Elementary questions on gain.

[katopan]

It takes a long thread to analyse a whole amp in a way that no one has ever posted before!

Sorry Matt, but I'm also going to respectfully disagree somewhat. There are plenty of amps that clip the power stage before anything else. Ones I can say for sure because I've had my CRO on them watching what's going on are my 18 Watt clone, my friend's JCM800 clone and another friend's Fender Deluxe Reverb. They all power grid clip before any PI clipping. Of course the JCM800 can be set to preamp clip with the power stage clean, but it overdrives the output stage beautifully with the 'gain' down low and the 'master' up high. The 18 Watt is especially good because of the very high sensitivity of EL84's. To quote Zaphod Phil, he's said many times that the EL84s are having the living snot driven out of them before PI clipping comes into play. And with those amps unless you are boosting the signal into the preamp, the pre is totally clean as it is only one stage. The PI clipping only comes in at the top 1/4 of the volume knob and you can hear as well as see on a CRO its effect.

I'm not sure where the idea that the PI is clipping first is coming from. I've seen a few people mis-interpret the clipped PI plate signals as PI clipping. In actual fact unless it's cutting off and getting right up close to the PI power node supply rail voltage, it's actually the power grid clipping taking the tops off the PI plate signals. Proof of PI clipping contributing to the overdriven distortion is when the PI grids reach clipping or cutoff. You see it at the PI grids and on the plate of the previous stage, rather than on the PI plates.

[matt h]

(deleted)

[tonequester]

Greetings to ALL.

Wow ! What info to absorb. I really appreciate my fellow forum folks willingness to reply. I have read that the pi contributes to the over driven sound, but have never heard that it was the greater source of distortion than the output tubes. I have been checking out buying a small custom built amplifier. I don't guess that it is proper to name the company, but I had a lengthy discussion with the owner and he claims that his low gain amps are designed to allow power tube clipping first, then the phase inverter, and finally the pre amp section.
The fact that after 35+ years of playing, I am finally going to invest a larger amount of cash(for me anyway) is one reason that I am trying to get a handle on just what happens when an amp is pushed. The amp I am looking at buying has some interesting and unique features. It is 100% all tube and hand wired in the U.S. Of course it SHOULD be a wonder to behold with a price tag of $1,500.00 sans cab and speakers.
Anyway, once again I thank you ALL for your interest and helpful advice. Best Regards,..................tonequester.

[pdf64]

I'm not sure where the idea that the PI is clipping first is coming from. I've seen a few people mis-interpret the clipped PI plate signals as PI clipping. In actual fact unless it's cutting off and getting right up close to the PI power node supply rail voltage, it's actually the power grid clipping taking the tops off the PI plate signals. Proof of PI clipping contributing to the overdriven distortion is when the PI grids reach clipping or cutoff. You see it at the PI grids and on the plate of the previous stage, rather than on the PI plates.

katopan, thanks for clarifying the above, that has been my conclusion also. Kudos regarding the amazing work in the Express thread.

I'd be grateful if you could clarify what exactly you mean by 'mixed mode distortion'?
I'm not quite sure whether it refers to the Express' unusual characteristic you revealed, in that the upper half-cycle is power tube, the lower half is PI, clipping; or whether it's a more general term for amps in which more than one stage clip at broadly similar signal levels.

matt, please be assured that my query was made with the best of motives and I'm not a 'flame war' type of chap; rather I'm grateful for being educated if corrected in a misapprehension, I lose interest if a technically correct position isn't accepted, or I'm happy to 'agree to dis-agree' if the matter is one of sensory perception (esp 'golden ears') or technical interpretation.
Regarding the terminology covering the effect of non-linearities in amplitude response, I like to use a cover-all of 'overdrive' to include distortion, clipping, compression etc.
I agree that from a player's perspective, generally clipping has to be quite significant before the tone becomes noticeably distorted.
Personally I don't enjoy playing an electric guitar that has a totally unclipped signal path; even for folky strumming, I find that the added 'chiming / jangley' harmonics and dynamic compression resulting from a non-obvious degree of clipping are necessary to mimic the response of a nice acoustic guitar.

