Biasing for harmonic distortion with inverting stages...

[vibrovox7reverb]

I know I just made a topic prior to this, but this is an unrelated question. When you purposely bias a preamp stage hot or cold to get it to clip on one side or the other and then it's inverted would biasing on to clip the same side give a more even distortion signal? For example, (and I may have this backwards) if I were to bias a stage really cold to clip the top of the waveform, and then the signal was inverted would biasing the next stage (assuming it's a similarly designed stage of the same type of tube) give the same type of clipping on the other side of the waveform making a mostly symmetrical distortion? I am not even sure it would sound good to the ear, but I'm just curious from a design standpoint since I'm under the belief that generally hi gain amps are designed hot, cold, hot, cold...when if it seems that maybe they should be hot, hot, cold, cold or something like that.

I am aware that the plate and cathode give off signals of opposite polarity, so does the mean that if the signal came in to the grid and out of the cathode that the signal would never invert?

Sorry if these questions are a little hard to understand...I'm kind of having trouble finding the right way to ask them. If I need to elaborate I will try.

[exclamationmark]

If the signal always came off the cathode of a stage, then yes, no inversion of the signal would occur. Unfortunately, gain is always less than unity when taken from the cathode, so is of little use in overdriving tubes upstream.

In regards to your first question... I suppose clipping would be mostly symmetrical if you biased the stages the same way, but only if you attenuated the signal back down to the original levels of the input to the first tube. Otherwise the gain of the first stage would push the second stage into greater overdrive, leading to asymmetry. Symmetrical clipping probably wouldn't sound too great, as a lot of the 2nd order harmonics would be eliminated, but you never really know until you try.

There is a huge difference in sound between hot biased and cold biased stages. To my ears, hot biased stages seem to have a more gainier, fuzzy sound with more high end harmonics (caused by grid current limiting). On the other hand, cold biased stages seemed to reduce the gain and emphasize mids and smoothness. I don't know about the absoluteness of my personal experiences however... I tested this by using fixed bias on the very last stage of a 3 stage pre-amp, with the first two stages being approximately centre biased. The tubes were 12au7's.

People usually try to alternate hot/cold stages to get a different spectrum of sounds, or even just keep all their stages centre biased to get clipping on both sides of the waveform from both grid current limiting and the other type of clipping (forgot the name of it ).

[vibrovox7reverb]

Really good point...I hadn't thought about having to attenuate the signal back down. I have certainly noticed the gain reduction of cold biased stages before. I am mainly interested in this because I would eventually like to make a highish gain amp, and because it's a good mental exercise for me when dealing with things like reverb, negative feedback loops, and split signals and making sure they come back in phase with the original signal.

[Chuck H]

Your idea is very valid. As noted, cold stages do tend to sound sort of smooth.?. Which seems odd to me considering that the cutoff end of the waveform under clipping is sharper and flatter than the saturated end. BUT... What if you took a signal that was clipped in cutoff and then clipped the other side in cutoff??? (I think this is your Q. right?) This is something I have wondered about too. In fact I plan to proto a preamp based on this principal. My ultimate goal is to get a harder clipped sound, not unlike diode clipping, from tubes. To what end I don't know. But if waveform topography holds true then by cutoff clipping both sides of a waveform you should be able to create a harder clipped sound rather than the smoother sound that seems to result from cutoff clipping only one side. To avoid too much asymetry in the intended waveform I plan to both regulate the B+ to the cold clipped stages and operate them fixed bias. I expect the results to be interesteing at least. If not utterly valid.

The way I see it, the idea is to get the signal level high enough to clip the first stage in the clipping daisy chain to the level you want. Then attenuate the following cold biased triode so that it clips in a similar way.

JM2C

[Amp Kat]

You can set the tube (preamp triode) to be inverting or not so it just depends on how you do it. As for getting a more harmonic content there are several ways to achieve a more even order content and the more symetric the tubes are the less you get. There are ways to clip the signal so it's not so square which really adds a lot of fuzz and odd order into the mix which is always going to be present anyway but the more 4,8 and higher you get the better sounding clip you'll get. unbalanced phase splitters and unmatched power tubes are a plus for even order THD but as far as strickly tied to preamp tubes you may try wiring a 12AX7 as a diode to get a more rounded clipping however one of the problems is the amplitube will modulate so tying that to a +/- reference voltage otherwise known as compliance would help. I've done it and measured the wave on a scope and it works and has a very smooth sound. LEDS and diodes wired back to back clip it symetrically and square it adding more odd than even.

[Gingertube]

Basic theory (maths) will show that symmetrical distortions/clipping etc. generate mostly odd harmonics and asymmetric distortions generate even harmonics. The "cold" stage has widely different positive and negative going signal swing capability and and clipping will be asymmetrical. That means even harmonic distortion and a "smoother" sound.
Cheers,
Ian

[loudthud]

A lot depends on how hard you hit the first stage. Once you clip off a large portion on one side, there is a base line shift to the next stage. This is caused by cap coupling. The cap creates a new baseline where 50% of the area under the wave is above and 50% below. This happends after an initial attack that might favor one polarity. The second stage will clip the other side but because of the base line shift, you wouldn't get a 50-50 square wave unless you hit that first stage really hard. The baseline will creap up to a narrower portion of the signal. A 50-50 square wave sounds hollow, not very good. A 40-60 square wave is more interesting.

