Distortion in tubed amplifiers
- Thread starter analag
- Start date
Help Support GroupDIY Audio Forum:
Try grid leak bias, sound like what your are describing. There is info on it in the Radiotron Designers handbook, and PPR and a few othe poeple answered my questions on it not to long ago. a search should pull that up for you.
adam
adam
> We all know how tubes don't clip the way transistors do,
I don't know that.
Hard-clipping implies excess gain. Gain is expensive in hollow-state, cheap in solid-state.
Naked transistors will actually make MORE rounded waves than tubes; so grossly rounded that we "always" use the cheap gain for huge negative feedback.
Wire-up a cathodyne and slam it: it will clip about as hard as any common BJT stage.
Crank a Macintosh MC-50 into clipping: some sand-state amps sound worse, but an MC-50 can't be confused for a Fender Twin or a Philco. It rises to the limits and lies there, negligible rounding.
Another thing about tubes: the grid is infinite-Z up to a point, and then low-Z. Use capacitor-coupling, it de-biases itself on strong input signals. This may be the most important "grey trick" that is natural to tubes and harder to do with BJT.
> I'm interested in the "grey area".
I think the "white versus black" areas were pretty well understood 50 years ago, but the "grey area" is much bigger and may never be understood.
I don't know that.
Hard-clipping implies excess gain. Gain is expensive in hollow-state, cheap in solid-state.
Naked transistors will actually make MORE rounded waves than tubes; so grossly rounded that we "always" use the cheap gain for huge negative feedback.
Wire-up a cathodyne and slam it: it will clip about as hard as any common BJT stage.
Crank a Macintosh MC-50 into clipping: some sand-state amps sound worse, but an MC-50 can't be confused for a Fender Twin or a Philco. It rises to the limits and lies there, negligible rounding.
Another thing about tubes: the grid is infinite-Z up to a point, and then low-Z. Use capacitor-coupling, it de-biases itself on strong input signals. This may be the most important "grey trick" that is natural to tubes and harder to do with BJT.
> I'm interested in the "grey area".
I think the "white versus black" areas were pretty well understood 50 years ago, but the "grey area" is much bigger and may never be understood.
AMZ-FX
Well-known member
I'm sure you have probably seen this article called The Cool Sound of Tubes...
In the section called Distortion Under Test there are some graphs of the harmonic spectra. The transistors have much higher levels of even and odd harmonics than do tubes. The really interesting part is that when you run a biploar or mosfet on high V+ supplies, you get fewer harmonics at lower levels just like tubes. The test conditions were slanted in favor of tubes (doh!) and are not specifically designed to reveal clipping characteristics. Also of particular interest is the graph of the transformer which has a very low noise floor but predominately 3rd order distortion.
regards, Jack
VacuumVoodoo
Well-known member
An interesting item on tube amp distortion by John Murphy of True Audio
bcarso
Well-known member
[quote author="AMZ-FX"]
In the section called Distortion Under Test there are some graphs of the harmonic spectra. The transistors have much higher levels of even and odd harmonics than do tubes. The really interesting part is that when you run a biploar or mosfet on high V+ supplies, you get fewer harmonics at lower levels just like tubes. The test conditions were slanted in favor of tubes (doh!) and are not specifically designed to reveal clipping characteristics. Also of particular interest is the graph of the transformer which has a very low noise floor but predominately 3rd order distortion.
regards, Jack[/quote]
It's really great that Eric B. got this forum, and did this reasonably well-balanced piece---a lot of useful infromation that should cause a few engineers, who wouldn't otherwise bother to investigate, to question their knee-jerk opinions. Thanks for linking this Jack.
In an article of its size there are, naturally, a few things not mentioned, but overall it is pretty thorough.
I'm sure you have probably seen this article called The Cool Sound of Tubes...
In the section called Distortion Under Test there are some graphs of the harmonic spectra. The transistors have much higher levels of even and odd harmonics than do tubes. The really interesting part is that when you run a biploar or mosfet on high V+ supplies, you get fewer harmonics at lower levels just like tubes. The test conditions were slanted in favor of tubes (doh!) and are not specifically designed to reveal clipping characteristics. Also of particular interest is the graph of the transformer which has a very low noise floor but predominately 3rd order distortion.
regards, Jack[/quote]
It's really great that Eric B. got this forum, and did this reasonably well-balanced piece---a lot of useful infromation that should cause a few engineers, who wouldn't otherwise bother to investigate, to question their knee-jerk opinions. Thanks for linking this Jack.
In an article of its size there are, naturally, a few things not mentioned, but overall it is pretty thorough.
pstamler
Well-known member
[quote author="VacuumVoodoo"]An interesting item on tube amp distortion by John Murphy of True Audio[/quote]
I read it; interesting, but when he said "As for class A tube preamps, Fourier analysis helps reveal the harmonic structure of the clipped waveforms, Murphy said, noting that the unclipped waves have no harmonics, except for residual distortion" I got off the boat. What he's basically saying is that there's no distortion present except the distortion which is present. That's a little too much Zen for me.
Peace,
Paul
I read it; interesting, but when he said "As for class A tube preamps, Fourier analysis helps reveal the harmonic structure of the clipped waveforms, Murphy said, noting that the unclipped waves have no harmonics, except for residual distortion" I got off the boat. What he's basically saying is that there's no distortion present except the distortion which is present. That's a little too much Zen for me.
