We put 1 liter pet-bottle filled with water (the weight is estimated at 1000 grams) on a Morph in 3 ways.
The value is in a following image.
The picture and the value is in following 2 images.
The picture and the value is in following 2 images.
The values seems too different. How could we fix this problem? Thank you.
Assuming there isn’t an issue with collecting the numbers, here’s a few suggestions:
You may want to place the whole Morph on a flat & stable surface. Your pictures show it mostly on a stack of paper-stuffs, but with the edge hovering in the air. If the Morph’s frame was perfectly stiff it might not make a difference. But it can bend, and the Morph might be detecting this in error.
Experiment with a constant Morph interface but a varying weight. It’s possible Sensel developed the Morph as a human interfacing device and not a weighing scale. As such it’s possible that their hardware+firmware combination ends up giving a non-linear force scale. If that is the case then if you can find the conversion of Reported Force to Actual Force then you could precompensate each grid value before summing. It’s also possible that it’s linear within a certain force range, but when you concentrate the forces then it gets pushed into a non-linear region.
You might want to try summing the entire surface for your initial experiments to see if you aren’t loosing some force to many low-force values that are being trimmed out of the calculation. If there’s a threshold for how high a force needs to be before being reported, then you could have a large number of cells that aren’t being reported when you have the same mass over different surface areas.
The Morph appears to have a grid of sensors, but not with perfect overlap. If you switch into the 3D graph method and slide a small footprinted-object across the surface you’ll likely see the force bump up and down as it transitions from one cell to the next. Assuming you only care about total force and not shape, you might try placing something between the Morph and your object to be weighed - Specifically something somewhat elastic and thicker on the Morph, then something stiff on top of that. That way when the object is placed on top of the two-layer stack, the top layer will spread out the force “evenly”, and the elastic material will try to spread out the force more.
As mentioned, if you move an object around you may see the “bumpiness” of the grid show up. You might try measuring the object’s force total from multiple locations, but same orientation. Then you could average it and see if that gives a more regular measurement.
While not likely, it’s possible that the auto-zeroing of the Morph might be negating some of the force measurements. Not that likely, but perhaps possible.
We tried several experiments and figured out that a sensel sensor returns non-linear value. The heavier the real force, the bigger the value than that should be. We would like to know this non-linear curve function. Thank you.
While you can wait for someone at Sensel to respond, you should probably map out the function on your own just in case they never give you the answer you need.
A cheap electronic scale will let you produce some calibrated weights from stacks of coins with perhaps a metal or plastic disk as the base to give you a smooth bottom surface. Then you could easily compute your own curve from that.
