I'm looking for some sonic comparisons between these much-used octal outputs wired in triode. Plus any fairly close octal equivalents like KT66, KT77, KT120, KT150 etc. Not including the lower power octals like the 6V6 family which we know to be good - just those with dissipation of 18W or over in triode.
Which of these have you used, and which would you recommend on the basis of sound quality? Thanks.
Which of these have you used, and which would you recommend on the basis of sound quality? Thanks.
I'm using a SE amp, but it could be push-pull also. Just looking for the best sound quality used in triode.
andyjevans,
In my experience:
Single Ended Triode Wired:
6L6GC and KT66 . . . KT77 and EL34 . . . 7591.
They all "sounded" real good.
I grouped them into 3 groups, the ones that could just "Plug and Play" substitute for each other. (Without adjusting self bias, driver, output transformer, etc.).
Push Pull Triode wired:
6L6GC and KT66 and 5881 . . . KT77 and EL34 . . . 7591.
They all "sounded" real good.
I grouped them into 3 groups, the ones that could just "Plug and Play" substitute for each other. (Without adjusting self bias, driver, output transformer, etc.). Don't forget to wire pin 1 to pin 2 for the EL34.
Caution: the 7591 has different pin out. Ask me how I know . . . I forgot that fact, and had to instantly turn the amplifier off.
Design Decisions:
Please know that I never operate any of the above tubes at or near their maximum ratings.
All my amplifiers are low power or moderate power.
No global negative feedback. Using Triode Wired plus (*) eliminates the need for global negative feedback
(*) I use output transformers with higher than normal primary impedance than many designers would use versus the Triode Wired plate impedance, rp.
The above is a lot different than how some designers and builders run their tubes and output transformers.
Sorry that I can not tell you which of the above tubes "sound" I like best.
. . . Perhaps my design decisions and topologies make them "sound" similar.
Keep on designing, building, listening, and enjoying!
(I do not have to tell You that, it is just a message for some of the other readers).
Have Fun Everyone!
In my experience:
Single Ended Triode Wired:
6L6GC and KT66 . . . KT77 and EL34 . . . 7591.
They all "sounded" real good.
I grouped them into 3 groups, the ones that could just "Plug and Play" substitute for each other. (Without adjusting self bias, driver, output transformer, etc.).
Push Pull Triode wired:
6L6GC and KT66 and 5881 . . . KT77 and EL34 . . . 7591.
They all "sounded" real good.
I grouped them into 3 groups, the ones that could just "Plug and Play" substitute for each other. (Without adjusting self bias, driver, output transformer, etc.). Don't forget to wire pin 1 to pin 2 for the EL34.
Caution: the 7591 has different pin out. Ask me how I know . . . I forgot that fact, and had to instantly turn the amplifier off.
Design Decisions:
Please know that I never operate any of the above tubes at or near their maximum ratings.
All my amplifiers are low power or moderate power.
No global negative feedback. Using Triode Wired plus (*) eliminates the need for global negative feedback
(*) I use output transformers with higher than normal primary impedance than many designers would use versus the Triode Wired plate impedance, rp.
The above is a lot different than how some designers and builders run their tubes and output transformers.
Sorry that I can not tell you which of the above tubes "sound" I like best.
. . . Perhaps my design decisions and topologies make them "sound" similar.
Keep on designing, building, listening, and enjoying!
(I do not have to tell You that, it is just a message for some of the other readers).
Have Fun Everyone!
I went through a phase of triode-wiring beam outputs and power pentodes. I tried 7591, 6550A, EL34, 6L6GC, 6P3S-E, 6P3S, and 6V6GTA/B. I never did try EL84 in triode, for some reason. I'd like to try EL86 in triode. That's supposed to be very good.
Of course the circuit and OPT will be the most important factors for final sound quality. But remembering that, I can only say that I liked this or that one better than the other. Purely subjective. Just my own personal preferences, built up over several years trying this and that.
In order of my preference from back then, I liked:
The above is to be taken with a large grain of salt because all of this taste-testing was done back in the 1990s, about 30 years ago. I have no idea what I'd prefer now.
