I believe I've copied the Caldwell high current bidirectional current source design but my design tops out at 4.5mA in the 10k load (R7) for any input over 4.5V

I would very much welcome another pair of eyes on this or some insight into what I've done to upset Multisim.

 

What is U1?

 

U1 is an INA149, I didn't intend to stray from the Caldwell design.

 

Are you sure V_in is 10Vdc?
I see no measurement of that voltage.

 

The colour of U1 suggests the model was included by you and is not from the Multisim main database. Probably something wrong with the model. Try an appropriate difference amplifier from the main database (e.g. AD629 or similar) or build a discrete one yourself with a +/-18V op amp.

I used an INA117 in the attached example. Output current tops out around +/- 9 mA because the output of the INA117 has to supply quite some current and is simply limited by how close it can get to the supply voltage rails.

 

Output current tops out around +/- 9 mA because the output of the INA117 has to supply quite some current and is simply limited by how close it can get to the supply voltage rails.

The data sheet says it should be able to output +20mA when the supply voltage is 5V above the output voltage.
(Below from data sheet with a ±15V supply.


Your sim show a difference of only 3.5V between the plus supply and the output voltage which is thus what's likely limiting the current.

Increasing the value of the 10kΩ resistors would increase that voltage difference and thus the maximum output current, but that also would increase the supply voltages to U1.

Alternately reducing the 10kΩ resistor values to about 6.5k and adding 5V Zeners in series with each resistor would keep the difference from ever dropping below 5V without affecting the supply voltage across U1.

 

Thanks to 0ri0n - yes the model is one I imported, your suggestion will allow me to move forwards.
Thanks to crutschow for peeling the next layer of the onion for me.
I'm grateful to both and am now going to the shed to mutter about simulators for a while...

 

now going to the shed to mutter about simulators for a while...

Only if you are sure the simulator is actually at fault.
I find that happens very infrequently.

You didn't answer my question in post #4.