he´ll need 4x JBL2220B as I have and he knows that, so experimenting in the waiting is the recipi
SET amp owners thread
- Thread starter bonzo75
- Start date
Are you trying to say that your amp makes no 2nd harmonic but does 3rd, 4th, 5th, 6th etc? How does that work when a push pull circuit cancels even orders. Of course if that is imperfect you will get some 2nd, 4th, 6th etc but at greatly reduced levels. You don't seem to get it. The proximity to the fundamental tone is critical for masking.
You are simply incorrect. We've been using a direct-coupled output called the Circlotron for the last 45 years. No feedback and no output coupling capacitor.
Here is a schematic diagram- no output coupling cap.
I have- we have also engineered a class D of our own design, which has the same distortion signature as an excellent tube amp, and so sounds like an excellent tube amp, but more transparent and without the menace of brightness and harshness.
More to the point, distortion obscures detail- things like what's going on in the rear of the sound stage, the palpability of the instruments themselves. This is easy to hear.
Apparently you did not understand what I said earlier. SETs have a high output impedance as do our OTLs, both on account of zero feedback. When you have any tube amplifier that is zero feedback, it requires a special speaker for everything to pan out. You can read more about how this works in this article about the difference between the Voltage Paradigm and the Power Paradigm. SETs and our OTLs are Power Paradigm devices.
No- not 'trying to say that'. My surmise is you misunderstood something. Yes, even orders are cancelled throughout the circuit of any fully differential amplifier. How this is a benefit is that the harmonics that are generated are at a fundamentally lower level since there is overall less distortion passed from one stage in the amp to the next. So if we were to look at the 5th harmonic (as an example) in an SET at a given power, and then look at it in one of our OTLs at the same percentage of power (for example each 6dB below clipping) what you would see is that the 5th is actually at a lower absolute level in the OTL than it is in the SET despite the enormous power difference (the SET might be at 2 Watts while the OTL might be at 35 Watts). This isn't trivial- this is happening across the entire spectrum. At 1 Watt, the difference in distortion can be as much as 2 orders of magnitude lower.
IOW, overall, the SET is generating a lot more higher ordered harmonics and they are falling off at a slower rate as the order of the harmonic is increased. That is why the OTL sounds smoother. This is easy to hear: if you've noticed how 'dynamic' SETs are, that is caused by distortion and nothing else. You may think 7 or 15 Watts is enough power for you because when you turn it up it sounds loud- maybe you even get comments from your S.O. to turn it down. But when you clean up the distortion that sense of loudness goes away because its caused by distortion rather than sound pressure, due to how our ears use the higher orders to sense sound pressure. Literally the SET is messing with our ears to create this illusion. A sound pressure level meter reveals this fact easily enough.
So:
Your last statement above (emphasis added) is something we both agree on. This is what Harry Olsen says as well; but he did not delineate between the 2nd or the 3rd, just the proximity to the fundamental tone. As it turns out, the 3rd is the only odd ordered harmonic that sounds musical to the human ear and is thus treated by it the same way as the 2nd. I think I've pointed out to you in the past that a properly functioning tape recorder generates a 3rd harmonic as its dominant distortion product and no-one seems to have a problem with that
SETs and zero feedback OTLs make distortion, both of whose harmonics fall off on an exponential curve. The exponent is the key modifier here- that is why I talk about a quadratic non-linearity (SETs) as opposed to a cubic non-linearity (fully differential circuits like our OTLs). The exponent of the cubic function allows the higher harmonics to fall off at a faster rate (on an exponential curve) than seen with an SET (where the harmonics fall off according to a quadratic function). So the OTL in this case is far better able to mask the higher orders than seen in an SET. From an engineering perspective (where math shows how this is so) this is not hard to understand.
I hope you get that engineering is fundamental to amplifier design. Its not magic or dogma. I think the problem a lot of the SET community has is they have been comparing apples to oranges in amps for such a long time that its difficult to see that there are a ton of variables they traditionally never take into account. I've mentioned them before: different output tubes, different class of operation, feedback, etc.
