Bias and hfe in a fuzzface

[brett]

Hi.  I've been wondering if the hfe we measure in a leakage test rig or in a DMM is close to what we get in a FF. 
In particular, Q1 is starved of base and collector current, and like many other devices, the hfe is a lot less when the collector current is low (the datasheet shows the hfe for 150mA for the device I'm considering using, whereas the Ic in a fuzzface is more like 150uA - a thousand times less)
To simplify matters, I started simple, with a leak-free Si transistor that Id used years ago in Si fuzzfaces. I thought they sounded ok. 
It's a BD139, which on my DMM measured hfe of 140 - too high to be much use.
So I set it up like Q1 in a fuzzface.  Usual supply (8V), collector voltage (6.6V, and base bias to achieve rhat (Vbe = 0.52V).
I measured just 1.1 uA of base current and 55 uA of collector current for a hfe of just 48.
For a device that measured 140 on my DMM.  That's much lower than I expected (I'd have guessed half of the DMM value, not one third).

Moral of the story is that hfe doesn't mean very much for low Ic, unless the Ic is quoted with it.

[antonis]

IMHO, h_FE should matter only for BJT saturation (e.g. ON-OFF switch) or current sourcing (e.g. LED driver) configurations..
For small signal amplification purpose, v_be (Base-Emitter AC voltage swing) is what matters..
(in the mean of BJTs actually are voltage operated devices and Base current is the result - and not the cause - of the Base - Emitter voltage, according to semiconductor physics and Ebers-Moll equation..)

[brett]

Yes, I should have written Hfe, not hfe, as I measured, and we all measure DC gain.
I don't know much about electronics and like to take shortcuts.
In the fuzzface, Hfe and instantaneous hfe are very different because Q1 operates close to cutoff, and hfe falls to zero (at cutoff) and rises to far higher than Hfe.  But nobody measures it, so we chat about Hfe.  And realistically, it's going to have some association with what we want to hear from the pedal (which is probably a nice amount of cutoff and variation in hfe).
As I said, these are just my guesses and rules-of-thumb.
Now tnat I've thought about it a little longer, I realise that for Hfe quoted at far higher Ic (and Ib and Vbe) the less we should be worried that a high Hfe indicates too much gain for a fuzzface Q1.  Mid power devices like the BD139 with Hfe of 120 to 200 at 100mA are likely to have the same gain as small signal devices with Hfe of 120 at 0.1mA. 
Back in the day (2005?) a few people gave favourable reports of such devices in Si fuzzfaces.  From memory, Gus praised them, and suggested their high capacitance might have a role. 
The same principles will apply in Ge fuzzfaces. This also explains why low-leakage Ge BJTs, when tested in a DMM and give a good Hfe (such as the magic 70 and 120) produce very "tame" fuzzfaces.  Their Hfe, measured in RG Keen's more realistic rig at Ib = 4 uA, might be 10, 30, or 50 less than the DMM value.