How Can Torsional Vibration Absorbers Be Tested on Small Scale AC Motors?

[jakedehaas]

Hi All,

First time poster on physics forums, but have been looking around for a while now. Very helpful site.

I'm currently undertaking a final year Mech Eng project studying the feasibility of installing a pendulum type vibration absorber onto small scale AC induction motors (1-2kw).

The problem I'm currently faced with is determining a way to test my designs in the lab. To do this, i need a way of obtaining 4 torque pulses per rev. I have come up with a couple of alternatives and need help with determining the suitability of each of the methods.

- The first method is using a stepper motor and providing it with a sinusoidal input to achieve as many torque pulses as i need per rev. Although the uni does have a number of signal generators and stepper motors, they are currently in use and hard to get a hold of.

- The other method is putting two universal joints in series with the yoke spaced at 90degrees from each other. I've done some work in Maple with the equations for the universal joints although I am not getting the results expected. Instead of getting four acceleration pulses per rev, i am only getting two on the final output shaft. I can post up the graphs and equations if required.

I would appreciate any input into this that anyone has. Apologies if this should have been posted in the homework section. I thought the engineering section was more appropriate.

Regards,
Jake

 

[DickL]

If my memory is true, drive shafts have their U-joints phased 90 degrees apart to smooth out the rotation, which is probably why your results are only showing 2 pluses per rev. Try phasing them at 45 degrees, I'm guessing on this.

 

[jakedehaas]

Thanks for the reply.

You are correct in that the joints should be 90 deg out of phase, but that is only if both the joints are angled the same with respect to the intermediate shaft they will cancell the accelleration effects, which in my case they are not.

I will try phasing them at 45degrees in the morning and play around with it a bit more to see what i get. But for now i think i will just sleep on it and look at it with fresh eyes in the morning. its 12:30am here in Aus.

Thanks again for the input.

 

[AlephZero]

You could convert between 2 and 4 pulses/rev with a 2:1 gear ratio.

 

[jakedehaas]

Thats a good idea. Using just one uni joint and then a 2:1 gear ratio would indeed give the required 4 pulses per rev.

However, gears would need to be used, as using belts or chains would offer too much damping. The university should have gears, or i might have to get some machined up. Although, adding gears does add some complexity to the test rig.

I think i might just make the double U-joint set up and measure the output to see what happens. I'll discuss the gearing possibility with my supervisor today and see what he thinks.

Thanks for your help