Hello everyone!
What are the differences between a velometer and an accelerometer? is the accuracy of one greater than the other? and how does the velometer acquire its readings ?
Thanks all

VibeHi,

A velometer is a standard ICP-type acceleroemter with an analog integration circuit inside the case that provides a voltage proportional to acceleration. The frequency range is typicallly less than an accelerometer, because of the integrator cicuit. I believe the accuracy is similar. I suggest you look at specifications on the manufacturers' web sites. The are many manufacturers including: PCB/IMI, CTC, Metrix, and Wilcoxon.

Walt

Thanks for the info,
We have a confusing situation here , on this particular feed pump we have a velometer mounted on the pump IB and OB bearings, connected to the control room DCS system, so we are monitoring 24/7, recently the readings started increased to 0.3ips, however our CSI 2130 says the overall vibration is 0.2 ips, our transducer was checked and verified to be good( 160 Hz at 1g). The hand held vibration meter also corresponds to the velometer readings, which value should I believe? Can anyone explain this.

Did you check the complete chain into the DCS system? Anything is possible. I think normally that the thingy connected to the DCS would be 10-1000Hz filtered but you never know. It may normally also be so with the handheld, that may not be the case of the CSI 2130 pending the settings in umpteen places. It does not make sense as I would normally think that the CSI would indicate higher if there is something outside 10-1000Hz. Have you seen any change in the frequency contents of CSI data? There is also the problem of transducer placement, measuring on a little different place may give a very different reading. Third, do the systems use the same setting, RMS, Peak etc? Sensor temperature sensitivity may also be a factor. So there are many reasons why 2 systems rarely give the same reading. I am surprised every time I get similar readings. On the other hand a trend change is a change and should be investigated. My guess is that it´s a frequency range problem so your change in the machine is either a hi or a lo freq. problem. Olov

Vibehigh, exactly what kind of "velometer" do you have there? (Mfg. & model number)

Is this perhaps an IRD 544 type pickup?

VibeHigh, there are two types of "Velocity Sensors", that are commonly used today. One is a true Velocity sensor, usually a large case with a mass that moves through a coil and generates it's own signal. Generally a very good sensor for lower frequency measurements. It is also fragile, compared to and IECP style accelerometer. This type if sensor is physically smaller, because it is an accelerometer with and integrator circuit (as mentioned in other posts), that converts the signal to a velocity signal. This type of sensor is more robust for portable use, but not as good at the low frequencies, as the coil design.
You need to verify what type of sensor it is, to make sure.
As mentioned in other posts, have you verified that the DCS is monitoring velocity "PEAK" of "RMS" and that the CSI unit is set for the same type of values?.
The trend increase is still a concern, but without real data (not the overall signal increase from the DCS) it is difficult to determine the cause. Is it driven at a low frequency due to imbalance, or coupling issues? Or is it higher frequency, generally associated with cavitation, or bearing defects?

This is the issue when sending "Vibration Data" to a control room from an "Overall meter". The proper unit, along with an understanding of what signal it is sending to a DCS/PLC type system, is crucial, in understanding what the change is related too.

Gary K.

Thanks guys for all the excellent info.
OLI

I will have to verify that the settings on the DCS system are the same as the CSI 2130.

Rustythevibeguy : I am not sure of the make or manufacturer of the velometer, I will get that info to you as soon as I check.

I have inserted signatures, I shall be grateful for some opinions

east_frac_botts.doc (173 Kb, 33 downloads) Fractionator Pump

If it is a Bently system, the DCS reading is almost certainly going to be in velocity 0-Pk. Do you have your CSI set up to read velocity in rms? If so your CSI will likely read about 70% of the monitor reading assuming that the vibration is mostly running speed.

Make sure the CSI is in 0-pk and try it again.

Vibehigh,
If your velocity sensor is connected directly to the DCS, then it is a velocity transmitter with 4-20 milli-amp output. Some manufacturers offer three output flavors: RMS, Peak (sine wave), and True Peak. Your waveforms have a Crest Factor greater than 3, so there will be a difference between Peak (sine wave equivalent) and True Peak. Another issue is to make your analyzer F-min and F-max settings the same as the permanent sensor's frequency range.

Walt

Thanks everyone , basically what you guys are saying it is most likely a difference in settings between the CSI 2130 and the velometer.
In any event why would we use a velometer over an accelerometer? does a velometer have special applications?
Walt I am glad you mentioned the crest factor, in my waveforms, what exactly does it represent? what values are acceptable?

Vibehigh,

Creat Factor is the ratio of Peak/RMS. For a single frequency sine wave the ratio is 1.4. Any crest factor that is larger tan 1.4 simply means that multiple frequencies are present. Vibration meters and analyzers that display the peak value often calculate the peak from the RMS x 1.4. This is only correct for a single frequency (sinewave) and under-estimates the true peak value when the Crest Factor is high.

