We've discussed balancing quality standards before, but I wanted to share an indicent that illustrates that "standards" are not worth the paper they're written on if not understood and properly used.

One of my customers has a large 3-head drum sander used to finish sand hardwood flooring. They've had "chatter" from time to time and would call me in to 'trim balance' the rolls in place.

I finally convinced them we had to go a lot deeper if they wanted to put this problem to rest.

I went with them to visit the "roll shop" where the rubber rolls are recoated and balanced. The balance machine was an older Hines solid-bearing machine, running EasyBalance software.

We wanted to balance the roll with the bearing in place, but they couldn't figure out how to do that. I said just take the V-block assemblies off and clamp the bearings to the pedestal. "We can't do that." (apparently the concept of balancing a roll in its bearings was totally new to them). I removed a cover plate and showed them how it all worked and they realized that yes, they could do that after all.

They set it in the machine and spun it and the balance was much worse than when they had balanced the same roll (without bearings) several days before. (probably due to bearing fit anomalies)

I asked the operator what he normally does about the keyway (this was a 'driven' roll) and he said "Nothing"... his exact words, I swear.

I then measured the keyway and showed that lack of a key was creating a 70 gram-inch imbalance. They were using a G2.5 spec and thought they were leaving about 10 gm-in in the roll which is actually a G1 grade (156 lb. roll, 2756 rpm). However, add the 70 gm-in keyway imbalance, and the roll was actually at G8.

Next I asked if they'd checked the runnout on the shaft where the sheave rides. Blank stares. We measured 0.007" TIR. For the 40 lb. sheave this roll uses, that's potentially another 63 gm-in of imbalance, which could take us all the way to 143 gm-in. That calculates out to G16.

So the nicely formated "Test Certificate" says they are well below the G2.5 tolerance level (G1 actually), but the actual grade is between G8 and G16, depending on how the shaft runnout aligns (or not) with the keyway.

It was not their intention to do a poor job, but they have no understanding of what balance really involves, no training in balancing theory, minimal training in the use of the machine (he knew that "keyway compensation" was there, but had no idea why or how to use it), and no understanding of the functional characteristics of the piece they were balancing.

This is a "roll shop" that has been in business for probably 50 years. The salesman said he'd been there 24 years and had never seen runnout checked on a shaft like that. He didn't know it even used a sheave.

So when someone tells you it's been "shop balanced" what does that mean exactly?

IRD Balance Quality Paper Calculations are especially helpful.

And there is also a Practical Guide by Earl Halfen that is very useful.

How have they stayed in business for fifty years. These are basic set up requirments for most rotating equipment that is balanced in a shop balance machine. We ran in to a semilar problem balancing 2 pole motor rotors on the seal fits to protect the sleave bearing fits. We were actually causing more problems than we were correcting due to the differances in center lines of the shaft.

In general people are more familiar with ISO 1940 part 1 on balance quality, but part 2 is on balance errors and should be on the self and/or mind if one has a balance shop.

As for 2-pole motors, they may not be tools from hell as some have said, but many are flexible rotors.

No, 2-pole motors really are tools from hell! Perhaps "flexible rotors" might explain some of the odd problems I've seen over the years.

This company only started balancing rolls about 4 years ago; before that, rolls were sent out for balancing. The machine they have is 'used' I think, but the installation looks to be well done. The EasyBalance software looks interesting.

On the machine, the upper pedestals are attached to the lower frame by means of a thin metal plate on either end. It's about 3/16" thick and 4" wide. Essentially, the plates act as stiff springs, allowing the upper pedestal to move side to side. I'm not sure how the software "chooses" a balance speed. I assume it's based on rotor weight and normal operating speed.

It gives a "one run" solution, just from the rotor physical data entered, and a single spin run. To do this, doesn't there have to be some sort of assumption made about the "lag angle"? How can it "know" where the heavy spot is? There are no force transducers on this stand. Just some sort of LVDT that looks at the movement of the upper pedestal.

quote:

Originally posted by rustythevibeguy:
...doesn't there have to be some sort of assumption made about the "lag angle"?

Why any assumptions should be made at all? I think that by knowing mass of the rotor and machine influence coefficients obtained during machine calibration tests, correction weight could be sized and placed after the first run. The reference point is never moved.

quote:

machine influence coefficients

OK, that makes sense. Of course the machine has to be calibrated for a given setup. If you change the mounts (V-blocks, Bearing-blocks, roll with bearings mounted) you'd need to recalibrate the machine, right? Would the influence coefficient change if the roll weight changes significantly, say from 50 lbs. to 500 lbs? Will the turning speed affect the influence coefficient?

And we have to assume the roll is rigid, so there is no resonance involved, right?

Perhaps the balance machine operates at such a slow speed that it is assumed to be far below resonance and the lag angle always 0?

Rusty - the balancing machine you describe was probably made years ago by Gordon Hines and is probably a "hard bearing" or subcritical machine, whereby it can be permanently calibrated through the use of a test rotor and application of known weights. I do not know if it is linear over the speed range they are using to balance the rolls, however. EasyBalance has a bode' plot function that can easily establish linearity over a speed range. It sounds like these guys could do with some training, if they did not know how to apply the key comp feature or to check for eccentricity. It seems that a lot of users know just enough about balancing tolerances to be dangerous, since they also misapply them on the low end, as well. Unfortunately the instrumentation, whether it is IRD, Schenck, Hofmann or EasyBalance cannot out think the operator - it will just state that the hunk of metal that it sees spinning in the balancing machine is at a certain level. As to the machine itself, it is probably using a coil type pickup to measure the deflection of the pedestal and balancing speed should be determined by the operator to be at a speed well below the critical speed of the suspension - thus the term subcritical. This machine probably does not use influence coefficients in the way we do in the field or on a "soft bearing" suspension - rather it is calibrated for sensing the force from a calibrated heavy spot and uses calibration constants which are stored in the EasyBalance software. With this in mind, changing practically anything except the pedestal stiffness would probably NOT affect the calibration. All of this assumes linearity of the Hines suspensions, of course. Did you test the machine with a known weight?

Earl, thanks for the info. Yes, it uses a coil-type pickup to measure deflection. This is the first machine like this I've ever seen and the 1st time I've ever seen EasyBalance. I thumbed through the manual and saw right away there was a lot they didn't know about what was possible with this machine.

Since I was just a 'visitor' I didn't want to be too pushy. They seemed genuinely thankful for the insight I was able to give them, once they figured out that I really did know something about balancing. They gave me an open invitation to help them further, so I think I'll try and get a copy of the EasyBalance manual and study up on it, and then go down one day and work with the machine and see what it will do, maybe on a Saturday when they're not using it. I'd love to do some bump tests on it to see where it's actually running. And I'd also like to hook up some accels and see how the machine data/balance compares to my 2130. It'd be educational for sure.

Rusty - drop me an email to bsg@grandecom.net and I will send you a .pdf manual.

This shop redid 3 sanding rolls and I went out to trim balance them once installed in the sander. They did a better job than before, but still left 0.008" runnout in the drive end shaft of one of the rolls. They were supposed to have repaired this. I assumed they would "flame spray" the shaft and turn it down, but apparently they don't have that capability. They simply welded up the shafts (all three rolls had similar problems) with a wire welder and turned them down... should have seen the pits from the weld porosity (can you say "stress riser").