[sdiy] matching discreete transistors can be so easy ...

[John L Marshall]

John's technique for matching diodes. I suppose that it could work for
transistors too: Connect all of the diodes in series, one long string. I
solder them, just a quick tack. Add a current limiting resistors in series
and connect to a sufficiently high power source. Let the string stablize
then measure the voltage drop one at a time. I record the reading on paper
directly below each diode.


Take care,
John
---------------------------------------------------------------------------
Pacific Northwest DIY Synthesizer meeting, July 20, 2002
See: www.sound-photo.com
---------------------------------------------------------------------------

----- Original Message -----
From: <jhaible at debitel.net>
To: <synth-diy at dropmix.xs4all.nl>
Cc: <jhaible at debitel.net>
Sent: Wednesday, August 21, 2002 7:11 AM
Subject: [sdiy] matching discreete transistors can be so easy ...


> ... with one of these solderless breadboards.
>
> Last time I selected transistor pairs for low offset voltage, it was a
nightmare.
> When I touched a transistor with my hands for a few seconds, it would take
> 2 minutes ofterwards to get a valid Vbe reading. I remember an evening
with
> a tiny hourglass and styrofoam, last time I did it. Gloves didn't help
much, and
> My hands are too clumsy for tweezers. (Especially when the transistor
leads
> don't fit into the test socket easily.)
>
> Now I don't have to select pairs - I have to select "11-tuples". The
PS-3200 clone
> needs groups of 11 transitors that are tightly matched. Time to try the
breadboard
> idea.
>
> I've built the well-known Moog test circuit (basically a 100uA current
source), but
> instead of using a transitor test socket, I connected three flexible wires
and built
> a little probe with 3 contacts that can be plugged into the breadboard
parallel
> to a transistor.
>
> Now matching is much more fun: Plug 30 transistors into a small
breadboard,
> then wait 2 minutes for the _whole batch_, and probe them one by one by
> inserting the 3 contact-probe next to a transitor. No chnge in temperature
> from this probing, and the whole procedure is really _fast_. I didn't even
> have to draw a big table and sort the transistors. Of the 30 in the board,
> 8 were close enough, so these stay in the board, and the others are
replaced
> from the drawer. Anther 2 minutes of stabilizing, and then it was fine.
>
> I didn't know selecting transistors could be so much fun. I have a *lot*
> of such groups of 11 yet to select, and I don't know how I could do it
otherwise.
>
> JH.