The viewpoint I've come to regarding this topic is that of all the various points in the signal chain of a regular rig at which overdrive can occur (eg pedals, pre-amp, phase splitter, power tubes), it's overdrive of the power tubes which has the most significant effect.
The reasons for this are -

1/ power supply sag is all but essential for a dynamic response that appeals to me, and the power amp needs pushing hard for a good degree of sag (whether plate or screen node).
2/ similar to a car passing through several paint shops, its the final one that probably has the most significant effect, so each overdriving stage in the path may affect the perceived overdriven tone / response but the last one is what the listener hears, and so is generally the most significant.
3/ my perception is that it is the electro / magnetic / acoustic resonant circuit created by pentode / beam tetrode power tubes being driven into cut off with the particular frequency / impedance load and SPL / frequency characteristics of the generic guitar speaker, that is necessary for a tone that appeals to me. To such an extent that I use a Fluxtone speaker to preserve what's left of my hearing.

It's not feasible or perhaps even desirable to run all preceding stages clean and only have power amp overdrive; to avoid blocking distortion at the power tube grids, and for sufficient gain needed with rock lead, one or (probably) more of the earlier stages will necessarily overdrive.

An overdriven LTP can sound quite similar to overdriven power tubes in terms of frequency balance but is let down by a flat dynamic response (probably due to point 1 above).
However, gain staging can be tweaked such that once the power tubes are overdriven, it doesn't seem to make any difference to the tone whether the necessary gain (and likely overdrive) is achieved by the LTP or a preceding stage.

Using an amp with a PPIMV, a good overdriven tone can be had with a 12AX7 in the LTP, whereas swapping with a 12AT7 is, to my ears, much less pleasent. However, turn up the PPIMV to overdrive the power tubes and the difference is barely perceptible. This has helped to lead me to point 2 above.
(The SE gain of the same LTP with 12AX7 is ~30, and ~20 with a 12AT7.)

Sorry for the long post, I hope the above clarifies my viewpoint.
Pete

[matt h]

(deleted)

[katopan]

Thanks Pete. "Mixed mode distortion" are my words describing the upper half-cycle is power tube, the lower half is PI, clipping. I've never seen any other amp where one side of the waveform is predominantly clipped from one stage and the other side from somewhere else in the amp. This behaviour is unique.

similar to a car passing through several paint shops

[pdf64]

matt, re the SE & B+ sag thing I agree, I'm fairly sure that current draw almost always will increase significantly when overdriven.
From https://tubeamparchive.com/viewtopic.ph ... c&start=30 see post by jml77

epiphone VJ with reduced interstage attenuation, with a 500mv pk-pk 500 hz i/p, with the amp cranked into an 8ohm dummy load, the increase in current draw from quiescent was 73%

So the dynamic response should be lively.
Pete

[matt h]

(deleted)

[Jana]

I've been really busy lately with my move into a house but felt the need to respond to this thread. I have skimmed through it but I think it is safe to say that I agree with matt h. The PI in the majority of amps is what is distorting. Technically, this is the "power stage" of the amp since the PI is a part of the power amp section. Matt's description of power tube distortion as a "fart in a whirlwind" is an apt term. I would have said, "an elephant that ate too many beans."

I caught hell on another well-known forum a few years ago for suggesting that it was the PI that was doing the clipping and not the power tubes. I would have been better received if I had taken a crap on a crucifix at the Vatican.

Alrighty, I have to get back to work cleaning cat hair and cat piss out of this house. I fucking hate cats!

[eniam rognab]

Alrighty, I have to get back to work cleaning cat hair and cat piss out of this house. I fucking hate cats!

me too!!! bark and woof are all i speak, cats are worthless....

good luck jana!!

[matt h]

(deleted)

[pdf64]

The PI in the majority of amps is what is distorting

jana, what draws you to that conclusion?
As described by katopan above, my finding is that the general case is the power tube control grids clip the phase splitter outputs well before the phase splitter outputs reach their max positive swing.
So in the same way that with an MXR distortion pedal, the diodes clip the signal much earlier than the opamp hits the rails, so the power tube control grids are the first point that clipping occurs in a regular tube amp.

