Sign up Calendar Latest Topics Donate
 
 
 


Reply
  Author   Comment  
trapper

Sr. Member
Registered:
Posts: 75
Reply with quote  #1 
I have several Banbury mixers with double-reduction Lufkin gearboxes with double helical gearsets and dual output shafts and 2000 HP motors. All of them are critical to production. 

When I first arrived at the plant, we were only taking horizontal readings on the gearboxes. After a following around a Lufkin rep doing internal inspections on all the gearboxes in the plant I asked whether I should be taking vibration on all three axes or is the horizontal sufficient. He recommended taking all three axes if possible. On some of the smaller ones (not Lufkins) it is impossible to get all three axes so I get what I can. For all the critical machines, I added the missing directions under the theory that it's better to have more data and not miss a problem than have something fail that I didn't see until too late.

This is a picture showing one side of the gearbox and where I get the data:
#3 Mixer Motor Side of Gearbox.jpg
At the outboard bearings (input and output) I can get a direct orthogonal reading in all three axes on the bearing/gearbox case. For intermediate shaft(s), I resort to getting the horizontal reading on the side of the bearing access cover, the verticals are on the bottom of the bearing that is supported by the casing, and axials are either at the 5:00 or 7:00 position on the lower casing body rather than the cover. It takes forever (well, typically about 45 minutes at these speeds) to get the entire gearbox.

I was recently injured at work and so have the other guys catch up what I missed before the end of the month which included 2 of these mixers. I'm doing the analysis from home.

When I looked at the data, I see they got all three axes on the input shaft bearings but only the horizontal (the least reliable in my mind) data for the rest of the bearings.

What do you think? Am I wasting too much time collecting all this data? Should I only take the horizontals? It would certainly make my life easier and I could actually finish all my routes with time to spare every month. Further looking showed they did the same thing on a few other machines also. Grrrrr!

I'm of the mindset that I'd rather have too much data than miss something until a time when we have a bearing failure and it's proven to me that I wouldn't have missed it with only one direction

Your thoughts ...

OLi

Sr. Member
Registered:
Posts: 1,874
Reply with quote  #2 
I always try to get 3-axis data whenever practically possible, when you are at the machine reducing collection of data is just a waste of time. Collecting a few direction extra do not take so much time compared to getting there.
I have a customer that got 10-15 "cloth" mixers, run for around 20 years w/o any data collection. We collected data on the worst that is "screaming" (for help?) and asked if it could run another 6-8 months and I said yes it can so we will see.
They had a competitor there on another machine that claimed the motor was bad when it was the gbx. We got called since I once previously once upon a time pointed out the correct bad bearing on the correct part of the machine and even if it was inner or outer and I am intrigued they remembered that.

__________________
Good Vibrations since early 1950's, first patented vibrometer 1956 in the US.
http://www.vtab.se
RustyCas

Avatar / Picture

Admin
Registered:
Posts: 1,795
Reply with quote  #3 
Banburys were one if my least favorite machines. I monitored several for about 15 years. I took orthogonal data on the high speed and intermediate shafts and horizontals and axials on the outputs. I can’t ever recall seeing a problem on the outputs, except the bearing cap studs snapped off on one of the outputs on the mixer side where no one ever set foot - I could see the upper half gapping open. Didn’t show up in the data.

The bigger issue is constantly varying speed. Mine were batch mixers with about a 3 minute cycle time. Within that 3-min cycle, there were three (or more) different speeds/loads (dry ingredients, oil/carbon injection, high-speed drop), and the cycles, speeds, and loads varied depending on the “recipe” in use that day. Basically it was vibration analyst hell, especially in the early years when everything on that floor was covered in carbon black and around the mixer was the most tenacious mixture of carbon and oil in existence. I would collect data at the same point (in the cycle)

Trends were extremely hard to interpret as was the spectral data since you might have a scary-looking peak at one speed, but 10 secs later it was gone. If yours is similar, what good are all those points if the speeds vary so much?

