I've always heard that the EL84 and EL34 clip more gradually than the 6V6 and 6L6, but that may vary with frequency, and is probably very dependent on the circuit they're in.
Since peak levels in music are often 10dB higher than average, many Hi-Fi tube amps may hit clipping more often than the user realizes, so I not only worry about how clipping sounds but also look closely at crossover distortion in push-pull amps. Certain topologies appear to have significantly less crossover distortion than others when jammed into clipping.
When people "roll" or compare tubes, I've never heard of a case where they really re-normalized the circuit variables for each change of tube legitimately, so I take any opinions with a big grain of salt. One guy who did do what might have been a good re-normalization said he could hear no difference at all (6V6 vs EL84). He didn't mention if the circuit had significant negative feedback, so who knows what to read into that.
Even tubes of the same number (12AX7's for example) will vary in a way that's not necessarily consistent. Tube tolerances have always been relatively loose, which makes comparisons more difficult. This can lead tube rollers to believe a certain brand or vintage tube is better sounding, when it seems just as likely that one tube worked better with the given circuit values and output loading than the others.
Output transformer loading appears to always cause some compression in a push-pull circuit or I.M. distortion and compression in a single ended topology, which will get largely corrected for if the circuit uses significant negative feedback.
There are those who claim less feedback sounds better. In many circuits I've seen negative feedback cause the higher order harmonic distortion products to extend further from the fundamental, which I don't like the sound of. With no feedback you get a weak output impedance driving the resonant and reactive load of a speaker, and nothing corrects for the compression due to the output tranny loading of the plates. I haven't found any hard core audiophiles who will admit it, but some compression IMO makes music more listenable over time. It's less irksome.
Personally I have settled on about 12dB of negative feedback in my Hi-Fi tube amp, a sort of compromise between ability to drive a speaker properly, and still able to have a relatively fast roll-off of harmonic distortion products. This sounds more natural and musical, and also gives the amp reasonable control over the speaker resonance and reactivity in general.
Since peak levels in music are often 10dB higher than average, many Hi-Fi tube amps may hit clipping more often than the user realizes, so I not only worry about how clipping sounds but also look closely at crossover distortion in push-pull amps. Certain topologies appear to have significantly less crossover distortion than others when jammed into clipping.
When people "roll" or compare tubes, I've never heard of a case where they really re-normalized the circuit variables for each change of tube legitimately, so I take any opinions with a big grain of salt. One guy who did do what might have been a good re-normalization said he could hear no difference at all (6V6 vs EL84). He didn't mention if the circuit had significant negative feedback, so who knows what to read into that.
Even tubes of the same number (12AX7's for example) will vary in a way that's not necessarily consistent. Tube tolerances have always been relatively loose, which makes comparisons more difficult. This can lead tube rollers to believe a certain brand or vintage tube is better sounding, when it seems just as likely that one tube worked better with the given circuit values and output loading than the others.
Output transformer loading appears to always cause some compression in a push-pull circuit or I.M. distortion and compression in a single ended topology, which will get largely corrected for if the circuit uses significant negative feedback.
There are those who claim less feedback sounds better. In many circuits I've seen negative feedback cause the higher order harmonic distortion products to extend further from the fundamental, which I don't like the sound of. With no feedback you get a weak output impedance driving the resonant and reactive load of a speaker, and nothing corrects for the compression due to the output tranny loading of the plates. I haven't found any hard core audiophiles who will admit it, but some compression IMO makes music more listenable over time. It's less irksome.
Personally I have settled on about 12dB of negative feedback in my Hi-Fi tube amp, a sort of compromise between ability to drive a speaker properly, and still able to have a relatively fast roll-off of harmonic distortion products. This sounds more natural and musical, and also gives the amp reasonable control over the speaker resonance and reactivity in general.
Last edited:
From Bob Richards, all good including some often overlooked points.
It is a question of what compromise one will settle for. E.g. negative feedback (what I think is common, some call quite a lot!) will cause immediate on-set of high order distortion after clipping which little or zero feedback will not to a bothersome extent. I reluctantly have to concede that overloading is a factor in medium output amplifier, although I am reluctant to design too much for that, to the detriment of desired features with normal listening. (Where does my music spend most of its time?)
I have used both tube types extensively in designs, though 6V6 before EL84. (As an aside I have also found 6V6 characteristics more consistent than for EL84 - but that may be because of brand.) So I cannot immediately jump to a personal preference - as said drive considerations; horses for courses to use a cliché. I have used 6V6s under higher voltages than EL84s (but the real limitation of electrode voltages provided dissipation is kept in line, is another story. There seems to be different explanations.)
