I had seen Morgan's design before I started my work, and I think I made reference to his amp somewhere in this thread or in my universal driver thread that preceeded this one. I have tinkered with similar LTP feeding LTP before when playing with screen drive. That design had coupling caps in both places because I was trying to extract maximum drive voltage from a limited B+ supply. The PC board was laid out for 4 coupling caps, and I simply jumpered one set and tested, then tried it the other way. I preferred the caps at the mosfet inputs since it allows one less stage to affect the bias on the output stage. Morgan preferred it the other way, and both ways are valid.
That's pretty much the way I see it too. The input signal is split between the two input triodes. I measure the output with my scope from ground to the plate. After all we spec the desired grid drive from grid to ground.
If you looked from plate to plate of the LTP you would indeed see the full swing, but one plate goes to each output tube grid, and each tube grid needs the same drive voltage swing. Each individual output tube grid sees only one driver tube plate.
So each 6SN7 LTP is capable of an amplification factor of only about 8?
To figure out gain structure of this type of amp, you look at half of the push-pull pairs?
So if you have a 6L6GC with a -40V bias, and you want to drive it to full power...
The peak input voltage at the driver stage grids needs to be at least 5V (5V * amp factor of 8 = 40), and before that, the first 6SN7 needs to see 625 mV or so (0.625 * 8 = 5).
Right?
Now add more gain for feedback (if desired). That would explain why you'd need a 6SL7 in the first stage. We'd expect to get only about 30X gain from a 6SL7 LTP phase splitter, right?
I hope I'm finally thinking about this the right way, 'cause I've been trying to understand this for some time now...
--
To figure out gain structure of this type of amp, you look at half of the push-pull pairs?
So if you have a 6L6GC with a -40V bias, and you want to drive it to full power...
The peak input voltage at the driver stage grids needs to be at least 5V (5V * amp factor of 8 = 40), and before that, the first 6SN7 needs to see 625 mV or so (0.625 * 8 = 5).
Right?
Now add more gain for feedback (if desired). That would explain why you'd need a 6SL7 in the first stage. We'd expect to get only about 30X gain from a 6SL7 LTP phase splitter, right?
I hope I'm finally thinking about this the right way, 'cause I've been trying to understand this for some time now...
--
In general an LTP phase splitter can be viewed as explained above. The first stage in the amps mentioned do indeed function this way.
The second stage in the amps mentioned does not. It is not a true LTP. If the first stage LTP was absolutely perfect the signals on its two outputs would be identical, but 180 degrees out of phase for all audio frequencies. If these perfect signals were applied to a second LTP there would be perfect cancellation at the cathodes. The second stage would not be an LTP at all. It would function like a common cathode amplifier with the cathode bypassed, and be capable of its full gain (about 15 with a 6SN7 or 6J5). In reality the first stage is not perfect especially at high frequency and when feedback is applied to the second grid. The second LTP stage will iron out the remaining wrinkles giving up a small amount of gain in the process. Attaching a scope probe to the cathode of the second stage usually reveals a small anount of distortion which is being cancelled. I made measurements on my amp when I was developing it but that was a few years ago. I plan to fire it up again soon for more experiments, but that won't happen for a few weeks yet.
If my source (dac or phono stage) supplied a balanced output with equal but out of phase outputs the first stage would also have full gain. This is not common, but can be done.
The grid drive issue needs some more clarification. If we had a 6L6GC with -40 volts on the grid we would need exactly 80 volts peak to peak of drive swinging from -80 volts (well into cutoff) to zero volts to remain in class AB1 (avoiding positive grid operation). More drive will attempt to push the grid positive causing it to draw current. The typical capacitor coupled drivers distort, clip, create blocking, and all sorts of other nasty things when you do this, so it is often avoided.
This thread is all about class AB2, so we will allow the grid to go positive to extract more power from the output tubes. How far can we go? We can go positive until we reach a point where more positive bias causes no more increase in plate current (just more distortion). This depends on the tube being used, its condition, and the operating point. We will assume +10 volts here to make the math easy. A 6L6GC can easilly eat +20 volts on its grid, in some cases more. Some tubes like the 845 can eat +50 volts on the grid and require 400 volts P-P of drive.
Now in AB2 the grid can swing from -90 volts to +10 volts requiring 100 volts P-P of drive. Now we know that the second stage will have a gain of slightly less than 15. Lets just assume its 14.5 for now. We know the first stage gain is about 7 if it is driven from a single ended source. The total gain from the input jack to the output stage grid is 7 * 14.5 or 105. We need 100 volts P-P at each output stage grid making the amp sensitivity about 950 mV P-P in this example.
