How to increase Harmonic distortion for clean sounds

[kg]

generally speaking (ie no clipping) tubes will exhibit more even order distortion at low cathode currents. This is due to the way the control grid pinches off current gradually.

[Chuck H]

generally speaking (ie no clipping) tubes will exhibit more even order distortion at low cathode currents. This is due to the way the control grid pinches off current gradually.

Ah, but that should be clarified! The implication being that if we run lower currents we get more even order harmonics. Great! BUT... In a push/pull output stage these distortions are somewhat cancelled. Getting an amp into some compression can detriment it's ability to accurately cancel these "distortions" as intended. So sometimes MORE cathode current can get you there. More harmonics doesn't count if we don't get to hear them.

[Dave H]

..Consider some of those great AC30 tones from the 60's that don't sound clipped at all, but don't sound clean either!?! This would be an amp that possesses all of the above qualities in spades.

I’ve always thought the AC30 has the best unclipped ‘clean’ tone of any amplifier so you’d expect it to have lots of lovely even harmonics right? Wrong! When I put it on the scope and wind it up to just below clipping it has a lot more 3rd, 5th and 7th harmonics than it does 2nd, 4th and 6th.

[Chuck H]

Since when do people (especially guitar players) really know what they want?

[Alan0354]

Ah, but that should be clarified! The implication being that if we run lower currents we get more even order harmonics. Great! BUT... In a push/pull output stage these distortions are somewhat cancelled. Getting an amp into some compression can detriment it's ability to accurately cancel these "distortions" as intended. So sometimes MORE cathode current can get you there. More harmonics doesn't count if we don't get to hear them.

But we are talking about creating even harmonics in the preamp tubes. Once you created the even harmonics, the push pull power amp will not cancel it no matter what.

To me, when a preamp have even number of cathode biased triodes like a typical amp, the even stages being an inverting stage, will introducing even harmonic in the opposite end of the wave ( more like stretching and compressing the other side of the wave) and cancelling the effect of the distortion of the first stage!!!

Maybe that's the reason why the Normal channel and the Vibrato channel of the old Fender BF and SF sound very different. The Normal stage has two preamp stage, the Vibrato channel has three stages because it needs to sum the reverb and vibrato in.

[Alan0354]

I’ve always thought the AC30 has the best unclipped ‘clean’ tone of any amplifier so you’d expect it to have lots of lovely even harmonics right? Wrong! When I put it on the scope and wind it up to just below clipping it has a lot more 3rd, 5th and 7th harmonics than it does 2nd, 4th and 6th.

AC 30 uses 220K plate resistance, as you can see in my calculation in post #43, you get much lower distortion with load resistance much higher than the plate resistance of the tube.

As I described in post #50, if you have even number of preamp triodes in the signal chain, the even harmonics cancel out.

[Chuck H]

But we are talking about creating even harmonics in the preamp tubes. Once you created the even harmonics, the push pull power amp will not cancel it no matter what.

Correct! And an important distinction. But the really fun stuff happens at the meeting place for compression, harmonic distortions, resonances and acoustic feedback. And that's the power amp.

To me, when a preamp have two stages of cathode biased preamp tubes like a typical amp, the second stage being an inverting stage, will introducing even harmonic in the opposite end of the wave ( more like stretching and compressing the other side of the wave) and cancelling the effect of the distortion of the first stage!!!

Nope. Because the stages are in series, not parallel. If the first stage creates a harmonic the second stage flips it and amplifies it. They (ideally) never appear side by side with each other. One relevant design note is that preamp stages often share a cathode circuit. If a suitable bypass cap isn't used there may indeed be some cancellation. I'm not sure if the shared current can have an affect on this. Then there is the small resistances in the decoupling circuits due to filter impedance, ground leads and the chassis itself. Though I've always figured the chassis to be an insignificant issue there are cases where enough current can occur across it to cause noise problems. So why not small cancellations as well? The best way to be sure is to use separate cathode circuits, don't daisy chain the grounds for these circuits, locate the decoupling cap very near the circuits it will decouple and choose a decoupling cap with a low impedance.