PS - I agree with 6A3sUMMER about running push-pull triodes without gNFB into an OPT with higher than usual primary impedance. You get less power out, but you also get lower distortion at typical listening volumes, along with a lower output impedance to better drive your speakers. I had great results running push-pull 300Bs into a 6k plate-to-plate load (UTC LS63 trafo's). I also liked push-pull 2A3s into a 5k p-p primary (Tango XE-45-5). I'd like to try a pair of EL34-triode into one of those 8k ohm primary Edcor 25W OPTs. Someday, maybe...
Of course the circuit and OPT will be the most important factors for final sound quality. But remembering that, I can only say that I liked this or that one better than the other. Purely subjective. Just my own personal preferences, built up over several years trying this and that.
In order of my preference from back then, I liked:
- 6V6GTA/B - Sweetest sound in triode. Low power limits its usefulness.
- 6P3S - The Russian 6L6GB clone with the larger octal base. I thought that sounded something like the 6V6GT types. Still not a lot of power, though.
- EL34 - Good compromise between more power and that triode 'sweetness' we like (or think we like).
- It was a tie between 6L6GC and 6550A, depending on which brand, etc. They sound pretty similar, but of course the 6550A will get you more power.
- 6P3S-E - The Sovtek '5881' from back in the day, with the wafer base and the thick glass. It sounded a bit dull and less 'detailed' than an EL34 or 6V6 in triode. One neat thing is that with cathode bias, all the Sovtek '5881' I tried biased up almost exactly the same as EL34.
- My least favorite in triode was 7591. I just thought it had a bit of an electronic 'edge' to its sound. That could have been the circuit I tried it in. I'd have to do it all over again to know for sure.
The above is to be taken with a large grain of salt because all of this taste-testing was done back in the 1990s, about 30 years ago. I have no idea what I'd prefer now.
PS - I agree with 6A3sUMMER about running push-pull triodes without gNFB into an OPT with higher than usual primary impedance. You get less power out, but you also get lower distortion at typical listening volumes, along with a lower output impedance to better drive your speakers. I had great results running push-pull 300Bs into a 6k plate-to-plate load (UTC LS63 trafo's). I also liked push-pull 2A3s into a 5k p-p primary (Tango XE-45-5). I'd like to try a pair of EL34-triode into one of those 8k ohm primary Edcor 25W OPTs. Someday, maybe...
Last edited:
The 6L6-oids are not known for producing much power in trioded SE, but IMO they can sound quite good.
I have a pair of mono blocks that runs 807s in SE with 6E5P cathode followers to push the grids positive for an extra watt or so and I'm quite happy with the sound. The tubes operate at 370V 50mA with a high-ish 8k plate load.
I have a pair of mono blocks that runs 807s in SE with 6E5P cathode followers to push the grids positive for an extra watt or so and I'm quite happy with the sound. The tubes operate at 370V 50mA with a high-ish 8k plate load.
I've been trying out what I have in SE, no global feedback.
- 6L6 marked Ruby.... not tried yet
- EL34 were marked International. A light, airy sound with good treble. The downside was a slightly grainy and coarse sound. Voices were OK but piano was rather wooden. This was into 3.5K.
- 6P3S, marked Sovtek 5881, were much better tonally with good voices and piano sound. This into the same 3.5K OPT but on the 4 ohm tap. But I have a problem with the high treble - it sounds rather down. Can anyone confirm if they had problems with the highs when using this tube in triode? I want to pursue this since the tone of this tube is excellent.
Does your Sovtek 5881 look like this? That's actually a 63PS-E
Or does it look like this? This is a plain 6P3S (no "-E").
If your "5881" is a 6P3S-E, then yes, I noticed that the high frequencies sounded kind of dull, but they're not measurably rolled off. I even noticed this in Fender guitar amps. The 6P3S-E would take the 'sparkle' out of the sound. Or you could say the 6P3S-E reduces some of the 'edge' and overly bright attack some.
From memory, I agree with the EL34-triode 'airy' assessment. I found the Russian 6P3S (no "-E") gave some of that, but without the edginess of the EL34 (or the 7591). The Russian 6P3S-E ("5881") was anything but 'airy'. At least, that's what my impressions were 30 years ago.
Or does it look like this? This is a plain 6P3S (no "-E").
