This is my best attempt at a "vacuum tube" style overdrive.
(Edited)
Look below for the latest version. This relies on Zexy ( for simple single sample delay) and Else. The only thing i could not easily build in vanilla was the Hysteresis emulation, which relies on lag2~ and glide2~.

This older version is quite different and messy. Mostly experimentation.
ThurboDrive~.pd ThurboDriveExample.mp3

@Booberg Can you list the dependencies?

... couldn't create
tanh~
... couldn't create
z~
... couldn't create
range~
... couldn't create
dA
... couldn't create
number~
... couldn't create
array: no method for 'saved'
number~
... couldn't create
array: no method for 'saved'
brickwall~ 18500
... couldn't create
brown~
... couldn't create
bandstop~ 1050 0.3
... couldn't create
lag~ 384 768
... couldn't create
>~ 0.5
... couldn't create
slide~ 10 -0.5
... couldn't create```

I don't really keep track of that im affraid.. usually keep to else, cyclone and maybe brickwall is from audiolab? If i wasn't lazy i think all those tanh~ objects are in there for writing tables for the transfer functions.. im not so sure how precise or heavy they are to run so i usually stay clear of them.. it's very possible to change those for a mathematical equivalent/approximation. And the brickwall~ is just for anti aliasing when downsampling so you might use other filters for that.

I actually continued working on this for some time, specifically for guitar. I didn't really reach all the way i was hoping to, i found that it's hard to induce "shimmer" digitally... alas, i might upload the results during the week.
Ps. A big facepalm moment when i realized that my 'pseudo-integration' method is really just a naive lopass filter with a bunch of gain :,) so that saves a lot of cpu for V2 atleast..

Cheers

Oh now i see "dA" and some "number~"s in there.. that stuff shouldnt be there so i might have beefed it and uploaded the wrong file or something. I'll take a look at it tomorrow.

@Booberg :+1 no rush, thanks for sharing!

I was in a rush last night.. I removed some junk, fixed some filter phasing and corrected a mistake i made with the Hysteresis subpatch. Sounds pretty tight now IMHO.
After a quick rundown i think its just dependant on two externals, Zexy and Else.
z~ , one sample delay, from Zexy. And lag2~ + glide2~ from Else.

I hope someone finds this entertaining I recommend using a cabinet convolution with this. I've used partconv from AudioLab (loaded with an ownhammer alnico silver IR) on the recording of the attached sound sample.

ThurboDriveGuitarV2.mp3 ThurboDriveGuitar2.2~.pd