Testing Germanium Transistors For Leakage and Gain

Now that I have become Transistor Boy, I am learning all about them.  Well, let me restate that.  I am learning some basics about them, and what I am learning is pretty much 30-40 year old technology.

Anyway, what I do know is that germanium transistors have a reputation for good tone in fuzz/distortion devices.  But we also know that germanium was a short-lived semiconductor material in transistors after high-purity silicon was developed.  Germanium is still used in electronics applications (night vision applications for one), but not in transistors.

And it turns out that virtually all germanium transistors are electronically 'leaky' over time, and the majority are no longer usable.  However, we can cook up a device to test germaniums and figure out which ones will still work in our retro-fuzz.

There is an excellent article, The Technology of the Fuzz Face, by R.G. Keen, that inspired me to try out a germanium fuzz build.  In the article, R.G. includes the schematic for a testing rig, and lays out a formula to calculate leakage and actual gain of germanium transistors.

Armed with the article, and brushing off my algebra knowledge, I cooked up a testing rig.

I won't go into all of the details here, since the article explains it very clearly.  One of the key pieces in the circuit are two resistors that need to have precise values in order to have accurate results.  One of them is a
2.2 Meg.  I measured a handful of my 2.2 M resistors and found one at 2.001 M.  So far so good.

The other resistor needs to be 2.472 k.  I thought I might find a close one out of my stash of 2.4 k 10% resistors.

I was wrong.  All of mine measured a bit less than 2.4 k; the closest was 2.397.  So I found a 68 ohm resistor that measured 68.1 ohms, and put it in series with the "2.4" k resistor.  Poof!  I now have a 2.465 k resistor...within .3% of the value I need.  I think that will work.

Update 11 March 2014:  A visitor wrote to correct my language above.  I previously had said "meg" instead of "k" above for the second resistor.  This is corrected.  I'm also of the old school so I tend to say things like "2.465 K" (even my meter says this).  The modern notation is "2k465."  

Then it's just a matter of wiring it up following R.G.'s drawing.  I put a socket in the circuit so I can just plug in a transistor, test it, note the results, and move on.  (And hopefully find enough out of the 100 or so I have to make some fuzzes).

I still need to make the arrangement a little more permanent - I may just tack the thing down to a piece of board.

Here's how it works:  put the DMM across the 2.472 collector resistor.  Put a transistor in the socket.  Read the voltage and calculate the leakage.

Then, throw the switch, and apply 9 volts to the circuit and again measure the voltage.  Subtract the leakage from the second reading, then multiply the result by 100, which gives the gain (within a couple percent) of the transistor.

I'm looking for two transistors with a relatively low leakage (hopefully 100 uA or less) and one with a gain of about 70 and one about 100 or so.  I have a bag of 100 Tungsram AC128s.  Surely I will find a couple dozen?