Amps with warm biased gain stages? Clipping theories?

[Mr. Lime]

There are designs like the Fortin Cali mod or the Ceriatone Chupacabra with 820R + 680n on the cathodes of the first two gain stages.
It's quite common to have a hot biased gain stage after a cold clipping one in Marshall amps which are clipping the same signal half as the 10k cold clipping stage. This drives the preamp even more into asymmetrical clipping.
Designs like Laney GH or some other hot rodded Marshall mods from the 80s add a gain stage which basically clips the same signal half again so "good" high gain seem to depend on clipping one half of signal swing extensively while keeping the other half relatively clean. The iconic cathode follower driven hard clips the already distorted half and softens the corners of the waveshape.
10-39k often find places in high gain amps like Soldano, Peavey, Recto and many Marshall knock offs.
In other articles it's been said that the symmetric clipping seems much more fluid and not as harsh. Some say it's preferred for playing solies and lead guitar. 1k5 center biased equals most of headroom and gain. Mesa or Kitty Hawk is known for that.
Many threads have been written about this but I would like to know it there are 4 gain stage amps that have more than one 820R on the cathodes?

Has someone tried something like that before?

2k7 on the first two cathodes is quite typical but 820R is kind of exotic.
I find myself ditching extreme cold clipping gain stages most of the time.

What are your guys theories and experiences?

[Stevem]

The harmonic overtone content that gets added to a tone / note when a tube clips is the main factor in what our ear precives as good and bad sounding distortion!

Do a search for the 1974 paper call Tubes vs Transistors ( is there a audible difference ) and read thru it a few times ( takes about one hour ) and I think it will give you a much better base line on which to work off of!

[Tony Bones]

Guitar Amplifier Overdrive, by Ulrich Neumann and Malcolm Irving explores that topic. I believe Malcolm posts here occasionally.

[roberto]

Diezel goes down to 680 Ohm on most cathodes, I tried as low as 470 Ohms, while nowadays I have developed some soecific stages (and frequencies) with Rps from 33 to 220k and Rks from 470 Ohm to 47k.

Don’t stop at classic configurations, install some fleas on the pcb and experiment.
Even vs Odd Harmonics is an old error that the web spread around.

[Mr. Lime]

The harmonic overtone content that gets added to a tone / note when a tube clips is the main factor in what our ear precives as good and bad sounding distortion!

Do a search for the 1974 paper call Tubes vs Transistors ( is there a audible difference ) and read thru it a few times ( takes about one hour ) and I think it will give you a much better base line on which to work off of!

Thanks, I will look for it!

Guitar Amplifier Overdrive, by Ulrich Neumann and Malcolm Irving explores that topic. I believe Malcolm posts here occasionally.

Nice, this one is new to me. I have read Blencowe several times and I got all TUT books so I'm open for new input!
On the other hand I tried many approaches from the books which turned into fail because theory and practice can very different in my experience and most of the time it's more promising just to tune by ears.

Diezel goes down to 680 Ohm on most cathodes, I tried as low as 470 Ohms, while nowadays I have developed some soecific stages (and frequencies) with Rps from 33 to 220k and Rks from 470 Ohm to 47k.

Don’t stop at classic configurations, install some fleas on the pcb and experiment.
Even vs Odd Harmonics is an old error that the web spread around.

Good to know that, I have to try 680 Ohm as well!
How about LEDs on cathodes?

[roberto]

Leds are ok on clean stages, not ODs.

[Malcolm Irving]

Our book 'Guitar Amplifier Overdrive' does go into a lot of stuff that is relevant to this, in particular bias-shift that occurs during overdrive conditions. Generally, we tried to avoid subjective assessments (one persons 'nice' distortion is another person's 'horrible') and stick to looking at what happens to waveforms and harmonic spectra.

I would agree that trying different component values and listening for the tone you want is a good way forward, but even that is difficult because there are so many variables (grid stopper values, attenuation between stages, coupling caps, cathode bypass caps, etc.). For cathode resistors (which set the bias for simple stages) a suggestion is to build with preset pots, tune to find a good sound and then measure the presets to set the final fixed values.

[roberto]

IMHO waveforms and harmonic spectra is as dangerous as subjectivity, because of the too many odd legends about even harmonics.
Samples, like the ones you have done in the online extension of your book, as one of the best way to compare circuits soundwise.

[Reeltarded]

I think Roberto has at least a little understanding of this topic as visible in his excellent (but broken for me) thread up top in this very section of the forum.

[roberto]

I see now that links are again broken. Thanks for the head-up Reeltarded.
I have to find the original files and upload them to this forum instead of using external links.
It's an old work that needs to be implemented with IMD and other details.

Recording a straight guitar track and process it through different stages (or different stages on the same well known circuit) could help to get a more universal and practical understanding of the topic.

I have to say that simulation is a good thing when you have no time to test directly (I'm often abroad worldwide), or when you want to focus your time ( I like this site: http://www.trioda.com/tools/triode.html ), but then you'll always find something interesting by going out of the standard choices (I came out of one of my favourite lead channels by lowering Rp way below what I expected).

[R.G.]

There's a large body of information about harmonics and their effects. Historically, it was a concern of musicology when scientific research into tone, pitch, and other facets of the musical experience started. It's dated, but Carle Seashore's "The Psychology of Music" ( (c) 1919! ) is a great place to start reading about this. It is probably a little premature to say that "Even vs Odd Harmonics is an old error that the web spread around"; the perceptual differences in harmonic content is a strong thread in the psychology of music.

