Struggling to find a way to measure rotary torque

In summary, the conversation discusses the struggle to find a way to measure instantaneous torque from a servo motor in an accurate and inexpensive manner. Various methods are proposed and discussed, including using strain gauges and a modified arm, hanging known weights from the arm, or using an accelerometer. The idea of using a six-point kinematic mount is also mentioned. However, the use of strain gauges on the output arm or servo mounting brackets seems to be the most feasible solution.

Homework Statement

I have been struggling, for over a month now, to find a way to measure instantaneous torque from a servo motor. Here's a servo that look similar to the one I'm using: https://www.servocity.com/html/spg805a-bm_standard_rotation.html
I need to find a way to measure the torque in an accurate, but inexpensive manner-which means a torque transducer wouldn't work for me. In my project, there will be times where a nonconstant force will be opposing the motion of the servo (just imagine that a rod, mounted horizontally, is attached to the shaft of the servo, and someone applies a force against the motion of the rod) i.e., the torque of the servo/shaft isn't always constant. Moreover, the speed of the rotation will be different at times as well. There has to be a way to do this, right? I am reading data in through an arduino, connected to the servo, and am free to use sensors. At first I wanted to measure the current uptake but don't think it's going to be accurate. Furthermore, even if I did go with using current, I'd like to validate my results through other/more direct means. Is it possible to somehow incorporate a strain gauge or something, and attach it to the shaft somehow to measure the torque directly (I'm not really sure how strain gauges work)? I asked my friend about this and he talked about using compressible "blocks", with a spring attached in the middle, and have the ends of the springs attached to two strain gauges. The spring might compress/relax differently according to differing levels of torque apparently. I couldn't really understand his idea, nor could I understand how such a thing could attach to the shaft, but he said it's something of a classic/old technique used for measuring rotary torque. I also pondered of measuring the acceleration (to calculate torque) using an accelerometer would work, but I guess I wouldn't be measuring the electrical torque of the actual servo at that point? I don't even know..

I'm really new to all this stuff (and still bad at it) so I apologize if I'm not being clear about anything. I'm really just looking for a clear/logical, sensible, and most direct way/idea I can implement to measure torque-even if it's not easy. I'm willing to learn and put in the effort.

The Attempt at a Solution

accurate, but inexpensive
The killer specification with which every project begins.
free to use sensors
How much room have you got for a motor mount? If there is room enough for an honest six-point kinematic mount, you can tie
a strain gauge
to whichever of the six constrains rotation.

Perhaps turn the servo on it's side and hang known weights from the arm. That would provide a known torque with which to calibrate the current. However you would have to do that at various speeds. If the load is light speed might be under your control (eg dictated by the signal sent to the servo) or will it be a high load, in which case the speed might be dictated by the load?

You could use strain gauges on a modified arm but they would also need to be calibrated (but at least you wouldn't have to compensate for speed).

Bystander said:
The killer specification with which every project begins.

How much room have you got for a motor mount? If there is room enough for an honest six-point kinematic mount, you can tie

to whichever of the six constrains rotation.

Well I'm not sure what an "honest six-point kinematic mount" is (sorry if that's dumb)? I tried looking this up on google and am not sure what to look for. Can you post a link of a place that sells these so that I can see what they look like/what the dimensions are? I may have room but I need to see it to be sure. Also, when you say tie the strain gauge, would the way I actually tie it not matter?

CWatters said:
Perhaps turn the servo on it's side and hang known weights from the arm. That would provide a known torque with which to calibrate the current. However you would have to do that at various speeds. If the load is light speed might be under your control (eg dictated by the signal sent to the servo) or will it be a high load, in which case the speed might be dictated by the load?

You could use strain gauges on a modified arm but they would also need to be calibrated (but at least you wouldn't have to compensate for speed).

That's what I thought about doing with the current myself. But I want to check it via another method because I'm not sure how accurate that would be.
I'm not sure about the load. I think it may end up being significant, but I'm also using a really powerful servo (222 oz-inch with a 3:1 gearing ratio). Shouldn't most of the servo's speed be maintained unless it's being countered by a powerful torque? I would estimate the load wouldn't exceed 5-7 kg max.

Can you clarify what you mean by "use strain gauges on a modified arm"? I'm having a hard time picturing this (I don't have experience with strain gauges, unfortunately). Calibrating may be annoying but isn't a problem. At this point I'm looking for anything.

Most strain gauges in effect respond to bending or stretching of the object they are mounted on. One way to measure torque would be to mount strain gauge on the sides of the output arm like this or put them on the servo mounting brackets.

"Kinematic mount?" Do you have room enough to allow "rotation" of the motor mount, and provide the reaction force through a lever constrained/clamped between two strain gauges?

CWatters said:
Most strain gauges in effect respond to bending or stretching of the object they are mounted on. One way to measure torque would be to mount strain gauge on the sides of the output arm like this or put them on the servo mounting brackets.
View attachment 77762

Well I'm a bit confused by your picture. Are you proposing I attach some sort of rod or arm to the shaft of the servo, and then put strain gauges on the sides of that? And by servo mounting brackets, do you mean the areas that the arrows are pointing to in the picture I uploaded? I just want to make sure I'm fully understanding you, and all this terminology is a bit new to me, sorry.

Attachments

• servo_mount.png
7.1 KB · Views: 533
Bystander said:
"Kinematic mount?" Do you have room enough to allow "rotation" of the motor mount, and provide the reaction force through a lever constrained/clamped between two strain gauges?

Oh, I think I get what you're saying. So you want me to attach a rod or some lever to both sides of the bottom mount of the servo, and then have a strain gauge on each side. But how would this give torque? When would the lever stretch?

Well I'm a bit confused by your picture. Are you proposing I attach some sort of rod or arm to the shaft of the servo, and then put strain gauges on the sides of that?

Yes. At least that's if your application needs an arm.

And by servo mounting brackets, do you mean the areas that the arrows are pointing to in the picture I uploaded? I just want to make sure I'm fully understanding you, and all this terminology is a bit new to me, sorry.

Yes that general area although not exactly position... One way to measure the torque of an electric motor is to measure the equal and opposite torque that the motor casing puts on it's mounting. In this example...

http://www.modelflying.co.uk/news/article/measuring-motor-output/2363

... the author has mounted the motor so the casing can rotate a few degrees (the propeller turns one way and the motor casing tries to turn the other way). He has fixed an arm to the motor casing and arranged for the other end of that arm to press on a set of scales. Torque = force * distance (where force is measured by the scales and the distance is the length of the arm).

I'm not suggesting you do it exactly like that but the principle is the same. Instead of scales you would need to measure the stress in the mounting using a strain gauge. The problem is you can't just stick strain gauges on the servo brackets because they would be in the wrong orientation.