I am trying to model a AD8227 InAmp sensing a differential input from a bridge pressure sensor in NImultisim (see attachments).
I set the gain resistor for the AD8227 to 487ohms and calculated the gain be 169.3. The REF voltage to the InAmp is about 0.481V.
But when I simulate this in NImultisim, the gain seems to vary a lot more than what I expected over my input voltage range of 0 - 25mV - (from <160 to about 168, see spreadsheet).
The CM voltage is about Vcc/2 and input and output voltages are all within range as far as I can tell from the data sheet, and fig 41 quotes a worst case gain non-linearity of about 120ppm.
I do not have any experience using InAmps, so could somebody please tell me if I am doing something wrong or if this is what I could expect to see if I built this circuit, i.e. is this typical behaviour?
Your circuit looks ok, but I suspect the problem is in your calculation of the gain. Gain is defined as the slope in the transfer function between the input and the output. Therefore, to calculate the gain from circuit measurements (or simulation, in this case), you must divide the change in output voltage by the change in input voltage.
Any amplifier has an offset voltage, which is defined as the voltage required at the input to make the output voltage exactly zero (or in reality as close as possible). This is effect is included in the simulation model for AD8227. If you simply divide the output voltage by the input voltage at any given point, you cannot eliminate such offset. If you do the math, you'll realize that your calculation will give you a larger "gain error" when you are close to zero, just like your spread sheet shows.
I corrected my spreadsheet calculations as you suggested and the gain figures are now what I originally expected. Thanks for setting me straight on that.