Load cell setup, weight deviates to much in time

Hello everybody,

For an experiment I had to make a scale that weighs a soil sample of around 80 kilo’s on a long term. I used 4 x 50kg load cells (3 wired) and a HX711 amp. For this experiment it is important that the measurements are pretty exact. Preferably a deviation of max 20 grams. Yesterday I did a test for 8 hours straight and I got the results as shown in the attachment. In this test I loaded the load cells with 45 kilo’s and set that state to zero. In the graph you can see that it deviated to -630 grams. The setup was placed in a closed room. If I want to go on with my experiment I have to get the deviation out or at least an explanation so I can use that as an argument. I really hope that there is someone that could help me with this problem.

Cheers,

Daniël

Ps. Pictures of the setup are also in the atachments.

Log the temperature each time you take a weight measurement and see if there is a correlation.
Temperature compensation is often required when dealing with sensitive transducers.

Hi Mik,

Thanks for your response, today my temperature sensor is arriving. I was thinking of the same, however I thought that the full Wheatstone bridge would eliminate that error. I will still be checking this. Could it also be that the power source is the problem? I am now using a power bank or an android cellphone adapter to feed the power. Can it be that the power supply has a non-constant voltage? Is there a way to get a more constant supply?

Greetings,

Daniël

Can it be that the power supply has a non-constant voltage? Is there a way to get a more constant supply?

Yes and yes. Use a precision voltage reference for the bridge.

Please describe the rest of your circuit.