(Another aside, if I can do so without flaming, is the often fixation that "negative feedback is bad". Where does this urban legend come from? 20 - 26 dB can be quite in order provided the design is kosher to begin with. Generation of higher order harmonics? Not unless .... but there, I am off topic!)
So just to support Bob, one must be careful of drawing conclusions from comparison of individual situations. Such do not represent valid statistics.
It is a question of what compromise one will settle for. E.g. negative feedback (what I think is common, some call quite a lot!) will cause immediate on-set of high order distortion after clipping which little or zero feedback will not to a bothersome extent. I reluctantly have to concede that overloading is a factor in medium output amplifier, although I am reluctant to design too much for that, to the detriment of desired features with normal listening. (Where does my music spend most of its time?)
I have used both tube types extensively in designs, though 6V6 before EL84. (As an aside I have also found 6V6 characteristics more consistent than for EL84 - but that may be because of brand.) So I cannot immediately jump to a personal preference - as said drive considerations; horses for courses to use a cliché. I have used 6V6s under higher voltages than EL84s (but the real limitation of electrode voltages provided dissipation is kept in line, is another story. There seems to be different explanations.)
(Another aside, if I can do so without flaming, is the often fixation that "negative feedback is bad". Where does this urban legend come from? 20 - 26 dB can be quite in order provided the design is kosher to begin with. Generation of higher order harmonics? Not unless .... but there, I am off topic!)
So just to support Bob, one must be careful of drawing conclusions from comparison of individual situations. Such do not represent valid statistics.
Where indeed does the urban legend "negative feedback is bad" come from?
In the 1940's and 1950's, the pre-eminent electronics magazine for home constructors and even some professionals was the English magazine Wireless World, in which the famous Williamson amp was published.
In the late 1940's, the respected author "Cathode Ray" (Markus Scroggie) had an article in Wireless World in which he analysed the effect on a triode amplifier of feedback ranging from zero feedback thru to moderate amounts. He showed very sucinctly and clearly with simple math that an amplifier that with no feedback had only 2nd harmonic distortion, low amounts of feedback would reduce the 2nd harmonic but produce significant 3rd harmonic distortion and trace amounts of higher harmonics. With greater feedback, the 3rd goes down as well, but the 5th starts coming up. Since 3rd and higher orders are nastier than the 2nd, he concluded that low amounts of feedback was bad. Note that with cost saving in capacitors and transformers, you can only have low amounts of feedback, or the amp will be unstable.
His math was sound, but his conclusion was wrong. ANY amount of feedback SOUNDS better than none at all. Many tube radios were made with only a few dB of neg feedback for just that reason. His mistake is that he only considered harmonic distortion. If he considered intermodulation distortion, he would have concluded differently, as negative feedback ALWAYS lowers intermodulation, and it is intermod that our ears and brain object to.
Applying neg feedback can, unless you know what you are doing, make an amp go into supersonic oscillation. You don't hear the oscillation, but the distortion caused by the oscillation overloading the amp sounds horrible.
And we must not loose sight of the fact that for a particular person, what sounds good to him may only be good for him. If you like 2nd harmonic distortion, and many do, then go for it. I once had experience with a chap who claimed a certain loudspeaker was better than another, known to have a flatter response and lower distortion, because with his preferred system he could tell the difference between two violins. I could too, when he demonstrated. Actually, a peak in his speaker response happened to coincide with an odd resonance in one violin. This is how you can trick yourself in audio.
Keit
In the 1940's and 1950's, the pre-eminent electronics magazine for home constructors and even some professionals was the English magazine Wireless World, in which the famous Williamson amp was published.
In the late 1940's, the respected author "Cathode Ray" (Markus Scroggie) had an article in Wireless World in which he analysed the effect on a triode amplifier of feedback ranging from zero feedback thru to moderate amounts. He showed very sucinctly and clearly with simple math that an amplifier that with no feedback had only 2nd harmonic distortion, low amounts of feedback would reduce the 2nd harmonic but produce significant 3rd harmonic distortion and trace amounts of higher harmonics. With greater feedback, the 3rd goes down as well, but the 5th starts coming up. Since 3rd and higher orders are nastier than the 2nd, he concluded that low amounts of feedback was bad. Note that with cost saving in capacitors and transformers, you can only have low amounts of feedback, or the amp will be unstable.
