Differential amplifier question

[EBK]

For some reason (lack of sleep, probably), I'm not trusting myself to answer this question.

Say I have a TL07x op amp configured as a single-supply (9V) difference amp, something like this:

In fact, let's simplify it by saying all resistors are equal.
If Vin1 and Vin2 each had a common 8.65 VDC component, i.e., pretty close to the supply voltage level, would this cause any problems (other than limiting headroom)?  I'm wondering about things like saturation, latch-up, and their relatives.

Later, I can explain in more detail what I'm trying to do, but I don't want to get distracted away from this academic question at the moment if you'll indulge me.   

[amptramp]

The input impedance seen from V_IN1 is R_G.  The input impedance seen from V_IN2 is R1 plus R2.  You could run into a problem if your source requires a defined load impedance.

Amplifier headroom, latch-up and common mode characteristics vary with the amplifier.  Some have rail-to-rail input common mode where nothing bad happens, even if the output is not rail-to-rail (some op amps are both).  Some op amps like the LF358 have the output go to the opposite polarity if you exceed the input common mode range.  Even the workhorse TL071 does not have input characteristics guaranteed within 3 volts of either rail with a 10K load and within 5 volts with a 2K load.

[EBK]

With all resistors equal, the non-inverting terminal would see half of Vin2 (assuming infinite input impedance....), which would be 4.325 VDC.  The inverting terminal would see 8.65 VDC if open loop. Common-mode input voltage would be 6.4875 VDC, right?

Looks like, for TL07x, common-mode input would have to be between 4 and 5V for a single 9V supply.

[EBK]

The input impedance seen from V_IN1 is R_G.  The input impedance seen from V_IN2 is R1 plus R2.  You could run into a problem if your source requires a defined load impedance.

I'm glad you mentioned this.  My source (modulator chip) does expect a certain load, or rather, my biasing calculations for it do.   

As a bit of background, this is for my SMOG circuit.  I want to convert the differential output of a MC1496 (not the easiest data sheet to navigate) to a single-ended voltage without AC coupling it first (that would be simple, but I actually only care about the DC component....).  Running it through a difference amp lets me neatly remove the bias voltage at the output while preserving the differential DC component.  I have given up with the single-supply approach.

[PRR]

> Vin1 and Vin2 each had a common 8.65 VDC component, i.e., pretty close to the supply voltage

If they are literally the same, then the right output is Zero.

Can your single-supply opamp swing its output to Zero?

If they are slightly different either way, then sometimes the right output is Below Zero (assumed negative). I'm sure your single-supply opamp will not do this properly.

[EBK]

If they are slightly different either way, then sometimes the right output is Below Zero (assumed negative). I'm sure your single-supply opamp will not do this properly.

Me too.  (I'm better rested now.) 

[anotherjim]

Assuming the grounds on the signal ins and R2 are actually at 1/2 reference : not supply ground...
-input won't see all of Vin1, it will depend on if the output will swing counterwise to pull the -input via Rf to equal the +input reference. Working and behaving properly, -input should only ever see the reference volts.