Load Cell Check

try calibrating by removing a single load cell a time

You might try contacting Crown Instrumentation. They manufacture load cells and also calibrate them. Someone over there should be able to provide the necessary information to solve your problem.

1. As I understand it [not being a weighing specialist], a weighing system consists of a plate on which the item to be weighed is placed. The plate is supported at 4 or more points (8 are mentioned in the "packer", but I cannot tell if a packer has two plates, each on 4 points or one plate on 8 points.
2. Each support point has a force transducer (cell). The total weight of plate + item must be mechanically equal to the sum of the forces upon the force transducers supporting them [and, hopefully, the sum of the cell readings will be the total weight].
3. If each force transducer has been calibrated accurately for kg per mV [and they are linear] and the sum of their readings has had the correct number deducted, the scale reading with the plate fitted (without weigh item) will be zero. The weight of an item added will be correctly shown.
4. This is all OK in theory. In practice, all the load cells must have their sensing axis vertical. They should all be touching the plate with no item on the scales (although this may not be essential for accuracy, impact loads are not desirable).
5. But if the dimensions of the plate distort, the force at each support point will not be equal. Similarly, if the bottom supports of the cells sink unevenly under load, the weight distribution between cells will change.
6. Is the plate being removed to test the cell concerned? Is it being distorted by being unequally heated while it is off? The the distortion would settle out again as temperature settled to normal operation.
7. But if 5. and 6. apply, the readings of the cells in other positions will change, hopefully such that the sum is correct.
8. Steel may be nearly an ideal elastic material, which instantly changes its dimension to its final value when load is changed. But I do not believe this ideal applies to compressed wet rust or concrete, which may be under the cells . I am sure concrete/rust compression/recovery will have hysteresis and time lags.
9. So my essential question is - how are the readings of the other cells changing while you are focused on this one cell position with a 17 kg creeping value??
   

Hi Zain, You probably need to monitor temperature changes during calibration and afterwards. Temperature variations cause loading on the tank when either the pipeworks or the tank is not allowed to freely expand. Cheers

Load cells are temperature compensated which means that even if temperature changes the signal does not change. In this case it appears that the sensitive part of the load cell (the strain gage bridge which is the bridge impedance?) is temperature sensitive since only time is enough to change output. Load cell being under supply the current through the strain gages heats them. If the temperature compensation is not working properly then you get the effect you mention. A first check would be to take the LC out of the working fixture and see if output varies versus time same way. This would confirm the failure of the compensation. There can be an other reason too. The question is has the concerned LC same sensitivity as all other LC or does it have a different one ? The last question is about the type is is a LC with integrated conditioner or not ? The best would be you supply the schematics of the connection and a picture (or better several) of the way the LC are put in place.

If it is just temperature, one should warm each cell, in turn, with a hair dryer and compare responses.

Behaviour is NOT the same if the strain gage bridge is heated by the supply current or if the whole LC is heated by an outside source. You have to think about where the heat is generated and how it spreads in the support. In fact it is what happens, LC are under supply without load and the local heat generates local elongations which are not as the one present when the compensation was done.