CA3096 OTA (was:AW: 2040 filter clone with unmatched transistors ?)

[Haible Juergen]

Hi,

	>Last time I built a 2040-type filter I used EN2016 and EP2015
arrays
	>(from Elantec) with delta Vbe up to 4 mV, i.e. not perfect
matching,
	>but good thermal tracking.
	>
	>The resulting CV feedthrough, without trimming each stage, was
around
	>1.5 volts using the resistor values shown in the data sheet
application
	>circuit. Quite a lot, but in other respects the filter performed
well.

Last weekend I finally breadboarded a CA3096E for use in a VCA,
VCF, or VCADSR. I configured the two pnp's as current mirror, and used
two of the 3 npns as differential pair. An LF355 was used as active
integrator (i.e. *not* 2040-style configuration). Feedback resistors
were 100k / 200R.
I switched the tail current of the OTA section from 15uA to 450uA
and checked the change of output offset voltage.
The simple 4-transistor cell had an offset voltage change of 100mV
at the output of the integrator.
I think this is not bad, but I might have been lucky with this very 3096
chip. (Must try several ones later.)
Then I made a few experiments, with surprising results.
Decreasing the positive supply voltage of the pnp mirror to 1.5V
did not change anything. (I expected an improovement because
of a more constant CE voltage across each transistor pair.) 
Building a 3-pnp current mirror with the two internal ones, plus an
external BC550C, made things *worse*, not better.
Using the 3rd internal npn to build the "floating cascode" I suggested
last week didn't work at all.

Thought I should share this. It leaves a few questions open, no doubt.

JH.