Now if you're being astute, you'll notice that bit about different output tubes. What this means is that if you built a fully differential PP amplifier using directly heated triodes and ran it class A as per the SET, you stand a chance to advance the art. I've built such amps (based on the 45, so make about 7 Watts) and the improvement over the 45 SET I used for comparison was pretty obvious. It measured better too: lower distortion, wider bandwidth, lower noise etc.
Last edited:
There are no capacitors, but there are feedbacks. Can you tell me what role the 2M resistors play?
As we say, "there are no friends for taste and color."
You have feedback in the amplifiers. Ask someone to explain your schematic diagram to you.
You show a circuit diagram with feedback and for so long assure that it is not there. What are you counting on?
The M-60 employs 2 dB of feedback which is miniscule. Sometimes people order it without. The amp does not need it, especially if the speaker impedance is higher, above about 10 Ohms. Its helps linearize the frequency response with lower impedance loads; something SETs could do with as well.
The larger amps, the MA-1 and MA-2, do not have any feedback at all. You were concerned that DC in the speaker would be a problem for an OTL without feedback, are you honestly saying 2dB would be enough to correct for that?? Especially in a circuit that has interstage coupling caps- the feedback has no effect on that at all.
Perhaps you could get someone to explain the diagram to you since you seemed to have missed that aspect of the design.
Hi Ralph, M-60 is something i might be interested in at some point.. Anyway, can you show harmonic distortion measurement of M-60?
(with it's 2db feedback model)
-I thought even a little feedback always introduces odd spectrum?
I can see everything well. You wrote about the lack of feedback, but there is. It does not affect DC modes, you are right. But you entrusted the user with the function of constantly monitoring the milliammeter and adjusting its zero. I would never do that. The user only has to turn on the amplifier and listen to music. Automation should do the rest.
Ralph, you still don’t get it. The absence of even order harmonics from the pattern is BAD. Only odd order pattern sounds worse than any other pattern. The lowish levels without feedback helps your OTLs sound transparent…but thin. Application of significant feedback suppresses low orders and creates new higher order components, albeit at lower levels than the benign low oder harmonics.You are simply incorrect. We've been using a direct-coupled output called the Circlotron for the last 45 years. No feedback and no output coupling capacitor.
Here is a schematic diagram- no output coupling cap.
I have- we have also engineered a class D of our own design, which has the same distortion signature as an excellent tube amp, and so sounds like an excellent tube amp, but more transparent and without the menace of brightness and harshness.
More to the point, distortion obscures detail- things like what's going on in the rear of the sound stage, the palpability of the instruments themselves. This is easy to hear.
Apparently you did not understand what I said earlier. SETs have a high output impedance as do our OTLs, both on account of zero feedback. When you have any tube amplifier that is zero feedback, it requires a special speaker for everything to pan out. You can read more about how this works in this article about the difference between the Voltage Paradigm and the Power Paradigm. SETs and our OTLs are Power Paradigm devices.
No- not 'trying to say that'. My surmise is you misunderstood something. Yes, even orders are cancelled throughout the circuit of any fully differential amplifier. How this is a benefit is that the harmonics that are generated are at a fundamentally lower level since there is overall less distortion passed from one stage in the amp to the next. So if we were to look at the 5th harmonic (as an example) in an SET at a given power, and then look at it in one of our OTLs at the same percentage of power (for example each 6dB below clipping) what you would see is that the 5th is actually at a lower absolute level in the OTL than it is in the SET despite the enormous power difference (the SET might be at 2 Watts while the OTL might be at 35 Watts). This isn't trivial- this is happening across the entire spectrum. At 1 Watt, the difference in distortion can be as much as 2 orders of magnitude lower.
IOW, overall, the SET is generating a lot more higher ordered harmonics and they are falling off at a slower rate as the order of the harmonic is increased. That is why the OTL sounds smoother. This is easy to hear: if you've noticed how 'dynamic' SETs are, that is caused by distortion and nothing else. You may think 7 or 15 Watts is enough power for you because when you turn it up it sounds loud- maybe you even get comments from your S.O. to turn it down. But when you clean up the distortion that sense of loudness goes away because its caused by distortion rather than sound pressure, due to how our ears use the higher orders to sense sound pressure. Literally the SET is messing with our ears to create this illusion. A sound pressure level meter reveals this fact easily enough.
