I recently put together a standard (or mostly so ;)) cantilever load cell for some tests I want to do (basically I am making a screw driven tensile machine). The cantilever has two strain gages mounted on it. These strain gages are wired in a half bridge configuration. In addition to this circuit I have a nulling potentiometer. The two standard outputs from the bridge circuit are taken and put into an instrumentation amp with a gain of about 750. I understand that typically these types of setups are unforgiving in terms of noise, but this is not my problem. The output of the circuit is not so noisy (at least with the level of precision I am concerned with), but it has an annoying drift(which varies in rate). Additionally there seems to be a significant delay in the settling of the voltage after I apply the load. When I jostle around the lead wires connected to the circuit typically the voltage will settle within a reasonable amount of what it was before, though occasionally it shifts by a much higher amount. From the tests I have done and loads I have applied (just static weights), it seems that the output is linear minus the drift. It might also be useful to know that this is a very thin Al-cantilever with the strain gages not covered and mounted with epoxy (I've been thinking that maybe I should cover the entire surface of the strain gage w/ epoxy, but this would mean that I would potentially lose a gage and have to remount one(again) if a lead broke or crossed or whatever). Any advice, speculation, and questions are very welcome.