Load cells and overload

I plan to do some experiments with load cells to measure force. It seems that you can get some of them very cheap of about 1.5 euro each for load cells in 1 to 20 kg full scale range.

When I look at specifications, I see, that it is normal to specify the full scale load. But I know, that in an application with mechanics, you can easily apply some short term overloads due to dynamics and accidental blows. In specifications, you see specified "Ultimate overload", that is typical 150% of full scale. I have also seen 200%. I have also seen designs, that include a mechanical stop to prevent overload. Therefore I think, that it is the possible overload of the mechanics, that will provide input to the specified full scale load of the load cell you want for the application.

I have seen quite many threads in this forum about load cells and the HX711 AD converter interface. So I guess that many of you got some experience with load cells.

What is the typical failure you see, if a load cell have been overloaded?

Don't think there is a typical failure.
I seen one that could no longer be calibrated

Thanks.
Perhaps there is a typical calibration then?

My guess would be, that you record zero point and two or more know weights applied to the load cell. Then everything should still be in the dynamic range, and on a XY-chart the results should be on a straight line.

Mechanical overload bends the aluminium out of shape.
Then the bridge doesn't output zero voltage anymore, and the HX711 doesn't work properly.
I had to bend back a load cell once in a vice to get it to zero again.
Leo..