rotary torque sensor

Hi!

I want to measure torque on rotating shaft. After hours of searching i could not find any option which doesnt cost a fortune. Existing sensors with data loggers, as a complete solution, are expensive.

My calculations shows following measurements:

max. torque: 75 Nm
max. rotary speed: 100 rpm
max. power: 300

i want to build my own system for these requirements and i want to use arduino as a data logger. But the problem is, i have no idea about what i should be doing. Here are my questions:

1-The output voltage is linear to torque input. And the output voltage of sensor is +/- 5 V DC. But how am i going to know, which torque value gives which output voltage?

2-Sensor has inner resistance and some other specifications. How do they affect my end results?

3-Did you have any other idea instead of using a rotary shaft to shaft torque sensor?

Thanks a lot, every feedback and every idea means so much to me!

I've considered such a problem recently, but have not researched. One of my college classmates used this problem for his senior design project, but that was 15 years ago. He had the same problem, the only commercially available solution was very expensive. I don't remember what he used in his final solution.

My thoughts: Use a strain gauge on the shaft. Circuit to measure the strain gauge is also on the shaft--both rotate with the shaft. Power the circuit using a wireless charger. Communicate the measurement with a bluetooth module or some other wireless device. All of this should be sufficiently small so it does not create a balancing problem and to prevent damage to the torque measurement hardware.

This is all hypothesis and I have done absolutely NO testing or prototyping of any kind.

Maybe this will help jumpstart some ideas for you.

If this is anywhere near correct: Online calculator: Power into torque, and vice versa, you are way off on the power value.

My solution would be to couple a DC electric motor to the shaft. Apply power to the DC motor to rotate in the opposite direction to your shaft. Adjust the current to the electric motor until it is able to slow your test shaft. Measure the motor's current and voltage at that point and then compute the power and then the torque needed to slow your shaft.

Heat generation will be a problem, so the test needs to be done rather quickly.

Paul

You didn't say if it was rotating or stationary torque reading?
Also, what kind of torque are you talking about?
A car engine puts out torque as does a servo.
One would use completely different methods.
If it is a static torque, you could use the part they use in a
cheap bathroom scale.
Rotating torque can be measured by having two pickups at the
ends of a shaft that has some spring. If there is torque one end
rotates such that there is an increase in phase angle between the
to pickups. Magnets and hall effects make sense here.
Calibration is easy, torque is measure as some weigh at some
distance. Put an arm on your measuring setup and hang a
weight on it.
If measuring inch-pounds, say 10 inch-pounds, use a 5 inch
arm and 2 pound weight.
Dwight

@flyboy:

i have found something similar. It is called DMS - "Dehnmessstreifen" and they use usually 1 or 2 channel telemetry for that. But the application seems a bit difficult. I will look into it. I have doubts. If it is really cheap and still accurate, why dont we see DMS applications everywhere? Are you familiar with DMS applications? What is best for my case? whats the difference between half bridge and full bridge strain gauges? And still, calibration is a main problem.

@Paul_KD7HB:

Actually i have thought about it, but i am still not sure how i measure it correctly. I mean DC electric motor doesnt have equal input and output. I could measure the rotation speed and do some calculations instead of adding another motor to do same calculations. Do you have any idea how i eliminate the deviation and possible measuring errors? I have seen some DC eletric bike motors with integrated torque sensor but i have no idea how do they work. What do you think, are they usuable for this task?

By the way, only one shaft has that rotational speed value. Most of the sensors have limited rotational speed values. For this reason i wrote only maximum values of different shafts which i want to test.

@dwightthinker:

i am talking about rotating torque readings. Your idea sounds interesting. But i have several possible problems about this one. All these shafts are rotating on water. It is kinda hydro electric generator which floats on water. It isnt stable because of that. I cant imagine if i can eliminate it. But i will look into that too, maybe i can find a way. Otherwise i can use it while i test it in my garage.

I haven't found a solution yet, but i am thankfull for all these ideas. I will look into these. But i can use every new perspectives. Don't hesitate to write or ask.

If you are using a generator, measure the voltage and current
output. You'll be able to back calculate the torque.
Dwight

I want to measure torque on rotating shaft. After hours of searching i could not find any option which doesnt cost a fortune. Existing sensors with data loggers, as a complete solution, are expensive.

That doesn't surprise me - its hard measuring the force on something you cannot connect to.
You can get fairly cheap wireless modules, but you'll have to do the design yourself and probably
need strain guage(s), amps, battery and wireless TX all in miniature packages...