I acknowledge that alternative conclusions may be drawn, eg the phase splitter output is being clipped, therefore that's where the overdrive is occurring; such that if the phase splitter could source sufficient current then its output would pull the power tube control grids positive and true power tube cathode current saturation would occur.

However, it seems to me that the amp would then be a class AB2 and so a rather different type of type of amp.

It may therefore be inherent in the definition of a class AB1 amp that its max output is that which occurs when the instantaneous sum of the bias voltage and the signal voltage rises to make Vg-k = 0.
Hope that makes sense - Pete

[cxx]

3. It all depends on circuit design. Old non-master volume amps will typically overdrive the power tubes first, then the phase inverter and then the preamp stages, as the volume is turned up.

I agree that it all depends on circuit design. However, the overdriving of power tubes first older non-MV amps is mostly a myth--and it's been largely debunked. In most cases, what people are claiming to be power tube distortion is actually PI distortion (typically in an LTPI arrangment. By the time most power tubes are distorting (thd rather than ear figure, which is much earlier), the earlier stages are audibly distorting to such a degree that the power tube's contribution is minor. There are always a few exceptions here and there, but they're few and far between.

One particularly interesting case that does come to mind with actual power tube distortion is the (or rather, "are the") tweed deluxe(s). This isn't to say the second (series) gain stage in them isn't overdriving long before the power tubes, though... but what that second stage is doing is interesting (is it the pre-concertina driver or both halves in the form of the earlier paraphrase models).

I cannot think of an amp which has distorting power tubes providing 100% (or anything approaching a majority, to think of it) of the distortion in the amp. Oh, wait, yes I suppose the early Vox AC5 did. A single EF86 driving an EL84... the EL84 would definitely be contributing audible distortion in that case, though the EF86 would be hitting the rails and contributing about as much. It'd be hard to say which was actually more responsible for the tone at that point.

I'm not sure that power tube distortion is a myth. Most every single ended amp is pretty much one gain stage after another with the power tube taking the brunt of the abuse and distorting first.

PP Amps without negative feedback are pretty much in the same boat. The power tubes will compress and distort before the PI unless you arrange the limits of each stage to prevent it. Most historical guitar amps aren't like that.

When you get negative feedback around the PI PA it's a different story. The PI produces a funny looking waveform as a result of the PA feedback canceling some of the input and amplifying the negative of the error, attempting to correct the distortion of the PA. The error is the difference between the input to the PI and the output of the PA. At some point it cannot compensate. There has been a lot of discussion on how this last part evolves.

So I don't think power tube distortion is a myth. I do think that the character of the distorted sound of an amp has a lot do do with where the low frequencies are eliminated. Fenders typically let most of the bass frequencies through all of the preamp and pinch it off just before the PI. Too much gets through for my taste. The lead marshalls with the .68uF bypass caps in various places, eliminate a lot of bass frequencies up front. The character of the resulting distorted sound is very different. The express looks to have a fendery tone stack early with a significant low filter just before the third stage and produces an overall brighter tone that the Marshalls.

[jazbo8]

Technically, this is the "power stage" of the amp since the PI is a part of the power amp section.

I'd avoid saying "power stage" because then the implication is that a PI is a power tube. Yikes.

+1, while technically correct once the NFB loop is closed, it is still a good idea to keep the PI separated from the power stage. As kapotan already pointed out, most of the "PI distortion" observed is actually due to the NFB signal, i.e., that the odd looking PI output waveform is trying to correct the output tubes' distortion, if you break the loop, then PI's output would be pretty clean even when the power tubes are clipped.

+1 also on gain vs. distortion, these two are often mixed up, in this thread, the question is largely on distortion, not gain per se.

It is also worth remembering that in the older amps, before the "hi-gain" designs came into vogue, the power tubes were all designed to be clipped first since you want to achieve the maximum output power using the most expensive tubes in the amps, otherwise, it would be a total waste of money - no manufacturers' bean counters would allow the companies to do that - just simple economics at work Of course, these days, any number of design options are available, it just depends on what the players prefer, which is usually tied to the music genre, some amps are just more suitable for certain genre...

[matt h]

(deleted)