Filter DC conditions sanity check please

[ElectricDruid]

Could you all have a quick look at this for me?



I think I've done something stupid, but despite that it seems fine.
It seems to me that these output filters have no solid DC conditions set up at all. They're essentially "floating" (L Out and R out go to output caps too). Nonetheless, the prototype I built works fine. Which I don't get.

Am I right that this is a mistake? And how come I got away with it?

Thanks!

[Rob Strand]

Yes looks like no bias path on the opamp + inputs.  Measure what DC voltage you are getting on the output of the opamps.  That will tell you what DC level the opamps are biasing at.     If we assume the caps are at 0V on start-up then that means the opamp will bias at the DC voltage on the FV-1 outputs.   I suspect there is a small amount of leakage through the electrolytic coupling caps and since the TL07x opamps are JFET input they don't need much leakage.   That would mean the opamps are biasing at something very close to the DC voltage on the FV-1 outputs.  With such large caps leakage is going to take *hours* maybe days to stabilize from the 0V across the caps to some tiny deviation from that.

With such low currents you could even get leakage on the PCB from nearby tracks at different voltages!

[Dormammu]

How can anyone draw any conclusions without seeing how the power supply of the circuit is organized? If this is a standard bipolar power supply, the IC does not require biasing.
Typical circuits have been developed for filter circuits — use them, there is nothing stupid in them.

[antonis]

If this is a standard bipolar power supply, the IC does not require biasing.

But C30 & C35 DO need a path to ground, no..??

[anotherjim]

Rob's explanation makes sense. If it works with that, you can probably use the 3v3 supply as an input DC reference source with =>10k resistors after the 1uF caps.

[Rob Strand]

I guess you could argue if it works OK as is it should work OK without the 1uF caps, just let the FV-1 bias voltage feed through to the opamp.  Not optimal for swing though if the opamp isn't rail-to-rail and the FV-1 is.

[ElectricDruid]

For the record, no, it's not a bipolar supply, it's typical stompbox 9V supply with a virtual ground at 4.5V.

Which makes me realise *another* thing wrong about it, which is that the 1uF caps at the front of the filters are the wrong way around. If I'd got it like I wanted it, there'd be 3.3/2=1.65V bias at the FV-1 end of those, and 4.5V at the filter end, so the positive end would definitely be towards the filter.

Since I screwed up the biasing, I got away with that one too. S'mazing...

[Rob Strand]

Just one of those not enough coffee days.  In the actual circuit without the bias resistors the caps would be OK, but I know what you mean, it's more about the intent.   When you do too many 9V pedals there is a tendency to put the cap + to the outputs, especially when your mind goes on autopilot. (Today it's a "no pilot" day for me, 4hrs sleep.)