vibration analysis

[JK1]

Hai,

I am involved in compressor (used for ac kits) noise testing.  Task is to frame some standard or setpoint to declare a compressor as noisy.

Various tests are performed on the compressor and the vibration data is collected sensor: accelerometer module: Ni DSA 4472 and RT 8145.  Iam new to vibration and i would be happy if there are handful ideas how to carry out this.  I have been monitoring the comp and trying to find out the most actively responding region in the frequency spectrum but i couldnt conclude anything out of it.  The data is collected admidst of environmental noise and i need ideas on eliminating its impact on the DAQ made.

Give me the ideas on how to process and the ways in which DAQ and analysis can be effectivly carried out.

With regards

Jk

[xseadog]

Hi

have you checked out the sound and vibration toolkit provide by NI?

https://www.ni.com/en-us/shop/product/labview-sound-and-vibration-toolkit.html

It is not in standard package, thus you have to shell out for it.

Also suggest you try a hit for vibration analysis on entire site.

xseadog

[xseadog]

Hi again

check out NI web on demand video:

http://digital.ni.com/demo.nsf/websearch/631CF3B809F088C886256D2100780C4A?OpenDocument&node=12049_US 

xseadog

[johnsold]

I built some production vibration test instruments before the days of computer-based DAQ. We took two sets of products as judged by the Quality Control people using their old methods (screwdriver handle against inspector's ear and blade against product case). One set was the "good" group and the other set was the "bad" group. We took wide range spectra of each in the lab and determined the frequency bands which changed the most for each of several types of failure. We then built bandpass filters and detectors for each band and indicators for the inspectors. Because the inspectors were not always consistent, we had to adjust the thresholds a bit during the first week or two of operation. (Several units in each set were determined to have been incorrectly classified by the first inspection). Six months later we received a frantic call from the production manager to recalibrate our tester because the failure rate had dropped from about 40% to <10%. Lab tests showed the tester had drifted less than 0.5 dB. The production had gotten much better because every unit returned for rework actually had something wrong and the tester correctly identified the type of problem. Once they started getting good feedback, they built them better the first time.

Some experimentation was required to find the best place to attach the accelerometer to the housing. This system worked on a noisy production line for many years.

Watch for ground loops if your accelerometer case is grounded. You don't want a faulty compressor to kill your DAQ system and computer.

Lynn