Voltage Regulators 102, Op-Amps 101: AC-coupling!

I built this power supply several years ago, always with the intention of adding an amplifier for the current-sense circuitry.

I used a .01ohm resistor for current-sensing, 'cause I didn't want it to interfere with the output voltage. The resistor was tied in series with the +Vout path.

The voltage across this resistor is measured, currently, directly. So, measurements are 1/100th of the actual value (kinda confusing to read on a multimeter).

So I've been fighting with analog circuitry for a few days to try to amplify this 100x. Seems simple enough with the "ideal Op-Amp." Yeah, right. First things first, I have to consider CMRR (Common-Mode-Rejection-Ratio)... something I hadn't considered an issue.

The differential voltage (voltage across the resistor) varies from 0 to 0.04V (with a 4A load, max for my supply, I want to amplify this to 0-4V). The Common Mode voltage (voltage at the resistor, WRT ground) is 0-24V. We're talking 100's of times difference between CMV and differential voltage. I don't remember my dB conversions at the moment, but I'm pretty sure a 100dB CMRR (which is *excellent*, and nowhere near the ability of standard Op-Amps and resistors) is not nearly enough for this circuit.

So, what's happening? As the output voltage of the power-supply varies, so does the output voltage of my Op-Amp, dramatically. The differential voltage (across the resistor) is being amplified in there, somewhere, I think. But it's completely lost.

This is why Analog circuitry has always let me down. It seems like a simple circuit. Even 2 years' training in EE lead me to believe it would be. Instead, it's a battle at every step. The resistors have to be matched to 0.01% tolerance (HAH! 1% is alot to ask for, even in my vast collection of assorted parts.). The op-amp itself has to be of high-quality... (so much for all these standard LM series chips I'd been meaning to do something with). I wanted to do it with a single-supply (+/GND as opposed to +/-/GND) which was the first difficulty which even a "single-supply op-amp" like the LM324 makes difficult. The list goes on.

Here are a couple references I came across to verify my experimental results: http://www.maxim-ic.com/app-notes/index.mvp/id/1180 http://www.maxim-ic.com/app-notes/index.mvp/id/746

So, apparently, high-side current monitoring is difficult. Let's switch to low-side! Still having issues... Not sure why, exactly. But I'm tired.

So: Op-Amps 101: they work best with AC signals (audio, etc), coupled through a capacitor. Anything else and you're setting yourself up for disaster.
Voltage Regulators 102: Measuring current on the high-side is difficult, to say the least.