I’m designing a rocket payload. I want to put 10 strain gauges in the payload, so I’m working on a circuit I can replicate 10 times.
I want to take strain gauges that are 350 Ohms each and can vary by ±5% over the full compression/expansion usage, and I want to measure them with the 10-bit ADCs that measure from 0V-5V on an Arduino Mega.
The trivial solution would be to use a voltage divider, but that only gets me 50 bits of resolution, and I’d prefer to use as much of the 1024 bit range as possible. A Wheatstone bridge is another popular solution, but it still has the same sensitivity range despite using more components and needing two analog sensors.
Thus I would like to use an Op Amp to magnify my sensor readings. I’d also like to minimize the number of parts I’m using. Does the circuit I’ve outlined below function as I’ve described?
I will have two power supplies, one providing 5V (V1 in the diagram) to all the gauge circuits and the op-amps, the other will be for supplying the 2.5V (V2) for the reference voltage on all 10 circuits. To minimize the number of different types of components I’m using, I set one of the gain resistors (R4) to be the same as the resistor in series with the strain gauge (R1). I’m estimating this design will take less than 100mA total across the 10 circuits (mostly flowing through the strain gauge and R1), though the power supply spec sheet says it should be able to source up to 1A total for all 10 circuits in a pinch.
My main concern at present is if one or more strain gauges gets ripped out of the screw terminals (our last rocket exploded a mile into the air and the last payload came down in pieces, but if the battery had stayed connected, we could have still recorded data). That would ground one side of the op-amp and I’m concerned that would put a large drain on the output of the op amp and the system as a whole. Is there some way I could guard against a strain gauge malfunction? Is there another simple circuit that would guard against failures while still offering the same performance (and uses a minimum number of chips)?
I simulated the circuit through CircuitLab with a panel of resistances for R5 and was able to confirm the equation shown correctly backs out the R5 value given VD as an input, so I’m fairly certain the equation properly represents the circuit I have drawn.