The purpose of learning is to expose the lies we have accepted in our model of the universe. To pursue truth is to feel more acutely its absence. If anyone would learn, they must remember what they have 0.00005forgotten that they did not know.

Deep.

The MAX9923 is not going to work. Actually, none of the amplifiers that I picked are going to work. While looking for a spice model of the MAX9923 I came across a good article by maxim about estimating error sources for current sense amplifiers.

They identified five different sources of error, explained what they mean and how to calculate them, and linked to a useful online calculator. As I expected, a large portion of my error came from the offset voltage of the amplfier, but the error due to the common mode rejection ratio was actually way higher (228uV vs a conservative 50uV on the output side).

The common mode rejection ratio described the ability of the amplifier to reject variation on the input signals that is common to both. They way that they say to estimate the error from this term is to compare the common mode voltage that you expect your system to use to the common mode voltage at which it was characterized. The datasheet says that tthe MAX9923 was characterized at a common mode voltage of 12V, and I am expecting use voltages between 4.2V and 2.9V. According the maxim article, unless I use common mode voltages close to 12V, I am going to get errors that completely drown out my signal. Even though the datasheet says that common mode voltages can range from 1.9V to 28V, the further the common mode voltage is from the voltage it was optimized for, the larger the effects of the common mode rejection error are going to be. None of the amplifiers I looked at were optimized for the range in which I want to work.

On top of that, the error calculator revealed to me that the MAX9923 has a specified output range. According to the datasheet, the min output voltage is between 1mV(typical) and 10mV(max). Using the highest gain amplifier, I would have to use a sense resistor of ~8Ohms to product satisfactory output voltages at my minimum current spec of 5uA. At my max current spec of 100mA, this would mean a voltage drop of 0.8V - way too much.

What does this mean? 1) common mode voltages are going to make it very hard to measure small currents. 2) I am either going to need higher gains or an amplifier with a lower minimum output voltage.

Practically, I think this means that I have to give up on using a high-side measurement system. A low side system will get rid of the common mode voltage problem and will open up more amplifier options to me.