Bipolar input cmoy problem

[BlazerFRS]

From the title you might think my problem is occilation, but no, it's much stupider than that.

I've been working on a bi-polar input cmoy/a47 for the past few days and i've finished the power section and gain stage, haven't moved onto the second psudeo buffer stage yet; I figured I'd pop a chip in and make sure everyting was good before moving on.

As a precaution I like to check my voltages and everything before plugging in a chip, so I hooked it up to an old 9v I had on my desk (reads 8.5v). My initial measurements were good- 4.20V from positive to ground and -4.30V from negative to ground. The odd thing is, after the battery is hooked up for a few seconds the split starts to drift, and after about 10-15 seconds I'm left with about +7.8V on one side and -.7V on the other. it continues to get closer and closer to zero on one side the longer it is hooked up. I've also observed that the split can move in either direction (i.e. sometimes it will split to a large + and small -, other times a small + and large -)

The voltage dividor is made from 2 3k33 resistors as I'm out of 4k7's, could this be the cause?

I checked both resistors, they are the same value. I've also gone at the thing with my multimeter to check for shorts and what not with no results.

The caps on the power section are 2x 680uf 10V Panasonic FM's, 2x 10uf tantalums, 4x .22uf polyester, and 2x .1uf ceramics close the the op-amp pins.

Any thoughts?

Thanks,
-Blazerfrs

 

[Pars]

Does the chip get warm or hot at all? It could still be oscillating...

Other things you could try would be lower value resistors (penalty... higher current draw) such as 1K. Or a TLE rail splitter. Or a discrete Sijosae (sp?) rail splitter (check Tangent's site for his ground article... it has the Sijosae splitter in it that can be built from Radioshack parts (a couple of trannies, resistors and diodes).

Are the inputs grounded when you are testing? Something tells me that with a bipolar chip, this could be important. I assume you have reworked the gain resistors, etc. to provide the balance between the inverting and non-inverting inputs required by a cranky opamp?

 

[BlazerFRS]

I haven't even gotten to plugging the chip in yet; I stopped when such weird things were happening with the voltage split.

The more I think about it the more ridiculous it seems... it a freaking voltage divider, how do I screw that up?

I've got a few chips around to try, AD823, OPA2227, and the infamous LM6172; I'll probably throw them in tomorrow and see what happens.

 

[mono]

I'd wonder if the resistors aren't a close (tolerance) match. Might also (temporarily) put a couple resistors between 1) -V & Gnd and 2) +V & Gnd for a load to see if that stabilizes it.

 

[Pars]

Quote:

Originally Posted by BlazerFRS I haven't even gotten to plugging the chip in yet; I stopped when such weird things were happening with the voltage split. The more I think about it the more ridiculous it seems... it a freaking voltage divider, how do I screw that up? I've got a few chips around to try, AD823, OPA2227, and the infamous LM6172; I'll probably throw them in tomorrow and see what happens.

Sorry, I misread your Quote:

Originally Posted by BlazerFRS i've finished the power section and gain stage, haven't moved onto the second psudeo buffer stage yet; I figured I'd pop a chip in and make sure everyting was good before moving on.

as meaning you had a chip in. Running off a battery, this shouldn't drift at all with or without a chip in. I don't think you need a load, as the divider will pull some current draw through the reistors... you are really only measuring a resistor in parallel with a cap(s). Without the power, ohming out the V+ and V- pins to VGnd should show the cap charging (0 to ~infinity). I would check out your bypass caps... do you have V+ bypassed to V-, or are these just in parallel with the Pana FMs? Polarity of all polarized caps correct?

-Chris