If your "5881" is a 6P3S-E, then yes, I noticed that the high frequencies sounded kind of dull, but they're not measurably rolled off. I even noticed this in Fender guitar amps. The 6P3S-E would take the 'sparkle' out of the sound. Or you could say the 6P3S-E reduces some of the 'edge' and overly bright attack some.
From memory, I agree with the EL34-triode 'airy' assessment. I found the Russian 6P3S (no "-E") gave some of that, but without the edginess of the EL34 (or the 7591). The Russian 6P3S-E ("5881") was anything but 'airy'. At least, that's what my impressions were 30 years ago.
Thanks Rongon - that agrees completely with what I'm finding. The EL34 as slightly "edgy", agree there. Are there any EL34 types that don't have this slight edginess and have a better piano tone?
My tubes are the coin base 6P3S-E. And yes - dull in the extreme treble, noticeably on drumkit with snare rimshots and cymbals. This is a real shame, since the tone is lovely.
I'll have to try the Russian 6P3S (no "-E"). Maybe the Ruby 6L6 is the same tube?
My tubes are the coin base 6P3S-E. And yes - dull in the extreme treble, noticeably on drumkit with snare rimshots and cymbals. This is a real shame, since the tone is lovely.
I'll have to try the Russian 6P3S (no "-E"). Maybe the Ruby 6L6 is the same tube?
andyjevans,
Ideas . . .
1. A triode wired pentode, and a triode wired beam power tube . . .
There is the Miller Effect Capacitance of screen g2 to grid g1. Also, g1 is no longer shielded from the plate, so Miller Effect there too.
With the low u of triode wired output tubes, it may not be a factor, but it is worth checking (versus the driver output impedance).
2. Reducing the plate load on triode wired pentodes and beam power output tubes, reduces the 2nd harmonic distortion, which leads us to # 3 . . .
3. Perhaps there is less harmonic distortion, so the higher frequencies seem to be rolled off (the distortion products you heard, are no longer there).
In that case, congratulations to Triodes (or triode wired).
4. A pentode or beam power tube, Not triode wired, and without negative feedback, can overcome part of the high frequency roll off of the primary to secondary, due to the leakage inductance. That is especially true if the primary distributed capacitance is low.
The plate is effectively a current source, its voltage swings further when the primary impedance increases because of the lighter load impedance at high frequencies . . . due to the leakage inductance that no longer couples the output load back to the primary, and so does not couple it back to the plate.
I remember hearing about this effect, and I seem to remember that [old] radio designers used this as a trick to make up for the loss of high frequencies of the speaker driver.
5. For single ended, I generally prefer 5k primaries for Triode Wired pentode and beam power output tubes.
6. The parts and topology of a tube amplifier is a system, with either Positive Synergy, or with Negative Synergy.
I apologize to those who do not want me to use the word 'Synergy' that way, but I am defended by the 1968 Random House Dictionary, College Edition.
I admit, part of me is still locked into 1968, instead of our modern day. Oh, vacuum tubes go back much further than 1968!
Hang in there, you will get some real good "sounding" SE amplifiers with triode wired output tubes, and without global negative feedback.
Ideas . . .
1. A triode wired pentode, and a triode wired beam power tube . . .
There is the Miller Effect Capacitance of screen g2 to grid g1. Also, g1 is no longer shielded from the plate, so Miller Effect there too.
With the low u of triode wired output tubes, it may not be a factor, but it is worth checking (versus the driver output impedance).
2. Reducing the plate load on triode wired pentodes and beam power output tubes, reduces the 2nd harmonic distortion, which leads us to # 3 . . .
3. Perhaps there is less harmonic distortion, so the higher frequencies seem to be rolled off (the distortion products you heard, are no longer there).
In that case, congratulations to Triodes (or triode wired).
4. A pentode or beam power tube, Not triode wired, and without negative feedback, can overcome part of the high frequency roll off of the primary to secondary, due to the leakage inductance. That is especially true if the primary distributed capacitance is low.
The plate is effectively a current source, its voltage swings further when the primary impedance increases because of the lighter load impedance at high frequencies . . . due to the leakage inductance that no longer couples the output load back to the primary, and so does not couple it back to the plate.
I remember hearing about this effect, and I seem to remember that [old] radio designers used this as a trick to make up for the loss of high frequencies of the speaker driver.
