[sdiy] ARP VCO drift

[rob]

Hi folks,
I need some expert guidance and ideas, a rather boring subject but this one
has me stumped.
I recently built a couple of ARP style VCO's (Odyssey/2600) with new
components but the same design.

The VCO waveshapes and frequency range are excellent, but the frequency
drifts a lot more than I expected or have measured with a real ARP 2600.
At 200Hz the ARP clone VCO varies by 1-2Hz, whilst my E-mu Systems VCO
varies by less than 0.1Hz, as does the original ARP.

I have tried - but with no impropvements:

Different makes of 1K87 3500ppm Tempcos
Adjusting the Tempco ppm with a series resistor
Odyssey and ARP2600 front end designs before the expo pair
2N3906/3904 transistors
BC550C/560C transistors
Integrator capacitor is 1% polystyrene
Comparator voltage is stable to at least 0.5mV
Frequency adjustment voltage rails come from a voltage reference chip, at
least better than 0.5mV

The transistor pairs are not matched, but are in contact with the Tempco.
There is some improvement by getting this contact.
The original ARP 2600 VCO's are of course encapsulated....

Where am I going wrong?! Any ideas?

Regards and Many Thanks for any help...I really would like to get these
VCO's to be usable.
Rob
www.emulatorarchive.com