Any way to fix this motor?

[calzap]

The motor is a 3 HP, 240 VAC, 1 phase, TEFC motor made by WEG.   It powered a two-stage centrifugal water pump.  The pump/motor started making a loud racket …  like gears grinding.  Noise seemed to emanate from the motor rather than the pump.  Externally nothing looked amiss.

Unfortunately, the design is that an extension of the motor output shaft is the pump shaft.  The only way to get a thorough look inside the motor was to remove the pump from the shaft, which meant the pump had to be completely disassembled.  This was done, and with a bit of study, the probable culprit emerged.

The centrifugal starting switch assembly spins freely on the motor shaft when it should be fixed to the shaft.  Figure 1 shows the shaft and parts on it.  The shaft in the pics turns the cooling fan; it is not the output shaft, which is at the other end of the rotor.  Looking at the shaft starting on the left is a flattened area for fan attachment (fan has been removed).  Next is the bearing assembly.  Next is the centrifugal switch assembly, which I believe is in the position in which it is supposed to be fixed to the shaft.  Next is an external retaining ring.  Finally, there is the rotor.  No loose parts or pieces of parts were found inside the housing.

I believe the metal ring of the centrifugal switch was force-fitted on the shaft, and the role of the retaining ring was to limit how far the centrifugal switch could be pushed on to the shaft during assembly.  With time, thousands of starts and about 3000 hours total running time, the metal ring of the centrifugal switch started slipping.  Now the whole assembly spins freely on the shaft, has a slight wobble, and can move back and forth on the shaft.  See Figure 2.   The motor still starts because the default still-condition is for the starting coil and capacitor to be engaged.

The loud racket is because there is some friction between the metal ring of the centrifugal switch.  So, it spins as the shaft turns, but as it does, its flyweights hit the tabs on the retaining ring.  See Figure 3.  The tabs show evidence of such collisions.  Because the centrifugal switch assembly can move back and forth along the shaft, it bounces between the bearing assembly and the retaining ring (ouch!).

If I could remove the bearing assembly, I could slide the centrifugal switch off the shaft.  Then drill a couple of small holes in its back plate and wire it to the tab eyes of the retaining ring.   The bearing assembly also appears to have been force fitted to the shaft.  Even if I could remove it without damage, getting it secured in position again would be problematic.  I don’t know of an adhesive that I would trust to secure the metal ring of the centrifugal switch assembly to the shaft.   Any ideas?

  Mike in California
 

[calzap]

I just realized that fixing the centrifugal switch assembly to the retaining ring might not work because the retaining ring might start spinning freely on the shaft.    How about removing the centrifugal switch assembly completely and using an external timing circuit to cutout the starting coil and capacitor a few seconds after power is applied?

Mike in California

[BrokenYugo]

Perhaps a tack weld or two? That or epoxy it on with a really good prep job, probably best to press the bearing off in that case, I'd just weld it.

[calzap]

Perhaps a tack weld or two? That or epoxy it on with a really good prep job, probably best to press the bearing off in that case, I'd just weld it.

I thought about welding the metal ring to the shaft with my MIG welder.  But doing that with the centrifugal switch assembly on the shaft might result in heat damage to the pusher plate (it's plastic).   If I could get the bearing off the shaft, then pusher plate could be removed temporarily and welding could be done then.   The bearing would have to be pulled off.  Perhaps a steel disk with a slot could be put behind it and then I could use a gear puller.

Mike in California

[james_s]

I had this happen with the metal cooling fan in a dishwasher pump motor, dribbling some superglue on it would help for a while but eventually it would break loose again. I ended up getting a nearly new dishwasher for free so I gave up trying to fix that one.

I can think of a couple of ways you might be able to fix this. Since you have a MIG welder probably the easiest is to weld it, just a quick tack weld in a couple of places ought to do the job so heat shouldn't really be an issue. Another thing you could try is take a sharp punch and hammer and punch a series of dings around the press fit lip to tighten it up on the shaft. You might also be able to drill and tap a hole and put a small screw in it, but the shaft may be pretty hard steel and the space is cramped.

