I'm building an open source project called Exbow (http://www.ecs.umass.edu/exbow/). This exoskeleton uses a servo to flex and extend the user's elbow. The user can control the angle of elbow flexion using a flex sensor on the back of their index finger. I want to replace this flex sensor with strain gauges. Another project called EduExo (https://www.eduexo.com/) uses a load cell attached to the forearm to accomplish this sort of user interaction.
Would it be better/easier to slap a load cell on there or wire in a full wheatstone bridge strain gauge circuit?
If you look at the two projects the one using a load cell has a break in the forearm that the load cell runs across. The Exbow just has a solid aluminum bar so if I were to use a load cell would I have to cut that aluminum bar? Instead of using the straight bar load cell (https://www.sparkfun.com/products/13329) could I use the s-type load cell (https://www.sparkfun.com/products/14282) to make it more compact?
Is the 10 kg load cell necessary? Could I go for a smaller one or larger one?
The load cell is a wheatstone bridge with 2 strain gauges.
From the description: "These load cells have four strain gauges that are hooked up in a wheatstone bridge formation."
I've never set up strain gauges before so I don't know how difficult it would be, in what orientation they need to be placed on the bar or if the bar needs to be cut. My own strain gauge circuit I think would be more compact, which is better for this type of application but if it takes too much work then I think the advantage of a pre-made load cell would outweigh the advantage of a compact circuit.
First get it working, then start optimizing.
Well it's pretty hard to argue with that logic I went ahead and ordered a load cell and a hx711 breakout board from SparkFun yesterday. I was debating whether or not to purchase some strain gauges as well but decided against it.
If I do decide to go for them in the future I'd still like some advice on how to place them. Should they all be facing the same direction, do they need to be placed at right angles to each other to help with different directions of strain? The forearm piece is a 1/8 inch thick by 1 inch wide aluminum bar. Would I need to modify it for the gauges to pick up strain better?
You need spectial glue to apply the gauges (the producer of the gauges sells this, too). You can use a 1/4 or 1/2 bridge and some precision resistors to complete the bridge. But I would davice to use fullbridge gauges - they cost more, but work as intended. When you look at the gauge, you see a lot of lines. Forces applied in parallel to these lines are detected. When you buy a fillbridge, you'll see 4 gauges oriented like a cross. With this assembly you get rid of temperature effects and can calculate the force vector in 2D. You can glie the gauge directly onto he object you want to measure, just watch out for the wires, they are brittle to say at least (embed them into some flexible material)
That's helpful. I wasn't sure in which direction the strain gauges measured force. If I build or buy my own I would definitely go with a full bridge. I saw a video of a guy hold a gauge near a lighter and the resistance changed a lot. I can imagine what a sunny day would do to a quarter Bridge circuit.
While I'm waiting for my shiny new load cell to arrive I've been trying to find code examples that'll let me know which dimension strain is being applied in. I want to make sure I can tell if the user is flexing or extending the elbow.