__________________
"The trend is your friend"
OLi

Sr. Member
Registered:
Posts: 1,874
Reply with quote  #4 
Carbon black is special I have spent a few days in a factory for that crap. If your box collect data at the same time for all 3 directions it take no extra time. My example now was a nice wash cloth plant, that is much nicer. 
__________________
Good Vibrations since early 1950's, first patented vibrometer 1956 in the US.
http://www.vtab.se
Danny Harvey

Sr. Member / Moderator / Supporter
Registered:
Posts: 1,394
Reply with quote  #5 
I have monitored three Banburys all with Lufkin drives for 24 years.  For the most part, it is really boring. You can literally watch gears wear for years and not need to do anything other than monitor it. It was really hard to show a trend of wear using CSI with their limit of 18 spectra in a waterfall.  18 months is not very long in the life of one of these in my experience.

The one time that there has been a problem with the reducer, it was a suddenly snapped high speed shaft late at night.  Nobody saw anything fall into the mixer, though. Great big shafts just snap all the time, I guess [wink] There was no hint in the data or during the annual inspection that had been performed just weeks earlier. 

I can only recall two high speed bearing calls and they were both bearing looseness, not wear.  We found them by analysis of the high speed gearmesh in the axial direction.  There were no significant harmonics of running speed that would have indicated looseness but an increase in 2 x hs gmf with running speed sidebands told me (probably with some help from some of the people here) that the alignment of the herringbone gearing was off.

Mine are on a vfd but once they are loaded, it's always the same speed.  Materials of differing characteristics (plastics) definitely make the load vary and the only way to compensate for that is time. 

Because ours are batch loaded, time is of the essence. By defining my parameters to limit us to about 8 revs or so, we are usually able to get through the entire drive in time for one cycle.  We get a vertical and an axial on each point. It takes twenty minutes if you hit it right at the start of the batch.

Our reducers all have two output shafts within the Lufkin drive.  The used to have a set of gears on the Banbury but those have been eliminated over the years. The best thing about these drives is that they are really tough stuff. If they were properly applied and are being used according to the design, they will last for decades.

Combine oil analysis with vibration analysis for the best trends. And don't forget the oil pump.

The motors are where I find the stuff.
trapper

Sr. Member
Registered:
Posts: 75
Reply with quote  #6 
Rusty, you've described it exactly. I don't know how many times I've had to break a foot (or feet) free from the accumulated carbon black, grease, oil, sulfur and saturated absorbent pads just so I could walk to the next bearing. The frequently changing speed is frustrating and the cycle time is short, as Danny mentioned, and I too struggle. I do spent an inordinate amount of time watching the time waveform samples as they are collected looking for some type of repetitive impacting rather than the nice clean gear mesh other peaks in the spectrum. The smearing of the spectrums (spectra?) due to significant speed change is obvious and then wait until the next batch is loaded to redo it.

Danny, I agree -- boring -- but I keep reminding myself that one going down for repairs could easily send a good portion of the operators home for a while until the can get the material production back up to speed. I too have found the motor bearings the most prone to problems. We're changing the 4th bearing the next shutdown. One other thing our sister plant has found was the apparent not following the PM schedule for greasing the gear couplings. One seized up and some of the others were found really worn.

Oli, 20 years with no data collection and still running. That certainly makes you think for a minute.

So it appears I may have been too cautious. With all that experience, I feel more comfortable reducing some of the points taken. Thanks for all the comments.

RustyCas

Avatar / Picture

Admin
Registered:
Posts: 1,795
Reply with quote  #7 
Oil pump (5 hp) once had an “impact” and I recommended they replace it. Next survey (2 months later ) the impacting was worse. I told them it needed to be replaced ASAP. Couple of weeks later they called and asked if I’d seen any issues on the high-speed pinion shaft. “Other than the oil pump about to crap out, no. Why?”  He said the high-speed bearings had locked up the night before. And that was the last time I ever heard that failure mentioned.
__________________
"The trend is your friend"
Previous Topic | Next Topic
Print
Reply

Quick Navigation:

Easily create a Forum Website with Website Toolbox.