[JohnR]

FWIW, the manual says "my newly patented current-multiplying DC coupling circuitry allows the 300B output driver tube to directly operate into a fixed ratio-reduced speaker output load." So yes, it look like the 300B is a driver for the output stage. Presumably more information could be gleaned from the patent, presumably (if anyone cared to search for it)?

[Steve Eddy]

FWIW, the manual says "my newly patented current-multiplying DC coupling circuitry allows the 300B output driver tube to directly operate into a fixed ratio-reduced speaker output load." So yes, it look like the 300B is a driver for the output stage. Presumably more information could be gleaned from the patent, presumably (if anyone cared to search for it)?

6,507,240.

se

[jon_010101]

1) Feedback sucks.  One thing that often makes tube amps sound better is greatly reduced or no negative feedback.  All transistor amps need moderate to high levels of feedback.  Transistors are inherently less linear amplifying devices and need artificial linearization to perform acceptably.

2) That said, a typical moderate-feedback PP tube amp sounds *a lot* like a typical moderate-feedback SS amp *if the sand amp has much greater power reserves and is not forced to clip*.  One thing I do *not* understand completely is how or why tubes clip "differently" or "more gracefully" but this is *true* in the real world.

Just wanted to stir this pot a bit -- Ever hear a Citation II with modern parts, or a perfectly-tweaked Williamson amp?  20-30dB feedback can sound very nice, if the hardware is up to task.  The modern trend towards reduced feedback in tube amps stems partly from the use of lower-grade output transformers.   

IMHO, the problem with a lot of transistor amps isn't the feedback, it's the class AB circuit with excessive crossover distortion.  Bias transistors (and push-pull tubes) in class A and things get a lot better.  A tell-tale sign of an annoying power amp is a distortion figure that is minimized at near-max power, but is an order of magnitude or so worse where it counts -- i.e., under 1W.  Feedback has a lot less work to do when the circuit's open loop distortion is reduced.

That said, I think single ended can sound great when given a friendly load.  I once had the pleasure of hearing a pair of AN Japan Gaku-On monoblocks, and it was a revelatory experience.  They were the first SE amps I'd heard that I'd prefer over a pair of good push-pull feedback amps  .  For the sake of experiment, I will soon be building a single-ended 813 ultra-linear amp -- but I do plan to use feedback, since an UL-SE output stage pretty much requires it.

[PaulFolbrecht]

Just wanted to stir this pot a bit -- Ever hear a Citation II with modern parts, or a perfectly-tweaked Williamson amp?  20-30dB feedback can sound very nice, if the hardware is up to task.  The modern trend towards reduced feedback in tube amps stems partly from the use of lower-grade output transformers.   

Hi Jon.  I also think feedback can sound great.  I haven't heard a Citation but I've heard the very good Chinese clones and similar amps - they can sound great!  But take away the feedback and it's even better.  Amps with settable feedback can be a real eye-opener.

IMHO, the problem with a lot of transistor amps isn't the feedback, it's the class AB circuit with excessive crossover distortion.  Bias transistors (and push-pull tubes) in class A and things get a lot better.  A tell-tale sign of an annoying power amp is a distortion figure that is minimized at near-max power, but is an order of magnitude or so worse where it counts -- i.e., under 1W.  Feedback has a lot less work to do when the circuit's open loop distortion is reduced.

Yes; excellent point.  I'd forgotten about that since I stopped seriously listening to AB SS amps long ago.  I believe there are, in fact, high odd-order products created in this situation too.  And, of course, distortion being inversely proportional to power is what robs detail, nauance, truncates decays (along with feedback), etc.

That said, I think single ended can sound great when given a friendly load.  I once had the pleasure of hearing a pair of AN Japan Gaku-On monoblocks, and it was a revelatory experience.  They were the first SE amps I'd heard that I'd prefer over a pair of good push-pull feedback amps  .  For the sake of experiment, I will soon be building a single-ended 813 ultra-linear amp -- but I do plan to use feedback, since an UL-SE output stage pretty much requires it.

Well, the speaker is always a critical part of the equation with SE tubes.

How much feedback?  6dB, maybe 10, seems to be very tolerable under most circumstances.

[aerius]

Let's think about this for a moment.

If you take a triode, and add a screen grid and a suppressor grid, you end up with a tube that has higher gain but is somewhat less linear.

If we work backwards from this, it would seem that a triode has some sort of internal feedback mechanism which linearizes the tube at the expense of reducing its gain.

Now, a transistor also generally has much more gain than a triode tube and is also generally less linear.