My calculations shows following measurements:

max. torque: 75 Nm
max. rotary speed: 100 rpm
max. power: 300

75Nm x 10.5 rad/s = 705W, not 300W

If it has a shaft of some length, two magnets and two hall zero crossing sensors
could be calibrated to measure running torque.
The twist of the mounting can be used to measure torque.
Dwight

hi wiilii

im working on a similar project, to calculate torsion, you might find the following links useful:

http://www.amazon.com/Seeedstudio-Grove-Ir-Distance-Interrupter/dp/B00VB2ELQU

Have a look...which arduino board are you using??

Emmanuel

Hi,
I know its a while since any posts, but has anybody thought of load sensors in the motor mounts.

Four mounting points, four load-cells, motor torque will cause differential loads due to reaction torque of motor casing against its rotor.

Just a thought.

Tom..... :slight_smile:

The magnets idea is interesting, but I doubt very sensitive.

Perhaps having a reasonably high resolution encoder on each end of the shaft would
be much more sensitive to small torsional strains (a large torsional strain might be
a bit risky, allowing for shaft failure, unwanted torsion oscillations).

You would then read from both encoders and record the difference in counts, which would
provide a discrete indication of torsion angle and hence torque.

The down side is that any shaft capable of 75 Nm is probably too fat for any cheap encoder
(5mm and 6.35mm are common sizes for cheap encoders).

I am facing with this problem too.

MarkT:
The down side is that any shaft capable of 75 Nm is probably too fat for any cheap encoder
(5mm and 6.35mm are common sizes for cheap encoders).

For that you can always use a gear and do a indirectamente reading with the encoder attached to the smaller shaft

Backlash and rigidity considerations would make that problematic I think.

MarkT:
The magnets idea is interesting, but I doubt very sensitive.

Perhaps having a reasonably high resolution encoder on each end of the shaft would
be much more sensitive to small torsional strains (a large torsional strain might be
a bit risky, allowing for shaft failure, unwanted torsion oscillations).

You would then read from both encoders and record the difference in counts, which would
provide a discrete indication of torsion angle and hence torque.

The down side is that any shaft capable of 75 Nm is probably too fat for any cheap encoder
(5mm and 6.35mm are common sizes for cheap encoders).

So just reading this I just wanted to add something. The magnets idea is not a bad idea and they are very sensitive. I know this since I am also researching torque measuring devices and the magnets are used quite often. The are used in a device called hall sensor which are used for measuring change in angular position of a shaft by determining the change in magnetic flux produced by the magnetized shaft or magnets attached to the shaft.

I have read through this topic and I also am faced with a problem.

Just to paint a picture. I am a fourth year mechanical engineering student from Stellenbosch University (South Africa). I am busy with my final year project which is as follows: "Develop a system to measure the torque on the shaft of a turbocharger".

While in essence it doesn't sound to hard, there are quite a few obstacles. Firstly the systems sound preferably work up to 30 000 rpm which eliminate using slip rings or any contact system. Secondly there isn't much free space on an assembled turbocharger for mounting any systems.

For this reason I have simplified the system by not going up to 30 000 rpm but maybe near 3000 rpm. Also I am going to have my own shaft machined and fitted with the compressor and turbine from and existing turbocharger.

So after all of this I actually need to get to my questions. I thought of using to gap sensors (or optical encoders), one of each end of the shaft, then by measuring the difference in time that it takes for each sensor to get a reading and having the rotational speed, one can determine the torsion angle of the shaft.

In theory this seems to be a viable option. Does anyone have any advice for me on how to do this. I do not have much programming skill and really want to use an arduino. so I need info on how fast arduinos can log data and what the max interrupt speed of and arduino is. Also what functions should I used to log the time at which the arduino receives the sensed information?

If anyone has any advice I would deeply appreciate it.

m.kapnoudhis@gmail.com

The shaft has to be slender to allow significant twisting at the torque levels used, which of course
adds elasticity in the system, which may have consequences. But it also has not to break or whip!

I suspect metal is a poor choice, perhaps delrin is a reasonable material - machinable, much lower
modulus, so can be a more reasonable proportions. However turbines run hot, so this may be
incompatible...

Another approach would be some sort of torsional spring arrangement joining the shaft ends.

we need some either pics or drawings

Sorry to grave dig.
Could you not use a liner effect?
Have a shaft off the rear of the motor through a bearing and out to a second bearing a specific distance on a swivel; then measure the length from the motor to the end of the shaft?

Hey. I'm going through exact the same situation here.

Have you got any solution for measuring rotary torque with a gap sensor or an optical encoder? Or did got follow any other method for measuring this?

Looking forward to your reply/suggestion.