Non-Local Closed Loop OpAmp Gain?

[CaptnYellowShirt]

I'm in the process of designing a high voltage constant-current supply for a project I'm working on.

For this project, I'm going to need currents in the range of 0.1 - 10 mA supplied at around 225 V.

Rather than designing the supply from scratch I'm opting to combine a Burleigh High Voltage OpAmp (typically used for controlling precision piezo devices) with a current sensing opamp setup. The Burleigh will supply the current at the high voltages, while a low-side shunt resistor will provide voltage feedback to some type of local precision opamp that ultimately feeds its output into the Burleigh box.

The Burleigh box is basically a non-inverting amplifier with an adjustable gain from 0 - 100. The precision sensing opamp will be something like an LT1100/1150 with a wirewound shut resistor and LM399 as a Vref.

My question here is on local feedback for the precision opamp. I'm concerned that, even though the Burleigh opamp closes the loop for the sensing opamp, without local feedback (e.g. << 1m of hook-up cables) the sensing opamp will find itself in some type of instability. That is to say, its open-loop behavior will dominate for some part of the spectrum.

I'm fairly week on my practical opamp Fu, so I was hoping someone might be able to set my intiution straight here before I build a test circuit.

[T3sl4co1l]

The voltage amp* inside a loop to regulate current?

*I don't suppose they make a current mode (Gm?) amp.  Which is basically any normal amp, but without the compensation and low impedance follower inside.  A voltage amp is normally made by intentionally slowing down a fast amp.  The excuse is, the dominant pole compensation is easy to use?

It follows, to have good current accuracy you need stupid high gain outside the volt amp (if you're talking LM399, it better be a damned fine amp and current sensor!).  Which means that amp has to be at least as slow as the volt amp, plus some headroom (expect maybe 1/3rd bandwidth overall).

The volt amp is dominant pole, but the pole moves up by the gain reduction.  So at a gain of 10, it's GBW/10, etc...  This makes it relatively easy to calculate the outer compensation too (if you were so inclined).

Tim

[CaptnYellowShirt]

I had a few LM399's laying around.