Sign up Calendar Latest Topics Donate
 
 
 


Reply
  Author   Comment   Page 2 of 2      Prev   1   2
RustyCas

Avatar / Picture

Admin
Registered:
Posts: 1,809
Reply with quote  #16 
Quote:
At least with my set up.


Curran, what setup are you using?  (FYI, we have no problem with anyone describing exactly what system they are using -- we don't see it as trying to sell something).  Do you "run routes" or are you mostly trouble-shooting, doing detailed analysis of particular machines/problems?  

I am with Ralph.  You just sort of know what you are interested in looking more closely at.  I take most all route date at 1600 lines.  If there are a small number of machines and time is not a factor (what I call "high-margin" accounts), I'll use 3200 lines.  But I am mostly a "something has changed" sort of analyst and I can usually see it in the spectrum, feel it, or hear it -- that determines what I'm looking for and what my 'extra data' entails.  And the 2140 has a plot of the last 24 values for each parameter I have defined (usually 8-12 parameters) so a quick look there tells me what exactly has changed.

__________________
"The trend is your friend"
John from PA

Sr. Member
Registered:
Posts: 963
Reply with quote  #17 
Quote:
Originally Posted by Curran919



Only gigabytes? I can fit 128 of those on a card the size of my fingernail. 

So when I was using ADRE for the first time to analyse someone else's data, I was shocked that it didn't record raw data (by default anyway) and that just by the analyst putting the wrong channel as the reference, all the data was void. I'd never have that problem with raw data.



Again, I was pointing out that the answer to the original question may vary.  The original question (in part) read "Do you use the formula for resolution and do the calculations or you simply collect samples with a few setups and then use the best based on your attempts?" and I simply "Maybe...it depends on whether you address the original question with respect to a walk around box or we go to a full blown diagnostic tool.  I have seen instruments where a RAW database can be measured in gigabytes.

You may have been shocked that the ADRE (meaning the 408) does not by default record RAW data but I suspect many more users would be shocked if RAW recording was enabled by default and the hard drive filled up rapidly.  Keep in mind the 408 can be a 32 channel device, it is not a walk around box. 

David Eason

Member
Registered:
Posts: 38
Reply with quote  #18 
Using the formula can create a host of answers, especially if the equipment is on a vfd. Your question is what do "I" do in my actual experience. Even though the formula can create lots of answers, the analyzers that I have used in my experience truly are only capable of just a few answers: 800, 1600, 3200, etc...... So it depends on what I'm looking for in the special diagnosis. If our routine resolution called for 800 lines and I want finer resolution, I don't bother to calculate 1140 lines or some such. I know I can't choose that in my analyzer, I have to choose 1600 or 3200. I may not always be looking for finer resolution, sometimes I may be looking for something that is changing rapidly. I may change from our routine 800 lines to 400 lines to get a faster response. If this change is too drastic, I can keep the 800 lines and increase my Fmax by 50% or so.
Does my answer fit your question, or am I off base?

David Eason
Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 599
Reply with quote  #19 

Quote:
Originally Posted by RustyCas

Curran, what setup are you using?  (FYI, we have no problem with anyone describing exactly what system they are using -- we don't see it as trying to sell something).  Do you "run routes" or are you mostly trouble-shooting, doing detailed analysis of particular machines/problems?


I've talked about my setup before. It is National Instruments and Labview based, programmed by my group in the past and now maintained by myself. It has some processing capabilities in real time (spectra and overalls) but is kept light to not clutter the interface. We then do the final analysis in Matlab, which is also programmed by myself. I use a few commercial products to validate my results. Suffice to say, I enter sample rate and frame length. The Fmax, df and number of lines all stem directly from this, with the latter being a byproduct for which I don't pay attention to, since it is never a limiting factor.

I don't run routes anymore. I did that for three years in the Alberta oilsands with a very similar NI setup. Now most of my collection is in our manufacturing/research testbeds. I often also accept data from other entitites of my company for analysis, from one of a dozen different systems: NI, Diadem, IMC Famos, Pr├╝ftechnik, ADRE, etc. So I must have an idea of how they all treat data storage. Really, its an even split on what is stored: raw or processed. I rarely get the opportunity to perform analysis on trended data that I have designed.

Quote:
Originally Posted by John
You may have been shocked that the ADRE (meaning the 408) does not by default record RAW data but I suspect many more users would be shocked if RAW recording was enabled by default and the hard drive filled up rapidly.  Keep in mind the 408 can be a 32 channel device, it is not a walk around box. 

That is fair. My point is mostly that there is no excuse for hardware designers to make non-walkaround boxes with a limited amount of storage as they always have. The 408i is not bad, but many are, especially when they try to 'encourage' users to use their cloud services and part with their data ownership. I feel a smarter design would be to have a global ring-buffer, so to speak, that would record raw data until the drive is full. New data would then record RAW, overwriting the RAW data from the oldest measurements.

Shurafa

Sr. Member / Supporter
Registered:
Posts: 310
Reply with quote  #20 
Excellent comments indeed.

I'm doing what many of you do, based on my sense of what is typical I attempt a few trials and take the best without real calculations. For sure, this is not what I do when I set a new machine in the data base or if I have a planned special test be conducted for machines that I cannot access after the data has been collected. Also, this is not what I do for configuring the online systems.

With respect to the data size after increasing the lines (or other parameters), these days I do not think we face a real threat of lack of memory (at least for most of the tests I did).

Regards- Ali M. Al-Shurafa
Previous Topic | Next Topic
Print
Reply

Quick Navigation:

Easily create a Forum Website with Website Toolbox.