Servo coupling testing method? Need help!

[jberge1]

Hi all,

I've been a lurker here for maybe 10 years. Run into an application that I'm struggling to even begin, and I think some of you have the expertise to point me in the right direction. I have a client who is looking to test a a line of prototype mechanical servo couplings. They want to see how much torque/RPM capacity each of their couplings have.

They had reached out to a number of outside testing companies but were quoted anywhere from $10k-100k which was way outside their price range. They are looking for something they can use in-house to test their prototypes but without the high price of an outside company doing it for them.

My background is mechanical design, and I've worked for a number of machine builders, but this is a little outside my comfort zone. I am considering having a variable speed motor with a known torque value which exceeds the estimated failure point of the coupling. Then I could attach the other end of the coupling to a pneumatic clutch of some sort, and regulate the torque with air pressure. It doesn't seem terribly scientific and I'm not even certain it would work. Can anybody recommend a low cost, possibly all-mechanical means to regulate torque with some type of user input, while also regulating RPMs with another user input? It would not need a fancy HMI but some way to repeat the same test would be fantastic.

Thanks,

Jberge1

 

[tbuelna]

jberge1-

You did not specify what capabilities your test rig requires (speed, torque, duty cycle, instrumentation, etc.) but the cleanest arrangement is usually some combination of electric motor and absorber. A DC motor and generator with the coupling located between them would allow precise control of load and speed, and would also be very efficient to operate since the absorbed power can be returned to the grid.

If mapping the dynamic characteristics of the coupling is important, then you will need to consider how variability in the motor/absorber dynamic performance might affect test results. For example, a DC motor and hydraulic absorber will have a different torsional vibration environment due to the dampening produced in the hydraulic absorber.

Lastly, most mechanical shaft couplings are designed to accommodate some amount of radial/angular misalignment between the mating shafts. You might consider making your test rig articulated, so that you can test couplings under various conditions of radial/angular shaft misalignment.

Hope that helps.
Terry

 

[MikeHalloran]

The first dynamic test of any interest to me would be the speed at which the coupling explodes. An air turbine and some kind of tachometer would do just fine for that. ... plus the armor plated test room, or at least a really good scattershield, or a setup in the middle of nowhere, provided that nowhere is at least a couple thousand metres wide.

Next I'd want to know the torsional stiffness, which could be measured with an arbitrarily large degree wheel and a torque wrench.

I'd want to know the torque capacity, which for low numbers of cycles would be the torque to yield, same apparatus as for stiffness.

For higher numbers of cycles, you might be able to estimate a life at a given torque from detailed knowledge of the design. If you want to measure it, then, yeah, you need multiple dynamometers to provide test data from a statistically significant sample of production units, or at least an array of motor/brake setups plus periodic measurement of stiffness and backlash.

The last number of interest is the moment of inertia, which can come out of a solid modeling system, or can be measured with test instruments, some of which can pump out a MOI unattended in a few seconds, I think.

That's the set of measurements I'd imagine would be important to users and manufacturers of servo couplings. There are probably more, but your customer should be able to provide a better list. Then you can estimate the cycle time for each measurement, how many test stations you need to get reliable data, and how big a staff you need to keep it going, stuff like that.





Mike Halloran
Pembroke Pines, FL, USA

 

[MintJulep]

They had reached out to a number of outside testing companies but were quoted anywhere from $10k-100k

A ten-fold range of quotes strongly suggests that the scope work the quotes were requesting is hopelessly ambiguous.

Mike's list of tests sounds about right to me.

There is absolutely no way that you will be able to design, build and validate the required test equipment from scratch, and then do any meaningful tests for less than $10K.