Best way to correct load cell output with large offset?

[Ahrenp]

I have a load cell with a 1.785mV/V output that I would like to amplify and connect to a data acquisition system. The load cell may have been damaged and now has a very large (relative to the output level, at least) offset at zero load, of around 135mV. Assuming that the load cell still has a linear output of 1.785mV/V above this 135mV offset (it appears to in my limited testing) what is the best way for me to subtract the offset prior to amplification? The load cell uses a standard bridge:



I built a test circuit (Crude schematic attached) that feeds the bridge output into a unity gain differential amplifier, which is then fed into an inverting summing amplifier along with a trimmable (and op-amp buffered) DC voltage that cancels the offset and has a gain of 100. This works and gives me the expected (inverted) output.

Is there a simpler/better way to achieve the same result?

P.S. In reality I can probably feed the signal directly (with no amplification) into my DAQ system and correct the offset in software, and my load cell may in fact be NFG, so the circuit design is more of a learning exercise.

[David Hess]

That is an awful lot of offset so the load cell is probably ruined.  Usually shunt resistances added between the excitation and signal pins can be used to bring the zero within range but the shunt resistance drift will contribute to the zero drift.

[Ahrenp]

I tend to agree that my load cell is probably no good. If we assume, for the sake of argument, that the load cell is fine aside from the offset, is my approach reasonable? Can shunt resistances be used to trim that much offset?

[David Hess]

What you are doing should work.  The zero correction needs to follow the excitation voltage which is the case with your schematic.  The amplifier buffering the offset trim is not needed.  The offset trim could also be applied directly to the load cell.

[Brutte]

Bridge is usually wired remotely so the unbalanced sensing wires might cause trouble.
I'd try to repair/balance the bridge itself by attaching some fixed resistors on it, keeping the sensing wires loaded with similar currents at all times.