Which flexible coupling is better?

[onemilimeter]

I wish to measure the vibration of a motor which is connected (via flexible coupling) to a DC generator as load. I found two types of flexible coupling as shown in attached image. Which one is better for this application (to minimize the vibration due to mis-alignment)?

Thanks

 

[itsmoked]

They would be the same as they both use the same principal.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[ScottyUK]

My guess is that A will have better dynamic balance but that is based purely on appearance.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[waross]

I wouldn't use either one. They appear to be suitable for a small amount of angular misalignment but no lateral misalignment. If you are anticipating misalignment you may wish to consider a type of coupling with a resilient insert.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[Skogsgurra]

A bellows coupling is usually better. But the torque range is rather limited. But, then again, so is torque for the couplings shown.
A tight fit to avoid unbalance is what you should look for.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[electricpete]

I agree with Bill's comment that offset misalignment could be troublesome for this coupling.

What is the size and speed of the machine?

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

If I had to choose between the two in the original post.

The one on the left appears to rely on a set screw to hold coupling onto the shaft. That tends to push the shaft off-center within the coupling hub which is not good (particularly as clearance increases).

The one on the right does not appear to have that set screw to fix the coupling to the shaft, one would presume that the dimensions ares such that when you tighten the the two halves together clamp onto the shaft, which is a good thing.

So if those are the only two choices I vote the one on the right. But Bill's concern is very valid imo. Also would seek out the advice of a knowledgeable distributor or application engineer. Users don't typically select couplings.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[sreid]

take a look at Zero-Max Shaft coupling especially the Servo Class couplings.

http://www.zero-max.com/productmain.html

 

[onemilimeter]

Thank you very much for everyone who has replied my thread.

Please kindly refer to attached image. In my current experimental setup, two couplings are used. One is used to connect the motor and the generator, another one is used between the generator and the encoder position. I think these couplings are "resilient" or "elastic" couplings. The rated torque and speed of this motor is 2Nm and 400rpm respectively. With this experimental setup, the encoder position reading (i.e. the relative "position" between the encoder reading and back-emf waveform of motor (a brushless permanent magnet motor)) is found to be not consistent. I suspect this inconsistency is due to the "not-so-perfect" couplings. I'm not sure this phenomenon can be refered to as "coupling's winding up" or "coupling's backlash". Please correct me if I'm wrong.

Therefore, I wish to move the position encoder to another end of the motor (as circled in the attached figure). I'm not sure if "clamp fixing" or "Set Screw Fixing" is better to be used with the position encoder. Do you think "Clamp Fixing" is good enough to hold the shafts and prevent "contact slip", for example between the coupling and the shaft of the position encoder?

I plan to buy one coupling which has a hole with diameter that matches the diameter of the position encoder shaft. If I'm doing so, then I need to machine a piece of metal to 'match' the size between the motor shaft and coupling hole.

Thank you.

 

[BobM3]

If the current couplings are torsionally resilient then, as the torque output from the motor changes, you would get some relative rotation between the generator shaft and the encoder shaft. I'd call that "windup". Depending on the coupling and its connection to the shaft there could also be "backlash".

Are you talking about several problems here? One being trying to minimize any vibration due to shaft alignment and the other between a phase change between encoder output and generator emf? Or do you think it is caused by the same problem?

 

[onemilimeter]

Hi BobM3,

May be let's forget about the vibration issue first.

Before doing any serious experiments, I need to align the position encoder signal to 'match' the back-emf waveform of the motor. Let's say the PM motor has a single pole-pair. Now the DC generator is operated as a DC motor that drives the PM motor. An oscilloscope is used to capture the back-emf of phase A of the PM motor. The out signals of the position encoder (i.e. A,B,INDEX) are fed to a microcontroller with built-in QEP module. The microcontroller calculates the rotor position (0-3999) and output a 12-bit data to a Digital-to-Analog converter (DAC). The output of the DAC is fed to another channel of the oscilloscope, and to compare with the back-emf as shown in attached image.

To align both signals, I need to add an offset value to the computed rotor position value (in microcontrol program) so that the THETA becomes zero. Several tests were done as follows:

[T1] PM motor was rotated in counter-clock-wise (CCW) direction. An offset value of "125" is required to make the THETA becomes zero.

