Descreet OTA from CA3096

[Edu Silva]

De: jhaible <jhaible at primus-online.de>

> Another thing I wanted to try but didn't find the time yet. Can you tell
us
> about the performance of the 3096-based version ? Offset variation
> over Iabc in particular? I mean, are the pnp's good enough to form 
> a precise current mirror that keeps its symmetry over several decades ?

Hi Juergen, all

Think I'm not from much help...  I did something several years ago, but
wasn't thinking about synhts at that time. I was looking for a better opamp
to replace the TL072's and 5532's used in my audio projects and tried the
array to build it (at the first time, 2 NPN's as the inp pair, a PNP as the
VAS and a complementar pair as the output stage; after, two 3096's, with
current sources, mirrors and stuff). Noise wasn't so little as expected, so
I ended up testing and building several full-discrete designs, from 3 to 12
hand-matched and noise selected transistors. Who needs those damned
monolithic opamps, after all?  :)

Back to synths:  The PNP's in the 3096 are known to be poorly matched (and
have other drawbacks), but I'm not sure if it makes the array unusable for
synth applications. Selecting the chips may help and I've read in the data
sheet that connecting pin 16 to the negative supply makes the trannies
happy...

Edu

_____________________
ES2 Audio - Som & Luz
Salvador-Bahia-Brasil