Friday afternoon (the only really good time for a problem on a critical fan) I found a Dodge Sleevoil bearing with multiple harmonics of running speed and a clear sharp impact at running speed. Checking things more closely, I found the cooling water piping to be vibrating so that the pipe nipples in the bearing housing were moving significantly enough to be of concern. I don't know what was found since I left shortly after reporting the problem. I'll find out soon enough, I'm sure.

My question is-How are these nipples held in the bearing housings? I couldn't see any bolts, just a rubber gasket and plate in a groove. Is there anything more than this? Does stress on the nipples cause bearing misalignment?

Danny,

The pipe nipples (water cooling) connect to the inner bearing housing and not the outer housing. A lot of vibration on the pipes indicates loose inner housing that is probably driven by rotor unbalance. There is a nut on top of the housing that preloads the inner housing and must be tightened to a specified torque value.

Walt

Walt,

I called to check up on things and found that nothing had been done The planner thought the tour foreman was taking card of it and vice versa. They checked back and found that there was a sharp increase in bearing temperature and vibration (overall from a mis-mounted PCM/Beta sensor) last Wednesday at 6 pm so Monday at 3 pm seemed like a good time to get on it. They have a 100 hour guaranteed response time to emergencies, I guess Is it free now?

Danny,
Usually the Cooling water supply and return lines are connected through flexible (SS braided) hoses to the nipples, otherwise due to high vibrations the nipples could be snapped off, the leaking CW contaminates lube oil instantaneously and ends up with a catastrophic lubrication failure.
Regards
Syed

I agree with Syed,

The ones I check all have flex lines connectred to the bearings. Seems strange one wouldn't.

Dave

Danny, Sam is exactly right. The nipples just thread into the bearing liner and are not attached to the housing at all (or at least they shouldn't be). I actually have a measurement point defined for each bearing where I measure the vibration on the cooling water nipple. My vertical point on the bearing is on the "stop" bolt or "plunger" bolt (not sure what the technical name is) that Sam mentions. Comparing the spectrum from the bolt point and the nipple tells me how loose the liner is. The liner will sometimes get loose from high vibration, from a loose locknut on the bolt, or from wear on the liner "runners" if the oil is dirty or acidic.

Your situation sounds serious... it's possible the nipple is cracked in the threads and is leaking water into the oil. They need to at least pull a sample of oil and check it for water. If the nipple is not broken off in the bearing, it can be changed out without disassembling the bearing. Just remove the rectangular seal plate and gasket and screw the nipple out.

Folks,
I always used this example for how spike energy 'demod spectra' worked much better than velocity on low speed impacting events.
The acceleration time waveform was a good indicator as well.
Ron Brook

Spike_Energy_on_Dodge_Sleevoil_Bearing_before_and_after_torqueing_locking_nut.doc (103 Kb, 35 downloads)

Yesterday when I told the maintenance planner he said he would get right on it. Then he told me at length about the new restaurant he tried in Richmond last week. I should get there between 7 and 9, get a window seat, order the 3rd bottle of chianti on the wine list and get the squid.

It sounds to me like everyone here agrees that this is a serious situation. Now if we can just get it put ahead of squid in the priorities of the owners maybe we can stop a really big unscheduled shutdown.

Rusty,

I would normally take the reading on the adjusting screw but it is covered by an insulating blanket so I get it on the housing. I remember your posting about using the water piping but have never picked up anything until this one. I actually picked it up from the velocity spectrum which had many harmonics and made me look at the acceleration twf which shows clear spiking at running speed. Curiously, Ron, this barely showed up in PeakVue. I'll look back at my settings, but I tried several different filters and never saw it. (BTW running speed is 893 rpm.)

Thanks all,

Danny,
I know what you mean about demod settings. They can totally miss an event if not set properly. You might want to change the filter to a lower setting. What did the time waveform look like?
Everything else considered, could you be looking at a bearing insert that is just badly shot?

