Newbie trying to design a tube overdrive

[exabrial]

Hey guys, I had a lot of success designing my preamp: https://123d.circuits.io/circuits/852190-jan-6418-vaccum-tube-valve-preamp . After hearing Effectrode's Tube Drive (http://www.effectrode.com/tube-drive/) I want to try designing something similar! Why? For the hell of it haha. I really want to learn this stuff.

The goal would to be having several stages cascaded and gently distortion the sound across 6-8 stages, rather than all at once. And possibly go off the beaten path a little bit and use JAN tubes that can survive an impact into a mountain. Why? Because I can't have nice things haha.

I'm looking at http://elektrotanya.com/PREVIEWS/63463243/23432455/egyeb/english_muffin_ect15_sch.pdf_1.png for a little inspiration. When designing a cascaded input stage for distortion/overdrive, what are some of the key takeaways I should think about?

[GibsonGM]

Hi dude,

Some good thoughts - maybe get started with ONE triode stage?  Mess around with B+ voltages, AND see how plate and cathode resistors, and bypassing, affects the tone.  Honestly ,while you CAN use that many stages to get smooth OD without real distortion....why?   That is a LOT of stages, man!!   And even if you want HARD distortion - you can do that with 4 stages...1 or 2 more for bells & whistles (tone stack...).  Find the "GTFO" posts...

You can do nice OD with ONE stage...or use 2 as the triode typically is a duo-triode like 12A*7....you could run in cascode or parallel...LOTS of neat things to try with just one tube!!!   

I recommend you grab Merlin's book (see your other post, I linked you there) and read....we can't really do more than copy existing designs unless we know how to make each stage, WHY we are making it that way...doing load lines to see what is happening, how much distortion each stage is creating and in what manner, how supply voltage affects headroom etc....great thoughts on the subject, just that it might be wise to start simple and work into bigger things!   The bonus is that Merlin's book covers GUITAR applications, not the old-time generic or hifi stuff...so you will learn those "takeaways" you are interested in!

That ONE little stage can generate some really nice sounds, man!!  And what you do to it is easily extended after that, to your 4, 6, or 8 stages...     

[PRR]

Peavey(?) had a 8-stage tube distorter. Sorry I don't have a link.

Generally when you beat a signal this severely, you want each stage to have a gain about 2. So if the tube is gain of 50, you want 25:1 attenuation between stages.

When you beat-up a signal 8 times, the intermodulation distortion gets crazy. First symptom is that single notes are tolerable but chords are intolerable. The usual approach is to limit the top frequency severely, then slope-away the lower freqs. Distort, then use a complementing filter to slope-off the highs and bring bass back into balance.

Grid R-C time constants are critical for interesting pluck-transient effects. They probably should not be all the same.

Just tinkering 4 parts in each of 8 stages gets 65 thousand things to "optimize". This is not a job for a putzer like me.

[Transmogrifox]

Sounds like fun.  I agree with GibsonGM -- start with 1 triode.  Figure out how it works with bias, and play with bypass capacitor, etd.

Since you won't get much (if any) distortion out of a single triode stage at high voltage (I'm assuming you're doing HV plate)...
Then add a second triode stage.

Of course, building something like the English Muffin exactly as you see is a good starting point...tinkering thing.  GTFO as also mentioned is another good example of several stages cascaded to create incremental distortion at each stage.

[teemuk]

Peavey(?) had a 8-stage tube distorter.

5150.   

Be it mentioned that at least two of those eight were a cathode follower for buffered FX loop and a gain stage for signal recovery from the FX loop. So in my book those don't really count. 

So Carvin's Quad-X easily tops that with eleven tube gain stages, all reserved for distortion and nothing but distortion. Yes, I know, those preamps were incrediably feature-packed... all THAT was solid-state.


Here is Quote from Quad-X manual that might shed more light to such seemingly crazy design:

6.2 The Tube Cascade™

At the outset of the Quad X-Amp project we knew we wanted to do something new with tubes.
We surveyed every significant tube guitar amp ever made and came to a surprising conclusion.
They were all basically the same. Yes, they had their minor differences in features and
packaging but by and large everybody had just copied the early Fender circuits. Most
importantly, there was little new design work being performed on the actual overdrive sound.
The overdrive sound of most amps was not much more than an accident. We were not content to
simply place another Fender clone in pretty packaging along with some extra features.