5. For single ended, I generally prefer 5k primaries for Triode Wired pentode and beam power output tubes.
6. The parts and topology of a tube amplifier is a system, with either Positive Synergy, or with Negative Synergy.
I apologize to those who do not want me to use the word 'Synergy' that way, but I am defended by the 1968 Random House Dictionary, College Edition.
I admit, part of me is still locked into 1968, instead of our modern day. Oh, vacuum tubes go back much further than 1968!
Hang in there, you will get some real good "sounding" SE amplifiers with triode wired output tubes, and without global negative feedback.
Last edited:
I actually found that when I compared US-made 12AX7, 6SN7 or 6DJ8 to Sovtek 12AX7WB or Russian 6N8S or 6N23P that the Russian tubes always seemed to have that sort of 'dark', mellowed treble response (subjectively). It works really well in some circuits that might sound a bit edgy with American tubes in them. Some circuits using 6DJ8 can sound quite 'zingy', so substituting a 6N23P (aka Sovtek 6922) can mellow them out nicely. Some circuits using 6SN7 can sound a bit 'hard', so subbing in a 6N8S can soften the blow a bit. But I think this will depend on what you like, subjectively.
Last edited:
Does that mean you generally like 10k plate-plate OPT primary impedance for triode-wired pentodes?
If so, I've tried that too. I remember I liked that for trioded 6550A (especially the Sovtek ones), but I remember it made EL34-triodes sound even more edgy and 'thin'. I had a pair of UTC LS63 back then, which had 10k primaries with taps for 6k operation. I switched back and forth to taste-test.
andyjevans - You might want to pick up a couple of Russian 6P3S and try those. They sound quite different from the coin-base 6P3S-E type. I like them better, but they tend to red-plate with >400V applied to plate and screen (which is where most guitar amps use 6L6s). I figure the 6P3S is more of an equivalent to the old 6L6GA type, with max plate volts = 360V and max screen volts = 270V. The 6P3S-E can take 450V on both plate and screen and last a good long time with no issues. Very rugged.
What make of EL34 do you have at the moment?
The Sovtek EL34 or 6CA7 might be worth a try.
I remember the old Mullard EL34 really did sound better to me than other EL34s. But that could have been confirmation bias or whatever. There was a bit of a mania among guitar players for collecting the 'best' output tubes, and EL34s were what came in those 1960s Marshall stacks. They were also original equipment in Dynaco Stereo 70 amplifiers and the like.
Another one that might be worth trying is the 807. A few members here say they sound very good. They might work well trioded.
Wavebourn and kodabmx used to say the 6P43P-E makes a very good triode. That's equivalent to EL86 (9 pin mini).
What make of EL34 do you have at the moment?
The Sovtek EL34 or 6CA7 might be worth a try.
I remember the old Mullard EL34 really did sound better to me than other EL34s. But that could have been confirmation bias or whatever. There was a bit of a mania among guitar players for collecting the 'best' output tubes, and EL34s were what came in those 1960s Marshall stacks. They were also original equipment in Dynaco Stereo 70 amplifiers and the like.
Another one that might be worth trying is the 807. A few members here say they sound very good. They might work well trioded.
Wavebourn and kodabmx used to say the 6P43P-E makes a very good triode. That's equivalent to EL86 (9 pin mini).
Last edited:
Some questions...
Re: Miller effect of g2 to g1 -- In a trioded pentode, g2 is shorted to the plate (anode). Doesn't that also short out the capacitance of g2 to g1? Now you'll have the Miller effect of a+g2 to g1. How can the capacitance of a to g2 be separate from the capacitance of g2 to g1 once you've triode-wired the pentode?
Re: Reducing plate load on a triode reduces H2 ... Really? Normally, THD goes up when the triode is more heavily loaded, both H2 and H3. The classical idea is that you increase the load impedance on the triode's plate to reduce THD (and increase the B+ voltage to compensate), as the triode will work into a flatter load line, for wider swing with lower THD. A steeper load line reduces the available voltage swing, putting the triode closer to grid current as it swings to 0V and closer to cutoff at the higher plate voltage end of the load line. I'm sure I must be missing something here...
rongon,
Thanks for the question!
Triode wired pentodes and beam power tubes:
For push pull, I use 5k, 6k, and 10k plate to plate.