So what if we add a bit of emitter degeneration to the transistor, bringing its gain down closer to that of the triode which would also linearize the transistor?

What would the fundamental difference here be?

Distribution of the various distortion harmonics.  A good triode will have mostly 2nd, a bit of 3rd, and practically nothing above that.  A pentode will have both more total distortion as well having more of the higher order harmonics present, it'll likely go up to the 5th or 6th and maybe even beyond.  Putting NFB on a pentode will indeed reduce the THD by cutting down the 2nd, 3rd, and possibly 4th harmonics, but the higher ones remain unaffected and may actually increase.  The distortion spectra of a pentode with NFB doesn't look anything like that of a triode.  Same thing with transistors, NFB will linearize a transistor and bring down its gain & distortion, but if you hooked up a spectrum analyzer to it the distribution of the various harmonics isn't going to look like what you'd get from a triode. 

Funnily enough, putting NFB on a triode will actually make it look more like a pentode, the 2nd & 3rd harmonics go down, but you actually create higher order harmonics which weren't there before.

[PaulFolbrecht]

1) Feedback sucks.  One thing that often makes tube amps sound better is greatly reduced or no negative feedback.  All transistor amps need moderate to high levels of feedback.  Transistors are inherently less linear amplifying devices and need artificial linearization to perform acceptably.

Let's think about this for a moment.

If you take a triode, and add a screen grid and a suppressor grid, you end up with a tube that has higher gain but is somewhat less linear.

If we work backwards from this, it would seem that a triode has some sort of internal feedback mechanism which linearizes the tube at the expense of reducing its gain.

Now, a transistor also generally has much more gain than a triode tube and is also generally less linear.

So what if we add a bit of emitter degeneration to the transistor, bringing its gain down closer to that of the triode which would also linearize the transistor?

What would the fundamental difference here be?

The difference, Steve, would be the loop length - if I take your word at its value and assume there *is* a feedback loop - I've definitely *not* heard things described like that before but allow that you may be right.  I'm honestly not looking for a peeing contest; you may have something.  But, the loop length is critical - as I'm sure you know - because a long loops means longer feedback times and the longer the feedback cycle time the more and *worse* distortion products created and the worse the sound.

One thing I do *not* understand completely is how or why tubes clip "differently" or "more gracefully" but this is *true* in the real world.

Care to show us? The Hamm paper seems to demonstrate in the real world that this is a myth.

I haven't read the paper yet but so far don't have too high a regard for it. 

Tell me this: does the paper go into detail on the effect of FEEDBACK on distortion products??  It is, in fact, the high NFB in the typical SS amp that creates the higher-order, most objectionable, nasty-sounding distortion.  And as was pointed out by Jon, in some ways at least this gets *worse* at lower signal levels due to crossover distortion which is a real recipe for disaster.

So, the very short answer to your question would be, and this is partly conjecture at this point, that the feedback-induced high-order products that SS amps produce when they clip is what causes this subject tube vs. sand "power" difference.  Typical PP tube amps with low feedback (<= 10dB) don't have this problem.

I've also seen at least one paper demonstrating - with scope results - that vacuum tubes do have the ability to swing *much* larger voltages for very short intervals than transistors - this is a real, fundamental difference between them as output devices, and this would certainly explain a very real difference in "clipping" behavior.

So, I'll read the paper and it might have been smarter to do so before replying.  But, if I were to read this morning a scientific paper in a peer-reviewed journal that explained to me in detailed, mathematical terms why the sky is red, I'd be inclined to keep believing it is blue.

4) Single-ended tube amps generally sound best of all because, in addition to running in class A (as do some PP amps as well) and having a natural distortion specra...

What exactly is a "natural" distortion spectra? Do you mean simply even-order dominant distortion? If so, what makes that somehow more "natural" than odd-order dominant distortion? Have you looked at the spectra of various acoustic instruments?

I will refer you to an excellent article by Nelson pass on this very subject.  It amazes me that he does make PP amps at all anymore (market demands, I guess), as he so obviously prefers SE circuits for these reasons. 

It's on the web - if you can't find it let me know.  Think I have it saved somewhere.

...we are using very linear output devices that allow zero feedback and also great simplicity in the circuit.

Is there really zero feedback? Or is it really just an illusion because any internal feedback mechanism in the tube isn't drawn out in the schematics?

Much more the former than the latter.

I love to flip over my Fi 45 monos and look at the circuit - it's the input tube, output tube, and output tranny!

Have you seen Nelson Pass' Zen amp? Doesn't even have an input transistor. Just an output transistor.

No voltage gain?  Like the F4?  Yeah, if you want a zero-gain amp you can make it even simpler.

