# Strain Gauge - Wheatstone bridge DRIFT

I have a 350 ohm bridge with 04 fixed resistors, later on one of them will be replaced by a strain gauge sensor. In the meantime I am trying to understand where the drift is coming from.

This bridge is connected to an in-amp with 1000x gain. In order to calibrate the bridge and unbalance it (so I can have, say 1V output at the in-amp output) generally I use a very high resistor (around 500k ohms) in parallel with one of the resistors on the bridge. These 350 resistors are 0.02% tolerance and 0.2ppm/C. The opamp is a AD8571 (extremely low drift and offset).

Unfortunately I have been struggling for months with a drift that shows in several different ways:

1. Temperature - Only by blowing some hot air briefly, the output varies +/-10mV
2. When the voltage output is increased by any external factor that is removed afterwards, it never goes back to its initial value
3. There is a voltage drift without any temperature change, I managed to have a very precise temperature sensor mounted on the top of the circuit to confirm.

All components are very high quality, spec'd for this particular application, low drift, low offset and really expensive. I do not know whatelse I can do to get rid of this annoying drift, unpredictability and lack of repeatability.

• What is your excitation voltage? Are you exceeding some max current spec? Oct 29, 2016 at 23:35
• That means the input varies by +/-10uV, right? (given 1000x gain). So, 0.2ppm/C * (guessing, you don't tell us) 5V excitation * (say) 10C = 10uV on the input ... isn't this drift exactly what you expect, or am I missing something?
– user16324
Oct 29, 2016 at 23:45
• Excitation voltage is 5V. The variation at the input given the temperature drift is correct so the question is, how to get rid of it since the lowest TC I found is 0.2ppm/C ? The feedback resistor is 200k ohm One more thing: When I force a heating up situation (from 25 to 45C with a hot air gun) it increase the output voltage but when the temperature goes back to its original value, the output voltage is still offset by a certain amount that never decrease. Oct 29, 2016 at 23:49
• An AD8571 is not an InAmp so what are you using? What generates the 5V also? Show your circuit! Oct 30, 2016 at 0:02
• The AD8571 inamp needs a feedback configuration to have its gain set. I am using a power supply to generate the 5V. So I have a 200k from out to IN- and same 200k from IN+ to GND Oct 30, 2016 at 0:05

You may be seeing thermal EMFs, as well as thermal drift of the resistors.

Blowing hot air with a hot air gun is not the way to test a circuit like that. Use a proper environmental chamber with a fan inside and control the air flow over the board (maybe with a bit of foam). You may have to wait an hour or two for the temperature to stabilize. The hot air gun will lead to wrong conclusions. You may see dynamic changes that even out as the gradients disappear.

Do not have anything on the board that dissipates a lot of power (I assume you are doing this).

Leaded resistors can be a bit less prone to change from external mechanical forces. Try flexing the board a bit and see if your reading changes. Chances are this is not your problem, but it's worth a check.

• It changes a little bit of the output voltage, it sits on a higher value and does not seem to go back. What on earth am I missing here? After so many years in the design industry, this is the first time I come across this issue. Oct 30, 2016 at 0:36
• Did you assemble the board yourself? How was it cleaned? 350 ohms is not a high impedance, but leakage of 50M would cause the effects you see. No clean flux is a particular problem. Oct 30, 2016 at 0:51
• Truth been said, I have not taken any further care regarding to cleaning the board. I populated the SMD components myself and due to it is a testing subject, components are replaced very often. Oct 30, 2016 at 0:55
• Check the solder you are using, this can be a tough problem to solve. Oct 30, 2016 at 0:56
• Should I ship the board to Toronto ? =) Dear Spehro, any solder you may recommend? Oct 30, 2016 at 0:59