Three key performance parameters for an instrumentation amplifier are common mode rejection, noise, and temperature drift. Common mode rejection simply measures how well the instrumentation amplifier subtracts. If the same signal is sent into each instrumentation amplifier input, an ideal instrumentation amplifier should have no signal at the output. In this case the differential signal is zero, and therefore the instrumentation amp output should be zero. Such an ideal instrumentation amplifier would have an infinite common mode rejection ratio. Real world instrumentation amps do not reject common mode signals perfectly. How well they reject the common mode signal is measured by the common mode rejection ratio (CMRR) which is simply the differential gain of the instrumentation amp divided by the common mode gain. A larger number is better. For many instrumentation amplifier architectures, common mode rejection increases with gain, so when comparing instrumentation amps, it is important to compare specifications at the gain customers are interested in. Another important specification is noise. The noise specification shows how well the instrumentation amplifier can resolve a tiny signal. This is typically expressed in microvolts peak-to-peak for low frequency noise, or in nanovolts per root Hertz when discussing higher frequency noise. Smaller numbers are better. An instrumentation amplifier’s noise performance typically improves with gain (referred to input), so again instrumentation amps should be compared at the appropriate gain. Finally, the drift specifications of an instrumentation amplifier indicate how well it performs over temperature. Gain drift is typically best when the gain of the instrumentation amp is only set by internal resistors. Offset drift is typically the best at high gain. In either case, smaller numbers are better.