Sign up Calendar Latest Topics Donate
 
 
 


Reply
  Author   Comment  
Shurafa2

Sr. Member
Registered:
Posts: 170
Reply with quote  #1 
I might install 4-20 mAmp overall vibration loops on some large fans. What consideration should I pay attention to based on your experience with these systems?

Most of my experience is on Bently Nevada high end products. Due to many reasons this option is less likely to be selected. One of the reasons is that the operation/shutdown of these fan is determined by the operators and auto trip is unlikely to be configured.

What are the limitations/options of the 4-20 mAmp systems that I need to consider during the design of the system? For example, do you recommend vibration and temperature combined sensors?

Regards- Ali M. Al-Shurafa
OLi

Sr. Member
Registered:
Posts: 1,912
Reply with quote  #2 
Vendor, we installed 1 sensor on about 60 bearings some years ago on fans in a diaper factory in Saudi where we delivered 3x4-20mA from each sensor so around 180 4-20mA signals. 1 vibration, 1 bearing condition and 1 temp. very successfull and just in time, quite a few bearings with problems was found to some surprise but obviously the damage was to be seen by anybody. So my suggestion is to not make it OA monitoring but filtered as it gives a better payback, if bearings are antifriction that is.
__________________
Good Vibrations since early 1950's, first patented vibrometer 1956 in the US.
http://www.vtab.se
vogel

Sr. Member
Registered:
Posts: 76
Reply with quote  #3 
What do you mean by a 4-20 mAmp overall vibration loops? Is it an accelerometer with a 4-20mA output or a standard 100mV/g accelerometer wired to a 4-20mA signal conditioner?

In both cases I've had not so nice experiences in the past due to the frequency response of these devices. Check the speed of your machine and the frequency response of the transmitter, note that usually the filters design of these type of devices is very poor and they are not filtering out everything that you would expect to be filtered out.

I would try to go with a continuous monitoring system that can measure waveforms and spectra, or at least several user-configurable spectral bands. If no auto-trip is needed, Bently and many others have systems with a price point close to 4-20mA loops (in case price is the issue).

If you really need to stick to a 4-20mA solution then I would choose something like Oli's suggestion: ISO velocity, bearing condition and temperature. If you can only have two, I'd skip temperature. For those machines or bearings that you cannot access with a portable anlyzer to diagnose a change in the 4-20mA trend, make sure to leave a BNC connector with the raw signal in an accessible area.
Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 599
Reply with quote  #4 
Definitely make sure the overall bandwidth is configurable. Then perform a little sensitivity analysis to ensure the overall value is stable when narrowing the bandwidth. Don't assume that the 4-20 sensor isn't picking up any high frequency content just because your standard ICP sensor doesn't see anything. We've limited our continuous monitoring sensors to 5-300 Hz.
OLi

Sr. Member
Registered:
Posts: 1,912
Reply with quote  #5 

Well, if you hit the mounted resonance of the accel with like gearmesh of the gbx, no filter will keep that out and accel will OL unless you swap to a MCV sensor :-)
We use classic analog filters, basically the same since 1985 and so far it worked with accel's, MCV's and eddy probes like 5500+ units later except one case as above with a
2xgearmesh at 18kHz, same as accel mounted resonance but swapping to MCV fixed that.
I don't like filtering if it is not handled so data are seen some other way specially I hate mechanical filtering so you can't even see anything in raw data, but that's only me.
In this digital world where the FFT is produced in the sensor IC, you also need to look out for the Aliasing ghost if the newbies don't check their sensor suppliers data.

 

 


__________________
Good Vibrations since early 1950's, first patented vibrometer 1956 in the US.
http://www.vtab.se
Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 599
Reply with quote  #6 
Quote:
Originally Posted by OLi

We use classic analog filters, basically the same since 1985 and so far it worked with accel's, MCV's and eddy probes like 5500+ units later except one case as above with a


There I go making assumptions that the digital state of the art is ubiquitous!

The sensors we use are simply oversampled at (I think) 12kHz and decimate. Then the overalls are calculated digitally (I assumed directly from discrete waveform).

Cool to see you are still competitive with setting the overall bandwidth with analog filters.
OLi

Sr. Member
Registered:
Posts: 1,912
Reply with quote  #7 
Well, they are not easily adjusted, you get what you order but they not crash and don't alias. Still if you oversample but not that much unless you have a delta sigma A/D and even then you may need some additional filtering. We have 2x4-20mA output, one 10-1000Hz or 1-1000Hz and one HP600Hz as standard.
__________________
Good Vibrations since early 1950's, first patented vibrometer 1956 in the US.
http://www.vtab.se
Previous Topic | Next Topic
Print
Reply

Quick Navigation:

Easily create a Forum Website with Website Toolbox.