I would assume that a 6L6GC would bias at about -35 volts unless you were running them above 450 volts, so the sensitivity may be a bit better than this. A KT88 in triode will bias to -50 volts or more so the amp would need a bit over 1 volt P-P to hit clipping. THis correlates with what I was seeing.
Add 6 db of feedback and you need to double the gain. 12 db of feedback requires 4 times the open loop gain. More feedback.....make a solid state amp cause that's what its going to sound like
The second stage in the amps mentioned does not. It is not a true LTP. If the first stage LTP was absolutely perfect the signals on its two outputs would be identical, but 180 degrees out of phase for all audio frequencies. If these perfect signals were applied to a second LTP there would be perfect cancellation at the cathodes. The second stage would not be an LTP at all. It would function like a common cathode amplifier with the cathode bypassed, and be capable of its full gain (about 15 with a 6SN7 or 6J5). In reality the first stage is not perfect especially at high frequency and when feedback is applied to the second grid. The second LTP stage will iron out the remaining wrinkles giving up a small amount of gain in the process. Attaching a scope probe to the cathode of the second stage usually reveals a small anount of distortion which is being cancelled. I made measurements on my amp when I was developing it but that was a few years ago. I plan to fire it up again soon for more experiments, but that won't happen for a few weeks yet.
If my source (dac or phono stage) supplied a balanced output with equal but out of phase outputs the first stage would also have full gain. This is not common, but can be done.
The grid drive issue needs some more clarification. If we had a 6L6GC with -40 volts on the grid we would need exactly 80 volts peak to peak of drive swinging from -80 volts (well into cutoff) to zero volts to remain in class AB1 (avoiding positive grid operation). More drive will attempt to push the grid positive causing it to draw current. The typical capacitor coupled drivers distort, clip, create blocking, and all sorts of other nasty things when you do this, so it is often avoided.
This thread is all about class AB2, so we will allow the grid to go positive to extract more power from the output tubes. How far can we go? We can go positive until we reach a point where more positive bias causes no more increase in plate current (just more distortion). This depends on the tube being used, its condition, and the operating point. We will assume +10 volts here to make the math easy. A 6L6GC can easilly eat +20 volts on its grid, in some cases more. Some tubes like the 845 can eat +50 volts on the grid and require 400 volts P-P of drive.
Now in AB2 the grid can swing from -90 volts to +10 volts requiring 100 volts P-P of drive. Now we know that the second stage will have a gain of slightly less than 15. Lets just assume its 14.5 for now. We know the first stage gain is about 7 if it is driven from a single ended source. The total gain from the input jack to the output stage grid is 7 * 14.5 or 105. We need 100 volts P-P at each output stage grid making the amp sensitivity about 950 mV P-P in this example.
I would assume that a 6L6GC would bias at about -35 volts unless you were running them above 450 volts, so the sensitivity may be a bit better than this. A KT88 in triode will bias to -50 volts or more so the amp would need a bit over 1 volt P-P to hit clipping. THis correlates with what I was seeing.
Add 6 db of feedback and you need to double the gain. 12 db of feedback requires 4 times the open loop gain. More feedback.....make a solid state amp cause that's what its going to sound like
Excellent George, thanks!
So, driving trioded KT88's a little into AB2 will result in an input sensitivity of about 1V or so using cascaded 6SN7 LTPs.......and if we add some feedback this sensitivity will be reduced. Sounds like I won't need a preamp at all, other than a passive volume control.
...and cascaded LTPs offer lower noise, and a bit more distortion cancellation than a "more conventional" front end......
I'm still a little fuzzy about the second stage, but my take-away is since the first stage is doing the phase splitting, we get a bit less than half the mu for gain, and since the second stage is not phase splitting, we get a bit less than the full mu for gain.
OK, next dumb question...why is it that most PP triode designs never use feedback? What's to be gained by applying feedback to a PP triode design? Is it that they are typically minimalist topologies as far as stages go with no gain to burn? Or that the Rp is already low so that there is no need to reduce the output impedance (compared to UL & pentode)? I understand the concept of the gain-bandwidth product so if we apply feedback to extend BW, then we will lose gain. Is it that folks like the distortion spectra from a PP triode just as it is? Or present trendy fashion? From playing with the feedback on my baby huey, I completely understand that too much feedback sounds rather lifeless.