Welding would be my vote as the easiest and most likely to last. Just remove the springs and slide the plastic part as far away as you can and give it 2 or 3 zaps, just little tack welds. Maybe a combination of the punch and tack welds, I don't think it will go anywhere then.

[BrokenYugo]

I didn't notice that part of the centrifugal assembly is plastic, that could be a problem, maybe wrap it in foil to prevent spatter damage? I would think it's some high temp plastic (nylon or better) that can handle the radiant heat from a 2-3 second tack, you don't need to run a bead here, the forces are pretty low.

Those bearings usually aren't on too tight, while it is very improper, you might get away with pulling it off by the outer race. They're usually pretty affordable push come to shove.

Drill and roll pin is another option, that comes to mind, though that weakens the shaft if this is the power take off end.

[highpower]

I don’t know of an adhesive that I would trust to secure the metal ring of the centrifugal switch assembly to the shaft.   Any ideas?

  Mike in California

Loctite 680 Sleeve Retainer Compound. Remove the switch assembly from the shaft and clean the ID with acetone. Also clean the shaft with acetone as well. After being cleaned and free from any dirt or oil apply a coat of the Loctite 680 on the shaft (only where the switch sits) and reinstall the switch. Note that you only have about 4 minutes to get the switch into position before the Loctite sets up and you will not be able to move it again. The 680 formula is designed to fill in the gap on slip fit assemblies. They have a different formula for augmenting press fits but yours being worn and loose will be better served with the 680.

So no welding needed and I can promise you the switch will not move again. In fact the only way to remove it again is to apply a couple hundred degrees of heat to break the bond of the adhesive. I've been using it for years. A bit pricey but it works. 

[CaptDon]

I vote for 680 or JB Weld also. Clean the shaft with a spray degreaser and apply 680 or JB liberally. Watch you don't glue the weights fast, they are pretty close to the surface you are trying to glue!!

[Ed.Kloonk]

I had this happen with the metal cooling fan in a dishwasher pump motor, dribbling some superglue on it would help for a while but eventually it would break loose again. I ended up getting a nearly new dishwasher for free so I gave up trying to fix that one.

I can think of a couple of ways you might be able to fix this. Since you have a MIG welder probably the easiest is to weld it, just a quick tack weld in a couple of places ought to do the job so heat shouldn't really be an issue. Another thing you could try is take a sharp punch and hammer and punch a series of dings around the press fit lip to tighten it up on the shaft. You might also be able to drill and tap a hole and put a small screw in it, but the shaft may be pretty hard steel and the space is cramped.

Welding would be my vote as the easiest and most likely to last. Just remove the springs and slide the plastic part as far away as you can and give it 2 or 3 zaps, just little tack welds. Maybe a combination of the punch and tack welds, I don't think it will go anywhere then.

I had a broken bracket on the mechanical part of a manual car window winder. Tried a few things. Welding it (too thin), a steel splint with cable ties (not ridged enough). Finally somebody showed up with this:

https://www.loctiteproducts.com/en/products/build/epoxies/loctite_epoxy_metalconcrete.html

Metal glue. 

[andy3055]

The last couple of posts have the best suggestions. I would not weld and it will be a whole lot easier than designing a time delay circuit and cheaper too!
Just keep it simple.

[james_s]

I don't think welding it would be difficult, all it needs is a couple of little tack welds. The plastic isn't a problem, just wrap a wet paper towel around it or use the heat blocking putty stuff sold for HVAC work. The epoxy is worth trying I suppose, I don't know if it will hold up but maybe combined with putting some dings in it with a punch will keep it in place.

[bill_c]

Above suggestions are good and most likely the cheapest option, but do a search for "Electronic Centrifugal Switch". They are great for dirty locations that would foul the standard contacts.

[BradC]

If I could get the bearing off the shaft, then pusher plate could be removed temporarily and welding could be done then.   The bearing would have to be pulled off.

