quote:

There is a reason u-joint manufacturer's, and OEMs that use U-joints, specify installation misalignment in the range of 1 - 3 degrees (but equal at each end, so input and output shafts are parallel).

I think most think is that of motor vehicle application where there is flexible mounting and thus having to offset to avoid passing through zero degrees.

I also think you are correct in stating the lubricant needing to be rolled into the needles to keep from going metal-to-metal.

In every instance where I've set up torque tubes to tight alignment tolerance I've experienced smoother operation and longer life. And some giving 2 yrs of good service.

Now, how long is the automotive drive shaft designed to last? 100,000 miles! Now given an electric motor running 24/7 at 1200 RPM that would be basically three months. Same number of rotatons. When doing R&D for drive shafts I had engineers tell me; don't think about it, replace your dirve shaft every 100,000 miles.

However given most industrial applications where each end is fixed you don't have to allow for flexible displacements and can have tighter tolerances. Or, at least my experience has been pleasant when I set the units to tighter tolerances with smoother operation and longer life. Try one and see.

An update.
We changed the motor and what a difference. The old motor had some unusual amp feedback signals coming from it and got unstable speed swings from time to time. The new motor is running much smoother BUT, the 15-16Hz speed oscilation is still there. Vibration wise there was a raised noise floor and thats gone. The alignment did not go well as there is a lot of congestion around the driveshaft. We ended up getting an optical survey done who could only get it to within 0.5deg, however there is no vibration or speed swing at 1,2 or 4 x turning speed.
The driveshaft is a GWB 390/65 x 3.7m long. Alowable angle at 360RPM, 15degrees. Actual angle 2.2degrees so any miss-alignment should have less effect??. No strain guage data yet. More to follow

Here is a chart from SAE "Advances in Engineering Number 7 - Universal joint and driveshaft design"
The maximum recommended angle of ~ 400 rad/sec2 is based on "disturbance". That is, the operator will be aware of noise or vibration

Here is a chart showing the motion characteristics of the driving yoke (input) and driven yoke (driveshaft tube) when operating at constant 3600 rpm, and over 4 degree angle. If there is another joint at the end of the driveshaft with equal angle then the output speed will be constant. The one I can visualize easiest is the speed variation, where the driveshaft is varying 3600 rpm plus/minus twice per rev, while the driving yoke is constant 3600 rpm.

Thanks Dan,
So if we are at 2.2degrees and 350rpm our angular displacement, velocity and acceleration of the torque tube is likely to be very small. And if the miss-alignment between input and output shafts is < 1 degree angulat velocity change would be neglegable?

We are exploring possible resonant frequencies. Does anyone have the calculation for torsional resonant freq? I am wanting to calculate this for the drive shaft/motor system.

This may be completely offbase, But I agree that alignment would be the first place I would look.

But can we get some more specifics on the drive system. Motor... VFD?

We ran into a problem with a drive shaft like this once. It also had speed stability issues, which ended up being caused by a faulty controller card on the motor control system. Use a strobe on the shaft while running and make sure your speed is holding steady. Is the motor studdering causing a great torsional force as it bounces back and forth between driving the roll and catching up with the roll?

This message has been edited. Last edited by: Torptorp,

Did a couple more calcs today.
At 1 deg miss-alignment the speed variation would be 0.06 rpm on 360rpm or 0.016%. We are getting 0.5rpm at 15Hz or 2.5 orders. This frequency moves between 14-17Hz and is not proportional to speed.

The calculated natural resonant frequency of the shaft is 14.5Hz!
So it appears to be a resonant frequency causing the speed oscilations but what is exciting the resonant freq?

Apart from the speed oscilation seen in the motor tacho all else seems to be running fine since we changed the motor, no vib/noise etc unusual. We are still awaiting the strain guage results, the vendor is having a few problems with their data analysis program.
Motor is a permanent magnet AC VFD

Is it sible that there is some sort of 'torque wind up' occurring? Just like a 4wd vehicle turning corners on concrete with the 4wd system engaged. I have seen similar on other cardan shafts where over torque has occurred.

quote:

Originally posted by Planty:

The calculated natural resonant frequency of the shaft is 14.5Hz!
So it appears to be a resonant frequency causing the speed oscilations but what is exciting the resonant freq?

Is that a calculated torsional resonant frequency? The calc should be for the system, and include stuff like reasonably accurate values of motor and driven equipment "I", u-joint stiffness, etc. Not always easy info to get.

Hi Dan,
Actually I was supprised that I could get this info reletively easily. Both the motor manufacturer and driveshaft manufacturer spec sheets had it. I could not easily find the inertia for the load but it is much larger than the motor so I tried several different values and it did not effect the calc much.
I used the formula
Fr= 1/(2*pi)*((1/Jm+1/Jl)*Ct)^0.5
Fr is resonant freq, Hz
Jm is Motor inertia, 113 kgm^2 from motor manufacturers tables
Jl is load inertia, kgm^2 guessed at 1000kgm^2 Ct Tortional stiffness driveshaft total 973,188Nm/rad, from GWB driveshaft tables (2,372,539 for tube and 1,650,000 for joints)

This gave me 14.85Hz

I then got a reply from the DS supplier that the DS stiffness is 1,480,000 Nm/rad.
This gives me 18.32Hz

Both frequencies are in the region of what we are measuring.

I have just found a paper that may help explain what is happening. I am trying to get my head around it (It is a little bit beyond my level of knowledge?) Here is the link for anyone who is interested.

http://minds.wisconsin.edu/handle/1793/10164

I have seen shaft failures in engine dyno labs due to the piston hits causing torsional vibration or worse exciting torsional resonance and turning drive shafts into pretzels. He also mentions speed control issues with a "softer" shaft. I would check the vfd tuning and drive system stiffness. Does the vfd compare its calculated speed to encoder feedback? Do you have a good feedback?

Mike,
Have you thought about looking at the current on the drive as relates directly to the torque. Also as this AC drive is a closed loop system control tuning this is important to dampening the response to torsional resonances and rapid speed changes, often the default settings are overly sensitive. How did you get on with the strain gauging? I've had good results on kiln fans and engines using this method. Also I've use slip rings (using conductive tape) and an accelerometer strapped to the shaft.
Regard
Bruce

I am not an electrical engineer so may not get these answers quite right but,
We have done extensive tests on the motor tuning. This drive is the speed master for this part of the machine. There are 2 other drives that share the load ( torque). We have a 2006 report by ABB engineers that show a freq similar to what we have now. Ive been told he could not tune it out and he thought it was a resonance. The motor has a 2000+ pulse tacho and continiously closed loop speed control. We have tried open loop control (to remove any tacho induced factors) and the speed oscilation got slightly worse. We can effect the oscillation amplitude by changing the controler gain. We have limited ability to change anything in the PID controler as I think this is an ABB "black box".
We are still awaiting strain guage results. The guy doing this is very busy at the moment at the mill next door. These results will be key to deciding what to do next, i.e. if stress is less than the driveshaft endurance limit we may decide to do nothing further. The first shaft had balance weights welded on the tube and next to the tube/knuckle weld. We will not accept any shafts with welded weights anymore.
Cheers