Question asked by @securelyfitz on Sep 9, 2013
Latest reply on Sep 9, 2013 by @securelyfitz

Hello,

I'm a very digital person by background, but I'm taking a stab at an analog circuit this time.

I'd like to use the AD8129 as part of an inexpensive differential probe for simple power analysis side channel attacks on cheap microcontrollers.

I've got the AD8129 powered by +5v and GND. Vin + and - are across a 10 ohm shunt resistor that goes between a 3.3v supply and the microcontroller in question. By observing the voltage across the shunt resistor I should be able to determine what instructions are being executed on the microcontroller.

I intentionally powered the AD8129 with 5v and the microcontroller with 3.3v so that the common mode voltage would be in the valid input range. I anticipated a delta Vin of up to 80mv (8mA peak draw * 10 ohm shunt) set a gain of ~11 (200 ohm and 2k ohm), tied REF (pin4) to ground, and expected it to have a Vout that would go from 0V to 0.88v, which i would feed to an AD9200 or similar expecting 0 to 1v input.

I quickly learned that's not how it works. It would seem that the only way to get a 0v output voltage is to have a negative -Vs.

Is this correct? What would i look for in the specs to tell me either 1) what -Vs i need to get a Vout of 0v or 2) what is the minimum Vout i could expect based on 0-5V supply

I took a step back to prototype a bit more so i'd better understand the chips. I'd like to avoid a negative supply because it means more parts, my understanding is that they can be noisy, and negative voltages just scare my otherwise digital brain.

It seems that with 0-5V supply, i really need to have a REF to center my output into the optimal range - I assume that would be (+Vs + -Vs)/2 = 2.5V. That works well for a very small gain, but with a gain of 10, that means that REF needs to be held at .25V - but it seems that REF (and FB too for that matter) seem to 'clip' and don't react to voltages below about .7V. Is this because they're out of range? Again, what should i look for in the spec to see this?

So I've tried a couple things that haven't worked as i've expected, and i think i understand what's wrong, but some validation of my assumptions and recommendations would be helpful before i move forward.

How would i go about doing G=10 on an AD8129 with 0-5V supply? If it's not feasible, should i be using a different part? should i break down and use a +-5v supply? Should i stick with the AD8130 and have a separate, discrete preamp before my A/D converter?

Thanks in advance for all input and help!!

-joe