I have two sites that collect data and I analyze it. Every now and then I'll get what looks like either the accel was popped on or not enough settling time was allowed for remote sensors. The FFT looks terrible on the low end, but viewed in accel, looks 'fine' in the high end. And since we're taking between 4-6 samples and CSI only saves the last sample, the TWF is still good, right? How good is the data I see? I know I'm out on a limb here--just trying to lemonade out of a lemon of a reading.

Wish there was a way to generate an FFT straight from the TWF you see.

Bad data are bad data and you can't make a silk purse from a sow's ear.

Don't use anything in question! An FFT on TWF is the spectrum. And why integrate? Velocity will never be better than acceleration unless it's a velocity transducer. Integration is relative, not absolute.

I integrate, therefore I am--relatively speaking.

I like velocity. Everybody likes velocity. 0.19 Gs at 1X doesn't sound as bad as 0.7 ips-pk. I still like acceleration and the people who use it, though.

I have always been a velocity fan by birth, real hi quality moving coil velocity probe. I was born holding a velocity probe as my father invented a design we now make the 3rd generation of since 1948. I have lately, like 20 years had to more and more and reluctantly admit that integrated acceleration works, good enough so I don´t agree on this with you Sam. For cost reason and adaption to defacto industrial standard I now supply hundreds of accel´s and use them daily and it works. Only hi quality places with hi demand on none failure like nuke plants and hydropower use velocity probes these days. And the data I get from the accels integrated signal is normally good enough or I retake data if I have been careless when putting down the accel. Note that I use flat rare earth magnet, not those rail things for curved surfaces. Only my thinking. Olov

quote:

Originally posted by OLI: . . .
Only hi quality places with hi demand on none failure like nuke plants and hydropower use velocity probes these days. And the data I get from the accels integrated signal is normally good enough or I retake data if I have been careless when putting down the accel.

Two very different areas of the power industry. Many hydro-turbines would have a very low fundamental for a traditional velocity transducer; the running speed could be 150 rpm or below, which makes it difficult for a velocity transducer (sub-harmonic vibration??).

I know I used many accelerometers in the nuclear business in some time ago. With Bently Nevada we did use velocity transducers on much of the equipment, prior to Bently using the integrated accelerometers routinely. I could see the integration electronics being a potental issue in a high radiation area, due to possible degradation over time, but velocity transducers also age.

Just curious (magenta).