System One was recently installed at our facility and the Bently-Nevada "Direct" value is approximately twice that of readings collected with our portable data collectors. This appears to be only on the Velomitor monitored points. I received a 36 page explanation from Bently-Nevada describing the Bently algorithm versus RMS x 1.414. Has anyone else experienced a discrepancy between Bently-Nevada "Direct" levels and their own portable data collectors overall levels?

If you alarm or trip due to the Bently-Nevada values, do you double the recommendations given by API, ISO, CSI or TA? (This may seem obvious, but seriously, could be important for insurance reasons, RCFA, etc).

Velomitor? How does one set up a velomitor for periodic vibration? I was trying to field-verify the Bently values with a velomitor and an SKF GX and ran into issues. It's not truly a velocity transducer as we know it, but an acceleromitor with an integration circuit, correct? Do we supply power to the transducer? Is it linear in response? If I set this in the data collector as a velocity probe, it doesn't seem to work. When I set it up as an accelerometer, it appears to work, but I loose linearity at approx. 15,000 CPM. Any Suggestions?

Jeremy Cole

See the article on this topic in the November 2005 issue of Orbit.
http://www.gepower.com/prod_serv/products/oc/en/orbit/d...crepancyreadings.pdf

Also check the frequency set ups in System1 and your portable. I have seen this when the pertable analyzer frequency range was larger than that of the on line system.

The Bently Velomitor is nothing special, basically the same as a Wilcoxin or a IMI.. Yes you need to supply power, it is an accel with an integrator on board.

If you are taking data from a Bently Rack do not supply power as the power is coming from the rack, but if you hook a Bently velomitor directly to a data collector then yes you need power. The scale factor off the monitor may be 500 mv/ips but hooked up direct I believe it is 100 mv/ips, or at least it used to be...

Defaults in System 1 will be in/sec 0-pk. I am pretty sure you can configure channels for rms if want to though. I like 0-pk.

Jeremy,
For sytem 1 only: The Direct value is taken from the highest PK value in the waveform. The direct value is set for a specific frequency, which may be different from the synchronous or asynchronous settings. For example, The direct value may be from 10 to 20,000Hz and the dynamic data at 3 to 1000 Hz. These settings can produce different values just from higher frequency settings, but they still only look at the highest waveform PK (direct). My personal opinion is that the direct value is best used with proximity probes, which sees relative displacement of a rotor.

The other main reason you are getting different values is the way your data collector is calculating the OVERALL PK value. I don't know which method your analyzer uses but one way of doing this is Digital overall Pk, which takes the spectrum 3 to Fmax data and calculates SQR(a1^2+b1^2.....)/Sqr(1.5). Check your manual to see how its calculating the overall PK.

The term direct does not describe the measurement fully. There should be another item, p-p, 0-p, or rms. With System Uno, I would think you could change this.

Check the signal timebase to see what is being measured on your second device as well. Check the sensor calibration and what Uno thinks it is.

hi jeremy2104,for system1 you have to give proper band pass filters.otherwise you will be getting raw data which will be higher than portable data collectors.you have low pass,high pass filters which will eliminate unwanted signals.also you should calibrate rack for correct full scale range having proper trip multiplier.