[sdiy] Humans are not Becoming Smaller

[Harry Bissell Jr]

Another issue is the matching of the transistors
is very important. In a monolytic process you get this
almost for free. In discretes it is very hard to both
match AND keep the temperature similar.

I tried a 'spice' model of a CA3080 from the
datasheet.

Using transistors and diodes... it did not work at
all. Offset was terrible.  Then I figured out that if
I diode connected transistors, it would work (and it
did).

eventually we will have to move on and use other
technologies to get the results we want. Maybe after
my lifetime (I hope).

There is (of course)... always the desert island :^P

H^) harry

--- Ian Fritz <ijfritz at earthlink.net> wrote:

> At 12:47 PM 6/2/05, laxt57 at aol.com wrote:
> 
> >   Well what I am planning is this. A OTA made with
> SMD that
> >  will fit the footprint of a ca3080. By input I
> guess I was not
> >  asking about the design of the OTA itself, all
> that is freely
> >  available on the web. I guess my question is with
> the available
> >  smd devices 847s,3904 etc do you think this sdiy
> do able?
> 
> Well, that's the question.  I haven't found any good
> SMD well-matched, 
> high-gain transistors.  I've been stocking up on the
> popular Japanese 
> matched pairs, some of which are quite good, but all
> discontinued.  In my 
> most recent discrete OTA I tried a pair of
> hand-matched ZTX795A's glued 
> together, but they didn't work very well.  They had
> too much LF noise, 
> presumably due to temperature fluctuations.
> 
> Maybe some of the non-monolithic SDMs could be made
> to work, I dunno.
> 
> Another question is whether it is even a good idea
> to try to reproduce the 
> 3080.  The Iabc mirror has always been a pain,
> requiring a mirror in the 
> expo converter.  Why not work with the simpler
> discrete designs that don't 
> have this "feature".
> 
>    Ian
> 
>