Maybe that's the reason why the Normal channel and the Vibrato channel of the old Fender BF and SF sound very different. The Normal stage has two preamp stage, the Vibrato channel has three stages because it needs to sum the reverb and vibrato in.

Probably not. There are many differences between those two channels. Not the least of which is indeed the phase. But I think that has more affect on what you feel as a player relative to whether you're facing the amp or the audience! If one channel is giving you grief with phase cancellation from a certain playing position you only need to turn around or change position to alter it. Carlos Santana is known to sound check a stage before a show and mark the stage in spots that give him the resonances he needs to achieve as part of his signature sound and technique.

[Alan0354]

Nope. Because the stages are in series, not parallel. If the first stage creates a harmonic the second stage flips it and amplifies it. They (ideally) never appear side by side with each other. One relevant design note is that preamp stages often share a cathode circuit. If a suitable bypass cap isn't used there may indeed be some cancellation. I'm not sure if the shared current can have an affect on this. Then there is the small resistances in the decoupling circuits due to filter impedance, ground leads and the chassis itself. Though I've always figured the chassis to be an insignificant issue there are cases where enough current can occur across it to cause noise problems. So why not small cancellations as well? The best way to be sure is to use separate cathode circuits, don't daisy chain the grounds for these circuits, locate the decoupling cap very near the circuits it will decouple and choose a decoupling cap with a low impedance.

But if you look at the waveform, the first stage stretches the lower part of the sine wave and compresses the top part. This gives the even harmonics in technical terms. Now, when this wave enter the second stage, the stretched part of the wave corresponds to the top part of the output wave of the second stage that is being compressed. Then the top compressed part of the wave from the first stage corresponds to the bottom part of the output wave which is stretched. So in all, they kind of cancelled the effect.

Now I am not say they cancelled out and become distortionless. Just a lot of the even harmonics got compensated. Still have a lot of distortion left. Maybe that's the reason Dave H. measured more odd harmonics.

[kg]

Ah, but that should be clarified! The implication being that if we run lower currents we get more even order harmonics. Great! BUT... In a push/pull output stage these distortions are somewhat cancelled. Getting an amp into some compression can detriment it's ability to accurately cancel these "distortions" as intended. So sometimes MORE cathode current can get you there. More harmonics doesn't count if we don't get to hear them.

clarified? nahh. the tube is a tube.

expounded on, perhaps.

[kg]

But if you look at the waveform, the first stage stretches the lower part of the sine wave and compresses the top part. This gives the even harmonics in technical terms. Now, when this wave enter the second stage, the stretched part of the wave corresponds to the top part of the output wave of the second stage that is being compressed. Then the top compressed part of the wave from the first stage corresponds to the bottom part of the output wave which is stretched. So in all, they kind of cancelled the effect.

Now I am not say they cancelled out and become distortionless. Just a lot of the even harmonics got compensated. Still have a lot of distortion left. Maybe that's the reason Dave H. measured more odd harmonics.

re: cancellation of distortion:

http://ken-gilbert.com/images/pdf/harmonic.pdf

[eschertron]

re: cancellation of distortion:

http://ken-gilbert.com/images/pdf/harmonic.pdf

Sidebands of sidebands!! Interesting to find out that distorting an already-distorted waveform adds more distortion!

[Leo_Gnardo]

Sidebands of sidebands!! Interesting to find out that distorting an already-distorted waveform adds more distortion!

It's what makes music interesting.

[Alan0354]

re: cancellation of distortion:

http://ken-gilbert.com/images/pdf/harmonic.pdf

That's exactly my point. I don't have simulation on the IMD the author has. But just by common sense rationing, I know the 2nd harmonics is cancelled if you have two cascaded 12AX7 stage.

That's exactly the reason David H notice odd harmonics only in Vox AC30. It has even number of triode stages on both channels.

