Help: simulating Ge leakage current

[mac]

I'd like to improve a little helper application I wrote to calculate the bias of a Fuzz Face, a Big Muff stage & a voltage divider stage.
I want to simulate the effect of germanium leakage current to make things more realistic and accurate.

Following RG http://www.geofex.com/Article_Folders/fuzzface/fffram.htm
in a leaky germanium transistor, the collector current ic seems to be:
ic = hfe*ib + ileak
where hfe is the real or corrected gain, ib the base current and ileak the leakage current.
What about ie, the emitter current? Is ie = (hfe + 1)*ic/hfe = (hfe + 1)*ib + (1 + 1/hfe)*ileak,
or ie = (hfe + 1)*ib +ileak ?
Am I correct?  Would you use another method?

Suggestions are welcome.

mac

http://www.geocities.com/guitarfxs

[John Lyons]

Bringing this back from the dead.
I'm curious how this can be done.
Have you come to any conclusion mac?

[Steben]

Reminds me RG mentions silicon JH-x fuzz pedals sounds horrible (IHHO). I have a JH-1 which is very musical sounding with BC108... Trick is to have fuzz set at 3/4 to 4/5 and the guitar volume does the rest. Might be a version issue (my pedal is post '98 with the trimpots)?

On topic: wasnt piggy bagging the solution to make silicon be more germanium?
https://www.diystompboxes.com/smfforum/index.php?topic=21801.0

[rutabaga bob]

https://i.postimg.cc/KYV1v66z/2222-fuzz-schematic.png

I came across an old schematic the other day, but I can't seem to pull it up on any search at the moment.  It is a 2N2222 Fuzz Face build with tweaks that the designer says will get more of a germanium sound.  Done by a Bill Petris in December 2010.

EDIT: Found it!

[teemuk]

On topic: wasnt piggy bagging the solution to make silicon be more germanium?

Depends in what regard. Simpler than "piggypacking" is just to connect a diode and series resistor in parallel to base-emitter diode. (For reference look for Vox VBM-1 or VR series amp circuits).
The parallel diode diverts some of the base current and what you get as result is - much lower gain - in the device. The lowerered gain, additionally, also reduces local feedback, which turns transition to clipping distortion softer. So yes, essentially you get behaviour similar to - low gain - germanium devices.
In other regards the device still functions like an ordinary silicon transistor.

[Rob Strand]

The problem with using silicon transistor models is the Vbe is totally wrong for a germanium.
If you adjust the Ics and Ies transistor parameters to get the correct VBE for a germanium (and VBC when BC forward biased),
you will find the leakage goes way up.  The voltage drop and the leakage are related.

https://electronicsarea.com/ebers-moll-model-bipolar-transistor/

If that's not enough then you would need to add a 'fake'  leakage current along the lines of mac's equations.

[Technically the band-gap voltage in the silicon model is wrong for germanium.   To some degree, you can tweak Ics and Ies without changing the band-gap voltage as well but it's a fudge.    IIRC, the reason you can get away with it is because the earlier version of spice normalizes some parameters.   If you use the correct band-gap voltage it should give correct temperature dependencies of Vbe etc.]

FWIW, this paper deals with the IES, ICS thing but side-steps the band-gap issue,
https://dafx17.eca.ed.ac.uk/papers/DAFx17_paper_28.pdf