glen herrmannsfeldt wrote:> Les Cargill<lcargill99@comcast.com> wrote: > >>> (snip on tube vs. transistor amplifiers) > >>>> I don't question the validity of solid-state amps creating mostly odd >>>> harmonics, but I'd like to know why. It must have something to do with >>>> the way that the amplifiers are designed. Totem-pole output stages might >>>> have more asymmetry when overdriven than a tube's transformer-coupled >>>> output, but H-bridge designs don't. What features of transistor >>>> amplifiers generate the odd harmonics? > > (snip, I wrote) > >>> I was trying to figure out the other way around. Why do tubes >>> generate even harmonics? > >> they don't. single ended circuits generate them because >> they tend to be less symmetric... > > I was remembering push-pull tube amplifiers. When I was in high > school, I inherited a Heathkit tube amp from my grandfather. > The one with the gold colored preamp that plugs into an octal > socket on the power amp. (Jerry might remember that we discussed > that amplifier here before.) I am pretty sure it was push-pull. >Most are. Guitar amps might not be, though. there's a "class A" mania in high-end audio, too. Sometimes that even means single ended.> However, unlike push-pull transistor amplifiers that use NPN > and PNP transistors and a center-tapped power supply, tubes > need a center tapped output transformer and single-ended > power supply. >power supplies vary...>>> It seems to me that if you clip the top and bottom equally, >>> you get odd harmonics. That would seem to me true for a push-pull >>> design in either tubes or transistors, but it might be that >>> transistors are better matched. > >> And you'd be right in all cases.... > >>> Thinking about it more, it is a nice flat clip that generates >>> the odd harmonics. If the supply sags while generating the >>> peak, I believe that will increase the even harmonics. > >> Depends on the duration... > > More specifically, different sag for the positive and negative > peaks. >Right.>>> In that case, the difference should be more in power supply >>> design than in the amplifier stage design. > >>> I do remember claims of transistor amplifiers having the >>> "tube sound" but don't know how they claimed to have done it. > >> FETs. Mainly MOSFETs. > > I suppose, but you also might be able to design a circuit > that had similar clip shape to a tube amplifier. >I forget what the story is with FETs. "Softer knee" or something - the saturation region was more gently curved, where an NPN would simply rail.> Also, the sag could tend to round off the clipped part > instead of the sharp corners from fast transistors and well > regulated power supplies. > > -- glen-- Les Cargill
RIP Jim Marshall
Started by ●April 5, 2012
Reply by ●April 10, 20122012-04-10
Reply by ●April 11, 20122012-04-11
Andreas Huennebeck wrote:> Well, these were 'tube watts'. Tube amplifiers create mostly even order > harmonics when overdriven which sounds nice while solid state amps create > mostly odd order harmonics which sound awful.I think its a little more complicated than that. Class A amplifers: https://en.wikipedia.org/wiki/Electronic_amplifier#Class_A especially ones made with tubes produce even order harmonics. Many of the smaller pre-1980 Fender amps of that were class A. Class B amplifiers (ie push-pull designs) are more likely to produce odd ordered harmonics. However since tubes are more non-linear than transistors *and* have a wider parameter spread for a given model tend to still produce more even order harmonics even in class B designs.> Having played as a drummer in rock bands I can tell though that the best amp > I've came across was the Vox AC30.Apparently the Vox AC30 is a class A: http://www.aikenamps.com/VoxAC30classA_2.html Erik -- ---------------------------------------------------------------------- Erik de Castro Lopo http://www.mega-nerd.com/
Reply by ●April 11, 20122012-04-11
On 11.04.2012 09:49, Erik de Castro Lopo wrote:> Andreas Huennebeck wrote: > >> Well, these were 'tube watts'. Tube amplifiers create mostly even order >> harmonics when overdriven which sounds nice while solid state amps create >> mostly odd order harmonics which sound awful. > > I think its a little more complicated than that. > > Class A amplifers: > > https://en.wikipedia.org/wiki/Electronic_amplifier#Class_A > > especially ones made with tubes produce even order harmonics. > Many of the smaller pre-1980 Fender amps of that were class A. > > Class B amplifiers (ie push-pull designs) are more likely to > produce odd ordered harmonics. However since tubes are more > non-linear than transistors *and* have a wider parameter spread > for a given model tend to still produce more even order harmonics > even in class B designs. > >> Having played as a drummer in rock bands I can tell though that the best amp >> I've came across was the Vox AC30. > > Apparently the Vox AC30 is a class A: > > http://www.aikenamps.com/VoxAC30classA_2.html > > ErikIt looks to me that the AC30 is a class A, but push-pull, so if all components are close to spec, it should also produce odd overtones... best regards, Andre
Reply by ●April 11, 20122012-04-11
