Almost certainly. :D

The micro finish on the bearing race was out of spec, so sufficient lubrication was not present given the increased friction on the roller bearing and the race.

Simple, really, given the facial plate distortion on the spiral planetary gear, driven by metal tectonic allchemy.

Four people got fired.

I'd say moisture was in the lube........

LOL!

The titanium melted at too low of a temp? Causing addl' promblems?

Dirty/inadequate lube leads to too much friction, roller bearing cr@pped out then Kaboom!

Did you sell them dirty lube or lube for dirty purposes?

After preventive maintenance, they forgot to re-install the cool collar, so the lube overheated!

Thats some funny *****!!! I love you guys.

Actually, the lube is from Shell - its an oil. The mill has 4 bearings on it all fed with the same oil from the same pump. Of the 4 bearings, only 1 fails every month - the thrust bearing. The tapered roller bearing fails about every 6 to 9 months and the other 2 (that you don't see) fail maybe every few years.

Lets get back to basics - failure occurs for the following reasons a) man b) method c) material d) machine - not much more than that and of thatm it can be broken down (sorry for the pun) even further. This bearing fails repeatedly because it is not the best choice 'machine' for this application. The lube is also pretty meager so 'material is at fault as well. I would even venture a claim that the 'method' is flawed - that being the maintenance practices as well as installation including balencing.

There are some tell tale signs but the first that one notices is the color of the lube - its way to dark. Mostly due to contamination and then oil oxidation. The oil out of the drum is amber and this stuff is pretty rank. The filtration system they are using is a joke. The challenge is the customer thinks it is OK becasue the other bearings do not fail at the same frequency as the thrust bearing so they think it isn't comtaminated - even after I showed them the oil being pumped to the bearings was tar black. The yellow metal didn't help but the oil theyare using is from a company focused on selling fuel not lubricants.

Another interesing thing was the rolling elements on the thrust bearing. The divots had a random pattern and also a straight pattern. This is rare. Normally the pattern is entirely random - to have a straight line divot pattern indicates something beyond pitting due to contamination, it really starts to speak of the balence issue.

Another indicator is the outer race of the roller bearing has interesting lines on it. I would be willing to bet that if I measured the width and divided by 60 I would have the speed of that shaft. Deviding by 60 because the 3 phase 460 volt motor is working at 60 cycles. I didn't think to look to see if the pattern worked its way all the way around the entire race - if so it is not only a vibration issue but an alignment issue.

The customer wanted to due some vibration analysis. I don't think it will tell them anything because while this system is running there will be so much incidental noise. It may not be the case - maybe it is worth a look.


We did sell them some oil to help but I made sure to let them know that they would continue to experience failures unless they took the following measures (that is if they wre going to continue on with the existing set up):
1) Install and balence the shaft and bearings proper
2) Clean the lines and sump
3) Install a weir (some baffles) in the sump
4) Install a better filtration system for the oil - centrifual filters and maybe even a Y strainer would certainly help. The paper spin-ons they are using are worthless.
5) Install a breather element and don't expose the sump and oil to the outside world.
6) Investigate rubber couplings to reduce the vibration (the loads maybe to great but its worth a look)
7) Check the 3 legs of the motor to make sure they are all in phase - I have seen where one of the hots is pulling 20% more juice indicating a phase issue. Strange as it may seem it can make the motor jump enough to set up a weird hormonic
8) Incorporate heat exchangers to keep that oil cool

These ideas may seem expensive but consider that the bearing alone is running them $3000 a throw not to mention that it takes 2 guys 4+ hours to pull and replace. Downtime, parts replace and labor costs add up each month to some serious cash - over $10,000 easy. Thats 10 large that might be avoided each month.

It also looks like it is trying to counteract a large moment force which bearings do not like.

Speedy:)

Looks like it got too hot.

They blow ten large a month on repairing this particular failure? Wow. I wish I was in their business that I could afford their overhead.

Never did say what it is on and what that machine does.

You didn't say where this is being used... I'd guess a paper mill, or chemical factory.

Would you elaborate? Is it an issue of getting down to micron filtration?

What you are going through is exactly what I do for a living. Every day, same thing, 1500 machines to babysit, dagnose, and monitor.
Have you calculated the thrust loads yet? Are they using laser alignment systems to align the machine?
High thrust loads are often associated with angular misalignment.

you well know that oil doesn't get black unless it is cooked or contaminated. My guess is cooked. What shell oil are they using and what is the rpm on the machine? If the speeds are under 3000 i'd recommend a shell meropa 150, if the speeds are under 1800 go with omala 220.

The process takes a titanium ring (about 3 feet in diameter) and spins it as a form is pressed on to shape it. Aslo know as metal spinning - How-its-Made had a program on making Woks, hubcaps and cymbals - used when the work peice is pretty rigid. Hydraulics apply the force and the two shafts - one the spin the piece and the other for the impression tool are held by the bearings. The drive end has the thrust bearing while the roller bearings are used for the 3 other ends.

IMO it is a faulty design. The customer is also looking into using additional bearings to relieve the load. I do not think it will solve the issues.

Here is what I was talking about before:

http://forums.pelicanparts.com/uploa...1166628688.jpg http://forums.pelicanparts.com/uploa...1166628704.jpg

As for the filter set-up, micron filtration would be great but without a pre-filter, a strainer or even a larger filter size element to take care of the fines, a <25 micron filter would clog in hours. Here is another idea:


http://forums.pelicanparts.com/uploa...1166629361.jpg

I'm supposed to get the loads from them in the next few weeks. They are supposed to do laser alignment but have put it off - needless to say they have some issues at hand!

The Shell product is the Omala but even the Meropa wouldn't have the wear resistance or oxidation resistance needed for this application. We got them an oil that will provide 5X the oxidation resistance and much more load and wear resistance as well. On top of that we will provide them independent oil analysis as well as several of the filtration opitions. The Shell distributor has never been in the plant even after repeated requests by the customer.

I'm just learning about this stuff but shouldn't a big 3 phase electric motor like this be smooth as silk. I thought they laid in the stator windings so you had an even continuous smoothly rotating magnetic field.

Those straight axial lines on the rollerand outer race look similar to what i've seen as false brinnelling.
Usually the cause is surrounding vibration from other machines or processes transfering to this machine through piping or foundation etc while this machine is at rest.
The vibration is extremely localized and creates small dents in the rollers and races. The same thing can happen from sudden thrust shock loads or improper installation techniques (BFH instead of an induction heater).

Another cause of straight lines is skidding. Are those lines on the active or inactive thrust bearing?

Excessive thrust clearance can cause the inactive (or unloaded) thrust bearing to stop rolling and skid. That wears flat spots in the roller and can give the wear pattern you are seeing. If this is a realtively low speed applkication i would recommend a medium preload on the thrust bearings, at least a couple hundred pounds.

I usually study races and rollers in a microscope, you can tell a great deal from the surface.

If you are going to do a serious vibration analysis make sure you count the number of rolling elements in the bearings or at least get the bearing numbers. That way you can calculate and plot the inner race, outer race, spin and cage frequency. Makes it so much easier to diagnose early stages of bearing failure as amplitude is basically meaningless in identifying stage 1 failures.

If you get some good spectrums let me know.
I look at em all day long (when I'm not on vacation like I am now).
Are you familiar with the vibration institute or rockwell automation?
I'm working on my level two certification for both organizations.

I use a CSI data collection machine but I learned on an IRD. Talk about an innexact science.