load cell output magnitude depends on which side force is applied

Also asked at electronics.stackexchange.com

I have a generic load cell (image) and am using Hx711 which is connected to arduino. The code is from Instructables.

When I fix and apply the load as in case 1, it shows positive force. But when I do as in case 2, it shows negative force. I haven't flipped the part just rotated about Z axis. Here is my question: Shouldn't both cases show same force considering region of 'strain gauge 1' is at tension and region of 'strain gauge 2' is at compression? Am I missing something?

You are correct IF the strain gauges are mounted where you think they are and they are strained as you think , it’s not so obvious if you look at the actual cell that it aligns with your drawing as the gauge may not be directly over the cutout where the beam bends

What you have is a beam type single point or platform load cell. The image link you posted shows a small label on one end with the capacity and the direction force is to be applied using a small arrow. While it will output both ways it is designed as a compression type. Anyway what you are seeing is not unusual.

Ron

It measures shear force, not bending moment. There are 4 strain guages altogether, one for each
thin section of the bar. Diagonally opposite guages experience the same sign of force.

Think of the bar as being a parallelogram linkage, with the thin sections being elastic hinges. Your
analysis would work if it was a solid bar with only 2 guages (that would indeed only measure bending
moment).

[ well technically it measures deformations (strains) due to shear force, not the force itself ]

Ron_Blain:
While it will output both ways it is designed as a compression type. Anyway what you are seeing is not unusual.

Ron

But why the change in magnitude depending on which side is being held?

It's not a compression type load cell. It works like this:

So when you switch support ends the gauges in tension change to compression and vice versa.

Thank you @DaveEvans.
@MarkT when you mentioned about shear, I looked a bit deeper. I did some simulation and here is my answer.

Shear vs bending

The main reason for the change in signal magnitude when held at left and force applied at right vs when held at right and force applied at left is because it is a shear beam rather than a bending beam.


This is a cad model of shear beam.


held at right and force applied on left


held at left and force applied on right

In the first simulation figure, left is held at stationary and force is applied at right, if we look at the upper region, the region above right hole has convex shape i.e. lengthening of the region and the region above left hole has concave shape i.e. compression of the region.

In the first simulation figure, right is held at stationary and force is applied at left, if we look at the upper region, the region above left hole has convex shape i.e. lengthening of the region and the region above right hole has convex shape i.e. compression of the region.


This is a cad model of bending beam.


held at right and force applied on left


held at left and force applied on right

Whereas, in case of beam with no holes, we can see that in both cases, the upper region is in tension i.e. elongation and lower region is in compression.

In case of shear beam, the strain gauge is above the hole region. The elongation and compression of region is dependent on which side it is held and force is applied. But for bending load cell, it is not the case.

Great images - yes, that shows it nicely - I had to think about it a bit but the simulation confirms my analysis nicely - note how there's a slight bending on top of the shear, but its secondary in magnitude and with the guages
wired up right it won't affect the electrical signal, which makes the cell pretty insensitive to torques due to clamping.

It's also the reason that scales can be accurate regardless where the load is placed on the tray or platform.

(Off center but in line with the cell changes the moment on the cell, off center but to the side changes the torque, but the cell is insensitive to those effects, because it reacts primarily to shear. )

@MarkT Glad you liked the pictures and have confirmed your suspicion.

DaveEvans:
It's also the reason that scales can be accurate regardless where the load is placed on the tray or platform.

(Off center but in line with the cell changes the moment on the cell, off center but to the side changes the torque, but the cell is insensitive to those effects, because it reacts primarily to shear. )

Being a new user of load cell, I hadn't really thought about the mechanical design of load cell. IDK why but I had this conception that it was based on bending beam rather than shear. Earlier I had this question about the difference in reading that would be caused by when load is applied 1cm from edge rather than 2cm from edge. But I thought that since it's calibrated so there would be no issue. But yeah finding about this issue and thinking a bit deep helped. It was only when you guys mentioned about shear, it clicked. I just made some faulty assumptions for no reason which gave me wrong impression of bending load cell.

The actual design using two drilled overlapping holes is the really neat detail, whoever thought of that
was inspired that day!