Sign up Calendar Latest Topics Donate
 
 
 


Reply
  Author   Comment  
cmblocker95

Member
Registered:
Posts: 2
Reply with quote  #1 
For various reasons, we would like to move our program from CSI to SKF for our vibration analysis program. This is a integrated pulp and paper mill with a mature vibration program that has evolved curiously for over 20 years.
I am familiar with both platforms. There are some things that I know that I don't know, like what filter settings will be "equivalent" or comparable for the two platforms. What are the limits of database translations, and such.
But there are things that I don't know that I don't know.
So, for those that have taken that journey, what do you wish you had known before taking the leap?
Vibe-Rater

Avatar / Picture

Sr. Member / Moderator / Supporter
Registered:
Posts: 1,397
Reply with quote  #2 
Word of warning - this is an experience 10 years ago.  Things might not be the same now. I worked for a small company that SKF acquired and inevitably they wanted us to change from CSI to SKF.  (BTW my experience is on about 20 different paper mills spanning 25+ years) At the time of facing a change 15 years.

There were a lot of shortfalls in SKF - then Microlog and Machine analyst.
Screen too small - important now that I wear glasses.
No X function - MAJOR (they had something like HAL but that is a joke)
No way of MPM - RPM conversion ratio. Paper machine rolls could only be order based collected with tach - not practical.

And a bunch more but can't remember them in detail.

We did make a long list which I'm told went to San Diego for implementation.  I think the X function might be there now.

Be careful!!  The sales people will promise the world. It's only afterwards that you might find out you are left with western Africa.

your decision was most likely based on cost and assumption that all is the same. SKF is not cheap and maintenance isn't either.  In pulp and paper CSI rules - no exceptions. rgds.
Vibe-Rater

Avatar / Picture

Sr. Member / Moderator / Supporter
Registered:
Posts: 1,397
Reply with quote  #3 
From memory, SKF have demodulation.  Env 1,2,3,4.  where filters are preset.  In some ways this is good.  It was explained to me in person by grandfather SKF that this was done because intention was to sell to masses who presumably did not really understand in detail.  CSI on the other hand provides much greater control but there are many.. many permutations and combinations of filter settings available so to find what works you have to trial and error a lot. For dryer cylinders for example a 500Hz HP filter works very well for picking up cracked inner races.  Not sure if that translates to anything SKF.  Must do somehow.

A very good friend of mine wrote a CSI to SKF data conversion program where spectra were converted. 9 months was not enough time to do waveforms as well.  Peakvue of course did not make sense to convert.  I am not sure if that program went anywhere within SKF.  Probably not. attitude in SKF is if it does not come from San Diego or Sweden then it does not exist. Shame really.  And there is virtually nothing left of the many acquisitions SKF squandered over the years. Very sad.  A lot of talent gone. The meat grinder that is large corporation.

I hope things are better now but in my day we were subject to 3 days of training to become Oracle database experts.  This was required even before installing Machine analyst. I believe they can do other sql databases now.  In any case I specialise in vibration analysis so don't want to be an SQL db administrator.  CSI, 1 disc and in 20 minutes you are up and running. (stand alone version).  The small company I worked for was a consultancy so we had a separate CSI database for each client and would copy those DB's to a laptop as required to take into the field and back again when back in office. With oracle a db was 1Gb minimum whether any data was in there or not.  So the "consultant" approach was not practical.  Hopefully things are better now with SKF.

Be careful.  I would run along side before handing over any cash to properly test in working environment so you understand it all before swapping over.  Devil is in the detail.  Your management have no idea about what you have to deal with.  They just entertain the business meetings / dinners laid on be sales people who also have no idea.  rgds
Danny Harvey

Sr. Member / Moderator / Supporter
Registered:
Posts: 1,402
Reply with quote  #4 
I'm making that change right now.

Here are a few things that might help you out.

I'm having difficulty getting used to the idea that wrt Demod, the fmax is not relevant to the filter used and the way they calculate the number of shaft revs to capture but I'm sure they have their reasons behind their practice.

I just posted a few more comments on the change in the vibe forum so you might get something out of that.

There's lots more on SKF's website but it's not always easy to navigate.

Let me know if I can help you with this change.