His math was sound, but his conclusion was wrong. ANY amount of feedback SOUNDS better than none at all. Many tube radios were made with only a few dB of neg feedback for just that reason. His mistake is that he only considered harmonic distortion. If he considered intermodulation distortion, he would have concluded differently, as negative feedback ALWAYS lowers intermodulation, and it is intermod that our ears and brain object to.
Applying neg feedback can, unless you know what you are doing, make an amp go into supersonic oscillation. You don't hear the oscillation, but the distortion caused by the oscillation overloading the amp sounds horrible.
And we must not loose sight of the fact that for a particular person, what sounds good to him may only be good for him. If you like 2nd harmonic distortion, and many do, then go for it. I once had experience with a chap who claimed a certain loudspeaker was better than another, known to have a flatter response and lower distortion, because with his preferred system he could tell the difference between two violins. I could too, when he demonstrated. Actually, a peak in his speaker response happened to coincide with an odd resonance in one violin. This is how you can trick yourself in audio.
Keit
Last edited:
The 6V6 was engineered by RCA to provide a few watts of reasonably clean audio in "All American Five" type radios. These radios were built to a price, but that doesn't mean they didn't puit a lot of thought into the design. They were built with no or only a few dB of feedback.
It is possible to engineer a tetrode to produce lower distortion than a pentode and that is just what the RCA tube engineers did. Without feedback, an audio power tetrode will produce less distortion than a pentode with no feedback.
And the 6V6 meant RCA avoided paying royalties to Philips, who owned key pentode patents. That's why tetrodes had more impact in USA than in Europe.
However, in a quality amplifer, there is more gain available, and more expensive capacitors and transformer are used. This lets you use more feedback, and then a pentode will generally do better. And pentodes need less drive, lowering distortion in the driver, as others have pointed out.
It is possible to engineer a tetrode to produce lower distortion than a pentode and that is just what the RCA tube engineers did. Without feedback, an audio power tetrode will produce less distortion than a pentode with no feedback.
And the 6V6 meant RCA avoided paying royalties to Philips, who owned key pentode patents. That's why tetrodes had more impact in USA than in Europe.
However, in a quality amplifer, there is more gain available, and more expensive capacitors and transformer are used. This lets you use more feedback, and then a pentode will generally do better. And pentodes need less drive, lowering distortion in the driver, as others have pointed out.
Last edited:
6F6 and 6V6 are not entirely comparable, and my understanding is that RCA developed the 6L6 beam power tetrode primarily to circumvent the Philips' Pentode patent and thereby avoid paying the licensing fees.
Compare the tetrode version of the 6CA7 with the pentode version EL34/6CA7 for a fairer comparison. (I have not made the comparison, but I believe this would be fairer since they are interchangeable)
Compare the tetrode version of the 6CA7 with the pentode version EL34/6CA7 for a fairer comparison. (I have not made the comparison, but I believe this would be fairer since they are interchangeable)
6CA7 was manufactured both as pentode and tetrode in this country which is why I specifically referenced it.
http://www.mif.pg.gda.pl/homepages/frank/sheets/127/6/6CA7.pdf
The RCA 6CA7 is also a pentode.
Sylvania as far as I know is the only company that manufactured a beam tetrode version of the 6CA7. (It's alleged that GE did as well, but the only ones I've seen were clearly rebranded Sylvanias) Even my Sylvania data book features only the pentode version. I have seen/owned bonafide Sylvania 6CA7 beam power tetrodes, and Sovtek makes a "copy" of the large bottle Sylvania 6CA7 which is popular in guitar applications.
http://www.mif.pg.gda.pl/homepages/frank/sheets/127/6/6CA7.pdf
The RCA 6CA7 is also a pentode.
Sylvania as far as I know is the only company that manufactured a beam tetrode version of the 6CA7. (It's alleged that GE did as well, but the only ones I've seen were clearly rebranded Sylvanias) Even my Sylvania data book features only the pentode version. I have seen/owned bonafide Sylvania 6CA7 beam power tetrodes, and Sovtek makes a "copy" of the large bottle Sylvania 6CA7 which is popular in guitar applications.
You have a very rich and technically advanced country, a lot of competition. Range of handsets and other devices is very large. It is hard to understand everything.
Difference in the electrical parameters of the ray beam tetrode and pentode not. These characteristics are achieved by the internal construction.
