HI fellows,
I’m looking for some help. I build a speed encoder with a contact tachometer that I use for torsion reading with a copp-tek converter. I tested the rig on a small balancing bench and got good result. I use it on two a ball mills on various shaft locations and I didn’t get any torsion value. I am sure now that the mill doesn’t have a torsion issue but I still would like to confirm with someone that has the hand on experience with torsion reading and converter. I must admit that I get a little confuse when the math from a FM torsion system comes in! I have to confess that that I was sure to get at least a little torsion since I measure a little load modulation on the motor’s current. I am willing to send the data base if you use Ams Suite 4.7. It is only 1570 Ko.

Best regard, Marcel

Marcel,

I have a similar system. I have used it once on a motor shaft driving a coal mill. How many pulses per revolution from the encoder? What is shaft speed? Did you verify the raw signal from encoder = #pulse/rev * rpm? Can you tell me more about your contact tachometer?

Walt
w_f_strong [at] msn [dot] com

Hello Marcel,

I would be willing to look over your torsion data. Send it to the 4xdiagnostics email address below.

Talk to you soon.

Thank you guys for your quick respond, I’m preparing a little file with pictures, calculations and spectrum and I will send it as soon that I’m done. Got some un-expected jobs.

Thanks again fop your interest I’ll send it over soon. Marcel

Hi guys,
Unfortunately, I didn’t get the result I was expecting. All the modulations are well over acceptable level. I will have to pursue the investigation further. I still send you the file with the graph and some picture of the instrumentation. It works really well! It’s too bad copp-tek doesn’t sell a torsion meter with all the electronic components build in one single torsion sensor. Took me a little while to figure out how to read speed without all the trouble of the reflecting tape.

Best regard, Marcel


Ball_Mill__8_torsion.PDF (3,953 Kb, 17 downloads)

Marcel,

Is is possible that the torsional converter output was connected to an accelerometer input on your vibration meter?
What input signal adapter did you use?

I used a SST adapter model 623 from CSI and it was connected to the volts input. I set set up the 2120 on velocity meter with a volts sensibility of 1V/ips.
Best regard , Marcel

Marcel,

I couldn't follow all of your data because of my language barrier. Your optical encoder looks like it produces about 12 pulses per revolution (from photo). This is rather low. The sensitivity is based on 100-pulses per revolution according to the Copp-Tek nameplate. Do you hand-hold the encoder or use a rigid bracket? Any relative motion (tangential direction) of the encoder will give a false measurement of torsional motion (angular velocity). The encoder will also produce some "apparent vibration" at multiples of encoder speed due to eccentricity of friction wheel and other sources. Your analyzer should be setup without low frequency adapter and not with velocity and not with 1V/ips. No wonder the results look goofy!
Try this setup:
Standard Voltage input adapter (no ICP power)
Units: nonstandard (AC volts) type-in "DPS"
Sensitivity: compute volts per DPS from CoppTek setup (for example: 122E-06 V/DPS for 100-pulses per revolution)
Integration: None (it is already a "velocity" signal)
Keep trying!!
Motor current analysis is a very good indicator of low frequency mechanical defects that produce torsional vibrations.

Walt

Hi Walt,
The encoder produce 10 pulses per revolution of the tach. The tach wheel has a smaller diameter than the shaft, which increases the pulse per rev. from 48.2 to 344.1 depending on shaft diameter. I expect to see 1X shaft speed of the encoder (tach) and machine’s from imperfection, which is visible on one of the graph. I keep your advise in mind but I won’t be able to perform more testing for a few weeks. I will try to uses a smaller tach’s wheel to increase sensitivity.
I’ll keep you inform, thanks, Marcel

Marcel,

I am not sure that I agree with you about the encoder pulses being multiplied by the ratio of the diameters. The encoder and shaft in contact shaft have the same angular velocity because their tangential velocity and therefore frequency content is the same. I would still use an encoder with more pulses, and the sensitivity should be based on actual encoder pulses and not "virtual" pulses becuase the encoder is rotating faster than the shaft it is contacting. I have not used a friction wheel encoder, but this is my approach.

Walt