I was given the task to make a new version of a board which have this little contraption for RS485 termination resistors:
simulate this circuit – Schematic created using CircuitLab
My question is, does this really work? I my concern is that the body diodes will mess things up.
I have seen that some manufacturers offer RS485 driver chips with a termination resistor that is possible to enable/disable, however theese doesn't match the required resistance in my application.
weird! without understanding the broader context, I don't know why you'd want 'dynamic' control over the termination & biasing resistors, but I guess someone had a good reason for it.
in 'ideal conditions' when there's little ground differential between each end of the line, this should work - the MOSFETs bring in the 910/160/820 resistors, and get you a typical ~120R termination & reasonable-looking bias resistors.
but one of the reasons for using RS485 is to tolerate differences in the ground potential between endpoints! '485 transceivers will have specs like "InputA/B can be +12 to -7V w.r.t. Gnd". If that happened to be the case, then yeah, not only will the body diodes come into circuit & make for 'fun', but even the gate voltages will shift too. Maybe that isn't a factor in your application, though.
I'm looking for a similar solution and I found this of TI proposes this reference design Controller Area Network (CAN) with Selectable Termination Reference Design.