So, driving trioded KT88's a little into AB2 will result in an input sensitivity of about 1V or so using cascaded 6SN7 LTPs.......and if we add some feedback this sensitivity will be reduced. Sounds like I won't need a preamp at all, other than a passive volume control.
...and cascaded LTPs offer lower noise, and a bit more distortion cancellation than a "more conventional" front end......
I'm still a little fuzzy about the second stage, but my take-away is since the first stage is doing the phase splitting, we get a bit less than half the mu for gain, and since the second stage is not phase splitting, we get a bit less than the full mu for gain.
OK, next dumb question...why is it that most PP triode designs never use feedback? What's to be gained by applying feedback to a PP triode design? Is it that they are typically minimalist topologies as far as stages go with no gain to burn? Or that the Rp is already low so that there is no need to reduce the output impedance (compared to UL & pentode)? I understand the concept of the gain-bandwidth product so if we apply feedback to extend BW, then we will lose gain. Is it that folks like the distortion spectra from a PP triode just as it is? Or present trendy fashion? From playing with the feedback on my baby huey, I completely understand that too much feedback sounds rather lifeless.
Ooh, ooh! Can I try?
Triodes have lower intrinsic distortion than pentodes, and what distortion they make is mostly 2nd harmonic, which is a lot easier on the ears than higher order harmonics. So the thinking is that you if you can get distortion low enough, you can do without negative feedback.
My limited experience with this so far is that you can certainly do away with NFB if you're using really good output transformers with directly heated triodes in the output stage (45, 2A3, 300B, etc.).
But if you're using more run-of-the-mill OPTs and/or using triode-wired pentodes (EL84, EL34) or beam tetrodes (6L6, KT88) then you'll probably want a little feedback to linearize the output stage. (Takes an electronic edge off the sound.)
Also, if you run those same triodes in class AB, or even AB2, you'll be running them at a less linear quiescent operating point than if they were biased with more current in "deep class A", so the distortion will be higher from the git-go. Again, a little NFB can help reduce the distortion, at the expense of reduced sensitivity.
I've been playing with a couple of very simple 2-stage power amps, one with PP 2A3's and one (a gutted Dyna ST70) with PP triode-wired EL34's. The 2A3 amp actually sounds worse as soon as I put any global NFB around it (loses some "life" and "air"). But it has really nice Tango OPT's and they're DHT's with very low intrinsic distortion. The EL34 amp usually has -3dB negative feedback, to smooth out a little edginess in the sound (3rd order HD?). It sounds better with -6dB NFB, but I don't have a driver stage in it that can deliver the voltage gain required. (I'll need to fix that.)
When I wire the EL34's back into UL, I can immediately hear a muddier sound. I assume this is the increased harmonic distortion. More NFB is the only way to make it sound acceptable in this config, to my ears at least.
That's all opinion, of course, and from a relative newbie to boot. So the usual grain of salt is indicated...
--
A couple holes in that. First, in a PP circuit, the even order products are strongly suppressed. When you actually look at the distortion (over frequency) of a triode stage, it's not particularly better than a pentode or UL stage, although it will have a lower source impedance and better load tolerance (a not inconsiderable adavantage). If there's an intrinsic superiority of so-called true triodes to triode-strapped pentodes, whether in distortion or source impedance, I'm unaware of it. For example, look at the distortion of PP 300B at 10 watts compared to triode-strapped 6L6GC at the same power.
And second, the gains are different: the pentode will have a higher open-loop gain, so if you compare distortion and source impedance at the same gain, the triode advantage doesn't look so good- and is sometimes worse.
And second, the gains are different: the pentode will have a higher open-loop gain, so if you compare distortion and source impedance at the same gain, the triode advantage doesn't look so good- and is sometimes worse.
Hi SY, thanks for correcting me -- again . Helps me to get a better understanding.
In a power amp, you're going to get all sorts of changing levels and loads, right? Don't you have to take a group of different relative level points to say whether this or that device has lower or higher distortion in this or that application? So talking about push-pull tube power amps...
At 3 watts output, assuming a properly designed push-pull amp with comparable OPT's, will a pair of 2A3's have higher distortion open loop than a pair of KT88 beam tubes? (at 3 watts output)
If yes, then wouldn't that imply that a PP pentode amp with KT88 outputs could be operated with no NFB if you kept the volume down so that it didn't put out more than let's say 5 watts?