Bearing pullers are available pretty much everywhere. Any mechanical workshop would have the ability to slide that bearing off the shaft. You might as well replace them while you're in there.

[james_s]

If the bearing is running smooth there is no reason to mess with it or replace it, I've seen electric motors run for many decades on the original bearings.

[BradC]

If the bearing is running smooth there is no reason to mess with it or replace it, I've seen electric motors run for many decades on the original bearings.

Me too, but if you remove the bearing, you've got half a chance of repairing the starter properly and not bodging it with epoxy. With the right tools a bearing removal / replacement is a 5 minute job making the broken bits completely accessible.

[elekorsi]

If the bearing is running smooth there is no reason to mess with it or replace it, I've seen electric motors run for many decades on the original bearings.

and i have seen many of them fail catastrophically in the worst time possible, beacuse somone said: Eh, it is fine, we will not replace it because it is running smooth...
Then you are searching for a new motor in the middle of the night, because there is no machinist to repair the bearing seats in the motor housings...

Such bearings are cheap and should be replaced at every disassembly

[james_s]

and i have seen many of them fail catastrophically in the worst time possible, beacuse somone said: Eh, it is fine, we will not replace it because it is running smooth...
Then you are searching for a new motor in the middle of the night, because there is no machinist to repair the bearing seats in the motor housings...

Such bearings are cheap and should be replaced at every disassembly

I've seen too many incidents where something was working fine and somebody replaced it and the new part turned out to be cheap crap and failed soon, this is why I typically don't replace something unless it's bad. I don't trust new mechanical parts to be of good quality, everything seems cheaper and lower quality these days.

[highpower]

Watch you don't glue the weights fast, they are pretty close to the surface you are trying to glue!!

That is a good point. You only want it under the metal ring at the rear of the 'frame'. Best to have the shaft vertical when reinstalling the switch so that any excess adhesive runs down the shaft away from the moving parts of the assembly. Also the assembly should be rotated a bit to help spread an even coating of the uncured adhesive around the parts for full coverage.

The question I have is why did the switch (ring) become loose on the shaft in the first place? It is designed to be an interference (press) fit onto the shaft. Has someone else had this pump/motor assembly apart in the past and caused the problem - or has the unit overheated at some point and cause the switch frame to expand and come loose from the shaft? Either way I would be replacing both bearings myself. Relatively cheap insurance against having the original bearings fail in the near future. If the motor got hot enough to expand the switch ring, I'm sure it cooked the grease in those bearings as well.

[james_s]

Probably just normal thermal cycles eventually stretched the metal, or maybe it was never quite as tight a fit as it should have been when it was built. The fan in my dishwasher was never messed with and it came loose on its own.

[highpower]

Very possible. For all we know vibration could have caused a stress fracture in the ring and it split. No way of knowing unless you've got it apart in front of you.

Somewhat like your dishwasher I had a plastic fan that was a press fit onto a universal motor that runs the pump in the back of our ice machine. The plastic hub had cracked and the fan blade rode up the shaft and started rubbing the sheet metal above it making all kinds of racket. The fan was made of one of those very slippery plastics that most adhesives won't adhere to and after a few tries I gave up on the glue idea. I ended up removing the pump assembly and taking the motor apart so I could chuck up the rotor in my lathe and drill & tap a hole in the end of the shaft. Drilled a hole through the end of the fan blade and installed a small screw with some thread locker to keep it on the shaft. Still going to this day.

Why? Because the manufacturer won't sell the fan blade alone. You have to buy the entire pump assembly to the tune of $300. Well worth the effort to fix (bodge) it. But I digress. 

[calzap]

Thanks for all the responses.  I removed the bearing.  I don’t have a set of bearing-pullers but do have a set of gear-pullers.  Definitely wanted the pull force along the motor shaft, not the outer ring of the bearing race … didn’t want ball bearings bouncing across the floor.  Motor has a 5/8 inch shaft, and I have some oversize 5/8 inch washers. So cut a slot in a washer and slid it behind the bearing.   And yeah, first measured where the bearing lies on the shaft.  Then applied a gear-puller to the washer.  Bearing came off easily and looks to be in good shape … turns smoothly, no signs of having been too hot.