I actually did designs in introducing even harmonics in the circuit and compared sonic wise. The one with even harmonics had a "singing" quality. I am not sure I would say it's an improvement, but it's a little different. Just like the comparison between the Vibrato and the Normal channels of the BF/SF Fender. I consider the Normal channel sounds better than the Vibrato channel. But this is subjective. You have to listen and judge yourself.

Here, the question is introducing harmonic distortion, whether it sounds good or not is not relevant to me here..............well it is important, but not what the question is.

[Alan0354]

Sidebands of sidebands!! Interesting to find out that distorting an already-distorted waveform adds more distortion!

That's how electronic signal is, you multiply the output power series of the first amp with the power series of the second amp, you get all sort of funny harmonics, the f1+f2 and f1-f2 IMD quickly become impossible to analyze like the author said. So, this kind of circuit ultimately is going to judge by ears.

This is only sine wave analysis of a SINGLE frequency. Look at the two tone test in the later part of the paper, look at how many IMD frequencies it produced just by two closely spaced frequency signal. The guitar signal itself has so many harmonics and can you imagine going through two stages that introduce it's only harmonics?

I have a suspicion if you do the FFT or look at it thought a Spectrum Analyzer, you see a continuous spread of frequency. It is my guess it's more informative to just look at the envelope of the spectrum than individual frequency.

[Gingertube]

While talking about AC30 amps note that the Top Boost preamp uses a DC coupled cathode follower to generate a large part of the sound.
Merlin covered this (the DCCF) here:
The Valve Wizard
The DCCF (DC coupled cathode follower) is often used to drive the tone stack. Marshalls in particular. It compresses one side of the signal to generate a far amount mostly even order harmonic distortion.
2nd 3rd 4th and 5th harmonics all add to the sound, the ones you want to get rid of (by limiting higher frequency response, or better yet, by not generating them in the first place) is the 7th , 9th and 11th harmonics which are musically "quint" (bearing no musical relationship to the original note) and quite objectionable. Note that reducing RL (even though it is in the cathode rather then the anode) increases the distortion just as for an ordinary common cathode gain stage.

If you draw a transfer function (that is a Vout vs Vin plot) you will see compression as a rounding off at the high output /high input part of the plot. The larger the curvature of this plot the larger the amount of distortion and the tighter the turn the higher the order of the harmonic distortion products.

And remember always:
The non-linearity which produces the harmonic distortion also produces intermodulation distortion, you can't have one without the other.
Also the higher the order of the distortion product the more intermodulation products are produced.

What do I mean by that?
Skip this part if maths "gets on your wick".
Take 2 notes (frequencies) and put them through a non-linear stage (lets use 3kHz and 4kHz for example) :
The maths works like this:
Take 2 sinusoidal signal (well we use cosine rather than sine to keep the maths simple)
Signal 1 = a1(cos x)
Signal 2 = a2(cos y)

From output = a1 cos(x) + a2 cos(y)
Expanding this and substituting in the trigonometric identity cos(x) + cos(y) = 1/2(cos(x+y) + cos(x-y)) and a lot of tedious algebra later we end up with an expression for:

1) the 2nd order term , For the superposition of 2 signals of x = 3KHz and y = 4kHz then four(4) new frequencies are created, 2 off 2nd harmonic terms and 2 off IM sidebands :
2x = 6kHz , 2nd harmonic of x
2y = 8kHz , 2nd harmonic of y
x + y = 7kHz , IM sideband
x - y = 1kHz , IM sideband

2) The 3rd order term gives six (6) new frequencies. 2 off 3rd harmonics and 4 off IM sidebands
3x= 9kHz , 3rd harmonic of x
3y = 12kHz , 3rd harmonic of y
2x + y = 10kHz , IM Sideband
2x - y = 2kHz , IM Sideband
2y + x = 11kHz , IM sideband
2y - x = 5kHz , IM Sideband

Note also that 2 of the 3rd order Intermodualtion products are very close to the original frequencies and are going to mess with your interpretation of the origial sound a little more.

Higher order terms produce even more Intermodulation products (more IM Sidebands) and can "clutter" the sound.

Cheers,
Ian