I am in need of some assistance in setting up the collection spec's for some of our vibratory conveyors. These are an all new animal to me as I'm used to looking at high-speed pumps, fans and motors. But I would like to branch out and start looking at our low speed vibratory conveyors. These conveyors are eccentric drives, that run in the range of 300-350 cpm. Most systems consist of a motor (10-30 HP, 1200 RPM) rigidly mounted to the frame, with a belt/sheave reduction to the eccentric drive. The drive is usually two pillow block style bearings on either end of an eccentric shaft, that has a flange (orbital) bearing in the middle that connects through an exciter spring and arm to the shaker deck, which is supported on a bank of reactor springs. We have a fair amount of problems with these orbital flange bearings. They get a bad "knock" in them and the maintenance guys end up changing out the entire drive. In most cases the mechanics are pretty conservative in their assessment of these drives and, in my opinion, they are changed out too early. And in a lot of cases, since they are replacing the entire drive, they are throwing out parts that are probably still good (because they have a hard time determining if the pillow block bearings are bad or if the orbital flange bearings are bad, so they just replace everything). This practice to me is wasteful, so I would like to try and help. And of course there are always a few cases where the problem does not get caught in time and the unit destroys itself and we suffer downtime. Also not a good situation, and again, I would like to try and help. My question is, what kind of readings should I be taking to get useful bearing data from these drives/motors? They are incredibly violent at 1x running speed (300-350 cpm), with velocity's as much as 10+ ips at the flange bearings. Even on the fixed pillow block bearings velocity's above .5 ips are not uncommon. What kind of collection spec should I be using? (# of avg's, overlap, windowing, filtering, lines of resolution, etc...) Should I filter out the 1x running speed altogether with a Hi Pass filter? Thanks in advance for the help.

Welcome Bill, glad you found your way here. I don't have any experience myself with these animals but you're sure to get some response. I have seen these types of conveyors at a saw mill in Dixfield MN. that I visited about a year ago. They were experiencing similar problems as you describe. Given the high amplitudes at low frequencies you're probably right to stay away from there and focus on the non-synchronous frequencies. Depending on your instrumentation, you may want to try using the HFD filters such as Peak Vue, gSE or SST whatever your instrument type is. This will help you determine if there are bearing defects or severe looseness (loss of shaft fit for example)Let us know what type of data collector you use and then we can talk about different set ups. Good luck.

As a first pass in monitoring these beasts, take your readings in ACCELERATION. This will essentlially filter out some of the high 1x. 10 IPS rms at 350 cpm is about 1 g RMS - a general purpose industrial accelerometer (100 mV/g) will handle around 50 g, so you don't need anything special for a sensor.

The dreaded dB/log scale will also be helpfull to show a wide range of levels, The big 1x peak will mask everything if a linear scale is used.

Jon
Spintelligent Labs

This message has been edited. Last edited by: Jon Chandler,

Bill,

A company that I used to work for performed the vibration analysis for a foundry and of course they had alot of these vibratory conveyors. What we found that worked the best there was to look at the level of the looseness in your spectrum along with the raising noise floor. Some of these bearings were buried under fine dust and we would watch the looseness indications and have them replaced when needed rather than as the maintenance guys saw fit. Putting your sensor on the eccentric bearings can be fun! We used poles with the sensor held at the end to put them on, and you still had to keep some presure on them to keep them on sometimes. One of our guys would routinely lose magnets on these. Good luck!

Thanks guys. I am using a dataPAC 1500 and the Entek Emonitor Odyssey software.

I don't try to put the accels on the eccentric bearing, but I have good luck with the pillowblock bearings. The ones I monitor have 3 pillowblocks with 2 eccentrics. I take a 3000 Hz waveform and trend the peak acceleration levels (waveform). If they are going to change the whole shaft (unit) out, I'd probably let them go until I can actually see a problem, or until the sound scares me.

I monitor simular conveyors here. As Jon pointed out measure in acceleration this gives a clear view of the bearings and eliminates the 1 times from your data. If bearings are failing often check with the manufacture the frequency range these conveyors are designed.Anything above or below it's frequency will destroy both bearings and eventually conveyor.
Hope this helps.
DaveL