I all
I recently collected some vibration data and took phase measurements as well
The vibration amplitudes are low but there are some odd phase readings “well they seem odd to me” can anybody shed any light or offer a suggestion as to why there are or why there seems to be differences between the phase readings? (167 degree phase shift between nde horizontal and vertical)
Is it a case of there will always be differences?
I have looked at the TA chart and although very helpful I cannot seem to get to grips with this.
As the vibration levels are relatively low could the resultant phase data be telling me that there is some unbalance preset or is there another problem?
Do phase readings vary this much when low level vibration is present? Sorry no axial measurements were taken.
Any comments greatly appreciated.
Mike

PHASE.xls (22 Kb, 58 downloads)

Mike, I haven't looked at your data yet, but remember phase is "triggered" by an integer event (1x, 2x, etc). If the amplitudes are very low, the phase readings may be somewhat erratic.

Thanks Rusty
I guess that not every response and result will be the same as the text books and charts show

Typically, phase is a measure of the time between a trigger event (strobe) and next positive peak in the filtered (i.e. sine wave form).

Measuring phase associated with low amplitude signals can be difficult. Try repeating the measurment several times with your current settings to see if you get the same results.

If you have a low 1X amplitude (say 0.5 mm/sec as in your data) your instrument might struggle to find the positive vibration peak if you have selected a large input signal range.
You might want to try a repeat of the measurement using a lower signal input range (if available)

quote:

If you have a low 1X amplitude (say 0.5 mm/sec as in your data) your instrument might struggle to find the positive vibration peak if you have selected a large input signal range.
You might want to try a repeat of the measurement using a lower signal input range (if available)

Thanks for the info things are starting to make sense, there is also a bearing fault frequency present in the spectrum, maybe I should have reduced my fmax to avoid this peak. Is this what you meant by the input signal range?
Mike

Not the Fmax, but the amplitude input range (if it is selectable).

Here's an example that might help:

Say you have selected an input range of 20mm/sec, then the +ve peaks on an actual 1X input signal of 19 mm/sec will be easily found. For the same input range of 20mm/sec, then an actual 1X input of 0.5 mm/sec will 'appear' to the box (depending on the quality) to be almost a straight line and picking out the peaks will be difficult.

For a low amplitude vibration signal, you need to select an appropriate/matching amplitude input range to best measure phase.

Thank you for a good explanation
I’m not sure if I can select the range with my equipment; when I collect data my equipment auto ranges to suit.
This is something I will look into
Getting there slowly tanks again

What make meter do you have?

After examinng the data that you posted - it is evidient the shaft is bent. The magnitude of the "bend" will be revealed by the amount amplitude of the vibration.

David D

Thats a bold call - could you explain how you arrived at your conclusion?

Without having axial readings, I wouldn't know how to use casing readings to differentiate between unbalance and a bent shaft and would love to learn something new.

Mike,

"Odd" phase pattern is not actually odd. It is just evident of other then imbalance forces (which might be dominant) acting upon this shaft and different response characteristics in H vs V axes. In your particular case it is likely presence of forces either from shafts or internal bearings misilighment (almost 180 and 0 deg difference in H-V respectfully).

As far as phase readings inconsistency is concerned it is not the instrument which is "struggling". The firmware algorithm can precisely and accurately detect waveform peak regardless of amplitude value as long as signal is higher then instrument noise. It is, again, a reflection of slight looseness in the mechanical system. Take more averages and result won't be much different.

[QUOTE]Originally posted by David_G:
Mike,

"Odd" phase pattern is not actually odd. It is just evident of other then imbalance forces (which might be dominant) acting upon this shaft and different response characteristics in H vs V axes. In your particular case it is likely presence of forces either from shafts or internal bearings misilighment (almost 180 and 0 deg difference in H-V respectfully).

David when I was asked to report on this I concluded that; Overall vibration levels were good but fundamental outer and inner race defect frequencies are present. I felt there was some sort of alignment issue between the bearings so in went in my report.
Low and behold the customer called to say that this motor had been recently overhauled and new bearings were fitted. Also the bearing housing was severely damaged so the bearing housing was metal sprayed and re-machined.
This is where I think that there is a run-out problem. The housings are not true to each other, similar to a cocked bearing situation.
Your above explanation of the phase readings; “now not odd” help to confirm this.
if Iget any more info I will post
Many thanks to all who replied