Peace,
Paul
It all comes down to listening. No matter how many equations you manipulate.
AMZ-FX
Well-known member
And therein lies the difference in the sound of the various boxes :grin:
-Jack
bcarso
Well-known member
[quote author="pstamler"][quote author="VacuumVoodoo"]An interesting item on tube amp distortion by John Murphy of True Audio[/quote]
I read it; interesting, but when he said "As for class A tube preamps, Fourier analysis helps reveal the harmonic structure of the clipped waveforms, Murphy said, noting that the unclipped waves have no harmonics, except for residual distortion" I got off the boat. What he's basically saying is that there's no distortion present except the distortion which is present. That's a little too much Zen for me.
Peace,
Paul[/quote]
Does that mean that the preference for single-ended stages corresponds to the sound of one hand clapping? :razz:
I read it; interesting, but when he said "As for class A tube preamps, Fourier analysis helps reveal the harmonic structure of the clipped waveforms, Murphy said, noting that the unclipped waves have no harmonics, except for residual distortion" I got off the boat. What he's basically saying is that there's no distortion present except the distortion which is present. That's a little too much Zen for me.
Peace,
Paul[/quote]
Does that mean that the preference for single-ended stages corresponds to the sound of one hand clapping? :razz:
> no distortion present except the distortion which is present. That's a little too much Zen for me.
No, no.
First: while he may have said some of these things, the write-down was done by an EE writer who does not fully grok tubes or tube-amps or tube-culture. EET is lightweight uncritical journalism anyway, and 99.9% of what they do is the latest/greatest stuff that nobody even understands yet.
The idea in that passage is: in a Guitar Amp which is routinely driven into and out of overload, the un-clipped waveforms are not too interesting. Yes they have a residual imperfection (whatever perfection could be on a signal that has no acoustic source) but it is mild. What IS interesting is the way it transistions from that tenth-percent residual to the 5% and 25% THD in the overload mode, and the various interesting ways distortion changes with depth of overload.
Murphy is no fool. He does use spectrum analysis where old-timers just used their ears, and he has studied the heavy-distortion regime whereas most audio workers stop when the wave comes out barely-bent. I do think he oversimplifies the problem of faking tube sound with sand, but he did get well beyond the gross diode-clippers of early sand-state "distortion effects" by aiming for asymmetry, droop, and time-constant action.
Barbour's old article is enthusiatic but dated. I would not agree that there are "new tubes". Everything we have is 1960s or older. The KT-90 is just a big KT-88; the JJ30 is admittedly an oversize 300B. The Russian monsters must date to the early 1970s, and are "new" only because they were State Secrets until the fall of the USSR. The 3c33 is clearly a 1949 6080 pass-tube that some Russian designer monsterized (and the 6080 itself ain't much more than a very-fat 1935 2A3). There is a twin-triode faddish in preamp designs that is just a 6BQ7 built for serious hard-ball (national defense instead of TV variety-show reception). All these tubes differ from their grandfathers less than the 6L6GC differs from an original 6L6.
No, no.
First: while he may have said some of these things, the write-down was done by an EE writer who does not fully grok tubes or tube-amps or tube-culture. EET is lightweight uncritical journalism anyway, and 99.9% of what they do is the latest/greatest stuff that nobody even understands yet.
The idea in that passage is: in a Guitar Amp which is routinely driven into and out of overload, the un-clipped waveforms are not too interesting. Yes they have a residual imperfection (whatever perfection could be on a signal that has no acoustic source) but it is mild. What IS interesting is the way it transistions from that tenth-percent residual to the 5% and 25% THD in the overload mode, and the various interesting ways distortion changes with depth of overload.
Murphy is no fool. He does use spectrum analysis where old-timers just used their ears, and he has studied the heavy-distortion regime whereas most audio workers stop when the wave comes out barely-bent. I do think he oversimplifies the problem of faking tube sound with sand, but he did get well beyond the gross diode-clippers of early sand-state "distortion effects" by aiming for asymmetry, droop, and time-constant action.
Barbour's old article is enthusiatic but dated. I would not agree that there are "new tubes". Everything we have is 1960s or older. The KT-90 is just a big KT-88; the JJ30 is admittedly an oversize 300B. The Russian monsters must date to the early 1970s, and are "new" only because they were State Secrets until the fall of the USSR. The 3c33 is clearly a 1949 6080 pass-tube that some Russian designer monsterized (and the 6080 itself ain't much more than a very-fat 1935 2A3). There is a twin-triode faddish in preamp designs that is just a 6BQ7 built for serious hard-ball (national defense instead of TV variety-show reception). All these tubes differ from their grandfathers less than the 6L6GC differs from an original 6L6.
AMZ-FX
Well-known member
Classic guitar amps have a moderate amount of negative feedback in the poweramp section... and most of the time they are not operated at volumes sufficiently loud to get into the poweramp distortion range. The distortion comes from either cascading preamp sections or boosting the guitar signal before it is sent to the amp. In addition, there is a lot of frequency shaping in the preamp and tone control sections of the preamp... the classic Marshall, for example, has a significant low frequency rolloff and a mid-range dip in the tone control. Tone shaping is quite important.
Is the transistion area from minimal preamp distortion to significant overload interesting... yes, and an area that is ripe for experimentation and testing.
regards, Jack
Similar threads