[T2] PM motor was rotated in clock-wise (CW) direction. With the same offset value of "125", the THETA was found not equal to zero! After adjusting the offset value to "120", the THETA becomes zero again.

[T3] PM motor was rotated in CCW direction. With offset value of "120", the THETA was not zero. After some adjustments, an offset value of "128" is required to make the THETA to be zero.

Obviously, the offset values required to make the THETA to be zero in T1, T2, and T3, are not consistent! Therefore, I suspect the problem is due to the couplings between the PM motor and DC generator, as well as between the DC generator and the position encoder. Please comment if I'm wrong.

Thank you very much

 

[waross]

Your signal offset is probably due to the angular position that the encoder has been installed with. To possile corrections. Loosen the set screw and align the encoder or use a software offset.
Now the signal may be aligned with zero software offset.
But when rotation is reversed, the encoder output may be inverted.
Without knowing more about the nature of your offsets, we cannot say whether the difference between offset 125 and offset 128 is important.
You may have torque peaks upsetting the position but probably not.
Two suggestions;
1> Apply a torque to a sample coupling and measure the windup. Does the coupling windup corelate with the offset error?
2> Move the encoder to the other end of the motor and forget about load torque induced errors.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[BobM3]

How is the existing coupling attached to the shaft? A set screw against a flat would be vulnerable to movement when the load or direction changes. Any friction would cause it to be non-repeatible. Coupling "B" shown in your original post would be good choice for generating a stong, non-slip clamp load on the shaft. Couplings with tapered bushings are best for concentricity and no slippage.

Are you running the same load (amps or torque) for each of the three tests?

Are the magnets secured tightly in the motor (or generator)?

 

[electricpete]

Our of pure curiosity (nothing to do with your question), what kind of experiment is it?

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[sreid]

The Zero-Max Servo Class couplings are Flex Disk couplings with zero backlash. Purchase them with the shaft sizes you need.

 

[onemilimeter]

Thanks for all replies.

To waross:
I use "software offset" method. In fact, the values "120", "125", and "128" were created for discussion in this thread. This phenomenon does happen in my real experimental setup, but I can't remember those exact offset values. But I can assure that the inconsistent position reading has caused several problem to my control algorithm, which depends very much on the rotor position to operate properly. Not only this happens in reversed direction, even in the same direction, different runs of experiment will also give different results and need different software offset values for the algorithm to work properly.


To BobM3:
The existing coupling is attached to the shaft (of position encoder) by "set screw". Three tests mentioned in my last post are open-circuit back-EMF tests, i.e. no-load is applied in the tests. Only friction load I think. In my opinion, the magnets are secured tightly in the motor. There is not magnet in the DC generator, which has both armature and field windings.


To electricpete:
Well, the tests mentioned in my last post are used to align the position encoder signal to match the back-emf of the PM motor.


To sreid:
I think I understand what "windup" is now. But I still do not really understand what "backlash" means. Would you please how the Zero-Max Servo Class coupling attached to the shaft and guarantee "zero backlash"? Using "set screw" or "clamp method"?


Thank you very much

 

[BobM3]

Any chance that high starting torque is causing the couplings to slip? Maybe try starting it with limited voltage?

I'm not quote sure what you are measuring. Is it the voltage produced on the wires of the generator? Or are you actually measuring the back emf of the driving motor?

Backlash is a "looseness" in the connection. Change the torque in magnitude or direction and there will actually be a movement between the shaft and coupling.

 

[onemilimeter]

Thanks BobM3. I agree with you that different "starting torque" will affect the outcome.

If you're referring to my experimental setup, the DC generator is now operated as DC motor to drive the PM motor. The terminals of the PM motor are left open, i.e. without connected to any load. The back-emf produced by the PM motor is measured and displayed using scope.

Now I understand what "backlash" is. Thanks BobM3. I think one of the causes to my problem is "backlash". May be I should consider using "zero backlash" coupling. But I don't know how this type of coupling can prevent "backlash" if the surface of the shaft is round and has not keyway. Interesting if someone can explain it. Thanks.

 

[waross]

Think of back lash as the space between the gears on a gear drive. Think of windup as a twisting shaft.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[BobM3]

Zero backlash is accomplished by tightening the coupling hub around the shaft so tightly that the friction between them carries the torque. There is no slippage then. It requires close tolerances on the shaft and of the bore of the coupling.