Spectrum analysis as quality control for gear reducers?

[geesamand]

Our company produces gear reducers. One challenge is that our customers are advancing into vibration spectrum analysis and PdM, yet we as manufacturers have not been actively developing this. (We do a quick unloaded spin test). Our equipment runs from .25 to 200HP (mostly 1-100) at speeds of 16.5 to 100RPM.

The goal of the vibration spectrum is to provide a useful reference spectrum for that machine to 1) verify smooth gear and bearing operation and 2) show relative quietness that will help isolate problems related to improper field installation (mounting, imbalanced rotating components, etc).

We'll need a load brake of some kind, which I expect will be the most costly capital investment. Has anyone found success in doing loaded tests at partial HP, or reduced RPM? How about running a flywheel brake and using ordered measurements?

Any suggestions or pointers are appreciated as we enter the initial planning stages.

 

[GregLocock]

On cars we test at all throttle openings and speeds, so the technology is out there.

For first steps a water brake probably makes the most sense, they are easy to set up and robust.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[MikeyP]

A water brake may struggle to generate a stable load at these low RPM, but a particle brake is probably a lot more expensive. You could run 2 gearboxes back to back to have a high speed at both motor and dyno ends. Preferably the increaser gear box should have different tooth numbers to the test gearbox, so the order responses from the two gearboxes do not get confused

M

--
Dr Michael F Platten

 

[SomptingGuy]

We measure geabox noise and vibration with a 3-room rig. In the first room there is a motor; in the second the gearbox; in the third a brake.

 

[geesamand]

Thanks for the suggestions.

Cost is a major factor in our approach.

These will be very short tests, and don't see the need major amounts of equipment to monitor the brake. Just the simplest way to apply a controllable power draw on the drive, do a vibration check, and dismount it. Ideally it would sit on the floor near the assembly area of our factory. Our factory once had such a hydraulic brake, but it was discarded at some point and not particularly missed.

 

[GregLocock]

Well, today's obviously dentist day. What was wrong with the system you had?

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[geesamand]

Testing under no load is of little or no value. So the first major upgrade in diagnostics is to put some torque on the units. Ideally full bhp and rpm would be used, but if a reduced power of some kind still yields useful results we'll consider that option.

The vibration spectrum would serve to verify the condition of the parts as-assembled, and in the case of something wrong it would be a first step toward locating the problem. As a bonus, it will serve as a very rough baseline to compare with field vibration reading.

Are any of these wants/needs out of line with practicality?

 

[Tmoose]

http://www.hq.nasa.gov/office/codej/codejx/Assets/Docs/RCMGuideMar2000.pdf

Section 9.4 appears quite similar to GM/s spec for new and rebuilt equipment.

Yes, an unloaded gearbox might be able to keep some secrets to itself, but not all of them. And the unloaded condition is at least somewhat repeatable. I predict you will be kept pretty busy 're-working" gearboxes if you adopt, or are forced to adopt, the limits in 9.4.1 and in 9.4.2 on page E-222 .

There have been more than a few variable speed machines whose gearboxes sing loudly (with attendant "high vibration at least on panels) at a few speeds somewhere in the range.

 

[geesamand]

Our gearboxes are designed to run at one speed, and our housings are such that things rarely sing unless it's the gear mesh itself. I can envision testing at reduced torque but keeping the operational RPM.

Our problem is really quite simple - we currently test a unit unloaded in the factory using a handheld total vibration meter. Our customers will sometimes install a unit and it proceeds to make major noise and vibration, and ask us how we didn't detect the problem. Post-mortems suggest that the problems are related to gear mesh errors and that the noise only appeared when loaded.

That's where this is going - to avoid looking foolish when the customer's vibration experts immediately detect a manufacturing defect. Whatever level of additional testing will be justified by the offset in warranty costs.

Thanks again.

 

[Tmoose]

"handheld total vibration meter"

What measurement points, vibration units, and frequency range?

The device may have an output available to process with fairly inexpensive hard/soft ware

 

[GregLocock]

A good example of this would be the diff, which distorts enough under laoding to substantially change both its efficiency and the vibration it produces.

So, it's obviously still dentist day: What was wrong with the hydraulic brake?


Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[geesamand]

The hydraulic brake had mechanical difficulties and was scrapped prior to my presence at the company.