Then took off the centrifugal switch assembly and removed springs.  Also made sure I could reassemble it on the shaft.   I intend to tack weld the metal ring to the shaft.   The output shaft is stainless steel (it continues as the pump shaft).  However, the back-end shaft, where the centrifugal switch is, is highly magnetic … so assume it’s a non-stainless steel alloy.  I’ll use the MIG setup that I use for mild steel … real MIG with C25, not flux-cored.

However, one minor complication appeared once I had the centrifugal switch off the shaft.   If you look at the pic below, you’ll see a thin fiber-board bumper pad.   Its purpose is to reduce the shock to the rectangular portion of the pusher plate when it moves to the metal ring during start-up.  This pad is cemented to the metal ring flange and will have to be removed with an X-acto knife prior to welding.  To replace it, I can make a gasket from gasket sheeting, maybe find the right size circular gasket, or use semi-liquid form-a-gasket material.

The pusher plate appears to be thermoset plastic (urea-formaldehyde).

I’m not impressed with WEG pump motors.  A previous one failed because the plastic fan blades deformed.   This is the second repair on this motor.   Earlier, the start capacitor failed.  It came with one rated for 125 VAC even though pump can be operated at 240 VAC.   I installed one rated 250 VAC.  But of course, it was too big to fit in the housing of the failed capacitor, so I had to make a new housing.    Not too surprised the centrifugal switch assembly came loose.

Mike in California

[calzap]

One additional note:  Because the rectangular portion of the pusher plate slides over the metal ring, the welding should probably be done to the backside of the metal ring flange. Otherwise, blobs of weld metal might prevent the pusher plate from moving all the way back during startup.  This means the metal ring flange, retaining ring and shaft will be welded together.  The space looks tight, and I might go with Loctite 680 after all.

Mike

[andy3055]

Even tack welding can deform the shaft!

[calzap]

Tried a power-off run with the MIG gun.  Can’t get a good angle.  I’m ordering some Loctite 680 adhesive and Loctite 7471 primer.   I’ll use naphtha to clean and degrease the shaft and metal ring.  The radial gap appears to be about 0.08 mm (0.003”) which is well within the filling spec of 680.

I’ll do a post after the job is done.

Thanks again for the advice and discussion.

Mike in California

[calzap]

Thanks for the many replies.  I had to put the project off for a while but am back on it.   Loctite 680 worked great for securing the centrifugal switch to the shaft.  First, I cleaned the shaft and mating surface of the switch with acetone, then applied Loctite 7471 primer to both surfaces and allowed to dry, then applied Loctite 680 to both surfaces, mated them and allowed to sit for a week.   Very solid attachment.   Have spun up the motor several times.  Switch has stayed fixed to the shaft and can hear it disengage when power is removed from the motor.

Unfortunately, the motor still has a problem that might or might not be related to the former centrifugal switch problem.   It still makes too much noise which can sound like a grinding sound or vibrational sound … it varies.   I suspected the bearings, but they look OK and feel smooth when rotated by hand.   I removed the rear one (near the centrifugal switch) and gave it a speed test with a drill … it passed.

Close inspection of the rotor and stator reveals that they have contact.  See the pics.  So is it an alignment problem or a warping problem?   I’m not sure.  The ends of the motor are held to the rather thin metal shell by 4 long, skinny through-bolts.  The shell is thinner than it looks in the pics because in the pics, you’re seeing the ends of the shell which are rolled.   In addition, the motor normally operates horizontally, but is supported entirely by a vertical metal plate that is bolted to the output shaft end.

Unless someone has a better suggestion, I’ll reassemble the motor with the through-bolts pretty tight, but not completely tightened, power it, then play with varying the tightness of different bolts until noise is minimized.  Other ideas?

Mike in California