Your 1st measurement point had a True Peak value in the waveform of about 0.6 in/sec and the indicated Crest Factor was 4.09. The RMS value would be 0.6/4.09 = 0.147 in/sec. The calculated Peak (based on single sine wave) is 1.4 x RMS = 0.2 in/sec. None of these values matches 0.3 in/sec on the permanent sensor, but you can see where some differences can occur.

Walt

A lot of talk about the type of transducer and advantages/disadvantages of each - all good discussions.
The fact that the readings have recently gone up is cause for concern. Even if the readings differ between DCS and CSI analyzer(for whatever reason), it seems that the monitoring system sees a change at both bearing locations.
"Assuming" your monitoring system is working OK then you really need to provide more detail on what has changed to solve the mystery. Provide some details on FFT, waveforms, process changes if any, etc. Then we may be able to help a bit more.

Regards
Jim P

Velomitor® is trade marked by GE. Other than that it is an integrally integrated accelerometer. Spell it differently, and you are probably ok. Marketing stuff.

"In any event why would we use a velometer over an accelerometer? does a velometer have special applications?"

The OP asked the above. You use a piezo-velocity transducer for several reasons, limitation of noise, no-lag velocity signal, limitations of diagnostic equipment, and maybe some other reasons.

One, you can get a velocity signal without integrating an acceleration signal in a monitoring system. This means that any noise in the signal between the mounting location and the monitoring system is not integrated from acceleration to velocity. It (the noise) still may be integrated from velocity to displacement, however.

Consider the signal to noise ratio. If you are measuring a 0.1 ips 0-pk, 15 Hz signal, then the acceleration in g's 0-pk will be 0.024. If the sensitivities of the two transducers are 100 mV/g and 100 mV/ips, then the velocity signal will be 10.0 mV and the acceleration signal will be 2.4 mV.

Another reason for using the piezo-velocity setup is that the velocity signal has virtually no transducer phase lag. If you choose to use a moving coil transducer, then the transducer has a frequency dependent lag. Some monitoring systems are not set up to handle transducer lag.

Another reason for using the piezo-velocity transducer is that some analysis systems, notably ADRE, don't allow double integration. If you want to view vibration in mils, then you must use a velocity transducer.

Michael Titone

One can find a plot of phase response at http://www.ge-energy.com/prod_serv/products/oc/en/downloads/330500_vpvs.pdf for a particular transducer in the last figure. This is not flat response.

A high pass filter is (should?) be used as part of such a device.

http://www.ge-energy.com/prod_serv/products/oc/en/downl...0750_330752_htvs.pdf is different. This has a 150 degree lag at 10 Hz.

Michael,

you mentionned signal to noise ratio in your latest comment.
What would be an acceptable range for a monitoring system for the full scale we want to monitor. For example, if we need to check from 0 to 10mm/sec peak, is it OK to have a signal from 0 to 10mVolt? What is typical?

Thank you.

Charles Auger

I was simply comparing the values of an acceleration or a velocity signal. I was attempting to point out that in the lower frequency ranges, for a given vibration, you have "more" velocity signal than acceleration. As many of us have studied in classes, acceleration tends to attenuate lower frequencies and accent higher frequencies, displacement is opposite, and velocity is more even across a wide band. This means that at lower frequencies, you are trying to measure a naturally smaller acceleration causing the signal to be smaller.

Obviously, your acceptable signal range will depend on the noise level present in your system. I'm not sure what typical noise levels are. For the applications I deal with, turbines, pumps, fans, etc., we mostly use standard piezo-velocity transducers, standard accelerometers, or proximity probes. Then we use good shielding and grounding to minimize the noise and we're OK.

Usually we are looking for values in the 0 - 0.5 ips or 0 - 1.0 ips range. For the case of 0 - 0.5 ips, the signal would be about 50 mV max for a 100 mV/ips transducer, right? I would expect accurate readings down to a value close to 0 ips, say 0.05 ips = 5 mV. Note these numbers are off the top of my head and may contain errors. I'm not sure how low we could accurately measure though.

Based on my possibly bad math, you are using a transducer with a sensitivity of less than 100 mV/ips, right? (sorry I don't speak metric that well - I'm getting there inch by inch) But you want to measure a low maximum vibration, right? If so, you may want to consider a more senstive transducer. This would give you more signal relative to whatever noise you have. Maybe someone else could chime in on what typical noise levels are.

Michael Titone

If you go to the nitty gritty lo signal you should look at HOW more output is produced. Piezo material produce noise, if you just make higher amplification and not higher raw signal you also amplify the noise. Moving coil velocity probes are better to amplify as the noise produced in the sensor is much smaller. To get 1V/mm/s is no problem with very good long term results. Olov