I have been involved with vibration analysis for about a year and a half. When I first took over this position our local "guru" decided to tweak our AP sets. What he ended up doing was creating one AP set for everything to use. When he first did this I didn't even know what an AP set was, but the more I've learned the more messed up this seems. We have equipment ranging from low speed to fairly high speed (6 - 54000 RPM). The AP set uses an Fmax of 5000hz with 3200 lines, and we use magnet-mounted accelerometers. Do you see any value in using something like this for periodic data collection?

Thanks,

You need to consider what this frequency range may miss. If any of your machines are gear driven, you'll likely miss gear mesh frequencies. Perhaps rotorbar frequencies. Turbine blade rates. Centrifugal compressor vane rates. All of these things are pretty important!

Too narrow of a frequency range will definitely miss problems, and may make vibration monitoring almost useless. There is no "one-size-fits-all" solution, particularly not based on a fixed frequency range for machines of all speeds. Better is an approach like DLI's, where they typically use two ranges of 10 and 100 orders, but even this can get you into trouble with missing gear mesh frequencies and blade rates on cnetrifugal compressors.

Best to make do a reality check on machines that may generate high frequencies. What is the gear mesh frequency? How many turbine blades are there? Missing important forcing frequencies is also a good reason to take a high frequency overall reading. A 10 Hz - 10 kHz reading will cover most machinery. Regardless of the frequency ranges used in the spectra, a high frequency vibration would be detected. Just keep in mind that the overall level should reflect what you see in the spectra. If not, take data at a wider range to see what you're missing.

Jon
Spintelligent Labs

I agree with Jon,
However, if you're talking AP sets, then you are using CSI, Right? And if so, then you are probably set to take digital overalls for speed.
If you have your overall set to digital , then you won't see any difference between the spectra and the overall. Make sure you take an analog overall first time or two around, and use it to check that you have your Fmax out to where it should be. Once you are satisfied with your Fmax settings, and sure you aren't missing something, then set up for speed if that is what you are after. I am no expert with the CSI software, and am assuming that you can compare the spectra overall with the "overall" overall value.

Just my opinion and MO

Dave