So:
Your last statement above (emphasis added) is something we both agree on. This is what Harry Olsen says as well; but he did not delineate between the 2nd or the 3rd, just the proximity to the fundamental tone. As it turns out, the 3rd is the only odd ordered harmonic that sounds musical to the human ear and is thus treated by it the same way as the 2nd. I think I've pointed out to you in the past that a properly functioning tape recorder generates a 3rd harmonic as its dominant distortion product and no-one seems to have a problem with that
SETs and zero feedback OTLs make distortion, both of whose harmonics fall off on an exponential curve. The exponent is the key modifier here- that is why I talk about a quadratic non-linearity (SETs) as opposed to a cubic non-linearity (fully differential circuits like our OTLs). The exponent of the cubic function allows the higher harmonics to fall off at a faster rate (on an exponential curve) than seen with an SET (where the harmonics fall off according to a quadratic function). So the OTL in this case is far better able to mask the higher orders than seen in an SET. From an engineering perspective (where math shows how this is so) this is not hard to understand.
I hope you get that engineering is fundamental to amplifier design. Its not magic or dogma. I think the problem a lot of the SET community has is they have been comparing apples to oranges in amps for such a long time that its difficult to see that there are a ton of variables they traditionally never take into account. I've mentioned them before: different output tubes, different class of operation, feedback, etc.
Now if you're being astute, you'll notice that bit about different output tubes. What this means is that if you built a fully differential PP amplifier using directly heated triodes and ran it class A as per the SET, you stand a chance to advance the art. I've built such amps (based on the 45, so make about 7 Watts) and the improvement over the 45 SET I used for comparison was pretty obvious. It measured better too: lower distortion, wider bandwidth, lower noise etc.
By not having even orders in your distortion spectrum, you eliminate masking and expose the listener to unmasked bad sounding odd orders (only 3rd is somewhat benign). It is actually WORSE if the amp has low 3rd harmonic then because that reduces even further the masking frequency range. 5th is now exposed…as are all higher orders.
Again, look at the experiment done by Keith Howard where he digitally added these distortion patterns to a recording file. He found that the all odd pattern was by far the most unpleasant sounding. The best was no added distortion and 2nd best was full spectrum with exponential decay at a lower than initial level.
It will if improperly applied!
Norman Crowhurst wrote about this problem 60 years ago. The usual problem is the feedback is applied to a non-linear feedback node; in a tube amp, that is usually the cathode of the input tube. We don't do that; instead the feedback is mixed with the incoming signal so its not distorted before its able to do its job. Obviously if you wanted to add feedback to an SET, the way to do it without causing harm would be the same way- rather than apply the feedback to the cathode of the input tube, where it would get distorted, you instead mix it using a resistive divider network and then send the resulting signal to the grid of the input tube. That way the distortion of the tube is compensated correctly by the feedback and so higher ordered harmonics are not generated.
In our later amps (the one of the schematic is from 20 years ago) we have automatic bias (which is a simple analog circuit). To your point about feedback, how do you explain the MA-1 , MA-2 or MA-3 ? They have no feedback at all. My prior explanation was sufficient. I showed you a schematic of the M-60 so you could see it has a direct-coupled output which you maintained it didn't. So you moved the goal posts: you beefed about 2 dB of feedback, but other OTLs we make which use the same circuit don't have it. Now you want to complain about manual bias, which we've not used in 20 years... are you interested in the truth or do you want to continue playing word games trying to make me wrong in spite of it?
And you still don't get it either- (and so we continue to butt heads for no reason). My prior explanation was sufficient. My surmise is you didn't read it.