We searched for the essence of the tube sound. Our earlier work had disproved the old truism
that says "tubes produce even order harmonics and transistors produce odd order harmonics".
That is a grossly simplistic view of the complex waveforms that can be produced by any
overdriven circuit. The mathematical method of Fourier analysis readily predicts the harmonic
structure of any waveform based on the shape of the wave. Using the method of Fourier
analysis it can be shown that tubes can be used to produce either odd or even harmonics or
combinations of the two, just as solid state circuits can. There is more to the tube overdrive
phenomenon than that often repeated claim of odds versus evens.

The Dance of the Harmonics!

Besides the ear, our primary tool for studying overdrive effects was a computer based spectrum
analyzer. As we studied the tube overdrive spectrum closely we realized that we were seeing not
only a mix of odd and even harmonics, but that the amplitudes of each of these two families of
harmonics were alternately rising and falling with the attack and decay of each note. In addition,
as each harmonic family would rise and fall a series of notches would sweep up and down
through that family of harmonics. Not only was the harmonic structure rich in content but there
was a lot of activity among the harmonics. In comparison, the harmonics from solid state
circuits would just rise and decay with each note. These simple circuits had none of the
harmonic agility of the tube circuits. There can be no doubt that this "harmonic dance" accounts
for most of the audible differences between otherwise similar tube and solid state circuits.

Using the methods of Fourier analysis we then related the harmonic activity of the tube circuits
to the modulation of the waveform that resulted from the coupling of at least two tube stages.
Typical amps always used at least two coupled stages. It seemed that if the sound of 2 or 3
coupled stages was good, maybe a system with more complex coupling could produce a more
complex harmonic dance and sound even more rich. More coupling means more tube stages.
We tried seven tube stages and saw a more complex modulation and heard more depth. Nine
stages gave more depth, and eleven stages even more. Beyond this point the extremely high
gain of the system became a problem, but we determined that eleven stages would be
manageable. The resulting overdrive sound has a depth and a degree of complexity never
achieved before.

With eleven tube stages operated in cascade there were quite a few variables to work with and it
took many weeks to finally arrive at the configuration that we provide in the Quad X-Amp. The
name for our new circuit was fairly obvious: the Tube Cascade™.

[PRR]

> Carvin's Quad-X

That may be what I was misremembering. Thanks.

Good blurb too. Self-serving, but gives a glimpse into the process.

So *two* possible plans to plagiarize from.

It is interesting that there are very-very few of these mega-munchers. With today's prices for Chinese tubes and PCBs, a flock of tubes does not add that much cost to the retail price of a high-end premium amplifier. A cynic might suppose that the difference in sound isn't worth its weight in beer.

[tca]

... Our earlier work had disproved the old truism
that says "tubes produce even order harmonics and transistors produce odd order harmonics".
That is a grossly simplistic view of the complex waveforms...

(...)

As we studied the tube overdrive spectrum closely we realized that we were seeing not
only a mix of odd and even harmonics, but that the amplitudes of each of these two families of
harmonics were alternately rising and falling with the attack and decay of each note. In addition,
as each harmonic family would rise and fall a series of notches would sweep up and down
through that family of harmonics. Not only was the harmonic structure rich in content but there
was a lot of activity among the harmonics. In comparison, the harmonics from solid state
circuits would just rise and decay with each note. These simple circuits had none of the
harmonic agility of the tube circuits. There can be no doubt that this "harmonic dance" accounts
for most of the audible differences between otherwise similar tube and solid state circuits.

That's really a funny quote from the Quad-X manual!

They kill the first myth

"tubes produce even order harmonics and transistors produce odd order harmonics".

and promote the other...

"harmonic dance" accounts for most of the audible differences between otherwise similar tube and solid state circuits."

That's really a new truism! Funny!

Cheers.

P.S.
The difference between vacuum tubes and SS devices is "harmonic agility" and "harmonic dance"!

P.P.S.
Quad-X manual talks about Fourier series... do they know Taylor series. I might suggest that agility and dance come from a polynomial expansion .

