15 HP 4 pole motor is sitting on a "shovel" type plate bolted down to a speed reducer flange. Coupling - tire type.

Original spectrum had dominating 2xRPM thus "suggesting" presence of misalignment. Gross misalignment was in fact found and corrected. Surprisingly, data collected afterwards did not change a thing. 2x with harmonics were still there and 1x orders were small.
Of course there is a possibility of resonance close to 2x, but...
1. this pattern is present in all directions
2. it appears that 2x is the fundamental frequency

Very doubtful it is a bearing fault (don’t have bearing info).

Of course, will perform a bump test when feasible and take hi res spectrum.

Any experience, speculations?

Thanks,
David

Top_screw_2xRPM_orders.doc (114 Kb, 50 downloads)

David,

Is there a sprocket on the reducer's output shaft by any chance? Their mesh sometimes falls close to running speed harmonics.

Hope this helps.

Sounds like the old scoop type design. Have seen several of these with the same symptoms. A permanent vertical brace from the floor to the bottom of the motor base took care of the problems I saw.

Dave,

The unsteady waveform amplitude in Figure 1 indicates either a close natural frequency or Beating with another vibration source. A simple impact test should clarify this. Also, if misalignment was severe, then the elastomer coupling element should be replaced, since that can get distorted. I once did a precision alignment, and the vibration was higher than when I started. It was a clearly distorted "Woods-type" coupling element.

Walt

Just about any kind of reducer with a sugar scoop or shovel type motor mount is prone to wagging it's behind.

Vertical bracing can cause a shift to horizontal movement and vice versa.

Being that this is only a 15 hp motor, you can try a simple test while it is running. Try putting a couple of sand bags on top of the motor and see if you don't change the mass enough to change the natural frequency. The looks from the bosses alone will be worth it and if it works (and I have seen it before), you'll be a god

Or you could do negative linear averaging or wait for it to be down and do a bump test.

myfirst call with a predominate 2X is usually loose base

D