Why is my opamp diff -> single ended circuit clipping?

[ve7xen]

I've built the below circuit as part of an audio DAC output.

In simulation it operates as expected, generating a clean output from the differential input. However in the actual built circuit, something goes wrong and it seems unable to swing below -1V. I've quintuple checked that the -12V supply is working as expected, and removed R21/22 and verified the DAC output is clean to that point. I thought perhaps the DAC outputs were getting pulled away from their DC bias and getting clamped there, but they seem fine and well above ground. The opamp is also behaving as expected here, both inputs are at identical potential and match the output. Note that R28 is not in circuit, so the mute circuit isn't in play. Not sure what I am missing here, probably something dumb that I will see a few minutes after posting here, but I'd appreciate some help!

Edit: Okay so inline attachments don't work...

[tszaboo]

My 2 cents: Your DAC has single supply. Probably it cannot sink the same amount of current as source.
Or do you have a current output DAC or a voltage output DAC? If it is current output, it might be worth to add an I/V stage on the output. Check the datasheet of PCM1794 to see what I'm talking about.

[magic]

Yes, this circuit loads DAC_LN on negative peaks more than it loads DAC_LP on positive peaks.

It seems unlikely that the DAC outputs are correct AND the opamps inputs are equal during clipping events, whatever causes them.

The other culprit could be the mute circuit (I think it could only work right if base drive is high impedance even below ground) but you supposedly disconnected it.

[SiliconWizard]

What is the DAC?

[ve7xen]

My 2 cents: Your DAC has single supply. Probably it cannot sink the same amount of current as source.

Hm, I think you may be on to something, it can probably only source current. I'd have thought this could only happen on the high side of the output though when the output voltage is above the DAC output (which is the opposite of what I'm seeing). I will see if I can measure if this is happening.

It seems unlikely that the DAC outputs are correct AND the opamps inputs are equal during clipping events, whatever causes them.

Sorry I didn't mean that the DAC outputs are correct. When the opamp circuit is connected, the distortion goes right back to the DAC outputs (CH2 trace in the scope shot is the DAC_LP). Just that the opamp seems to be doing what it's supposed to with the distorted signals.

The other culprit could be the mute circuit (I think it could only work right if base drive is high impedance even below ground) but you supposedly disconnected it.

It should be out of the circuit since I removed R28 but I will remove the rest of the components just to be sure...that was my first thought too. I'm pretty sure this won't work, it was a bit of an afterthought copied from the evaluation board.

What is the DAC?

It is CS42526. It's a voltage output DAC with specified 150R output impedance.

Really not very good at these analog circuits, thanks for the comments!

[ve7xen]

Seems like it may just be a loading problem, as suggested. The flat-topping happens when the DAC_LN is asked to source more than 2mA. Increasing the R15/16 to 2.4k (obviously other components will need to be scaled as well...) seems to have solved the problem.

Thanks!