And if yes to that, wouldn't that imply that a properly designed PP KT88 amp with no NFB would sound better than a properly designed PP 2A3 amp (again with no NFB), as long as the levels were kept low enough that only 3 watts output was required?
But why does the DHT amp sound better open loop than the comparable beam tetrode KT88 amp? Purely a subjective difference? Or a perceived difference that doesn't really exist in the real world? I thought the open loop distortion of a pair of KT88's was bound to be higher right from the gate than a pair of 2A3's. Wrong?
In a push-pull circuit, the even order distortion harmonics cancel, but don't the odd-order harmonic distortions add up? I thought pentodes (all else being equal) yield a higher proportion of odd-order harmonic distortions than triodes... So in the PP circuit, pentodes will leave behind more odd order harmonics. No?
--
Also... I thought that triode-strapped beam tubes have a different distortion spectra than DHT's. For instance, triode-strapped 6L6GC will have higher 3rd order harmonic distortion than 300B at low output powers. That's supposed to account for the subjective difference between the sound of the two... Total hooey?
. Helps me to get a better understanding. In a power amp, you're going to get all sorts of changing levels and loads, right? Don't you have to take a group of different relative level points to say whether this or that device has lower or higher distortion in this or that application? So talking about push-pull tube power amps...
At 3 watts output, assuming a properly designed push-pull amp with comparable OPT's, will a pair of 2A3's have higher distortion open loop than a pair of KT88 beam tubes? (at 3 watts output)
If yes, then wouldn't that imply that a PP pentode amp with KT88 outputs could be operated with no NFB if you kept the volume down so that it didn't put out more than let's say 5 watts?
And if yes to that, wouldn't that imply that a properly designed PP KT88 amp with no NFB would sound better than a properly designed PP 2A3 amp (again with no NFB), as long as the levels were kept low enough that only 3 watts output was required?
But why does the DHT amp sound better open loop than the comparable beam tetrode KT88 amp? Purely a subjective difference? Or a perceived difference that doesn't really exist in the real world? I thought the open loop distortion of a pair of KT88's was bound to be higher right from the gate than a pair of 2A3's. Wrong?
In a push-pull circuit, the even order distortion harmonics cancel, but don't the odd-order harmonic distortions add up? I thought pentodes (all else being equal) yield a higher proportion of odd-order harmonic distortions than triodes... So in the PP circuit, pentodes will leave behind more odd order harmonics. No?
--
Also... I thought that triode-strapped beam tubes have a different distortion spectra than DHT's. For instance, triode-strapped 6L6GC will have higher 3rd order harmonic distortion than 300B at low output powers. That's supposed to account for the subjective difference between the sound of the two... Total hooey?
Last edited:
Right, the load is not constant with frequency. It's pretty constant with time, though, unless you're dealing with PA systems and instrument amps. That's why you wouldn't use pentodes without feedback. But... as a reference point, you can compare distortion of a triode with a pentode, both with a fixed resistance load. That's how the tube manufacturer datasheets do it. Now, once you knock down the pentode gain to that of a triode, the load dependence similarly decreases. But the distortion does as well, so if you start off with similar distortion, the pentode will end up with MUCH lower distortion at the nominal impedance (by the feedback factor), and the distortion will not rise as quickly at frequencies where the load impedance is different.
The third harmonic stuff is hooey. Sure, you could find some triodes that are lower in third than some triode-strapped pentodes. But let's take a good pentode and a good triode, run them in class A1, and compare them. First, assume for both cases that we match the tubes well to eliminate second- third will then be the predominant harmonic. Push-pull KT88, triode strapped, will give roughly 1% distortion at 15W, dominated by third. It will take about 80V peak to peak to drive them. Push pull 300B will give about 2% at the same power, also dominated by third. And it will take about 130V peak to peak to drive. That will mean higher distortion from the driver.
thanks, that's food for thought. Of course, the obvious question is then, why does anyone bother with DHT's, or vacuum tubes for that matter? Wouldn't transistors be better yet? Same old question, I know. Off-topic...
--
Is it the lack of global negative feedback that tube people like?
--
--
Is it the lack of global negative feedback that tube people like?
--
Last edited:
OK, so take negative feedback out of the discussion. Let's say we like the sound of a circuit with no global negative feedback, at least. Now the challenge is to beat down intrinsic distortion. We like active loads, current sources and sinks, all that.
