AM26LS32A: -

Hi Ray,

The device itself has some internal resistances that you could include in your simulation if you'd like. Figure 7-5. The external resistors are so much smaller than the internal ones that I would just ignore them. 

Ray Pourjalilvand said:

Vb-Va= 0.39V (which is outside of the +/-200mV undefined region)

Ray Pourjalilvand said:

but the problem is that I won’t be able to achieve |Vb| > |Va|.

You do meet the requirement.

Using your values, Vb-Va = -390mV which is less than -200mV (Vth-) so you have around 190mV of noise margin if we use the min value here.

With Hysteresis the value becomes 240mV. In most systems this would be enough to provide a fail safe that ensures 1Y outputs a logic low during the open condition.

-Bobby

Ray Pourjalilvand said:

(1) I see the internal resistance in the datasheet is stated to be 15K Ohm typical. Does that sound right to you? Is that 15K across the A and B inputs I suppose?

That's the defined leakage across the common mode range. You can choose to model this as a worst case scenario.

Ray Pourjalilvand said:

(2) I’m a bit confused by what you said about not including the external resistors in the simulation, but wouldn’t that not give me the offset that I need?

The opposite, not including the internal registers (I'm saying ignore the internal leakage). 560 ohm << 100k ohm the eq resistance is basically 560 ohms anyways so 100k is not relevant. 

Ray Pourjalilvand said:

(I’m not simulating the circuit, just calculating the biasing resistor values).

Sorry, I assumed your circuit was taken from a simulation tool.

Example below where using 12k internal with worst case condition of GND shift doesn't do much to the biasing voltage across 100 ohms. This case You still have a fair bit of margin against noise. 

Below is your base case with a VoC of 2.38V

Ray Pourjalilvand said:

(3) It sounds like with these resistor values I’ll able to ensure a logic low output during the open condition but what about the concept that these biasing resistors are not equal in value? Would that create any unbalanced issues?

You're technically generating a much stronger Biasing voltage than someone who does not use the external biasing so the VoC is actually going to be more stable. If you're worried about this, then using split termination with a filtering cap will provide much more stable VoC for the AC frequency. You could also add in a common mode choke to further stablize the VoC. Then lastly to ensure VoC delta isn;t emitting too much, use a shielded twisted pair cable. Though, I think you may be over thinking this unless you absolutely need to pass some sort of EMI compliance. 

-Bobby