 
Attached Files
pdf @ptitude Analyst Hot Keys.pdf (358.95 KB, 14 views)
pdf SKF Order Tracking.pdf (358.88 KB, 16 views)
pdf skfenveloping theory and uses.pdf (750.06 KB, 13 views)

Vibe-Rater

Avatar / Picture

Sr. Member / Moderator / Supporter
Registered:
Posts: 1,397
Reply with quote  #5 
I'm not changing Danny.   re order tracking.  The reason this is close to heart is that I spend most of my time in pulp and paper.  And for those that are not in that game order tracking is important BUT a paper machine is simply a fixed speed machine that runs at different speeds for every grade of paper it makes.  So we have (in metric units) 10 - 15 gsm (grams per square metre) toilet paper or as y'all call it toilet tissue.  Here in Australia we can call them poo tickets.... then up a notch to newspaper - going out of fashion now. Then photo copy paper (fine paper ) 80 gsm.  Then up from there here in Australia 350gsm is probably the heaviest and all sorts of colours come from coast in NSW.  If you ever buy beer in a box than the cardboard will be from a board machine.  So we have Tissue - fine paper and board. I like to think i do quite a bit of that.  So last Monday start of day 445 MPM.  Then grade change down to 415 MPM.  Speed change to make a big difference.  You need to be able to cope with that Ralph will know. rgds

Vibe-Rater

Avatar / Picture

Sr. Member / Moderator / Supporter
Registered:
Posts: 1,397
Reply with quote  #6 
Oli too as wel as no doubt many others.  
OLi

Sr. Member
Registered:
Posts: 1,915
Reply with quote  #7 

Part from converting us we did convert 5-6+ CSI sites back in the day to SKF, early to mid 1990's and it was a few papermills... They still mostly use SKF many years later so they got used to it. I do support the comments above, SKF software was actually better in the DOS version..... We converted along towards new toys since then so I only have a tacho magnet foot from a CMVA10 kit that still are used with a 2120 when balancing. So I use the best from each. I also have one remaining job once every year where I use the Commtest VB3000 with CSI neck strap... I look forward to have the data in the cloud so I can sit on my but and see that alarms come down as rain?
System we currently use is so fast and machines so small so it will not change speed during the time it takes to collect a set of data ;-)? But I agree it is a pain not to have the CSI functionallity for papermachines.

 


__________________
Good Vibrations since early 1950's, first patented vibrometer 1956 in the US.
http://www.vtab.se
cmblocker95

Member
Registered:
Posts: 2
Reply with quote  #8 
We do not intend to convert the old database to the new, but rather to set it up similar to the ols CSI database. One of the good parts about having years of data is that you have a long history to draw on. One of the bad parts is the program is set up 20 years ago when the software was not so friendly and everyone was learning the technology at the same time. And so the program develops over the years on the initial setup, evolving to make changes to improve the setup, but unable to overcome the initial setup. Thus the many databases and users over the years act as "evolution" to the vibration program: You might get a swan or a platypus.
At this point, I'm willing to eat the old bird and start fresh. 
Vibe-Rater

Avatar / Picture

Sr. Member / Moderator / Supporter
Registered:
Posts: 1,397
Reply with quote  #9 
Hi cmblocker95,

Your mentioning platypus might mean you are Australian. Maybe not, I have only seen them in zoos and 1 or 2 times in the wild. in my life time. (think 50+) They are rare and especially beautiful creatures. Better than a swan.

I worked for SKF and my then boss, still friends, wrote a conversion program. He was given a Microdog and a 2120 and told off you go. Background is that we worked for a small company which was acquired by SKF. But were CSI users. Then eventually were forced to swap to SKF. We were 80% working in pulp and paper so definitely in OLI's words "But I agree it is a pain not to have the CSI functionallity for papermachines." And is fundamentally why I refused to swap.

The conversion program should have been sent to San Diego but in true SKF / large company style if it wasn't developed by some idiot far up in the corporate tree / structure it didn't exist.

The conversion program was able to convert CSI spectra to SKF. Machine hierarchy and all. 15 years ago now.

But the key point!!!
The client will always place importance on history.  But in fact will never go back to look at that, save a few maybe 4 - 8 surveys.
So it is always better to start afresh if forced to move away from CSI.  That way you can tailor everything in accordance with new system and probably use that opportunity to increase resulution etc.  Bit of a clean slate start and using your experience gained so far.

If you have no choice - start afresh. If you still have influence don't move away from CSI. rgds




Previous Topic | Next Topic
Print
Reply

Quick Navigation:

Easily create a Forum Website with Website Toolbox.