# How Can I Measure the torque applied by a servo motor?

I believe that there must be a tool that can measure $$oz-in$$ of torque. I do not trust what the servo manufacturers state on their site for torque values, so I want to test them for myself.

Can anyone suggest a tool which I can use to do this? I have used fishing scales before, but I need something more sensitive than that and my units are pretty small, such as around $$20\; oz-in$$.

Thanks.

• 20 oz-inch is equal to 1.44 kg-cm. Why are you considering that "small"? Can't you just place a weight on a scale and use the servo to lift it and then see how weight it subtracted? Just be careful of course not to leave the servo like that for too long. Commented Mar 21, 2014 at 20:17
• I don't have a bunch of different weights to check different size servos. And I really want to see how much torque the servo is pulling along its travel. Commented Mar 22, 2014 at 0:10
• A torque wrench? Commented Mar 22, 2014 at 3:27
• You forgot to add a smiley face. I was hoping to get a viable solution to this problem. I really need to be able to check to see if the servos I am using are up to the task. I would prefer to test my servos rather than relying on manuf. data. But then again, doesn't everyone want to verify their equip? Commented Mar 22, 2014 at 5:38
• @RCHobbyist, you don't need different weights, you just place one that is certainly heavier than the servo can lift, it will pull it up with the maximal force it can produce which will in turn reduce the weight information on the scale. Commented Mar 22, 2014 at 11:12

One option for you if you are willing to spend ~ 100 dollars is a particle brake. I just saw a 15 in-lb particle brake on ebay for 99 dollars. If you search, you may be able to find a better deal. You can couple a particle brake to your shaft and the torque that the brake applies is proportional to voltage. Their torque is usually pretty linear with respect to voltage.

• What's the tolerance on those brakes? It's hard to imagine you can make an accurate measurement. If you back-drive another motor you're relying on the second motor? Surely you can't determine torque from resistance, as you say it depends on the motors' torque constant which in this question is not trusted. Commented Mar 26, 2014 at 3:39
• Measureing resistance just to make sure it matches what the manufacturer says. Just a check. As long as you can get an accurate speed measurement with a strobe or tach, that's all you need. I'm not sure what the accuracy of a particle brake is. It probably depends on the size and manufacturer. For a cheap alternative to a dyno its not a bad choice in my opinion.
– Eric
Commented Mar 26, 2014 at 3:59

There are lots of ways of measuring torque from first principles. Obviously the first problem is that you need a sensor, you have lots of options:

• You can use a torque sensor. ( e.g. imada.com/torque-testers/torque-gauge ) couple it directly to your shaft. Drive your motor. Read the sensor. Your error would be the sensor's accuracy plus any errors introduced due to inaccurate coupling.
• You can use some sort of force sensor. Your scale is one example. You mount an arm on the motor and you read the force at a given distance. This gives you torque. With a scale you want to: idle your motor, zero the scale with the weight of your arm on it, and then energize the motor (pushing on the scale) and read the scale. Any zeroing error plus measurement error on the arm would eat into your accuracy. Variations on this is putting a large weight on and driving up. It's all the same principle.
• You can use a position sensor. Mount a large mass of known moment of inertia onto the motor and start driving it. Take position measurements and differentiate them to find velocity and then acceleration. From acceleration you can work back to torque given the moment of inertia inertia. This has the advantage of being able to measure torque at speed. This is basically an inertia dynamometer. You accuracy will depend on how accurately you know the moment of inertia of the system (including the shaft), you timing resolution, your position sensor resolution, friction potentially introduced by the sensor (e.g. an encoder) etc.
• velocity sensor. same princple as above.

I'm sure you know that torque is rotational force about an axis. It is a combination of distance from the center of the shaft and the force. It seems that you are being suggested all kinds of fancy tools (no disrespect to anyone), how about keeping it simple and cheap (and a bit inaccurate)? put tape around the shaft to grow the diameter (make it tight so it doesn't compress under the next step). Then put a string and a weight (pennies in a plastic bag). When your motor stalls, measure the weight of the pennies, and multiply it by the radius of the shaft. Since tape doesn't weigh much, and it's offset by the counterweight on the other side, it shouldn't skew your readings too much. It's one way to what you are looking for.

You didn't specify this in questions, but I am assuming you want starting torque, which is the highest anyway, and goes down with speed.

Good luck.

Measuring the force applied to a measured moment arm will give you what is needed to compute the torque.

I think the simplest way is to use a servo horn and a kitchen scale (if you already have one).

On the servo horn, measure the distance from the axis to the mounting hole you'll use. When that arm of the servo horn is horizontal, your measured distance will be the moment arm for the torque.

To measure the force, to use an electronic kitchen scale to measure how much lighter a weight gets when the servo tries to lift it.

Choose a weight a little too heavy for the servo to lift. With the weight on the scale, use string or something and rig the servo to try to lift it at "arm's length" -- that is, with the servo horn arm horizontal. With the weight on the scale, but servo not lifting, zero (tare) the scale. As the servo struggles to lift the weight, whatever force it exerts will reduce the scale's measured weight, directly reading the force you want, albeit negative.

If you are measuring your only servo, consider that some servo designs cannot tolerate stall. Overheating and broken gears are the symptoms of engineering to minimize weight and manufacturing cost.