I bet a zero-gain SET could be killer.

[Bill Baker]

However, you still didn't answer my question.

Does the 300B act as a driver tube for a solid state output stage?
A 300B tube is only capable of putting out about 9 watts...

His website does nothing to explain how it works.

Still interested though

 HI Jeff,

 I biggest thing with the Monads is not to make assumptions as the circuit is very unique. My understanding from my short time dealings with Butler is that the 300B tube is in the output stage but not in any way that can be considered conventional. I don't know if I would call it a driver as I believe it is on the other end of the circuit or somewhere within. I also believe it does not use B+ power as we would normally associate with it which may be the reason it would last 'forever'?

 I have never been inside nor heard a Monad (in fact they weren't even called Monads back then) so please do not quote my words as gospel. I could be way off base here. What I do know is that the Monads employ a very unique circuit and BK is one smart dude. We may never know exactly how the Monad's 300B operates but then again, for those who own these magnificent amps, they may not even care.
 Right DU

[PaulFolbrecht]

Have you seen Nelson Pass' Zen amp? Doesn't even have an input transistor. Just an output transistor.

Oh, by the way - there's no way it has *nothing* but an output transistor.  Meaning no other parts (except power supply).  I believe you'd be seeing a THD figure of around 70% if that were the case. 

When I said the Fi amps consist of input tube, output tube, and transformer, three parts (not counting the power supply, again), I was being quite literal.  There is nothing else.

[Steve Eddy]

Distribution of the various distortion harmonics.  A good triode will have mostly 2nd, a bit of 3rd, and practically nothing above that.

But as I said in a previous post referencing a piece that I'm still trying to find, when the transistor was run at the same 180 volts as the triode it was being compared to, the distribution of the distortion products were the same as that of the tube. The only difference was that the overall distortion was a bit higher with the transistor.

Putting NFB on a pentode will indeed reduce the THD by cutting down the 2nd, 3rd, and possibly 4th harmonics, but the higher ones remain unaffected and may actually increase.

Can you point to some evidence of that? I've never seen this phenomenon occur.

Same thing with transistors, NFB will linearize a transistor and bring down its gain & distortion, but if you hooked up a spectrum analyzer to it the distribution of the various harmonics isn't going to look like what you'd get from a triode.

I don't know that that's the case. As I've said, when run at the same voltage as the tube, the transistor has the same distribution as the tube, albeit higher overall. I don't see why adding a bit of emitter degeneration wouldn't bring the levels down to that of the tube.

Funnily enough, putting NFB on a triode will actually make it look more like a pentode, the 2nd & 3rd harmonics go down, but you actually create higher order harmonics which weren't there before.

Technically, yes. But they would be at exceedingly low levels. I mean, if they're down below -120dB, does it really matter?

se

[aerius]

But as I said in a previous post referencing a piece that I'm still trying to find, when the transistor was run at the same 180 volts as the triode it was being compared to, the distribution of the distortion products were the same as that of the tube. The only difference was that the overall distortion was a bit higher with the transistor.

No need to tell me, I actually saw it on a spectrum analyzer earlier this week and heard the results on a set of speakers.  Transistors at 160V sound about as good as Lynn Olson's Karna amps, which in turn are better than any other amp I've heard.

Putting NFB on a pentode will indeed reduce the THD by cutting down the 2nd, 3rd, and possibly 4th harmonics, but the higher ones remain unaffected and may actually increase

.Can you point to some evidence of that? I've never seen this phenomenon occur.

I saw it on an HP spectrum analyzer, pentode tubes were the 12W6, 6L6 and 47.  If I get to drop by the guy's place again, I'll bring my digital camera and take pictures of the instrumentation screens.

Same thing with transistors, NFB will linearize a transistor and bring down its gain & distortion, but if you hooked up a spectrum analyzer to it the distribution of the various harmonics isn't going to look like what you'd get from a triode.

I don't know that that's the case. As I've said, when run at the same voltage as the tube, the transistor has the same distribution as the tube, albeit higher overall. I don't see why adding a bit of emitter degeneration wouldn't bring the levels down to that of the tube.

To be honest, the maths are beyond me, but as I understand it and actually saw, degeneration works primarily on the lower order harmonics and it always creates higher order harmonics.

Funnily enough, putting NFB on a triode will actually make it look more like a pentode, the 2nd & 3rd harmonics go down, but you actually create higher order harmonics which weren't there before.

Technically, yes. But they would be at exceedingly low levels. I mean, if they're down below -120dB, does it really matter?

I don't know.  Personally I don't think it really matters if it's that far down, but at the same time I admit I'll probably feel better knowing that it's not there at all. 