[tca]

Now getting back on topic.

When designing a cascaded input stage for distortion/overdrive, what are some of the key takeaways I should think about?

Scale and proportion.

[GibsonGM]

Gotta add...warmth vs. distortion....how MUCH of each you set up each stage to have.  Grid current limiting vs. cutoff (assymetrical clipping thru bias point selection)

How much bass? (too much = blocking distortion, poo poo sound. Not enough = shrillness) 

Where do you want to do things with EQ, and what are the pros & cons?

There are enough of those questions to last a lifetime

[amptramp]

If you are doing one triode stage as has been suggested, instead of a 12AX7, try using a 6AV6 dual diode - triode.  The triode has identical characteristics to one half of a 12AX7 but since a 6AV6 is mainly a radio tube, you will probably not get the stratospheric pricing associated with audiophile tubes like the 12AX7.

As an added bonus, the diode sections can be used to implement a clipper so you can get fuzz / distortion effects if you need them.

[exabrial]

I will definitely start with a single stage and work up and till it's "as smooth as butternut squash pie" - one of my band members. Thanks for all the advice, and guys thanks for being positive and patient with my questions, this a great community. Thank you!

[J0K3RX]

You might want to start here... It's very simple and you can use the basic project outline for just about any preamp you choose.  You can also use transformer/s if you would rather go that route..   

http://www.diystompboxes.com/smfforum/index.php?topic=93936.0

I recently made a (Fortin) Randall Diavlo variant using basically this same design. 

[teemuk]

It is interesting that there are very-very few of these mega-munchers.

IMO, not so really. If you examine about every traditional "hi-gain" tube preamp you will note that the overdrive in them takes place in maybe just two or three gain stages max., all other are just for amplification to achieve that gain, or amplification for recovery from losses of all kinds of filters or to simply meet some required signal levels.

Hardly any traditional guitar amp "spreads the distortion evenly" across its entire preamp, so to speak. Most of them clip in very few gain stages, which are also typically configured to boost clipping asymmetry and therefore cause all those aforementioned DC offset shifts and dynamic changes in harmonic patterns of distortions.

All this simply doesn't need gigantic amount of gain stages. Maybe one, two or perhaps three. All other designs are clearly and intentionally "going over the top" from that principle. e.g. Quad-X.

A cynic might suppose that the difference in sound isn't worth its weight in beer.

Yes, and he might probably be right too. IMO, Quad-X sounds pretty much like every other high-gain amp, although one of its channels probably distorts about 6 - 8 more tube gain stages than a traditional high gain tube amp.

I'm pretty sure the effect on "feel" created by this additional harmonic complexity probably has equally diminishing returns too: Because of associated DC offset level shifts, and their effect on clipping's overall symmetry/asymmetry) two asymmetrically clipping and capacitively coupled stages are way more "touch sensitive" concerning distortion characteristics than, say, a basic shunt diode clipper in a single stage. That's the thing about "harmonic dance" and such... Will adding more coupled gain stages radically improve performance from that I don't know because I never experimented. Carvin's sales buff naturally claims so, personally I have doubts.

[tubegeek]

One distortion mechanism that is analyzed VERY carefully in Merlin Blencowe's HIGHLY recommended book (and not just because he's a good bloke who is likely to read this post - Hi Merlin!) is the compressive effect of grid current in a cathode follower stage after a gain stage.

The classic example is the Tweed Fender Bassman and that is exactly what Merlin looks at in his book. (Merlin Blencowe is user "merlinb" here at DIYStompboxes.)

Cutting to the chase, I am partway through a project which attempts to replicate this effect using subminiature tubes. I had it wired up, and then I broke a tube, and the power supply I was looking to use turned out to be too full of unsuitable junkbox parts to do the job, so I stalled on it. My Plan B is to use better parts - but maybe a SMPS power supply would be a better idea for a small pedal-sized gizmo. (I'm not looking to run this at low voltages, it's meant to be a proper tube circuit.)

Anyway, all my design notes are in a thread here and I hope they would make for interesting reading:

Designing a two-stage tube circuit with compressive cathode follower ala merlinb