[sdiy] matched transistors

[Terje Winther]

At 11:14 +0000 06.02.2002, Brown David wrote:
>There are several issues here.
>1.  Obviously an integrated transistor pair such as the MAT02 will have
>better thermal tracking than two separate transistors glued together.

Yes, but more expensive, and sometime more difficult to get hold of.

>2.  I'm not clear about which parameters are the most important for a
>matched pair.  I had assumed a good Hfe (gain) match was all that matters.
>However other contributors to the sdiy list has suggested matching Vbe to
>within a millivolt or so is important.  Anyway, you test procedure sounds
>OK.

Thank you.

>3.  Where did you obtain the Format Pro book from?  Is it in English or
>German?  I built (and am still completing) the original Formant.

>From Helm-Elektronik in Germany. It is in German only (ouch!) AND
BEVARE!!!!! it is FULL of errors. You really have to know what you are
doing, othervise it will be a desaster. There are several places on the
´net that have reported the errors.

I am building the Formant Pro (together with modules from Oakley,
digisound, doepfer and a moog clone) together with a friend of mine who is
very skilled in electronics. Even so, there are several times when we had
to seek help from REAL professionals to solve some of the errors in the
books. I don´t have the URL right here, but I believe it is:
http://www.helm-elektronik.de


>-----Original Message-----
>From: Terje Winther [mailto:terje.winther at notam02.no]
>Sent: 04 February 2002 20:56
>To: Brown David
>Subject: RE: [sdiy] matched transistors
>
>
>>For VCOs, the matching needs to be very close.  For example, the gains of
>>MAT02 pair are advertised to be within 2%.
>
>So far i have only used pair that are within 1% apart, but it takes forever
>to measure a batch. If I could go to 2%, it would significantly reduce
>preparation time.
>
>>By the way, how are you measuring the transistor gains?
>
>I really don´t know; I have a measuring instrument someone else have made;
>I just insert them properly, let them be in there for quite a while (many
>minutes each), note down the reported number and move on to the next.
>
>I am planning to build a gain tester for some transistors needed in the
>Formant Pro VCOs, based on what is detailed in the book.


tw.  :)