[Steve Eddy]

The difference, Steve, would be the loop length - if I take your word at its value and assume there *is* a feedback loop - I've definitely *not* heard things described like that before but allow that you may be right.

This might be worth a read:

Inherent Feedback in Triodes

As for loop length, how short can you make a resistor?

I haven't read the paper yet but so far don't have too high a regard for it.

Er, you haven't read the paper yet but so far don't have too high a regard for it?

How does that happen?

Tell me this: does the paper go into detail on the effect of FEEDBACK on distortion products??  It is, in fact, the high NFB in the typical SS amp that creates the higher-order, most objectionable, nasty-sounding distortion.  And as was pointed out by Jon, in some ways at least this gets *worse* at lower signal levels due to crossover distortion which is a real recipe for disaster.

Crossover distortion is a whole different issue from negative feedback.

So, the very short answer to your question would be, and this is partly conjecture at this point, that the feedback-induced high-order products that SS amps produce when they clip is what causes this subject tube vs. sand "power" difference.  Typical PP tube amps with low feedback (<= 10dB) don't have this problem.

But that's a matter of how much feedback is used, not the devices themselves. Whether tube or solid state, the more feedback you have the more harshly it will clip.

I've also seen at least one paper demonstrating - with scope results - that vacuum tubes do have the ability to swing *much* larger voltages for very short intervals than transistors - this is a real, fundamental difference between them as output devices, and this would certainly explain a very real difference in "clipping" behavior.

You wouldn't be referring to that piece that was published in Stereophile a while back would you? Where they showed that while the tube amp bottomed out on the bottom half of the waveform, it was able to swing much higher voltages on the positive half of the waveform compared to the solid state amp they were comparing?

If so, it's another of those dopey apples to oranges comparisons which don't impart any truly useful information.

I will refer you to an excellent article by Nelson pass on this very subject.  It amazes me that he does make PP amps at all anymore (market demands, I guess), as he so obviously prefers SE circuits for these reasons.

You talking about the one where he refers to "single-ended nature of air"?

No voltage gain?  Like the F4?

No, it has voltage gain.

se

[PaulFolbrecht]

This might be worth a read:

Inherent Feedback in Triodes

As for loop length, how short can you make a resistor?

I'll check out the article, and I am missing your point.  Certainly any global feedback loop in any SS amp is much longer than any such triode-local loop?

Er, you haven't read the paper yet but so far don't have too high a regard for it?

How does that happen?

By reading the conclusions being drawn from it in this thread.

Crossover distortion is a whole different issue from negative feedback.

Of course.  The point is that, in the real world, in real amplifiers, tubes and transistors have all kinds of important differences - many of them related to the fact that transistors *need* NFB in order to operate anything close to linearly.  I don't know if this article isn't point this out or people in this thread are ignoring it, but without considering things like this any point anybody's trying to make is rather academic.

And, yes, it is high time I read the article.    But is is 80d here in October and it's time to go outside and play now.

But that's a matter of how much feedback is used, not the devices themselves. Whether tube or solid state, the more feedback you have the more harshly it will clip.

Of course.  And in the real world, transistors need lots of feedback and tubes don't.  Do you agree with that?

You wouldn't be referring to that piece that was published in Stereophile a while back would you? Where they showed that while the tube amp bottomed out on the bottom half of the waveform, it was able to swing much higher voltages on the positive half of the waveform compared to the solid state amp they were comparing?

Was not in S-phile; I read it online and it was not s-phile's online format. 

If the phenomena is real, I'd expect it's been written about considerably.

You talking about the one where he refers to "single-ended nature of air"?

Yes.

[Double Ugly]

We may never know exactly how the Monad's 300B operates but then again, for those who own these magnificent amps, they may not even care.
 Right DU

That would be correct, sir! 

And FWIW, what you've said about the 300B being used unconventionally is (I think) on the money.  However, BK isn't forthcoming - at *ALL* - thus my hesitancy to try to explain something I know nothing about.

The Monad's design topology is but one of many, many things about which I'm ignorant, so as long as they make beautiful music, it's all good. 

[Bill Baker]

And FWIW, what you've said about the 300B being used unconventionally is (I think) on the money.  However, BK isn't forthcoming - at *ALL* - thus my hesitancy to try to explain something I know nothing about.

 Me too. That's why I didn't go into detail. I use to know a bit more about the design details but that was a few years back when they were a little bit more talkative about the design. Now my memory can not retrieve this information as it has moved on to other things

[Steve Eddy]

When I said the Fi amps consist of input tube, output tube, and transformer, three parts (not counting the power supply, again), I was being quite literal.  There is nothing else.

