RE: Virtual ground question
Originally Posted by Calroth
Question about the virtual ground option.
I'm hooking an Alien DAC up to a PIMETA. So I'm wiring up AV+, AV- and AG.
AV+ and AV- is fine. But I'm confused with AG. On the PIMETA, that's connected to input ground, which is connected to the Alien DAC OG. So if I wired that, then I would be effectively connecting AG to OG (and the rest of the ground plane).
Is that right, or am I missing something?
Right, I posted that a month ago, and in the meantime have been trying to wrap my head around virtual ground theory. As a primer: I have an Alien DAC, a PIMETA, and an OPA551 to act as the IC4 virtual ground buffer between the two. It's all to run on a 12V supply, split into +6V, 0 and -6V.
The first thing I tried was what I described above: connected the Alien DAC OG to PIMETA IG, and the Alien DAC AV+ to the same. It pulled the entire virtual ground on both sides up by 4V... so we were +6V, +4V and -6V. That ain't right, Calroth. So back to Tangent's virtual grounds page.
So, after some critical thought, I decided that the Alien DAC's ground should be buffered and isolated from the PIMETA's, though still at the same voltage. So, disconnected the Alien DAC OG to PIMETA IG (leaving the left and right channel wires there). Now, the Alien DAC's ground is buffered. In theory, its current draw etc. shouldn't disturb either ground level. And it didn't... much. The PIMETA's ground stayed at 0V, so the OPA551 was definitely isolating the Alien DAC. However, the Alien DAC's ground was offset by +170mV. Better!... but still unacceptable.
Scratched my head. Then did what every self-respecting DIY'er would do: I Mini³-ified it! Added a 600R ferrite on the output of the OPA551, outside the feedback loop - plenty of spares from the Alien DAC construction. I had to lift a few pins to do it. You don't want to see the board... between this and all the other things I tried, it ain't a pretty sight. But, the ground is now offset by about +19mV. Yeah, that's close enough for me! Hooked it up and got it playing music.
So. Having read all that... any comments? Did I do something completely stupid? Next thing to try is putting some 100nF bypass caps on the OPA551 V+ and V- (also left over from the Alien DAC build). The datasheet suggests it, but I think that's just a precautionary thing. Also, I suppose I could shift the DC blocking caps from the Alien DAC to the PIMETA to wipe off that last bit of offset.