I have a fixed bias PP EL34 output stage, in Blumlein/ultralinear. Let's say we do the best we can to reduce any open loop distortion. We have a dual-differential driver like the one under discussion here, which self-balances well, etc. We choose the OPT primary impedance to be at the sweet spot for lowest distortion, rather than highest output power delivery to a load, in line with the voltage rail we've chosen. We only expect it to run in class A.
We pick a musical selection that can be set to only demand a maximum 5 watts output per channel (peak), even though the amp is capable of a lot more than that.
Now we take a comparable fixed bias PP 2A3 amplifier, capable of only 7 watts output per channel. Say it's using the exact same OPT's as above, same driver circuitry, and that circuit has the same voltages available, same voltage swing capability, etc. But again, we limit our playback to 5 watt peaks.
We assume a really easy to drive loudspeaker for a load. Maybe a single driver with a smooth impedance curve. Again, we're not talking about playing loud here, or even demanding full range and high fidelity from the loudspeaker. We're just talking about the amps here.
They both have no global negative feedback, no Schade local feedback loops, none of that. Just open loop. (Let's leave out the UL OPT's feedback btwn screens and plates and consider that to be "intrinsic" to the design.)
Which one of the above amplifiers will be "cleaner sounding"?
Which one will have lower distortion within our 5 watt window?
Does any of this matter?
--
Should this be a different thread? Or is this an ages-old subject that I'm just dredging up again for the umpteenth time?
--
I have a fixed bias PP EL34 output stage, in Blumlein/ultralinear. Let's say we do the best we can to reduce any open loop distortion. We have a dual-differential driver like the one under discussion here, which self-balances well, etc. We choose the OPT primary impedance to be at the sweet spot for lowest distortion, rather than highest output power delivery to a load, in line with the voltage rail we've chosen. We only expect it to run in class A.
We pick a musical selection that can be set to only demand a maximum 5 watts output per channel (peak), even though the amp is capable of a lot more than that.
Now we take a comparable fixed bias PP 2A3 amplifier, capable of only 7 watts output per channel. Say it's using the exact same OPT's as above, same driver circuitry, and that circuit has the same voltages available, same voltage swing capability, etc. But again, we limit our playback to 5 watt peaks.
We assume a really easy to drive loudspeaker for a load. Maybe a single driver with a smooth impedance curve. Again, we're not talking about playing loud here, or even demanding full range and high fidelity from the loudspeaker. We're just talking about the amps here.
They both have no global negative feedback, no Schade local feedback loops, none of that. Just open loop. (Let's leave out the UL OPT's feedback btwn screens and plates and consider that to be "intrinsic" to the design.)
Which one of the above amplifiers will be "cleaner sounding"?
Which one will have lower distortion within our 5 watt window?
Does any of this matter?
--
Should this be a different thread? Or is this an ages-old subject that I'm just dredging up again for the umpteenth time?
--
No. As soon as you say, "Use a topology suitable for A but exclude topologies suitable for B, then compare them to decide which is better," you've lost the relevance of the comparison. That you'd include plate-to-grid feedback in one topology but not allow it in another is particularly odd.
SY,
What you say is correct, but I read that R. is postulating class-A here, not in general.
Rongon,
Your question is difficult because it will partly depend on the characteristics of the driver circuitry as well. Let me start by giving that for 2A3: µ=4,2, rp=800 ohm and gm=5.3 mS. For EL34 under UL conditions, exact data is not given, but with 43% screen taps the amplification will be close to 12 and rp close to 2K.
EL34 would require significantly lower signal drive, thus bringing in the pre-ciruitry. But at the low level you propose that need not make a significant contribution. Depending on the load (though you stipulate a benign one), the 2A3 might 'sound' better because of the lower rp, thus damping effect. The distortion of the EL34 topology is however going to be far lower, round 0,3%, compared to a 2A3 level of around 1,5 - 2% (?). (And not to open another can of worms; but the sound of an amplifier is very subjective. Quite a degree of 2nd harmonics can be tolerated; might even make the sound nicer - add to music everything an octave higher, and the result is probably the well-known 'richer', 'warmer' sound. But as said a different topic.)
But the playing field is not exactly level in that EL34s under these conditions are operating in a very low portion of their normal region, while the 2A3s are approaching maximum. To stop there; a far longer reply than you expected. Bottom line: I do not expect an audible difference in sound.