You're sure about that are you?

What's loading the plate of that input tube?

se

[Steve Eddy]

No need to tell me, I actually saw it on a spectrum analyzer earlier this week and heard the results on a set of speakers.  Transistors at 160V sound about as good as Lynn Olson's Karna amps, which in turn are better than any other amp I've heard.

Ok.

Now I'm wondering what we'd see if instead of a 6SN7 at 180 volts, we looked at say a 6GM8 at the same 15 volts used for the transistor in low voltage mode.

Hmmm...

I saw it on an HP spectrum analyzer, pentode tubes were the 12W6, 6L6 and 47.  If I get to drop by the guy's place again, I'll bring my digital camera and take pictures of the instrumentation screens.

Interesting. Can you give me a brief description of the amp topology and how the feedback was implemented?

To be honest, the maths are beyond me, but as I understand it and actually saw, degeneration works primarily on the lower order harmonics and it always creates higher order harmonics.

Well, mathematically speaking, feedback lowers distortion across the board, but does extend the harmonics to higher orders. But those higher orders will still be far below the lower orders.

I don't know.  Personally I don't think it really matters if it's that far down, but at the same time I admit I'll probably feel better knowing that it's not there at all.

I think we call that... "neurotic."

se

[Steve Eddy]

I'll check out the article, and I am missing your point.  Certainly any global feedback loop in any SS amp is much longer than any such triode-local loop?

I didn't know we were limiting the discussion of feedback to global feedback. I was referring to simple emitter degeneration, which is just a resistor at the transistor's emitter. In other words, it's local.

By reading the conclusions being drawn from it in this thread.

Not sure I follow. The only thing I cited from the paper was their saying that the notion of tubes clipping more softly than transistors was a myth.

Of course.  The point is that, in the real world, in real amplifiers, tubes and transistors have all kinds of important differences - many of them related to the fact that transistors *need* NFB in order to operate anything close to linearly.

Well there's evidence that this is not necessarily the case when the comparisons are more apples to apples.

But I still don't know what this has to do with crossover distortion.

You'd said "And as was pointed out by Jon, in some ways at least this gets *worse* at lower signal levels due to crossover distortion which is a real recipe for disaster."

The reason distortion gets worse as signal levels are reduced is because of the crossover distortion itself, not because of negative feedback. The discontinuity at the crossover region remains essentially the same regardless of signal level. So as the signal level increases, the crossover distortion becomes a smaller and smaller percentage of the signal and hence distortion goes down, and vice versa.

As I said, this is a whole other matter from negative feedback.

Of course.  And in the real world, transistors need lots of feedback and tubes don't.  Do you agree with that?

Not necessarily, no. Again, there is evidence that when you run transistors at similar voltages to tubes, they virtually identical distortion characteristics as tubes. And in this comparison, while the overall distortion was a bit higher, it wasn't dramatically so.

Also as I pointed out in a previous post, cascoding can dramatically linearize a transistor (or a tube for that matter) without using any negative feedback.

Was not in S-phile; I read it online and it was not s-phile's online format. 

If the phenomena is real, I'd expect it's been written about considerably.

Well without benefit of reading what has been written about it, I can't really comment on it.

Yes.

Ah, ok. Well suffice to say that I think the whole "single-ended characteristic of air" thing isn't really an argument. By the time we would start experiencing the single-ended characteristic of air, our eardrums would be slammed six inches into our skulls.

se

[Bob Reynolds]

And they are all dwarfed by the distortion products in the loudspeaker.   

Thanks for pointing that out Kevin. 

[aerius]

What do you mean by "odd-order *only* HD composition of PP amps"? Most all push-pull amps I've seen have also had significant amounts of even-order distortion products in addition to the odd-order products.

se

And they are all dwarfed by the distortion products in the loudspeaker.

Actually I'd say it's arguable in some cases, most of the SE tube amps will have a significant amount of distortion, in many cases the 2nd & 3rd harmonics are only 30-40dB down from the fundamental.  A good speaker on an open baffle can put those same harmonics about 40-50dB down.

[Freo-1]

Thanks, Freo! 


I've saved the .pdf, but I'm too wiped to digest anything tonight (spent half the day and night driving to my parents' home in TN).  I may not get to it for a few days, but I'll definitely spend some quality time with it when life allows.  My interest is definitely piqued! 

Thanks again! 

You are most welcome!
One of the more interesting items there is the fact the output stage uses no negative feedback, and exhibits  zero crossover distortion. Pretty cool for a MOS-FET output!