Is "Power Tube Distortion = Holy Grail" a myth?

[Andy Le Blanc]

All statements about push pull are idealized and assume "identical"
tubes and voltage conditions. If this is not the case, IE. real world conditions,
you can only realize favorable ratios of odd and even order distortion,
which are measured arithmetically based on the operating condition
of the circuit BEFORE clipping, and actually has little bearing on the behavior
of the circuit when it is over driven.

The best you can get is a favorable spectra of even and odd order harmonics.

This is where loading and the application of feedback becomes critical.
A wreck or say a bassman isn't quite the same with the wrong loading or feedback ratio.

A bit like tuning a wine glass with the right amount of liquid, get it to ring just so.

[Cliff Schecht]

Nice plots Dave! Yes, we aren't operating linearly with our amps but the harmonic spectra that you posted mean just as much.

One comment.. Before full on clipping, a Marshall style preamp will produce predominantly even order harmonics. It's a function of how much you push that second and third stage into distortion. The waveshapes that are achieved at the second stage show that the tube is operating past it's linear limit but isn't hitting its hard limit, hence the nice 3/2 law style soft clipping/waveshaping that sound great (and if designed right is usually pretty dynamic in nature, i.e. easy to go from clean to distortion). At full on preamp clipping, the predominantly even harmonics disappear and you're left with a whole lotta odd (the signal is completely squared before it hits the TS/PI/PA). The tonestack will alleviate the harshness of a completely square signal in a Marshall style circuit though.

I love the Express method of shoving the tonestack and volume after the first stage. You can still limit frequency response nicely but this is part of the reason the Express stays so dynamic. Setting the frequency response earlier will dominate the response of the rest of the system while the second and third stage generate all of your extra harmonics content. Once you tune in the TS and volume "correctly", you setup the amp to go from that nice clean (although an FFT will show a good amount of even harmonics still, hence the great chimeyness!) to a fully clipped wave that sounds...well you guys know how the Express sounds.

[dave g]

Here's another, this time a non-HRM Dumble.

I'm currently going to school in California, but my workshop, tools, amps, etc are all in Illinois. Once I get back for winter break I'll need to make some more plots; the ones I have on my laptop that I've been posting are all a few years old.

I agree that these types of plots are useful for getting an idea of the "character" of the distortion produced. Hit a single note at a certain amplitude and these plots show approximately the relative weights of the harmonics you're likely to see.

What they don't show is how the distortion changes with respect to the amplitude of the input waveform, how it changes when multiple notes are played together, etc...the fact is that the types of nonlinearities we're discussing here don't care about "frequency"; all they care about is signal amplitude with respect to time. It's just how the math works out that clipping a sine wave introduces integer multiples of that frequency. There are plenty of other types of non-harmonic distortion at work too.

Lately I've been interested in the early 70's Orange designs. Very Wreck-ish preamp, but with a split load PI rather than a LTP. The standard Wreck circuit has too much gain for my tastes, and I've been thinking about building myself an OR-120 style circuit with a fixed biased 2x6V6 output section and using it as a starting point to see if I can come up with something I really like.

[Cliff Schecht]

Once I get my HP 8903B setup through GPIB I plan on taking plots of a lot of the stuff I've built. But this is after I figure out how to write the requisite software in Labview.

[vibratoking]

Cliff,

There are VIs for that unit available on the web. Most of the LV code should already be written. I have used LV for years if you ever have any questions.

[Cliff Schecht]

Sweet! I'll look into that and see if I can figure it out. I know Pete Millett uses an 8903B and wrote some code to do FFT's and such for his but he warned me about how antiquated his software is.

[vibratoking]

The VIs of this type usually give you what you need to control and interface with the device. This is usually the hardest part of LV. After that, it is pretty easy. At least once you have come up the small learning curve.

[Ian444]

Hi, people, this is my first post here, I think... I have been pondering output stage distortion all week since getting my first simple guitar amp build up and running. I did some searching and stumbled across this thread. From looking at the scope, it is the split load PI that actually distorts first as the grid voltage of the ouputs approaches zero (this is a fixed bias amp), as was mentioned earlier in this thread a couple of times.

To my newby brain it sure looks like the output stage is simply amplifying a distorted signal and is not generating any distortion within itself. Does that make sense? Sure sounds good, very smooth transition from clean to crunch and beyond.

[daydreamer]

Hi Ian444,
Welcome mate!
what amp circuit did you build?
Andy

[selloutrr]

I'm sure it's more common then you might think.
While recording it's very common to run the amps wide open with the head in the control room and the speaker cabinets in another room. Then dial back to get the desired tone.

The Fender Twin Reverb remains the most tracked guitar amp or all time. If you want to hear what one sounds like wide open all you have to do is listen to Nirvana's Nevermind.

Tube output distortion / compression is a huge part of "the sweet spot" it's directly related to the volume relationship guitarist have with an amp.
"I can't turn it down that's my tone."

It's not always direct balls to the wall that makes the magic in the output tubes. somethings just turning the master volume up all the way so the outputs aren't starved and then using the gain to excite the preamps is even more zen.

[Ian444]

G'day Andy. It's a bitsa, bits and pieces, Chinese iron from hi-fi's, the PI is based on Merlin's ideas. It's a bit of a mess inside from tweaking, need to buy the final parts and tidy it up. Now it was supposed to be a clean amp for acoustic guitar for my niece, which it is, but after the volume goes past half way it grunges up very nicely and makes nice blues/rock tones, of course, I'm totally biased too schematic and a pic Still need to build a cabinet for it.

[TheGimp]

Hey Ian.

Small world.

Iron from the Meng Yue?

What kind of pick up will be used?

[Reeltarded]

The amps I like best have a balance thing going on where I run the master in a sweet spot, and turn the gain up until it squeels if you aren't tap dancing and hold your breath. There is a sweet spot there. Beyond it sounds like a toilet full of hammers. I like Marshall stuff though, mostly without masters?

Yeah!

[renshen1957]

Yes. It is Merlin's book that has brought the question to mind - specifically, the "Feedback Theory" chapter. His description of how distortion behaves with feedback is exactly the reason I ask the question.

I'm completely willing to risk being thrown off the island.

Of course, it could be that "power tube distortion" belongs on the same island as "Vox AC-30 Class A"...

Hi,

The Amps by Jim Kelley were designed on the idea of Power Amp distortion as opposed to pre-amp distortion (Dumbles and most High Gain amps).

Dave Funk makes a brief reference as how the Trainwreck distorts (PA then PI then Preamp) in his book on Tube amps volume 1 Fender amps.
(the only volume he published)

Vox AC-30 are Class A up to circa 22 1/2 Watts and then they slip into class AB. The preamp section of an AC-30 is always class A (

Best Regards,

Steve

PS You might ask the question about power distortion over at powerscaling.com (you have to join the forum, though). By controlling the Voltage in the PA only, one can distort the Power Amp at any volume down to a fraction of a Watt output.

[Reeltarded]

My favorite Marshall 50 was found to have some clipping diode network thingy added. I have heard many artful uses of that since then. I can like it.