I'll grant you that the web spreads a lot of ... um, unfounded rumors and errors. And that the issues about the harmonic content of a tone are not as simple as saying even versus odd harmonics, but just like dirt can sometimes conceal a gold nugget or a diamond, not >all< of the net is erroneous cr@p.

But I digress. It's really not useful to focus on just the cathode resistor in a triode stage. The cathode resistor doesn't tell you much until you also specify the grid voltage and source resistance, any cathode bypassing, the plate resistance and the source resistance. If you use a self-biased stage like the vast majority of guitar amps do, your grid will be at zero volts through a high resistance and the cathode resistor will be the major contributor to plate current. Plate >voltage< will then be heavily dependent on the value of the plate resistor. You can vary the output clipping by varying the plate resistor as well as the cathode resistor to move the tube's operating point from near "saturation" (although triodes don't really do that) to near cutoff (and they do cut off hard when driven off). That varies the gain of the stage, and that means the clipping point changes not only with DC bias from the plate and cathode resistors (and any bypasses) but with signal level.

That's just the output. You can clip at the input in typical stages with a bypassed cathode by overdriving the grid with a high impedance source signal. The changeover from non-conducting to conducting when Vgk goes positive can affect only one signal polarity.

Obviously, you can't get rid of even harmonics without symmetrical clipping, and you can't symmetrical clip with input clipping in a triode. Well, maybe you could contrive a very special setup where the amount of clipping on an input/grid of a triode was balanced with hard cutoff on the other polarity of signal at the plate, but I suspect that it would only be at a tiny range of input signals. You can get symmetrical clipping from two tube sections; the easiest is a diffamp set up to do so. But getting free of any even harmonics is going to be tricky.

The most musically useful distortions I've seen in preamps is when several tube stages are cascaded. The output stage will generally ciip first, since it gets the biggest signal. As signals (and gain) get larger, the clipping of earlier stages MAY show through if the distortion of the final gain stage that clips isn't so severe that it covers up earlier distortions. This by the way is a way to shape distortions: use dividers between tube stages to cut some of the signal level and postpone distortion of the last stage until you want it to happen.

I urge you to look into music psychology for the effects of the harmonics of musical tones, and into the how and why of stage clipping at both input and output. it gets complicated.

[roberto]

It's dated, but Carle Seashore's "The Psychology of Music" ( (c) 1919! ) is a great place to start reading about this.

Just bought one copy, thanks for the suggestion.

[roberto]

Mr. Lime,
the following lines quoted from R.G. applies not only to tube amps, but to stomp boxes as well (see how many multi-stage OD came out in the last decade).

The most musically useful distortions I've seen in preamps is when several tube stages are cascaded. The output stage will generally ciip first, since it gets the biggest signal. As signals (and gain) get larger, the clipping of earlier stages MAY show through if the distortion of the final gain stage that clips isn't so severe that it covers up earlier distortions. This by the way is a way to shape distortions: use dividers between tube stages to cut some of the signal level and postpone distortion of the last stage until you want it to happen.

Give a read here too: https://www.ampbooks.com/mobile/classic ... ldano-slo/
and go back to the modmodmod thread on sloclone too, to see how the single gain stages on the different mods are stacked together, how the attenuation between stages is achieved (voltage dividers? low load? both?) and how it's suggested to change it when you change a stage. So, how the different stages are send out of linearity by applying different amplitudes at the very beginning of the amp.

I've discussed alot over there and proposed alot of different solutions compared to the original SLO (but also other designs too):
- slightly higher Rp with lower Rk
- lower Rp with higher Rk
- lower Ck & higher Co vs higher Ck and lower Co
- voltage dividers of high values
- voltage dividers of low values
- low load with no voltage dividers
- split loads
etc...

[Mr. Lime]

Thanks a lot for that input!

The ampbook website has always been a good read to me.
Funny, at the moment I had tried to ditch as many dividers in the signal path and omit the cathode bypass caps as I can to get the most "punch" out of the preamp as suggested in the german TubeTown forum. Have to try things like a direct coupled LTPI on a gain stage which removes even the coupling cap as well.
On the other hand I got nice results with a local NFB stage and I would like to experiment more with these although they were not recommended to me to use instead of cathode followers. Rob Robinette suggests values up to 33M for feedback resistances.

Will take some time to work through the Mods Mods Mods thread but I'm sure it will get me some more insight!

I agree with R.G. that impedance matching is truly a key which I experienced especially in my Silver Jubilee conversion experiments.
My current problem child is a Komet (similar to Trainwreck) circuit which has lots of problem areas for such a simple amp and they are well known for output tube and PI distortion while the choose of components in the preamp seems pretty sensitive!
Generally lots of stuff I read seems to be controversial in practice. For example the Engl Blackmore circuit with high Ck and high Co.

Maybe we could even create some kind of "Rule of thumb" list to better predict what's going to happen if we use several conditions?

[pullshocks]

AX84_Blues_Preamp_Schematic.pdf

You mentioned LED bias. The AX84 Blues Preamp includes a stage with parallel 12AU7 triodes, one biased with red LED and one biased with green. This follows a 7199 pentode input and cathode follower. I have this preamp going into a SE 6V6 power amp. You can search the AX 84 forum on "LED bias" and find a fair amount of discussion on this topic.