On 4/10/2012 12:24 PM, robert bristow-johnson wrote:> On 4/10/12 10:53 AM, Jerry Avins wrote: >> >> I don't question the validity of solid-state amps creating mostly odd >> harmonics, but I'd like to know why. It must have something to do with >> the way that the amplifiers are designed. Totem-pole output stages might >> have more asymmetry when overdriven than a tube's transformer-coupled >> output, but H-bridge designs don't. What features of transistor >> amplifiers generate the odd harmonics? >> >> > > i think, Jerry, it's because about the only mechanism for distortion > with the solid-state amps is the simple clipping at the rails and that > was pretty symmetrical. tube amps had all sorts of other subtle > distortion sources, like the non-linear curve of the tubes, which may or > may not be symmetrical.Both the drivers and output stages of high-power amplifiers -- tube amplifiers especially -- are push-pull*. The only mechanism for asymmetry (and hence the production of odd harmonics) is mismatching tube characteristics, largely corrected by negative feedback. Transistors clip more abruptly than tubes, and solid-state amplifiers more often use regulated power supplies, but I don't see offhand how those differences affect the harmonics generated. Jerry ____________________________________ * Single-ended audio amplifiers must be Class A. -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●April 11, 20122012-04-11
On 4/10/2012 4:40 PM, glen herrmannsfeldt wrote:> Jerry Avins<jya@ieee.org> wrote: > > (snip on tube vs. transistor amplifiers) > >> I don't question the validity of solid-state amps creating mostly odd >> harmonics, but I'd like to know why. It must have something to do with >> the way that the amplifiers are designed. Totem-pole output stages might >> have more asymmetry when overdriven than a tube's transformer-coupled >> output, but H-bridge designs don't. What features of transistor >> amplifiers generate the odd harmonics? > > I was trying to figure out the other way around. Why do tubes > generate even harmonics? > > It seems to me that if you clip the top and bottom equally, > you get odd harmonics. That would seem to me true for a push-pull > design in either tubes or transistors, but it might be that > transistors are better matched.Arrgh! I got my odds and evenses reversed. What a [duh]! Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●April 11, 20122012-04-11
On 4/11/12 11:52 AM, Jerry Avins wrote:> On 4/10/2012 12:24 PM, robert bristow-johnson wrote: >> On 4/10/12 10:53 AM, Jerry Avins wrote: >>> >>> I don't question the validity of solid-state amps creating mostly odd >>> harmonics, but I'd like to know why. It must have something to do with >>> the way that the amplifiers are designed. Totem-pole output stages might >>> have more asymmetry when overdriven than a tube's transformer-coupled >>> output, but H-bridge designs don't. What features of transistor >>> amplifiers generate the odd harmonics? >>> >>> >> >> i think, Jerry, it's because about the only mechanism for distortion >> with the solid-state amps is the simple clipping at the rails and that >> was pretty symmetrical. tube amps had all sorts of other subtle >> distortion sources, like the non-linear curve of the tubes, which may or >> may not be symmetrical. > > Both the drivers and output stages of high-power amplifiers -- tube > amplifiers especially -- are push-pull*.i would agree with that regarding the output stages (and there has to be a center-tap output transformer involved, and there will be some hysteresis coming with that). i wouldn't expect push-pull (with the accompanying transformer) for the earlier stages.> The only mechanism for > asymmetry (and hence the production of odd harmonics) is mismatching > tube characteristics,again, you mean *even* harmonics, no? if you drive a non-linear system that has odd symmetry (which means you need matching of components for the push and pull sides) with a sinusoid, you get no even harmonics. but you're absolutely right: that since vacuum tubes don't come in NPN and PNP genders, the only way to get push-pull outa them is with a transformer on the output and, usually, we try to match the components which will eliminate asymmetry and the even harmonics. BTW, i can only remember Class A, Class B, and Class AB from the olden daze. someone will have to explain to me how them other classes work.> largely corrected by negative feedback.i thought that some of those single-ended (transformerless) tube stages had asymmetry which generated even harmonics.> Transistors clip more abruptly than tubes, and solid-state amplifiers > more often use regulated power supplies, but I don't see offhand how > those differences affect the harmonics generated.well, i can see how a transformerless transistor amp with Class A push-pull can get very funky if the power supply *does* droop. doesn't explain it exactly for tube amps with an output transformer, except for the earlier stages that may very well be single-ended Class A. so the kid hits a power chord, the power supply (with tube rectifiers) droops, the final stage clips more (but stays symmetrical), and the previous stages clips (softly) in a more asymmetrical manner. BTW, how're you been doing down there in NJ, Jerry? Shem? Davide? -- r b-j rbj@audioimagination.com "Imagination is more important than knowledge."
