Measuring force developed by a robot

I've been playing around with a Lego Mindstorms NXT robot, gearing it for speed and power. To my dismay, when geared for power, it wasn't any harder to stop than when geared for speed -- I suspect that it doesn't have very good traction (as there really isn't much mass pushing down on the wheels).

It would be nice to be able make quantitative statements about how much force (and thus, how much torque) a robot puts out. Does anyone know how I can do this? Would, for example, a spring scale or force gauge work well for this purpose, and if so, what is a good maximum rating to look for?

• If you're into building robots, don't forget to commit to the Robotics Proposal. Sep 30 '12 at 17:32

I've always used a standard spring kitchen scale rotated. Put the scale between the robot and a wall (or another fixed object) and let the robot push...

BTW: This is one of the tips that are also suggested in the robotic sumo chapter of "Building Robots With Lego Mindstorms" book by Mario and Giulio Ferrari (two italian RCX gurus). So, it's really "in topic" with the original question.

• I finally found our kitchen scale, and it looks workable. Didn't notice your edit until just now. Thanks for the idea. Dec 18 '09 at 0:38

A cheap way to do it would be to have the robot pulling on a string which runs over the edge of the the table (or to a pulley), and gradually add weights (such as washers) to the end of the string until the robot can't lift any more.

It would at least provide a starting point, but if your Robot has poor traction you don't want the weight of your Newton Balance to affect your results in either direction, if possible. There are devices which measure breaking strain on ropes etc (I'm not sure what they're called), they measure up to a value but don't decrease until you reset it, that might be more use, but I don't know whether they come in a small enough (or accurate enough) version.

Edit: Strain Gauges?

Edit(2): Maybe a Secondary School Physics Textbook might have some suggestions for experiments that could be adapted to measure this.