I have decided to try my own Peakvue AP/AL sets as opposed to letting RBMware set them up. In the "Periodic Analysis Parameter Set" window it asks for input "Upper Freq. (CPM/Order)" is the same convention used here as a conventional point. For example a typical medium sized electric motor having ball bearings..perhaps 70 orders of running speed?

Thanks
Alan

Alan,

One thing about using an order based AP set, if the order(s) forces the Fmax beyond the limits of your set HP Filter, the next highest HP Filter will be used.

Example:
If you use order based and the running speed happens to force the Fmax to say, 1001 Hz and you have chosen a HP Filter of 1000 Hz, the HP Filter will be changed to 2000 Hz and the Fmax will increase to the next step also.

It will take a number of different AP setups to cover different RPM machines where a fixed Fmax will always be there.

Only my opinion.

Ralph has got it spot on. The PeakVue Fmax values available are: 20, 50, 100, 200, 400, 500, 1000, 2000, 5000, 10000 (all Hz)

If you are primarily looking for bearing related issues, then a 40 order Fmax will usually be suitable with PeakVue.

This also reduces the tendency to push the HP filter up, as Ralph has pointed out above.

I would also strongly recommend that you only take one average.

Ian

The most important thing to remember about Peakvue is that it will process any energy over the high pass filter as stress waves. A common complaint I hear on motors with peakvue is the high 60 peak, usually due to setting the HP Filter below the rotor bar pass frequency.

I recommend taking a high frequency spectrum and setting the high pass filter above any energy you see in that spectrum.

Rich

I'm attaching a file that lists all of the discrete fmax values for the various types of averaging for PeakVue and normal spectra. The filter group (HV1, HV2, HV3) is shown on the data collector during startup I think.

fmax.xls (20 Kb, 35 downloads) discrete fmax

There is a very detailed paper available on the CSI website that described the recommended setup.

The suggested setup has changed over the yesrs, they used to recommend a fmax of about 70 orders, similar to regular measurement point setups. Now it is 40 or 30 orders for bearings, and 2 to 3 times gmf for gearboxes.

The two most important items in setting up peakvue are the fmax and the filter selection. About 98% of the time you will use a high pass filter. The filter should greater than or equal to the fmax. The vast majority of the time filters between 500hz and 5000hz should be used. Extra caution should used with 5000hz filter settings, it is very sensitive in terms of the method of mounting.

I see advantages of using a fixed set for non variable machines, I will still use orders for variable machines, some of the machines can have widely set speeds for different processes (fixed when set).the one case I am struggling with at 14 to 20 rpm with a 22228cc brg. If an HP of 500hz is used, will it not miss the expected freq range expected of defectts? Would it make sense to have two Peakvue points set up one with a BP of 20 to 150hz, and the other with an HP of 500hz? Instead of using a conventional SST point? The first data that I took showed a failing bearing in the Peakvue waveform, but did not seem to be out of line in the conventional SST point. (of course this was with RBMwares goofy settings that I have since changed).

Alan

Alan,
What f-max are you using for the slow stuff. When you get into very low speed machines you need to use a very low f-max to see defects in the spectrum. I correctly called an inner race defect on a bearing turning 1 rev in 150 seconds. PeakVue samples at over 100kHz but it stores the maximum peak sampled in the period determined by the inverse of the f-max*2.56. So if you used a f-max of 100 Hz it will store the maximum peak 256 times per second. This would be 15360 samples per minute. A shaft turning 10 rpm with a bearing defect at 9 orders (I know it would be non-syncronous but this is easier) you would get an impact 90 times per minute. The higher percentage of periods with the impact you have the better it will be seen (stand out above the noise floor) in the spectrum. With the 2130 you can use a 1Hz f-max with PeakVue. That is what I use on the 1/3 rpm shaft now. Takes over 20 minutes to take a reading but this is a critical machine. Also I used the 500 Hz f-max high pass filter on this reading.