I have a question about something I noticed while balancing rotors with an old Schenk balancing machine. Often when trying to trim out the last bit of residual unbalance the phase angle would start to creep or progress around the shaft. Note that the phase angle was still stable. For every ten degree move in weight the phase angle would move ten degrees and the computer would ask for ten degrees more. It wasn’t long before you’d be adding or removing weight on the opposite side of the rotor. My initial thoughts were that this was due to trying to solve a coupled problem on a single plane or opposite plane of the problem. Most likely it had something to do with user error in setting up the balancing machine.

Any time my trim balance shots don't give me the desired results, I start from that point with new reference and trial runs to recalibrate my balancing equipment for new influence coeffiencts. For example, if my starting imbalance is 20 mils and I get it down to 2 mils and start "chasing my tail", I'll make new reference and trial runs, effectively starting a new balance job with initial unbalance being 2 mils. Art Crawford told me to do this years ago and it has worked great for me. Hope this helps.

Waylon:
I've had similar experience with other equipment - particularly when you are trying to correct for the last little bit of imbalance. Joe makes a good suggestion. Reminds me of an old saying ... something about getting the elephant body thru a pin hole and all that's left to go thru is the tail (but don't elephants travel tail to trunk? So much for old sayings ... )

Regards, H

Quote from Joe,

quote:

I'll make new reference and trial runs, effectively starting a new balance job with initial unbalance being 2 mils.

I must admit, no one ever told me to do that, but many years ago when I couldn't get the final little bit out, I found that is the only way to go, start a new job with a new reference phase and amplitude. The system doesn't respond as it did when the first weights were added or removed, especially if the first amplitude was fairly high.

Of course this is only my opinion and I could be wrong.

Did your phase angles lock together on the 2 plane? Assuming a good distance between the planes and when both ends match in phase basically look elsewhere or a new plane to balance or remove an underlying fault.

I agree with the above on starting over. I am not sure whether it is the system coefficient that changes, or whether the human errors build up when trying to do EXACTLY what the machine tells you and assumes you do. In any case, starting over with a new run will take care of it if it is still unbalance.

Dave

Hi Waylon,

Please tell me if I'm wrong or right.
The old Schenk m/c you are using is one that you first calibrate by setting up dials etc. from physical measurements from distance to correction planes... to bearing support... etc. etc. as opposed to other machines that you set-up via trial weights ?

If I am right in assuming the first option, then I agree with what some have suggested.

Run with a new set-up and be extra careful with your measurements. The Older Schenk m/c's do not take into consideration where the nodal point of the rotor is really located other than by calculating it from the dial set-ups you have put in. From its calculations it assumes that the nodal point lies somewhere in the middle of the rotor and that is really a rare occurence. The nodal point is the point of transition of the right plane unbalance effect and the left plane somewhere in the rotor.

With balancing machines that uses trial weights method ( i.e wax for instance ) it is by far more acurate than the old Schenk method because you actually introduce an unbalance whose cross-over effect can readily be noted on the other plane and therefore can be nulled out electronically.

I think that the residual little bit of unbalance you can't get rid of comes from the other plane and the reverse also applies.
Thank God that Schenk have abandonned manufacturing this type of machine.

Good luck

Markoleo

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

Markoleo,
It sounds like you are very familiar with the old Schenk units. I worked in two deferent plants where they both used these machines. Both machines used the select shaft arrangement and measure procedure. These machines also used what I think would be described as transducers with mechanical amplifiers. The machine rotor configuration selection was probably where the errors got introduced into the system. We always eyeballed the measurements. Generally this would get you close enough to meet the balance requirements.

The setup description sounds like a typical hard bearing machine. Schenck still sells them. Schenck's HM series is still described as using " rigid rotor supports and a rugged mechanical design " that could well mean the dainty if not delicate lever arrangement as the ones from the 70s.

http://www.schenck-usa.com/ind_elec_motor_over.html

One thing that has lead to tail chasing for me with a few Schenck HMS was mis-interpreting zero. It can be configured as the lead or trailing edge of the shiney tape. With a small shaft and using the wrong edge of big tape it is easy to consistently apply correction at the wrong angle even with protractor. Eyeballing angles usually cost me a few extra runs.

Waylon - if you are using a hard suspension Schenck machine (which I suspect you are), then the good suggestions about recalibrating with smaller trial weights are moot. This is a "permanently" calibrated system and I suspect that it just simply needs to be recalibrated using care with accuracy of calibration procedure. Since you say the signal is steady and not "nervous", then I suspect that it is just a case where you are off on the angle calibration - something that would not show up so much on initial gross corrections.

Hard suspension Schenck balancing machines are also sensitive to shaft runout. Balancing to zero in the balancing stand is not necessarily the same as balancing to zero in the intended machine and in some cases such as when there's a large difference between the stand balancing stand and the machine running speed , the results in the machine and in the balancing stand can be quite different.

Thirty years ago ( I don't know about now ), the Schenck stands used a twin-T filter, the same type as the old IRD 600, 330, and 350 type analyzers. The amplitude and phase characteristics of the twin-T filter is not the most stable of filter designs.

I think the reason you often have to 'start over' when balancing a piece with a large imbalance, is the "system response" (and thus the influence coefficient) changes as you remove the imbalance. In other words, a large imbalance causes the system to be more non-linear because you have such large forces acting on it. Things flex, the shaft rattles around in the available bearing clearance, etc. As you remove the imbalance, the forces decrease, and the system becomes more linear.

Thanks for all the great answers.

The tape position was possibly another source of errors. The way our old machines were set up you had to hang the pickup over the small part of the shaft. In my book using the small part of a shaft is like asking for trouble. A 2 inch shaft would give you about 57 degrees per inch of rotation. A 1/8th of an inch error in setup would give you a just over 7 degrees of error. This would be a real easy mistake to make setting up this machine especially when dealing with a limited number of weight positions.

Wouldn't you reference the tape position back to the portion being balanced? This way your angle measurements would be made on a larger circumference.