[sdiy] MC1495 multiplier chip discontinued completely

[jhaible]

>http://www.oldcrows.net/~jhaible/tonline_stuff/jh1_ring.gif

>I saw a similar one OTA circuit in EN, also temperature
>offset problems.

Yes, this circuit was "invented independently" by Serge, an EN
contributor, yours truely, and I bet a few other people as well.
It's so *obvious* to do it that way.

Temperature drift is remarkably ok even in my 3080 version (though
a lot of drift is to expected - but I just never play it at extreme
temperatures), BUT more important: You get rid of it as soon as you're
making diode pre-distortion.


>The linearised 4 Q gain cell has a slightly similar behaviour.
>The log input suffers from distortion, but at least my
>discrete application shows some dead band. Maybe due to mismatch
>and low base current behaviour (recombination will destroy
>ideal expo behaviour for low currents), there is liess current then
>predcited, similar to dead band somehow.

I don't understand where the deadband would come from.
It's all class A stuff - no current goes down to zero.
(Referring to the ordinary, linearized 4Q Gilbert multiplier.)

JH.