Sign up Calendar Latest Topics Donate
 
 
 


Reply
  Author   Comment   Page 2 of 2      Prev   1   2
fburgos

Sr. Member
Registered:
Posts: 670
Reply with quote  #16 
Electricpete is about right, V-DG is the overall digital of the spectrum or waveform in the units aquired while PK/RMS (In my case I use RMS) is the RMS value of the spectrum inside the window, try changing units PK/RMS value also change.

attached

1) original spectrum in Velocity, it was aquired using analog integration.
1- Original - velocity (1).png 
2) spectum in acceleration g peak, digital integration just press 1 in your keyboard.
2- Acceleration.png 

3) total band energy trend (2X-3X ish), cursor and RMS are practicall same
3- Band energy velocity.png 
4) total band energy trend (2X-3X ish) in acceleration, cursor and PK are same.
4- Band energy Acceleration.png 

5) overall trend and spectrum, V-DG is same as overal digital, same value.

5- overall and spectrum.png 


Allen Plymon

Member
Registered:
Posts: 28
Reply with quote  #17 
electricpete,

AMS Suite recalculates the overall of the spectrum each time you change the horizontal axis.  The original overall value remains the same...  

Allen Plymon
electricpete

Sr. Member
Registered:
Posts: 647
Reply with quote  #18 
Thanks Allen, that was basically the behavior I was describing.  The displayed overall goes down when you zoom in, corresponding to the peaks remaining within the display window.  I have no problem with that much. 

The part that puzzles me is WHICH overall goes down...
  
We talked about two overalls:
  1. V-DG (which I assume is digital overall, limited to Fmax)
  2. Pk (which I assume is analog overall, not limited to Fmax).

Given my assumptions about the nature of these two overalls, I would have thought that band-limited V-DG (digital overall) would be the one that goes down when I zoom in. But in fact I observe the displayed V-DG remains the same and the displayed Pk (analog overall) goes down.  The "analog" overall terminology was based on the fact that an analog circuit was used  (at least in the older days) on the input of the analyser before any anti-aliasing or sampling.  Whatever single number came out of that surely can't be adjusted by calculation for a new maximum frequency.  On the other hand I could easily calculate a new digital overall based on a lower max displayed frequency upon zooming in (srss of the magnitudes of the frequency bins remaining in the window, times a window factor).   So it's not logical why Pk is the one that decreases rather than V-DG.  It leads me to question if I have understood the definitions of V-DG and Pk correctly (I don't think I have). 

 
Attached Files
pdf OverallWithZoom.pdf (62.91 KB, 7 views)

OLi

Sr. Member
Registered:
Posts: 1,912
Reply with quote  #19 
Or it is a "feature" eg something wrong, not unheard of maybe.
__________________
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  #20 

I'm sure the CSI terminology is logical to someone somewhere, just not me...

Attached in an excerpt from CSI knowledgeabase article NK-1000-0269. The top table suggests that since we have TWF in G’s and spectrum in ips, then we must be using Digital Integration Mode and Digital Overall mode. (I should be able to figure that out directly from the software, but sadly I'm not that savvy)

So right off the bat, I shouldn’t have been talking about analogue overall in my previous posts because we are not using analogue overall mode.

But it still leaves me cold in understanding what's going on. Regarding the relationship between Pk and V-DG in the unzoomed condition…. Pk is higher than V-DG. These are both digitally derived quantities. So I presume V-DG goes to the normal spectrum Fmax and Pk goes higher? If so how much higher…Is there another frequency specified somewhere (and if-so, where)?  If not, I guess the Pk must be based on digital samples just downstream of the input anti-alias filter /  ADC (upstream to software downsampling to lower effective sample rate 2.56*Fmax).  In any of these cases, how can we recalculate Pk from stored waveform or spectral data already downsampled to 2.56*Fmax.  Beats me.  I'm giving up on this question for awhile since I'm not accomplishing anything other than filling up the thread and confusing myself ;-)

 
Attached Files
pdf NK-1000-0269_Excerpt.pdf (369.71 KB, 11 views)

fburgos

Sr. Member
Registered:
Posts: 670
Reply with quote  #21 
I have all my points in overal mode digital, thats why my overall is same as V-DG, I have some offline acceleration spectrums with A-DG insted of V-DG, my 2140 the settings are analog overall and integration mode frequency.

from which waveform are calculated the "pk-pk waveform" and "crest"? 
Noknroll

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 843
Reply with quote  #22 
Quote:
Nok - is your spectrum fully zoomed out? 

Pete, Yes
electricpete

Sr. Member
Registered:
Posts: 647
Reply with quote  #23 
Sorry, it was a big distraction for me to worry about the zoomed behavior. Based on the definitions, it didn't seem logical to me which overall they chose to adjust during zoom (Pk)... but maybe they just preferred to show the energy of the reduced/zoomed frequency range in Pk parameter rather than V-dg in order preserve the original V-dg since V-dg is a Trended parameter (similar to what Allen said).

Where does that leave us:
  • V-dg is digital overall representing stuff below Fmax.
  • Pk (for me) or rms (for op = nok) when unzoomed represents all energy up to a very high frequency 20khz or more.  It's similar to what we used to call an analog overall, but it may come from the digital side of the instrument depending on settings.

The remaining question: why is op's V-dg higher than rms? That I don't know. (On mine the Vdg is lower than pk as expected).  Could be that Vdg is always a peak value as Oli suggested, even though that didn't quite seem to match the spectrum. Maybe someone else can clarify these last questions.
fburgos

Sr. Member
Registered:
Posts: 670
Reply with quote  #24 
I belive this V-DG might be from the waveform used for the spectrum, here is a spectrum with skyslope the rms is high, and special wave.

Capture.png 

Danny Harvey

Sr. Member / Moderator / Supporter
Registered:
Posts: 1,401
Reply with quote  #25 
I think that the twf being in velocity is the result of analog integration and the reason that you should use a stwf to get an acceleration twf.  The data collector will prompt you when it is taking a stwf, I think.  It's been a while since I used our 2130's and I never had a 2140.

That could account for some difference but I can't see it being really significant mechanically.  It may make a difference mathematically, though.
electricpete

Sr. Member
Registered:
Posts: 647
Reply with quote  #26 

I've found a few of our machines that have the same unexpected pattern as the op: V-dg > Pk.
I can't figure out exactly what the characteristics are that explain which machines have it and which don't.   
If I understood correctly, fburgos was suggesting this might result when you have ski slope. 

Slide 1 has the pattern. But only a small ski slope. 

Slide 2 does not have the pattern in spite of large ski slope. Although maybe there is high frequency content increasing Pk in this one. I need to look for one that has a ski slope without a lot of harmonics. 

 
Attached Files
pdf Vdg_VsPk.pdf (81.79 KB, 8 views)

Noknroll

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 843
Reply with quote  #27 
Thanks for all the input fella's

Previous Topic | Next Topic
Print
Reply

Quick Navigation:

Easily create a Forum Website with Website Toolbox.