Edit: I should add that grid current also causes the waveform to shift down. Grid stoppers and a low value grid leak resistor help moderate this effect.

[Chuck H]

I know about the waveform re centering and all. For a 50/50 clip this should be easy to solve with biasing, no? Anyhow, That "hollow" sound can be a useful sound too. Lot's of pedals do it. Lot's of guys like 'em. But your recognition of asymmetry has me thinking it would be fun to have a variable control for this. Thanks.

[exclamationmark]

There are also tube diodes, which you can use in the same way as solid state ones (ie, back to back), or alternatively wire a triode like one. There's a few threads floating around where people have used them. Another issue concerning asymmetrical clipping is that tube halves are never quite the same, and even if they were, would change over their life time.

[Steve Conner]

Never been that impressed with tube diodes to be honest. To me, the main difference between tube sound and solid-state is the varying duty cycle of the clipping as Loudthud just explained.

The duty cycle actually varies with the envelope of the guitar signal, and (insert math here ) the harmonic spectrum of a clipped wave depends quite strongly on the duty cycle, indeed the effect is used to add interest to analog synth waveforms, where it's called PWM. So you get a complex, swirling overdrive sound instead of the hollow buzz from a diode clipper.

Peavey's Transtube amps make quite a good effort of recreating it.

[Amp Kat]

Never been that impressed with tube diodes to be honest. To me, the main difference between tube sound and solid-state is the varying duty cycle of the clipping as Loudthud just explained.

The duty cycle actually varies with the envelope of the guitar signal, and (insert math here ) the harmonic spectrum of a clipped wave depends quite strongly on the duty cycle, indeed the effect is used to add interest to analog synth waveforms, where it's called PWM. So you get a complex, swirling overdrive sound instead of the hollow buzz from a diode clipper.

Peavey's Transtube amps make quite a good effort of recreating it.

You may be using the wrong type diodes. 6AL5's or comparable or very noisy as opposed to 12AX7's wired as diodes. The other thing as I noted is if you don't use compliance or reference voltages it won't sound very good at all as the signal will modulate. If you can get the signal on the scope good and rounded clipping analogous and not squared like the digital signal clipping from diodes it will sound pretty sweet.

The Peavey Amps are nice sounding for SS especially the Bandit and XXL but they not close to even order THD although they do a nice job with the odd order to get an almost tube sounding amp.

[J M Fahey]

Varying duty cycle all over the place *does* make for more interesting sound.
The (something) Murphy guy (if I remember well) who designed the clipping stages for Carvin SS amps explains it in an interview.
I have it saved somewhere.
The Crate distortion circuit does that too; it has an extra diode and RCR time constant added to change the clipping Op Amp (s) bias depending on signal strength.
Itīs either built under licence from Murphy/Carvin , or just a ripoff , because itīs too similar to be just a coincidence.
Transtubes do the same, but with plain bipolar transistors.
*Maybe* they did this so as to be "different enough" to have a patentable circuit, but the basic idea *is* the same.
Tubes, of course, do it all by themselves all of the time.
Fets, when heavily driven, too.
Interestingly, Fuzz Faces do the same !!!!
Just overload one of them and watch waveforms on a scope.
Itīs funny how Scope Technology, available for decades, is so *little* used by amp designers, even less so by Internet Gurus, who spread all kind of nonsense which can be proved fake in *seconds* just by watching the wiggly lines on the screen.
Oh well.

[Amp Kat]

Yes no doubt about that JM and in reality Marshall and Peavey both use SS diodes wired back to back as clippers in some of their amps but the principle is simulating or trying to simulate what a string of cascaded preamp tubes do naturally without the high tube count and space that it takes up. On the scope it looks the same however Peavey used this same technology in the JSX & XXX and called it a compressor which is what it really is because the diode is bounding or limiting the signal so to speak and in the JSX they even added a pot to each end of the diodes to vary the amount of compression or clipping and called it a variable compressor.

[tedmich]

Yes no doubt about that JM and in reality Marshall and Peavey both use SS diodes wired back to back as clippers in some of their amps but the principle is simulating or trying to simulate what a string of cascaded preamp tubes do naturally without the high tube count and space that it takes up. On the scope it looks the same however Peavey used this same technology in the JSX & XXX and called it a compressor which is what it really is because the diode is bounding or limiting the signal so to speak and in the JSX they even added a pot to each end of the diodes to vary the amount of compression or clipping and called it a variable compressor.

I thought the dual diode coring circuit in the PV Bravo->JSX was accepted as a variable (in the JSX) poor mans noise gate NOT a compressor? Some swear by jumpering it but others say it adds very little to the sound.

[loudthud]

Itīs funny how Scope Technology, available for decades, is so *little* used by amp designers, even less so by Internet Gurus, who spread all kind of nonsense which can be proved fake in *seconds* just by watching the wiggly lines on the screen.
Oh well.

I have spent many long hours playing through an amp and watching an oscilloscope. It helps to dim the lights so each sweep persists long enough to see it.

My boss at a semiconductor company was going on a recruiting trip and ask me what he should look for in a newly graduated engineer. I told him to find someone who owned an oscilloscope. He didn't like that suggestion because he didn't own one. But, he did own a nice boat.