[sdiy] CA3046 vs. CA3086 vs. LM394 vs. ...

[René Schmitz]

Hi Larry and all,

Larry Troth wrote:
> I know that some of this has been discussed before, but I don't
> recall anything specific about this groups of similar parts.  And I
> did not find anything comprehensive in the archives (maybe I just
> missed it).
> 
> CA3046 vs. CA3086 vs. LM394 vs. SSM2210 vs. MAT-02 vs. THAT 300 vs.
> ???

There are the super matched pairs, LM394, SSM2210, MAT-02 and so on.
They are large geometry devices, which you can think of as lots of small
transistors in parallel. What that means is that noise goes down, and
they have lower rBE, which helps with expo conformance (but some form
trimming/compensation is still required). You have high beta. Also the
transistors are often interdigitated that is they are arranged
internally in a checkboard fashion so that temperature gradients are
largely eliminated.
The CAs by comparison don't have high beta, are not interdigitated, have
comparitively high rBE (see the rBE comparison table on my homepage).



> Other then price, when dealing with SDIY circuits, is there really
> any audible difference?

Taking parts and trying to judge them solely on specs is nonsense.
A part will have a circuit its in, and that determines the performance.
Also that determines what youre really requireing from a part. Only then 
you can start deciding for that particular circuit.

You can quite easyly ruin all the low noise (or the stability, etc..)
you bought from an LM394 (or the like) by the surrounding circuitry. 
Don't be tempted to think that if you just buy the most expensive you'll 
get superior results, on rare occasions will that be true. On the other 
hand, if you know how to do it you can get the same results with more 
moderately priced components. (For instance the rBE can be compensated 
for even if its initially higher.) Investing only where it is really 
required.

> What parameters, from a specification sheet point of view, should I
> be looking at to see the difference?

That depends on what circuit we are talking about. For a VCA for example
I'd look into noise figures. THere the super matched pairs really excel
over the CAs...

Cheers,
  René


-- 
uzs159 at uni-bonn.de
http://www.uni-bonn.de/~uzs159