Fluctuation in 4 half bridge load cell via combinator

Question

I'm currently using a Sparkfun 4 Half-Bridge load sensor (50 Kg) with combinator and amplifier both are from the Sparkfun. Link:https://www.sparkfun.com/products/10245 (For Load Sensors)https://www.sparkfun.com/products/13878 (Combinator) https://www.sparkfun.com/products/13879 (HX711) Our application requires long-term use i.e 8 to 10 hours of load application and monitoring the weight change during this time. Is this possible to achieve the above requirement using these chips/amplifiers? As we have experienced a drift in the readings if a weight is placed for more than approx. 3 minutes. We are using 4 half-bridge sensors, a combinator, and an amplifier.The issues I am tackling right now are,

1. Zero balance: Load sensor value exponentially increases or decreases after calibration(this is a cyclic thing and keeps happening). The error ranges usually from -20g to +40g.
2. Drift in reading: When the load is placed and kept for more than 3 minutes a drift is observed which increases with time and rises up to 50 to 60g. Can you please assist in this matter? have used different libraries. Results are the same.

• jsotola: how much weight is placed on the sensors?
• Ayub: @jsotola 9.2kg.
• RowanP: Can you post the minimal code that demonstrates each of your issues? Also a photograph showing the setup?
• RowanP: What is your anticipated error range based on the sensor data sheet and amplification? e.g., creep, zero drift, noise

Answer 1

Consider a change in temperature either the cause or partially the cause for weight drift. This supplier web page talks about load cell temperature compensation. There they consider a 2000lb load cell. Their chart infers a 20 degree Fahrenheit change can cause an temperature uncompensated load cell to change about 1.5lb.

enter image description here

• Ayub: I took precautions and tried to maintain the same temperature throughout the test. So I don't think temperature is an issue in my case. Could it be the HX711? @st2000
• st2000: If this answer does not solve your problem consider un-accepting it so that others may offer alternative solutions. If you like this answer you can always up-vote it. There are many possible sources for fast and slow unexpected changes. Temperature is just one. Bad or under-rated power supplies is another. Bad voltage references. Air currents (many balances are built inside an enclosure). Vibrations (some balances are built with heavy bases). Also, you did not say over what period of time you saw variation. Or the total weight of the mass you tested with. This may help to answer you question.
• Ayub: Thanks for replying @st2000. Performed a couple of tests. The duration of the tests was between 40 to 60 minutes. 1)Zero balance test. 2)Full Scale for zero balance the fluctuation was from -2 to 9 within 5 minutes and the maximum and minimum peak experienced in an hour were -4 to +12. And the full-scale error was almost 50g over 1 hour.
• st2000: 1. Until you solve this problem, I do recommend un-accepting the my answer (and, if you like, vote it up) to encourage others to offer alternative solutions. 2. It is encouraged for you to edit your question to make it clearer as to what you are asking. I would add the test weight of 9.2kg and the above times. 3. Where the changes random (i.e. 0g, -4g, 8g. 2g ...) or were they slow showing a trend (i.e. 0g, 2g, 6g, 8g ...)?