Zonicbook/ eZ-TOMAS tachometer query

[Allan273]

I have collected transient data from a machine however the tachometer input channel was behaving very curiously. The proximity probe that was looking at a shaft keway did not give a consistent voltage trigger, i.e. the first peak was gave a value of -6V, 2nd peak -3V, 3rd peak 0.2V and then back to -6V. The pattern repeats. What this meant was that every third tachometer peak (pulse) was lost in noise and so couldn't get an accurate phase/ speed measurement. To combat this I used a different tachometer proximity probe input (which gave consitently -4V peak/ pulse) on a slower speed shaft and used the RPM Multiply and assigned this 'calculated' tacho to my high speed shaft channels. This gave me accurate speed however when I come to view polar or bode plots for all 12 channells the phase signal is all over the place, e.g. steady 1x amplitude but phase 124deg one measurement then 248deg, then 48deg, then 298deg. Have I set this calculated tacho up incorrectly? If so, can I retrospectively change this without resurveying the machine?

Any help appreciated.

Thanks,

Allan

[Rob@NIUK]

Hi Allan237,

I'm sorry to hear about your problem with the proximity probe not giving a consistent trigger. How are you getting on with it? It sounds as if you have taken steps in the right direction by using another probe on a seperate shaft.

To summarise your problem as I understand it, you wish to aquire both the speed and phase of the one shaft and plot this data on both polar and bode plots to analyse the vibration of the shaft, however the probe in use is giving unwanted values.

As you are using another probe on a secondary shaft, am I correct in assuming you do not have an alternative probe to use on the primary high speed shaft to determine if the hardware is at fault?

I would like to gather a little more information about the problem to further understand what is going wrong and how to go about finding a solution. Would it be possible for you to upload some sample data and the code you are using to plot the data?

Are the phases varying between the individual channels or are they occuring on seperate tests?

As the zero degree point of a polar plot is the angular positon of the proximity probe, am I right in assuming that this position is not varying throughout or per test?

If you are unable to upload sample data, can I ask wether there are any noticeable patterns between the phase differences that you have been noticing? If so this may point to RPM Multiplyer causing issues due to the number of rotations in must go through to represent a single roation on the high speed shaft.

Could you possibly provide me with the RPM Multiplier value you are using to convert the slower shaft rotations to the highspeed ones?

Thankyou for your detailed descrition, hopefully we can shed some light on the problem at hand!

Kind regards,

[Allan273]

Hi Rob,

I've tried to upload some files but the data is rather large and I'm unsure if I can upload all the relevent reference files also. The folder has 151MB of files. Maybe a FTP site?

The high speed shaft (HSS) tacho could not be trusted as every third revolution there wasno downward spike in DC voltage - very strange. From the waveform view you can identify the keyphaser peaks and the point where one is missing although data is continuous. It's as if every third rotation the probe only sees a shaft and not a keyway. Maybe a screenshot would better illustrate this better. Therefore I was trying to use the low speed shaft (LSS) and multiplying by the gearbox ratio of 7.365. HSS is fixed at 13,110rpm, LSS is fixed at 1780rpm.

The phase at the bearing I'm interested in between the X and Y probes is always 128degrees. However both phases move around randomly yet still within 128degrees. e.g. X is 100deg, Y is 228deg, next X is 300deg, Y is 68deg, next reading X is 40deg, Y is 168deg. The polar plot looks like a 2D ball of wool as the phase jumps around so much. This phase jumping is the same at all bearing locations though not 128degrees between X and Y.

Tacho set up used was LSS voltage mode and the HSS 'calculated' was voltage mode also but with the 7.365 multiplier. Both tachos were taken from the same panel output and split using a BNC T piece. This set up gave a very precise and stable speed measurement for both shafts but there is no 'phase marker' for the HSS except every 7.365 revolutions. Therefore I think eZ TOMAS is using the LSS reference to give me my phase data. Any solution to this would be appreciated. I can upload more data or screenshots if required.

Thanks,

Allan

[Rob@NIUK]

Hi Allan,

It sounds as though you are taking steps in the right direction by narrowing the problem down to the configuration of the tacho.

After looking into the problem further I have found some usefull information that may help you towards solving this issue.

Are you aware of the scan rate of your Zonicbook hardware?
If the tachometer trigger you are looking at on the HSS is equal to or greater than the scan rate divided by ten, then this is a possible cause of your curious lack of third spike.

With regards to the LSS, I would be curious to look at the data you have aquired. If the difference between measurements is a consistent value, specifically 131.4 degrees then this would suggest that the .365 fraction of a revolution is causing this difference.

In this case it may be usefull to upload some specific data you have aquired to see if we can pinpoint what is causing the unexpected phase data.

Thank you for your thorough response, hopefully we can pinpoint what is causing the eratic phase data.

Kind regards,

[Andy_Mellor]

Allan I have encountered this problem in the field and solved it. You need to increase your sample rate to capture the tachmeter signal with adequate definition. This is why the LSS gives you a reliable trigger and the HSS doesn't. Increase your fMax (and if you want to maintain spectral resolution, also increase the no. of lines.) and you should be OK - look at the values of max / min RPM you get in the configuration tab - I'll bet your HSS rpm was more than the max RPM your setup was suitable for. Unlike say, ADRE, the Zonic book does not cater for synchronous sampling, so you may run into issues such as this and you need to pay attention to the allowable speed range. Zonic and eZ TOMAS training provided on the basis of > 8 year's field experience. > 25 year's experience machinery troubleshooting. andy . mellor @ pmar . co . uk.

[Andy_Mellor]

One more thing - You can only get phase measurements (at least in my version of eZ-TOMAS) with the tachometer in 'Counter' mode. Voltage and multiplier mode won't give you phase. There is no cross-channel phase potion. This has driven me to frustration to the point where I ended up writing a VBA macro to calculate relative phase data from exported time waveforms. You can get cross-channel phase in eZ-Analyst. andy . mellor @ pmar . co. uk