You are ignoring how much reduced the distortion is (two orders of magnitude), and how as the order of the harmonic is increased, its amplitude falls off faster than in an SET. We don't need the feedback at all (see above links). Its just there in the M-60 because people insist on using lower impedance speakers it can't handle that well (same as if you tried to use a 4 Ohm speaker on an SET with the 16 Ohm tap...).
Obviously I'm getting sufficient masking. A perusal of our customer feedback (and reviews, awards, yada yada) shows that. In some cases, the amps are replacing respected SETs. Occam's Razor suggests your overly complex explanation is false. Keith H didn't do the same thing- he didn't add harmonics decreasing on a cubic exponential curve. He did something else; his findings have nothing to do with a fully differential amp.
You also have to explain why reel to reel tape works so well! Before you can lay any criticism at my door on account of the 3rd harmonic, you have to explain why reel to reel tape works so well despite the 3rd harmonic. Explain also why the 3rd harmonic in a speaker masks higher ordered harmonics. Good Luck with all that...
In short, your explanation above is simply incorrect, based on a false assumption or bad science.
It doesn’t mask simple as that, plus speakers don’t generate high order harmonics significantly like active electronics circuits do. As I sai, 3rd harmonic itself is not the issue…why are you being so dense about this? It is the lack of even order and subsequent lack of masking that is the issue.
Your OTLs do some things very well…natural tone is not one of them…at least not compared to top SET amps and real life. If they were better, trust me , I would own your OTLs.
The best recordings I have ever heard were not from tape but direct cut on lacquer driven by a tube cutter. Tape only sounds better than other formats when playing an original on the machine it was recorded. Otherwise, it sounds different than vinyl or digital but not necessarily better.
I agree with you on the cutter. As you might know, I ran an LP mastering operation (I had the Westerex 3D cutter mounted to a Scully lathe with the 1700 series electronics). Its quieter, lower distortion and wider bandwidth than tape. Ideally. If you have the cutter set up properly, the resulting lacquar cut can rival digital for noise floor. Most of the noise comes in during the plating and pressing process...
The reason tape sounds different on different machines has to do with calibration. Sounds like you didn't attend to that when you did whatever you did that caused your comment. Anecdotally we recorded a big project (3LP set) once using two machines- a refurbished Ampex tube machine and a nice Otari MX5050. So two master tapes- the Ampex tape sounded better on the Otari than the Otari's tape did and the Otari tape sounded better on the Ampex than it did on the Otari. But the best was the Ampex tape on the Ampex machine. So I get there are differences, but the way you describe it, the machines you played were out of calibration.
You're not going to convince me on this masking thing. So far, everything you've turn up has either not supported your claims or were red herrings outright (such as Keith Howard's thingy). Its not density, its education and experience. If you want to convice me you'll have to try a lot harder.
I have the additional benefit of having been at recording sessions and have the tapes, LPs and CDs thereof; having been there I know how they are supposed to sound. This is a rare tool but is one I recommend to anyone! Your statement about the OTLs not sounding natural is false. They do exceedingly well, but there is a variable of speaker. If you played one of our amps on a speaker that the amp was unhappy with (which, based on your comments at face value, apparently is what happened, and as best I can make out must have also been decades ago), who knows what it would sound like. I've always been very careful to vet speakers ahead of time with all of our customers to make sure I don't make a sale to someone that won't work. Everyone is the better for it and that's part of why we're still around over 45 years.
You see, you think I am trying to convince you…I am not. I am providing the forum with my analysis of available literature, which IMO, you misinterpret or twist to fit whatever product you make (now claiming Class D has distortion like a SET…but lower…indeed…wonders never cease). OTL pattern is just like a SET but lower…except it’s not as the even harmonics are cancelled.
So, I think you are wrong in your interpretation and try to present an alternative analysis. I realize you won’t be swayed and that’s a pity because it probably prevents you from making your best possible amp.
Huh? You're not trying to convince me?? why all the links and endless responses?? Are you trolling me? Sheesh.
I'm not twisting anything. This is the root of the head butting right here. As far as I can make out, your analysis has been in error or that of the links you've provided have, sometimes both (example: Keith Howard's stuff).