For push pull, that selection is 'allowed'/advised, because: the damping factor of push pull is symmetrical; the 2nd harmonic distortion is reduced by push pull output, and smaller wider bandwidth transformers can be used.
True, even if the Ratio of rp to primary plate to plate impedance for push pull . . .
is lower than the Ratio of rp to primary impedance for single ended.
Single ended is . . . so different than push pull (as noted above).
I hope that explains my reasoning.
Thanks for the question!
Triode wired pentodes and beam power tubes:
For push pull, I use 5k, 6k, and 10k plate to plate.
For push pull, that selection is 'allowed'/advised, because: the damping factor of push pull is symmetrical; the 2nd harmonic distortion is reduced by push pull output, and smaller wider bandwidth transformers can be used.
True, even if the Ratio of rp to primary plate to plate impedance for push pull . . .
is lower than the Ratio of rp to primary impedance for single ended.
Single ended is . . . so different than push pull (as noted above).
I hope that explains my reasoning.
Oh my gosh!
My compositions are often very unclear.
I apologize.
1. Reducing the load on the plate . . . I meant Increasing the impedance that the plate has to drive. The higher the impedance the plate "sees", the lower 'the loading' is on the plate (easier load, not worse load).
My English teacher would have given me an F grade.
Pentode / beam power modes:
The screen is at AC / Signal ground. It is a shield between g1 and the plate. g1 to plate capacitance is very, very small. Miller Effect Capacitance is very low, even if gain is very high.
g1 to g2 capacitance has no Miller Effect Capacitance, g2 is not moving.
Pentode / beam power tubes that are Triode Wired:
g1 to plate miller effect capacitance is a factor, the screen, g2 no longer shields g1 from the plate.
With g2 now moving with the plate, there is Miller Effect Capacitance from g1 to g2.
Miller Effect Capacitance is Real Capacitance x Gain. More accurately (1 + gain) x capacitance
Real Capacitances do not change. The only exceptions are:
connections that cause Shielding, versus connections that do not cause shielding,
Movement due to mechanical vibrations,
(*) And, electrostatic forces that change when the signal voltage changes, causing the electrostatic forces between two elements to change at a signal rate, thus vibrating the element(s) when the signal frequency is the same as the element(s) mechanical resonant frequency.
(*) this effect does not happen very often, but I had an incident that was very repeatable, a certain note would activate the mechanical resonance because of the varying electrostatic field strength. No, it was not speaker to tube activated. It was caused by Electrostatic forces that were varying as signal voltage varied.
My compositions are often very unclear.
I apologize.
1. Reducing the load on the plate . . . I meant Increasing the impedance that the plate has to drive. The higher the impedance the plate "sees", the lower 'the loading' is on the plate (easier load, not worse load).
My English teacher would have given me an F grade.
Pentode / beam power modes:
The screen is at AC / Signal ground. It is a shield between g1 and the plate. g1 to plate capacitance is very, very small. Miller Effect Capacitance is very low, even if gain is very high.
g1 to g2 capacitance has no Miller Effect Capacitance, g2 is not moving.
Pentode / beam power tubes that are Triode Wired:
g1 to plate miller effect capacitance is a factor, the screen, g2 no longer shields g1 from the plate.
With g2 now moving with the plate, there is Miller Effect Capacitance from g1 to g2.
Miller Effect Capacitance is Real Capacitance x Gain. More accurately (1 + gain) x capacitance
Real Capacitances do not change. The only exceptions are:
connections that cause Shielding, versus connections that do not cause shielding,
Movement due to mechanical vibrations,
(*) And, electrostatic forces that change when the signal voltage changes, causing the electrostatic forces between two elements to change at a signal rate, thus vibrating the element(s) when the signal frequency is the same as the element(s) mechanical resonant frequency.
(*) this effect does not happen very often, but I had an incident that was very repeatable, a certain note would activate the mechanical resonance because of the varying electrostatic field strength. No, it was not speaker to tube activated. It was caused by Electrostatic forces that were varying as signal voltage varied.
Last edited:
That IS a shame, as I've used them in my old Jolida 302b, triode strapped, and never thought they were dull sounding. They were my favorite over RFT EL34's and "super" 6BG6GA's. This is PP Class AB amp.
jeff