Danny,
One thing that I have found on Peakvue in a situation like this is that it may not show up regardelss of the settings. The reason being, if it is purely looseness or turning speed impacting that is not modulating above the HP filter setting, it will probably not be seen.

Scott

Gosh, what did we do before PeakVue? On this type bearing, I don't believe PeakVue will tell you anything at all, no matter what your settings are. As I recall, PeakVue looks at "stress waves"... here you have a journal running on an oil film in a liner that "floats" on lubricated runners in a large rigid housing. Where exactly are your stress waves going to come from?

Rusty,

Since there was motion that would seem to create metal-to-metal contact in the pipe nipples I would expect stress waves from that. Apparently there was not enough movement yet.

I wouldn't expect them in normal circumstances but we don't work in those very often, do we?

I caught it because I use all my tools. Something I learned from old timers like you.

Rusty,
I am older than trees. I used to laugh at the old timers that wanted everything in displacement. I knew that velocity was a much better indicator of vibration severity, but I gave the customer what he wanted. So, now I look at machines for 25 years, with velocity spectra, acceleration time waveforms, and all of a sudden people are asking me about demodulated spectra, spike energy, peak vue, etc. What?
Fortunately, todays data collectors don't adversely affect my time in the field to take all of this information. So, on the odd occasion where one of the measured parameters indicates a problem better than another, it's a perfect world! Pull out that data and present it to the customer.
I have been accused for years of 'shooting from the hip'. Fortunately, a very large majority of the shots find there mark. I am very thankful that the ability to take more than enough data is easier than the old days.

Danny, using "all your tools" is a real key to being truly successful in this business. But, as Ron points out, we keep getting more and better tools so we have to work harder at staying "current". Some of the tools really are better, but others are IMO simply "marketing" as companies keep trying to distinguish themselves from competitors or justify the cost of "upgrades" (often meaning, "stable revenue stream")

Don Rainey, another old-timer and a mentor of mine, is fond of saying "It doesn't take long to collect more data than you can explain." So I try not to collect just a whole lot more than I need. If anything, I probably collect data on a lot more points than is really necessary, but I have found you can't predict which measurement positions are going to give the clearest picture of a problem.

For Sleevoil bearings my first tool is my hand that I place on the bearing housing. It gives me a good indication of temperature, vibration level and looseness (impacting). I'd leave PeakVue in my tool box for another machine. I did not graduate from the Bently school of proximity probes, so it is easy for me to say forget about them for these fan bearings. Either bearing housing velocity or displacement using accelerometers or seismic velocity sensors is fine for measuring vibrations on large sleeve bearings on fans.

Walt

Rusty, I wish I had asked this. The housing may not always be too rigid.

quote:

As I recall, PeakVue looks at "stress waves"... here you have a journal running on an oil film in a liner that "floats" on lubricated runners in a large rigid housing. Where exactly are your stress waves going to come from?

Since stress has units of psi (N/m^2), how do we measure this? Rusty's later statement may have some insight with regard to marketing.
\

Hi guys,
I’d like to share my experience on this subject. I used to monitor 25 fans with sleeve oil and around 18 synchronous motors with sleeve brg. And only based on experiment, Peakvue revealed to be a good indicator of brg looseness when prox. were not available. The set-up is crucial and programmed with PeakVue assistant in the frequency calculator. The machine’s speed range from 164 rpm to 900 rpm and waveform was a good indicator of once per revolution instability. It was trend able and brg measurement was over spec. The trend increased over a period of time before temp start to increase. For some reason, PeakVue’s waveform on low speed application is a better indicator than the spectrum. Prox with traceable calibration is the perfect tool. Still, housing reading that shows big difference, confirm holding down bolt looseness. Of course it is all achieve by experience and I would not be able to explain the reasons behind those result but the results have been confirm with physical measurement of the brg.

Best regard, Marcel