Sign up Calendar Latest Topics Donate
 
 
 


Reply
  Author   Comment   Page 2 of 2      Prev   1   2
electricpete

Sr. Member
Registered:
Posts: 647
Reply with quote  #16 

Quote:
[edited into chronological order:]

27/07/2018 overall is 16.06 mm/s rms, spectrum 15.62 mm/s rms and wave 15.97 mm/s rms

03/08/2018 overall is 8.08 mm/s rms, spectrum 14.43 mm/s rms, and wave 8.23 mm/s rms

 

As others mentioned, 16.06 / 8.08 are “analog” overalls (includes all frequency content) and 15.62/14.43 are “digital” overalls (only shows content below Fmax).  A clue is the label V-DG where DG stands for DiGital.  And the 15.97 /8.23  is TWF rms which only captures frequencies below the sample rate.

The logical conclusion from examining those numbers is that you had an decrease in high frequency content above spectral Fmax and TWF sample frequency in the 8/3/18 data.   That decrease showed up in the analog overall  as expected since it captures all frequencies, but it didn’t affect the digital overall or TWF rms because those are only seeing frequencies below Fmax (digital overall) and sample rate (TWF).

Since the change occurs above the Fmax and sample rate, we don't particularly expect it to show up in our spectrum and twf. But we can also see that the true peaks are lower on 8/3. That seems consistent with a decrease in high frequency content.

Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 599
Reply with quote  #17 
Quote:
Originally Posted by electricpete

As others mentioned, 16.06 / 8.08 are “analog” overalls (includes all frequency content) and 15.62/14.43 are “digital” overalls (only shows content below Fmax). 


Hold the phone... do the CSIs really have a true analog, dissipation-based RMS unit? Or does it still happen after A/D? It would also make "sense" if the analog was just bandwidth'd 0-fmax and the digital was bandwidth'd 10-1000 Hz, for example.
electricpete

Sr. Member
Registered:
Posts: 647
Reply with quote  #18 
Everything happens after the A/D.    So maybe it is misleading terminology for people to call it "analog" overall or for me to say it includes "all" frequency content.  That "analog" overall includes all frequency content up to the bandwidth of the analyers which is 20khz or more for CSI (in contrast the "digital" overall is limited to a much lower frequency range below the spectrum Fmax). 

Previous thread on the old board
https://www.maintenance.org/topic/analog-overall-vs-digital-overall

...especially CSI Tech Note 95-00008 on the first page of the attachment posted by some guy named John (attached)



 
Attached Files
pdf CSI_2130_cautions.pdf (141.23 KB, 16 views)

Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 599
Reply with quote  #19 
Can I infer from that that CSI always samples at a fixed rate and just later downsamples as needed (after the analog overall is calculated, if applicable) to meet a given fmax?
fburgos

Sr. Member
Registered:
Posts: 670
Reply with quote  #20 

In the point setup I have,

integration mode override: analog
overall mode override: rms digital

This overall is after integration of the acceleration wave to velocity wave, then fft takes place on the velocity wave.

I believe then my overall trend and spectrum trend must be the same, as in 3/8 not like 27/7 where overall is lower than spectrum (both are digital rms)

Also my special waveform is specified in g (different acquisition time) and it’s similar to the overall trend when I forced a digital integration on mhm.

how do you setup your overal and integration mode?

OLi

Sr. Member
Registered:
Posts: 1,913
Reply with quote  #21 
Would this be related artifacts from the 2100, 2110 times when the box actually had a analog, not dissipation but logarithmic but still analog RMS calculation and there never was a ability to get the same number in the box as in the PC software unless you had the specific settings (now long forgotten) to generate the same or similar value? It was a lot of discussion at the time related to this.
__________________
Good Vibrations since early 1950's, first patented vibrometer 1956 in the US.
http://www.vtab.se
electricpete

Sr. Member
Registered:
Posts: 647
Reply with quote  #22 
We discussed earlier how the overall  can be digital or analogue (I gather that is overall mode override selection).  Apparently CSI also has a digital vs analog selection for integration mode.  I'm not familiar with that, but just found this link (again on the old board):
https://www.maintenance.org/topic/csi-2130-integration-mode


John from PA

Sr. Member
Registered:
Posts: 963
Reply with quote  #23 
Something likely worthwhile in all these threads.

https://www.maintenance.org/topic/csi-mathematical-procedure?reply=399590942964580184

https://www.maintenance.org/topic/analog-overall-vs-digital-overall

https://www.maintenance.org/topic/difference-between-digital-and-analog-integration

https://www.maintenance.org/topic/integration-and-overall-mode-override

https://www.maintenance.org/topic/integration-digital-vs-analog

Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 599
Reply with quote  #24 
Wow, I feel pretty annoyed at their flagrant misuse of analog/digital. You see this with all sorts of relic terms like 'unfiltered' which in no way is unfiltered. People get mad at mixing up damp and dampen, but this disregard of logical terms is downright misleading.

Quote:
Originally Posted by electricpete
We discussed earlier how the overall  can be digital or analogue (I gather that is overall mode override selection).  Apparently CSI also has a digital vs analog selection for integration mode.  I'm not familiar with that, but just found this link (again on the old board):
https://www.maintenance.org/topic/csi-2130-integration-mode


This was also plain in the pdf you had linked above. Integration mode just determines whether acceleration is converted to velocity in the (undecimated?) digital waveform or the digital spectrum. It also says that if digital integration is selection, only digital overall is possible, otherwise you get g's in your overall. At that point, you may as well just have a high frequency band instead of calling it an overall. All misleading... *sigh* 
electricpete

Sr. Member
Registered:
Posts: 647
Reply with quote  #25 

I share your frustration.  I have some background in digital signal processing, but I frequently don’t have a clue what is going on in the analyser.  It is often difficult to trace backwards from the terminology backwards to what is really going on in the analyser.  For CSI (which our plant switched to last Fall) there is a big manual with a lot of words… but I certainly haven’t read it.  It would be nice if the manual included some kind of block diagram with descriptions of how the signals are processed.

I’m not sure, but I'm leaving open the possibility that analog integration might truly be analog (unlike analog overall).  

Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 599
Reply with quote  #26 
Quote:
Originally Posted by electricpete

I’m not sure, but I'm leaving open the possibility that analog integration might truly be analog (unlike analog overall).  



Sure, unfortunately in your PDF, where they go through the order of operations of analog integration, they do not explicitly have a step for 'A/D conversion'. So it could be anywhere, but the fact they leave it out makes me think everything is indeed in the digital domain.

Previous Topic | Next Topic
Print
Reply

Quick Navigation:

Easily create a Forum Website with Website Toolbox.