I have 4 50 kg load sensors connected to a OpenScale. However, having trouble getting a good calibration. I have some weight readings but drifts have been significant.
Any help would be appreciated. Thanks!
What does your weighing apparatus look like?
Can you share a photo of your wiring?
What’s your calibration method? Here are some tips OpenScale Applications and Hookup Guide - SparkFun Learn
Thank you for your quick response!
Here is an image of the weighing apparatus and the associated wiring. It is then connected to my laptop via the USB cable.
I’ve been following the recommendations of:
- Tare the scale
- Add reference weight (I’ve tried both a 2 kg weight and a 23 kg weight)
- Wait 30 minutes
- Calibrate
Those three wire half-bridge(?) load cells need to be wired to make full bridges.
Can you put some numbers on the drifting?
How high, how low, how fast, what load?
The OpenScale already has a load cell combinator built in so I don’t believe that step would be required. (See OpenScale Applications and Hookup Guide - SparkFun Learn)
At the 23 kg load after 24 hours it had drifted almost to 90 kg. Here is a picture of when it was first working.
You need holes in your platform so the sensors have room to flex. Square holes a little smaller than the outside dimensions of the actual sensor will do.
Check this post for more info.
I’ve created 3D-printed load cell feet with an outer rim to allow flexing and address issues I faced with another scale, but I’m encountering similar problems.
I’m not seeing those feet in your image, you definitely need either the feet or holes in your platform.
Also, I do see a lot of flux residue where you soldered the load sensor wires to the openscale. Although it doesn’t seem so, flux is slightly conductive and with the teeny tiny voltages the load sensors produce, slightly conductive is enough to cause big issues calibrating.
I’d also clean all that flux off with some IPA and make sure the board is completely dry before attempting another calibration.
1 Like
Thanks for the recommendation! I will clean that up a bit more and then try another calibration.
My mistake, they have a bunch of scale boards.
As said, it’s probably mechanical, those half bridge planar beam load cells (surplus from bathroom scales, I think) are low profile and inexpensive but have some mounting headaches and they are never going to be great performers.
One technique you can use is to just flip the platform upside down and monitor the weight/AD counts. Make sure they’re stable and return to about the same start value after you manually press/load & release each load cell. If so, you can then put a second platter/boards across the feet and better see the cells flex when you apply weights.