This is a 480vac overhung fan motor. The dominant frequency seen on the axial direction is 6 * line frequency (magnitude 0.2 ips). This frequency is much smaller in the other directions and also has been much smaller historically.

I am used to hearing about 6*LF on dc since it is created on the output of 3-phase full-wave bridge rectifier type circuits.

I am not used to seeing it on ac machines.

1 - Could it be related to harmonics on the power system from another electronic source fed from the same ac supply? I know electronic sources can cause harmonics back on their inputs, but I thought these were predominantly 5th and 7th harmonics.

2 - Can it be related to electromagnetic noise from a nearby dc circuit? How can this be easily checked?

3 - any other ideas?

21600_On_VaneAxialFan.ppt (160 Kb, 39 downloads)

Hmm, interesting. The only time I've really run into sizeable harmonics of LF on an AC motor is when there is a VFD involved.

I could not speculate on your other theories. I guess the only "simple" way to check would be to hook up a power analyzer (such as a Dranetz) to determine the quality of your power.

Good luck on this one.

VFDs also perform full wave line voltage rectification prior to generating variable frequency AC output.

It is interesting as to why did 6xLF show up predominantly in axial direction?

David

Pete,

Is there an ac inverter present? What are the amplitudes in a linear scale?

Thanks for your ideas.

This not not fed from a vfd or inverter (although there may be electronics such as rectifier-inverters fed from the same ac supply as this).

I forgot to mention this is a 4-pole 60hz machine.

I'm not sure why only axial. We were thinking maybe that measurement source is close to something dc fed by electronics. We're going to try holding the accelerometer stationary in the air just off the axial measurement point and see if we get anything.

Danny - I added slide 3 in linear scale. Magnitude of the 6*lf is just over 0.2 ips.

21600_On_VaneAxialFanRev1.ppt (218 Kb, 24 downloads)

Coupling type?

There are no bearings for the fan, so either the fan is mounted on the motor shaft extension or else there is a very rigid coupling.

Note the frequency of the peak (21608cpm) is much closer to 6*LF (21,600) than to 12*running speed (21,415) . Also on log scale looking at double that frequency we see a tiny peak much closer to 12*LF than 24*running speed. Also on a number of other points on this motor we see tiny peaks always closer to 6*LF than 12* running speed. These frequencies come from Enteks "peak label" function which seems pretty accurate. All of this leads me to believe it is 6*LF and not 12*running speed.

Pete,

If you average the sideband spacing around 12.1083 orders as shown in the slide labeled frequency in orders you get almost exactly .2 orders.

I'm trying to figure how a shaft rub and scr firing frequency where there should be none could figure into this and am coming up blank.

Is this a new or chronic problem?

I'd say new problem.

This frequency has always been present on this motor but at extremely low levels. In the axial position in the past, magnitude always between 0.001 and 0.01 ips. On other positions lower or non-visible. Sister motors have that same behavior as well (present at very low levels, highest on the axial, always less than 0.01 ips).

This is the first reading ever above 0.01 ips at this particular frequency and it shot all the way up to 0.2 ips.

Pete,

In a past life I used to see this condition on Navy ships. If there were other machines fed from the same source, we would sometimes see high levels at 360 Hz on some machines. I'm trying to recall the exact situation....

I believe the problem was most common on vane-axial fans of a smallish size....20 - 25 HP perhaps (fan wheel installed directly on the motor shaft). As I dig a little deeper, and this part is a little fuzzy, I believe this was a problem on motors with a cast rotor, usually where the casting may have been off-center with all the little numbs used for trim balancing removed from one side of the rotor.

The problem can be difficult to track down as it depends on the state of other machines fed from the same power source. It might be high during one measurement and low during the next for no apparent reason.

Jon
Spintelligent Labs

UPDATE - A followup survey performed this morning (9 hours after the high vib found last night) showed that axial 6*LF is now back at 0.01 ips. The rest of the vib spectrum is unchanged.

That seems consistent with some kind of external interference (at least that's my story). Would feel better if we could find the source, but I'm not sure if that'll happen.

Any more thoughts?

I realize this is probably a pipe dream, but what about cycling loads on the same circuit while monitoring the offending test point? Kind of like Ye Olde Grounde Isolation Procedure. Other than that, and assuming this is purely an electrical problem, the only other thing I can come up with is recording load electrical parameters over time to see what, if anything is changing. Then compare the time/date stamp against what load may have cycled on/off.