You've pointed to me to Harry Olsen, yet when I look at his book you recommended, he didn't support your position.
This has happened over and over, at this point over a period of years. Sure looks like you're trying to convince someone- walks like a duck, quacks like one...
Never once have I said that the 'OTL pattern is just like a SET but lower'!! ... Please don't do the strawman thing, please? - when you do that, it means your argument is inherently false. I'd prefer to get to the bottom of this instead.
What I have been saying is the distortion signature of our OTLs (a fully differential amplifier, unlike many OTLs) is better than and different than that of SETs. I've explained how and why in very easy to understand terms, I suspect several times on this thread alone. If you were to actually read them you would be forced to agree with me, if you believe all that stuff about exponential harmonic decay. That is what you've been spouting at me and I've spouted it right back- we're on the same page with that but you don't seem to see it. Exponential decay is good, and can be improved upon by using a better exponent.
I've also never said that our class D has the distortion signature of an SET; I've said it has that of a very good tube amp but at a lower level.
There is no electrical DC feedback in your circuit diagram, but there is an option to do the feedback function manually. And this function is performed by the user. Huge milliammeters on the front of the devices prove the importance of these manipulations.
Maybe in the latest models you are using an analog zero adjustment on the output. But this is impossible without feedback. Will you show a schematic diagram that proves otherwise?
You firmly believe that things that are bad for you should be better for others. It is very self-assured, but has no foundation.
Now why would he do that - its his IP & you are an amp builder. Buy one that go till your hearts content...
What’s really happening is this is called the “SET amp owners thread”, and someone’s using up half the content lecturing everyone else that SET are inferior to PP, OTL, Class-D, whatnot, making all sorts of claims, including that he’s the one being patronized. Best to have your popcorn ready and enjoy the show!
Greetings from Switzerland, David.
Last edited:
Trolling? Sorry, but who is coming on a SET owners thread and telling them that their SETs are no good and that your amps make lower distortion (Duh!) and better distortion (highly questionable for all the reasons I have laid out before. Also, contra to most of the literature that is available around psychoacoustics of distortion)? You don't think this will result in pushback from experienced listeners who have come to prefer the sound of SET amplification? Many of us have used OTL in the past. I myself have owned two different kinds one of which used 6C33C and the other used some TV tube type that eludes me now (EL509 or something?). I have not owned your amps but heard them a number of times at other people's homes and at shows. Not convinced enough to buy.
I haven't heard your new Ga based Class D but feedback, while positive, doesn't confirm that they sound like tube amps...and what does sounding like a tube amp even mean? You are posting on a SET thread, so when you say tube amp, why would you mean a push/pull tube amp or an OTL? You do realize that SET amps generally sound very different from most push/pull type designs, right? I have head some that are close and those are Class A triode (like 300B or 2A3)...but still not the same sound.
Likewise, telling them that only 20% of the power is "usable" because of distortion, ignoring the fact that A) most SET users have high sensitivity speakers where only a few watts are needed to get high SPL and B) That the audibility of distortion is dependent on absolute SPL, which works perfectly then with amps that increase THD with increasing power because the masking holds up as the SPL increases...assuming the CHARACTER of that distortion doesn't change with SPL...in most cases it doesn't...there are always exceptions. As an example. Let's say you have 100dB speakers and an 8 watt (3%THD) 300B amp. At 8 watts the SPL level will be about 109 dB for a single speaker so a pair at 3meters will still be easily 105+ dB SPL or more. Do you think 3% THD, that is mostly 2nd and 3rd harmonic, will be at all audible with 105dB SPL? Most of the time that amp will be cruising at mW because most people listen at well below 100dB. Average levels of 85dB are already quite loud and i rarely listen that loud. Even if that amp makes 4W at 1% THD that is over 100dB SPL at the listening position. That is 50% the rated power. Do you think that 1% is audible with 100dB levels...especially if those are mS long peaks and not sustained levels? Get real and think about what you post!
Here I was thinking this was a SET amp appreciation thread. Hopefully we can go back to it being what it was created for. If only hinting would suffice.
Similar threads