Reply by ●April 11, 20122012-04-11
On Wed, 11 Apr 2012 11:55:07 -0400, Jerry Avins <jya@ieee.org> wrote:>On 4/10/2012 4:40 PM, glen herrmannsfeldt wrote: >> Jerry Avins<jya@ieee.org> wrote: >> >> (snip on tube vs. transistor amplifiers) >> >>> I don't question the validity of solid-state amps creating mostly odd >>> harmonics, but I'd like to know why. It must have something to do with >>> the way that the amplifiers are designed. Totem-pole output stages might >>> have more asymmetry when overdriven than a tube's transformer-coupled >>> output, but H-bridge designs don't. What features of transistor >>> amplifiers generate the odd harmonics? >> >> I was trying to figure out the other way around. Why do tubes >> generate even harmonics? >> >> It seems to me that if you clip the top and bottom equally, >> you get odd harmonics. That would seem to me true for a push-pull >> design in either tubes or transistors, but it might be that >> transistors are better matched. > >Arrgh! I got my odds and evenses reversed. What a [duh]! > >JerryI do stuff like that all the time. Left-right, odd-even, modulator-demodulator...something in my brain doesn't keep track of individual items in pairs very well. Eric Jacobsen Anchor Hill Communications www.anchorhill.com
Reply by ●April 11, 20122012-04-11
Eric Jacobsen <eric.jacobsen@ieee.org> wrote: (snip)>>Arrgh! I got my odds and evenses reversed. What a [duh]!> I do stuff like that all the time. Left-right, odd-even, > modulator-demodulator...something in my brain doesn't keep track of > individual items in pairs very well.Often called sign error (an odd or even power of negative 1), there is an old saying: "Good physicists make an even number of mistakes." When I was in high school, my physics teacher told about how the district bought a series of movies (16mm, usual for schools at the time) that were all someone lecturing at a chalk board. (We never got to see them.) As the story goes, there is one where it gets to the end just as he discovers that he made a mistake sometime earlier, but doesn't get to go back and find it. Oops. -- glen
Reply by ●April 13, 20122012-04-13
On 4/11/2012 1:08 PM, robert bristow-johnson wrote:> On 4/11/12 11:52 AM, Jerry Avins wrote: >> On 4/10/2012 12:24 PM, robert bristow-johnson wrote: >>> On 4/10/12 10:53 AM, Jerry Avins wrote: >>>> >>>> I don't question the validity of solid-state amps creating mostly odd >>>> harmonics, but I'd like to know why. It must have something to do with >>>> the way that the amplifiers are designed. Totem-pole output stages >>>> might >>>> have more asymmetry when overdriven than a tube's transformer-coupled >>>> output, but H-bridge designs don't. What features of transistor >>>> amplifiers generate the odd harmonics? >>>> >>>> >>> >>> i think, Jerry, it's because about the only mechanism for distortion >>> with the solid-state amps is the simple clipping at the rails and that >>> was pretty symmetrical. tube amps had all sorts of other subtle >>> distortion sources, like the non-linear curve of the tubes, which may or >>> may not be symmetrical. >> >> Both the drivers and output stages of high-power amplifiers -- tube >> amplifiers especially -- are push-pull*. > > i would agree with that regarding the output stages (and there has to be > a center-tap output transformer involved, and there will be some > hysteresis coming with that). i wouldn't expect push-pull (with the > accompanying transformer) for the earlier stages.Generally, only the high-power stages are push-pull, and only the output will be transformer coupled. (More than one transformer makes stable negative feedback difficult.) It is usually easy to provide enough headroom in the early single-ended circuits to keep their distortion contributions well below that of the power stage. Of course, my main experience is with amplifiers where distortion of any kind is considered a BAD THING. Even a ham Class B modulator or rough-and-ready PA amplifier with no feedback at all operates at no more than 10% THD.>> The only mechanism for >> asymmetry (and hence the production of odd harmonics) is mismatching >> tube characteristics, > > again, you mean *even* harmonics, no? if you drive a non-linear system > that has odd symmetry (which means you need matching of components for > the push and pull sides) with a sinusoid, you get no even harmonics.Ouch! That's so embarrassing, I'm beginning to wonder about incipient senility.> but you're absolutely right: that since vacuum tubes don't come in NPN > and PNP genders, the only way to get push-pull outa them is with a > transformer on the output and, usually, we try to match the components > which will eliminate asymmetry and the even harmonics. > > BTW, i can only remember Class A, Class B, and Class AB from the olden > daze. someone will have to explain to me how them other classes work.Well, there are Classes AB1 and AB2 depending on whether or not grid current is drawn. Older Class B (ans some Class AB2) drivers were loaded with "swamping resisters" to keep the load impedance reasonably constant. I started using negative voltage feedback instead, and eventually saw others doing the same.>> largely corrected by negative feedback. > > i thought that some of those single-ended (transformerless) tube stages > had asymmetry which generated even harmonics.They required double the supply voltage. I never dug into them.>> Transistors clip more abruptly than tubes, and solid-state amplifiers >> more often use regulated power supplies, but I don't see offhand how >> those differences affect the harmonics generated. > > well, i can see how a transformerless transistor amp with Class A > push-pull can get very funky if the power supply *does* droop. > > doesn't explain it exactly for tube amps with an output transformer, > except for the earlier stages that may very well be single-ended Class > A. so the kid hits a power chord, the power supply (with tube > rectifiers) droops, the final stage clips more (but stays symmetrical), > and the previous stages clips (softly) in a more asymmetrical manner.I'd have no trouble building a distorting amplifier if I wanted one. :-)> BTW, how're you been doing down there in NJ, Jerry? Shem? Davide?We're all fine. Thanks for asking. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