In the USSR produced tube 6P27S in references indicated as an analogue EL34. In fact, it was a copy of tetrode 6CA7. Which company I do not know, I can show photos.
Difference in the electrical parameters of the ray beam tetrode and pentode not. These characteristics are achieved by the internal construction.
In the USSR produced tube 6P27S in references indicated as an analogue EL34. In fact, it was a copy of tetrode 6CA7. Which company I do not know, I can show photos.
I would guess Reflektor in Saratov, now known as Expo Pul and owned by New Sensor Corp. They currently make both tetrode and pentode versions of the 6CA7/EL34 family.
One of the best I've encountered was the American made Amperex. (Holland version even better I'd suspect) They were OEM in Marantz 9..
One of the best I've encountered was the American made Amperex. (Holland version even better I'd suspect) They were OEM in Marantz 9..
From what I've read, it sounds like the audible difference between the 6V6 and 6BQ5/EL84 is pretty minimal, easily dominated by the circuit and speaker, and maybe the bigger difference is reliability. The 6V6 is said to have more space between certain elements, resulting in the lower gain, but suggesting both better reliability and consistency from tube to tube. If that's true, that may be the more significant difference. Tube gain is too easily dealt with upstream to be a significant issue IMO. The 6BQ5 filiment draws about 30% more current than the 6V6. The 6V6 being physically larger may get rid of heat better, suggesting a longer life. I'm thinking the 6V6 wins for use in a Hi-Fi amp, but not by much.
You're right, let's finish offtopic.Would be interested in seeing pix..
Wondering if we should start a new thread, what do you think? (My fault for going OT) I can obviously move the relevant posts.
Photo 6P27S easy to find in Google.
The EL84/6BQ5 and slightly later 7189/7189A were far more commonly used in hi-fi applications than the 6V6; Acrosound, Dynaco, Eico, EMI, Fisher, Heathkit, Leak, Pioneer, Sansui, and Realistic all used the 6BQ5, Scott used the 7189/7189A. Pilot used one or the other. The only stereo hi-fi amplifier using the 6V6 I have ever seen was made by Stromberg-Carlson, it was a nice amp however.
Scott employed fixed bias and pentode connection with the 7189/7189A, Dynaco cathode bias and UL. Most Heathkit integrated amps used cathode bias and UL connection.
The best sounding examples I have owned or heard all employed UL connection, and indeed the Mullard based designs from Leak, EMI and others also were UL.
The best sounding 6BQ5 amp I have heard was the Acrosound 2020 followed surprisingly closely by a Dynaco ST-35. The two really could not be more different topologically speaking.
I reinterpreted a classic 6J7/6V6 SE design here (switchable triode/tetrode connection, similar topology to the inspiration) some number of years ago, and I thought the end result did not entirely justify the effort I put into it; that said for some reason it works exceptionally well with some versions of the Altec 614 coaxial driver.
Scott employed fixed bias and pentode connection with the 7189/7189A, Dynaco cathode bias and UL. Most Heathkit integrated amps used cathode bias and UL connection.
The best sounding examples I have owned or heard all employed UL connection, and indeed the Mullard based designs from Leak, EMI and others also were UL.
The best sounding 6BQ5 amp I have heard was the Acrosound 2020 followed surprisingly closely by a Dynaco ST-35. The two really could not be more different topologically speaking.
I reinterpreted a classic 6J7/6V6 SE design here (switchable triode/tetrode connection, similar topology to the inspiration) some number of years ago, and I thought the end result did not entirely justify the effort I put into it; that said for some reason it works exceptionally well with some versions of the Altec 614 coaxial driver.
Certainly in the 20W and under per channel class of amplifiers this would be true. Octals were more expensive to make, and were quickly supplanted by the 7 and 9 pin types in cost / space sensitive applications, I find it somewhat interesting that these tubes arrived only shortly after the first octals. A decade before they were still using UX,UY, and related bases.
kevinkr thanks for the detailed response.
Transition from tubes oktal to noval technical development. 6V6 old tube.
Even in the backward Soviet Union after 1950 6V6 (6P6S) was not used.
Completely switched to EL84 (6P14P).
Noval was released version 6P6S - 6P1P.
But it was not popular, it was replaced by 6P14P.
Transition from tubes oktal to noval technical development. 6V6 old tube.
Even in the backward Soviet Union after 1950 6V6 (6P6S) was not used.
Completely switched to EL84 (6P14P).
Noval was released version 6P6S - 6P1P.
But it was not popular, it was replaced by 6P14P.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.