(Perhaps out of turn, I would also like to comment here on another matter on which you have previously touched. The 'negative feedback is bad' (per se) thing is an uninformed urban legend too often perpetuated. It is only bad in poorly designed circuits or where improperly applied or both. It is one of the the most powerful tools with which to reduce amplifier distortion, and consequently also one of the most abused. Again a different subject.)
What you say is correct, but I read that R. is postulating class-A here, not in general.
Rongon,
Your question is difficult because it will partly depend on the characteristics of the driver circuitry as well. Let me start by giving that for 2A3: µ=4,2, rp=800 ohm and gm=5.3 mS. For EL34 under UL conditions, exact data is not given, but with 43% screen taps the amplification will be close to 12 and rp close to 2K.
EL34 would require significantly lower signal drive, thus bringing in the pre-ciruitry. But at the low level you propose that need not make a significant contribution. Depending on the load (though you stipulate a benign one), the 2A3 might 'sound' better because of the lower rp, thus damping effect. The distortion of the EL34 topology is however going to be far lower, round 0,3%, compared to a 2A3 level of around 1,5 - 2% (?). (And not to open another can of worms; but the sound of an amplifier is very subjective. Quite a degree of 2nd harmonics can be tolerated; might even make the sound nicer - add to music everything an octave higher, and the result is probably the well-known 'richer', 'warmer' sound. But as said a different topic.)
But the playing field is not exactly level in that EL34s under these conditions are operating in a very low portion of their normal region, while the 2A3s are approaching maximum. To stop there; a far longer reply than you expected. Bottom line: I do not expect an audible difference in sound.
(Perhaps out of turn, I would also like to comment here on another matter on which you have previously touched. The 'negative feedback is bad' (per se) thing is an uninformed urban legend too often perpetuated. It is only bad in poorly designed circuits or where improperly applied or both. It is one of the the most powerful tools with which to reduce amplifier distortion, and consequently also one of the most abused. Again a different subject.)
Yes, good point. Well, the answer is that the UL configuration is part of the transformer, and so in a way defines the circuit beyond adding "external compensations."
But we could make it pure pentode operation vs. DHT. Sure. Let's make it so that the pentode's screen voltage equals its plate voltage (both rather low, in other words). We still exclude plate-to-grid feedback in both types. So the same OPT for both cases.
If I read him right, Johan feels that at lower levels no audible difference would result at low power between the two examples (UL vs. triode).
I suppose that could mean that if you could optimize a circuit for PP EL34's in UL, then take that same circuit and optimize it for triode, then at low-ish levels, you would not be able to hear the difference between the two.
My question is really this --
I want to believe that a PP UL or triode-wired pentode amp can be made to sound every bit as good as a PP DHT amp. I really do. It would save so much money. I've built or helped build a few amps with a driver circuit very similar to the one in this thread (it's a keeper!). I want to build one for myself, with the pair of Tango XE-45-5 (5k p-p w/ UL taps) I'm fortunate to own.
From personal experience, and in PP, the DHT's seem to yield something magical, especially 2A3's and 45's (I don't like 300B's as much). I don't know what it is. I thought it was that these tubes are capable of very low distortion output when in a reasonably well designed circuit, lower distortion than triode-wired EL34 or 6L6GC, or even triode-wired 6V6GT or EL84. Am I just plain wrong about that?
--
But we could make it pure pentode operation vs. DHT. Sure. Let's make it so that the pentode's screen voltage equals its plate voltage (both rather low, in other words). We still exclude plate-to-grid feedback in both types. So the same OPT for both cases.
If I read him right, Johan feels that at lower levels no audible difference would result at low power between the two examples (UL vs. triode).
I suppose that could mean that if you could optimize a circuit for PP EL34's in UL, then take that same circuit and optimize it for triode, then at low-ish levels, you would not be able to hear the difference between the two.
My question is really this --
I want to believe that a PP UL or triode-wired pentode amp can be made to sound every bit as good as a PP DHT amp. I really do. It would save so much money. I've built or helped build a few amps with a driver circuit very similar to the one in this thread (it's a keeper!). I want to build one for myself, with the pair of Tango XE-45-5 (5k p-p w/ UL taps) I'm fortunate to own.
From personal experience, and in PP, the DHT's seem to yield something magical, especially 2A3's and 45's (I don't like 300B's as much). I don't know what it is. I thought it was that these tubes are capable of very low distortion output when in a reasonably well designed circuit, lower distortion than triode-wired EL34 or 6L6GC, or even triode-wired 6V6GT or EL84. Am I just plain wrong about that?
--
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.