Karl,
Thank you for the additional information.
I used your filter parameters for a new design and the resulting design is in the attached report file.
As you reported, the first stage output has the correct swing, but is shifted by +0.5Vdc.
The second stage output has negligible swing and is shifted by -2.3Vdc.
The same circuit was built in TINA, and it seemed to run fine. It had the expected AC response and the pulse response was fine as well.
The mid-point voltage between the two stages did not have the DC offset observed in the WEBENCH output.
In the WEBENCH circuit, the inverting and noninverting pins are swapped when you compare their positions in the first and second stages.
As an experiment I tried swapping the VCC and VEE connections in the second stage of the TINA circuit.
The results were inconclusive. That resulted in no output swing, which would be expected for backwards supplies.
But the second stage output did not have the -2.3Vdc shift seen in the WEBENCH design.
I will look into this further and let you know what I find.
Regards,
John