[sdiy] Heads up on SSM2210

[Ian Fritz]

At 08:14 AM 11/18/2004, Grant Richter wrote:

>I guess the improved spec must be initial Vbe matching? Does this affect
>total temperature stability? I am still not clear on the significance of the
>spec for initial Vbe matching. For example, they all spec 10 mv. as a max.
>mismatch, but typical specs improve by price. What is the actual affect of
>Vbe mismatch in a VCO application?

Well, we've argued around and around about that over the years here.  If 
the Vbe offset is due to a benign effect, such as a difference in mesa 
area, then the offset is unrelated to temperature stability, and just 
produces an unimportant offset in the converter output.  If the offset is 
due to a difference in the reverse leakage current, then the 
temperature-dependent prefactors will not cancel and there will be 
drift.  This is the usual assumption, and is why Vbe matching is targeted.

I was just looking at the old AD nonlinear handbook.  For their early 
converters they used discrete transistors to get the lowest possible series 
resistance, and they carefully matched the transistors for reverse leakage 
current.  I believe this may still be the best approach.  But it requires a 
laboratory-grade measurement (controlled environment, electrometer pA 
meter) and would be quite difficult to do at home.  That's why Vbe matching 
is used instead.  It's easy to do, but IMO inconclusive.

   Ian