Stay tuned for Jerry Steele’s upcoming article on EDN entitled, “Simple, Precise, Bi-Directional Current Source”—here’s a teaser from Steele, ON Semiconductor's Applications Manager:
Any attempt to build bi-directional grounded load current sources has complications. The most classic attempt with the improved Howland requires either precision resistors or resistor network or expensive diff amp (and even then, requires trim).
My circuits complication: floating power supplies. But we live in an age where that is a readily and economically solved problem. In my circuit, a single resistor determines the precision.
Some final lab work this weekend should make my article about ready for review. I have been using the simplified version with my Analog Discovery 2 to measure amplifier output impedance over frequency with the network analyzer function for some time.
Interesting: I see how it works. But if you use both outputs at once, each supplying base current to a transistor, and then widen the feedback loop to include them, it becomes similar to the ol' LH0021/41 power amp. Some of the circuit tricks with that and similar parts/circuits surely apply too, like limiting and frequency compensation…. Jeez, when did I get OLD?