The strain gauge I'm using has a resistance of 120 ohms, so the other three resistors in the bridge have the same value (tolerance is +-0.01)
First the good news - it works! As I apply a load to the strain gauge the voltage changes exactly as expected.
The problem is this - the ADC reads positive values only. Unloaded, the bridge gives a positive voltage, as I apply a load the voltage goes down, and soon reaches zero - at which point the ADC stops working.
The apparent solution is make some very minor adjustments to the resistances in each arm of the bridge to ensure I have a positive voltage over the whole range of loads I'm interested in. The problem is that the adjustments I need to make are minuscule (in the order of 0.01 ohms)
My question is this, how do I make such tiny adjustments to resistance? I can't buy a 0.01 ohm resistor, and if I could the resistance of the solder joint would probably far exceed that of the resistor!
The good news: Only the resistor in series with the strain gauge needs to have the same resistance, and it does not need to be high accuracy. If you use a second identical strain gauge that is unloaded, it can act as temperature compensation.
For the other side of the Wheatstone Bridge, you can use any resistors that are low enough with reference to the input impedance of the ADC. So you could use, say a couple of 100k resistors with a 10k 10t trimpot between them. Just trim it for the output you want.
Since the indication is going in the opposite direction you want, just reverse the connections to the ADC differential input.
A op-amp would give me an offset, which would be very useful, but adding an op-amp would require a complete re-build from scratch (new circuit board etc). So I'd prefer to go for a simpler solution first. (if I can think of one)
The ADC I'm using is 24 bit with a gain of 128, so I don't think I actually need the extra gain an op-amp would provide.
Only the resistor in series with the strain gauge needs to have the same resistance, and it does not need to be high accuracy
Ok, didn't realize that, thanks.
If you use a second identical strain gauge that is unloaded, it can act as temperature compensation.
Yes - since my original post I've found a few references to this - it sounds like a good idea.
Since the indication is going in the opposite direction you want, just reverse the connections to the ADC differential input.
Yup -already done this one this only needed to couple of wires to be swapped on my circuit board.
So you could use, say a couple of 100k resistors with a 10k 10t trimpot between them. Just trim it for the output you want.
Good idea, I'll try this, but there may be an additional issue I didn't mention on my original post. The whole assembly is mounted on a drive shaft (ultimate objective is to measure power) so I'm not sure if there's a danger of the trim-pots 'rattling' about and causing unwanted resistance changes.
For the other side of the Wheatstone Bridge, you can use any resistors that are low enough with reference to the input impedance of the ADC. So you could use, say a couple of 100k resistors with a 10k 10t trimpot between them. Just trim it for the output you want.
Don't these resistors also need to be very accurate with a high tolerance? From what I've read elsewhere, I need to ensure that changes in supply (excitation) voltage to the bridge don't effect the output voltage, and to do this I need to ensure that both the resistors in the other side match each other exactly (although they don't need to match the values on the side with the strain gauges in).
The other resistors need to have very good temperature stability. You may have problems with vibration if the entire circuit is mounted on the drive shaft.
You should be powering the Wheatstone Bridge from a well-regulated, well-filtered source. However, you are interested in the difference between the two, so slight changes in the source voltage will tend to be minimized.