TINA-TI: How to simulate ac coupled differential amplifier

Actually, I think you only needed a ground on the V+ inputs. Does the ideal op amp not require power supplies? odd. 

Michael,

This would only provide a reference voltage for the output while leaving the input common-mode for both gain stages floating  - thus, in the actual circuit the inputs would collapse on one of the rails.

??????????????????

Running a DC sim meters added, they all look biased to me?

Yes, all the inputs seem biased at ground level but only in simulation because of the default transconductance  of 10^12ohm that connects them to ground.  In a real circuit,  however, there is nothing to keep all these nodes at 0V - the reference voltage does NOT provide biasing for Vcm.

I must be missing something, 

If you ground the V+ input, the LG holds the V- input at ground if there is DC feedback like here? 

If you ground the positive input, the feedback attempts to keep the negative input at virtual ground by delivering required current thru the feedback resistor but this current has nowhere to flow if you don’t provide the current path to system ground, so the negative input collapses on one of its rails.

This is new to me, in an actual op amp, this should work fine with no feedback current other than bias current and of course nothing back in the cap coupled path. 

Using an  actual op amp, new OPA211 model, this looks like it works ok - little bit of phase margin issue due to higher R and V- input C, 

And running a DC operating point sim, looks ok to me.