist that is so useful and so cheap, it is surprising it isn't seen more often. The noise is a bit a matter of luck with many triodes. You do give yourself a headstart though by choosing a triode, especially in ZNFB designs; after choosing the triode it is only a matter of determinedly finding its sweet spot, which will usually translate to all topologies suitable for that tube. Here's a tip for newbies, if any of them still come to this rat****ery: start near maximum spec sheet dissipation and work your way down; tubes are a lot sturdier than they might look. And supposedly dependant on the gm, so the higher the gm is with the Iadc, the lower the noise. It's what makes the 417A so good. One has to try things, but for MC amp and low input noise, the highish gm of a suitable j-fet costing $1.20 usually beats any triode you care to point a stick at. I'm decades past MC amps; I've used CD exclusively for so long now, I can hardly remember the frustration of those round shellac artifacts of an ancient civilization. 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 |
Reply |
|
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
Cathode Follower | Vacuum Tubes | |||
Need Help For Cathode Follower! | Vacuum Tubes | |||
Cathode follower thoughts... | Vacuum Tubes | |||
Cathode Follower | Vacuum Tubes | |||
Alternative to source or emitter follower | Vacuum Tubes |