I’m using the Qwiic Scale with a 1kg load cell and set it up with the complete scale example sketch on an ESP32 dev board. I tried it once before in the same configuration but took it apart and rebuilt the connections and found that it maxes out at a certain value, well below the rating of the load cell. When I initially tried the Qwiic Scale with the same load cell, it didn’t have this problem but after using it a second time, this started happening.
I don’t understand what could have caused it to behave like this. It’s the same load cell, fairly sure its the same code, but this time it seems to max out.
It’s hard to give an exact weight, but it’s somewhere around ~380g. The raw readings also seem really high, but I can’t recall what they were on my first use. The zero offset right now is 7738847, and the calibration factor is 1699.05.
It’s possible that the load cell has developed a high output where it appears to be loaded even at rest. Measure the signals/output of your load cell at no load, then at full capacity. If you happen to have another weight between those, log the output there too. These will be less than 50 millivolts. Make sure it’s proportional and repeatable/returns to zero in millivolts. If it’s not, look more carefully at the mounting hardware.
High output is sometimes caused by an overload of the load cell. If you bend one to the point it deforms, it remains bent when unloaded which looks like load/weight. Of course, it’s a lot easier to overload the 1 kg load cell than, say, a 100 kg version of the same.
I think this makes a lot of sense. I didn’t think anything had changed with the load cell, and there’s no noticeable deformation, so I didn’t try a different one but it’s definitely possible and it’s the only thing that would make sense.
You could but, if this were my project, I think I’d be most interested in finding out what changed/went wrong/whether anything wrong and this means careful measurements of the existing components as described. If the load cell ‘popped,’ I’ll want to know why before installing another one without some idea what happened. Even the cheap ones are kind of expensive to be tossed out like fuses.
It’s also completely possible (likely, even) that the load cell is fine and there just a problem with the mounting or other physical element. A tiny bit of grit under a load cell or flex in the mounts can cause all sorts of issues. The deformation for your 40% deflection is visually imperceptible.
If …IF… the load cell has a high output but still works ok (linear and repeatable) through its 1000 gram range, you could trim the output down the output with a fixed resistor. This is useful at times for common but inexpensive single-cell scales found in household & commercial kitchens, classrooms, shipping desks, etc.
I mean, the load cells that I’m using are just over $1 each. I’m okay with discarding a broken one, especially given how rare it is.
Having tried another load cell, the readings are perfectly fine now. The zero offset has gone from that enormous 7738847 down to 65986, and the weight is no longer maxing out within the rating of the load cell.