Ge leakage and negative feedback question

[mac]

Can Ge leakage current be considered as negative feedback? 

After all it is current going from the C to the B, like an imaginary resistor between C and B.
And it is exactly the model ic = hfe*ib + leakage used by RG to measure germs.

For example, consider these two cases:
let's say I have two germs, g1: (100; 0.0ma) and g2: (100; 0.1ma), wired as follows: 10k from Vcc to C, emiter to gnd, and a resistor Rf from C to B.
g1 - Vc: 2.6v, ic: 0.6335ma, ib :0.006335ma if Rf 395k (Vbe: 0.1v)
g2 - Vc: 2.6v, ic: 0.6335ma, ib :0.005335ma if Rf 470k

By the book, g2 voltage gain is bigger, so the question is, the 75k difference due to leakage must be taken into account?

mac

[PRR]

> like an imaginary resistor between C and B

Classic pure leakage is more like a current-source than a resistor. Won't give much NFB.

Real-factory production flaws "could" be more resistor-like.

If you know what to do with NFB, try it, and see if the results agree with NFB theory.

[mac]

Thanks Paul,

Classic pure leakage is more like a current-source than a resistor. Won't give much NFB.

So it is like a pipe for electrons isolated from the signal being generated at the collector...

Real-factory production flaws "could" be more resistor-like. [/quote]

... unless the N-P junction has some internal resistance due to doping or unwanted impurities?

I have to take a look at the various transistors models 

If you know what to do with NFB, try it, and see if the results agree with NFB theory.

Mh... it is hard to do this with just a cheap chinese DMM
But I have lots of jap germs. Maybe I can find some having the same gain but very different leakage, and trust my eroded ears... not very academical 

mac

[PRR]

> I have to take a look at the various transistors models

You have to be an old-school geek to find models of Ge leakage. Serious work came with Silicon, all serious SPICE models are for Silicon.

There _is_ a detailed discussion of Ge leaking in GE Transistor Manual around the 7th edition. You can get beat reader's copies for like 3 bucks. Worth having, but spotty.

> it is hard to do this with just a cheap chinese DMM

Any working DMM is more accurate and reliable than the common bench-meters of the Germanium Era. (Be sure the meter's battery is good.)