Well, something is wrong then since I don’t get a sign change when swapping the transistors.
But damned if I can tell what’s wrong. Double checked the circuit and it looks right to me…also tried rebuilding it on a newer breadboard and it’s acting the same on there:
Note this is the reversed form of the circuit since I’m now working on the PNP 2N3906’s.
The red wire from my power supply is +12, Brown is GND, and Grey is -12 (I know not the best choice of colors but they were the wires I had on hand when I rigged up a cable to go from my PS to the breadboard.)
I checked continuity to make sure the breadboard connection are valid and all seems good.
I can see the voltage drift when I first put a transistor in and then it slowly settles in after a minute or two as the temp stabilizes.
Oh…wait…now I see my mistake.
Works a lot better if you connect the collector of BOTH transistors. I’m seeing the sign change and getting a more reasonable .5mV on the first pair I shoved in there.
Guess I need to recheck those BC547’s on the Steiner filter now…