OpenServo.com

[john493]

One thing I want to do with Openservo is use it to measure the load on the various joints of a hexapod. As a minimum, this would give a way to find out whether a leg has met the ground yet. I am also hoping to do finer things than this, such as work out where the centre of gravity is and hence improve balance, especially when doing tricky things like walking on four legs.

I made a simple test rig with a balanced metal beam screwed to a servo (Hitec HS422), some holes at distances from 60-100 mm from the centre (in 10mm steps), and some 180g and 270g weights which I think were made for something to do with fishing. I was using the OSIF software to drive the servo.

The results were good but not as good as I'd hoped. I do see variation with load, not linear but not too far off. However the thing that surprised me is how much the current varies with previous activity. For example with a 270g weight at 80mm, I was seeing a reported current of about 65-70. But if I lifted the weight and gently released it, this would go down as low as 20 (and the ammeter in the circuit agreed).

It occurred to me that this MIGHT be due to the absence of the I term in the production s/w I'm using, meaning that in one case it is still trying to get to the right place, with a substantial P term, and in the other it is already there, thanks to my intervention, so the P term is less. If that's the case then a stronger servo should show a different result.

I could also try different software, though truth to tell I'm reluctant to try reflashing my one and only functioning Openservo board. So I'll probably leave that until I have another one.

John

[dustynrobots]

I'm trying to do the same thing, and just stumbled upon this open servo project. How does open servo sense current? I'm using a standard servo with a resistor in series between the black (ground) wire of the servo and ground. Then I sense the analog voltage on the high side of the 1 ohm resistor and read it through an analog pin on an Arduino. This works, but the current is super jumpy. What's I'd like is to sense it directly from the motor so I get a nice current vs. torque line (like in a regular DC motor), not just a current spike every time the motor moves. I've tried hacking into a standard servo to get right at the motor, but when I connect to ground through a resistor the motor goes nuts!