Sign up Calendar Latest Topics Donate
 
 
 


Reply
  Author   Comment  
marksl

Sr. Member
Registered:
Posts: 153
Reply with quote  #1 
Hi Guys,

I've collected two sets of data from a Toshiba 3MW slip ring motor. The first set of data identified a bearing defect which resulted in the drive end bearings being replaced. My question is there are two unknown forcing frequencies in both sets of data at 173.09Hz (10.734x) and 187.91Hz (11.266x) which don't appear to be related to the bearing frequencies nor does there appear to be a harmonic series. I haven't analysed many slip ring motors so I'm not sure it there is a known frequency I'm missing.

Any suggestions would be greatly appreciated
marksl

Sr. Member
Registered:
Posts: 153
Reply with quote  #2 
Quote:
Originally Posted by Noknroll
What about motor speed? Driven equipment? Defect frequencies? VFD? Nearby equipment? Etc.


Run speed is 1000cpm, there is no driven equipment as the motor is in a rewinder's workshop undergoing testing.
Noknroll

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 843
Reply with quote  #3 
So both of these frequencies appear to be 0.266X sidebands of 11X
Noknroll

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 843
Reply with quote  #4 
173.09Hz (10.734x) and 187.91Hz (11.266x) calculate to 2 different running speeds approx 967 & 1000 rpm
marksl

Sr. Member
Registered:
Posts: 153
Reply with quote  #5 
Toshiba 3MW Drive end.JPG
electricpete

Sr. Member
Registered:
Posts: 647
Reply with quote  #6 

Possibly one is from a BPFO family (or 1x sidebands) and the other is from a BPFI family (or 1x sidebands).  We know that BPFO + BPFI expressed in orders is an integer… so the same applies to any 1x sidebands of these frequencies.

If your marked fault frequencies are right, maybe we might be looking at 10.734X = BPFO+6x = 4.734+6X and 11.266x = BPFI+5X = 6.266X+5X.  Both sideband series might be peaking in this area due to resonance.  

I'd put it on a log scale and identify the other frequencies as precisely as possible in orders, and then look to see what patterns may be present (are there additional smaller peaks spaced exactly 1.000x away from these... are there harmonics of bpfo and bpfi... etc).

By the way - slip ring motor could be wound rotor induction motor or synchronous motor… since you mentioned a round number speed 1000rpm I'm guessing synchronous with 50hz supply.

marksl

Sr. Member
Registered:
Posts: 153
Reply with quote  #7 
These tones were present in two tests at both the DE and NDE locations. After the first test both drive end bearings (FAG) were replaced due to a defect. The bearings are now SKF brand.

NU244C3 6244C3
electricpete

Sr. Member
Registered:
Posts: 647
Reply with quote  #8 

 

 

      BRG ID   BEARING TYPE      #B/R     FTF       BSF      BPFO      BPFI

      ------   ------- ----      ----     ---       ---      ----      ----

     83216     FAG NU244-E-M1     19     .428     3.406     8.134    10.866

     11472     SKF NU244          19     .435     3.810     8.274    10.726

 

 

      BRG ID   BEARING TYPE      #B/R     FTF       BSF      BPFO      BPFI

      ------   ------- ----      ----     ---       ---      ----      ----

      3226     FAG 6244           11     .426     3.290     4.683     6.317

     81313     FAG 6244-M         11     .426     3.293     4.683     6.317

     77737     SKF 6244 M         12     .428     3.415     5.140     6.860

     24399     SKF 6244M          12     .428     3.415     5.140     6.860

10.734x and 11.266x could match for the SKF NU244 as BPFI and BPFO+3X.   

I agree my explanation is less likely since  the bearings show similar symptoms after change, especially with different style bearings.  I would still put it on log scale to check the pattern.  That is my way of figuring out oddball peaks… more peaks help to fill out the pattern. Even if it's not a bearing pattern you may see something that helps you recognize an underlying pattern.

Ralph Stewart

Avatar / Picture

Sr. Member
Registered:
Posts: 235
Reply with quote  #9 
Have you check closely to see if the 2 frequencies you are talking about are not sidebands off of Harmonics of motor speed?

Just asking. They look awfully close to being equally spaced to me. [smile]

See the attached plot.

Thanks and Have a Great Day,
Ralph
Sidebands.JPG 



__________________
Ralph Stewart
AlertAnalytical.com
marksl

Sr. Member
Registered:
Posts: 153
Reply with quote  #10 
Quote:
Originally Posted by Ralph Stewart
Have you check closely to see if the 2 frequencies you are talking about are not sidebands off of Harmonics of motor speed?

Just asking. They look awfully close to being equally spaced to me. [smile]

See the attached plot.

Thanks and Have a Great Day,
Ralph
Sidebands.JPG 



Yes it does look like they are part of a harmonic family, however the motor was running in a workshop and there wasn't any other machinery running nearby which would explain this


Noknroll

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 843
Reply with quote  #11 
Quote:
Yes it does look like they are part of a harmonic family, however the motor was running in a workshop and there wasn't any other machinery running nearby which would explain this

Mark
Why would the absence of other nearby running equipment exclude harmonic families? or have I misread your comment?
Ralph Stewart

Avatar / Picture

Sr. Member
Registered:
Posts: 235
Reply with quote  #12 
Quote:
Yes it does look like they are part of a harmonic family, however the motor was running in a workshop and there wasn't any other machinery running nearby which would explain this


They look like sidebands off running speed harmonics.

That would make these peaks a modulator of running speed.

In the spectrum, they look as though they are not a harmonic "family" but a frequency of approx. 4.44 Hz, modulating  the 11th, 10th, 9th, and 8th harmonic of running speed.

Why could it not be coming from the motor itself rather than from some "outside forcing function"?

Did you check to see what the "exact" left and right spacing is "off" the 11th, 10th, 9th, and 8th order of running speed harmonics is?

1000/60 = 16.6667 Hz = 1x

(16.6667)(10.734) = 178.90 Hz
I am not sure where you got the
Quote:
173.09Hz (10.734x)
with running speed of 1000 cpm.
(16.6667)(11.266) = 187.77 Hz
(16.6667)(11) =183.33 Hz
183.33 - 178.90 = 4.434 Hz
187.77 -183.33 = 4.44 Hz


Noknroll has already mention the opinion that they are or appear to be sidebands and not harmonics of something. True there may be a harmonic family with a fundamental frequency of 4.44 Hz on the low end of the spectrum, but these peak around 11x IMO are not being displayed as a harmonic family in your spectrum. IMO. [smile]

Just my opinion and I could be totally wrong. [smile]

Thanks and Have a Great Day,
Ralph

__________________
Ralph Stewart
AlertAnalytical.com
electricpete

Sr. Member
Registered:
Posts: 647
Reply with quote  #13 

I'll just repeat something which may or may not be obvious, but I think it's an interesting thing, and it's a tie between my post and nok's and Ralph's post.

When you have BOTH of these patterns:

1 - BPFI pattern with 1x sidebands

2  -  BPFO pattern with 1x sidebands

…. it will very often look like non-sync sidebands around a harmonic of running speed such as shown in op. The reasons are that BPFI + BPFO  (in orders) always adds to an integer (absent slipping) and the peaks from both families often peak in the same neighborhood (based on resonances).

 A specific example for this SKF NU244. 

* 19 rolling elements. 
* Fault frequencies in orders:
*
BPFO = 8.274
*
BPFI =  10.726
*
(8.274 + 10.726) = 19.000


Let's say there is a resonance associated with sensing these fault frequencies through the housing and sensor and that resonance is in the neighborhood of 11X.
Then the highest peaks that show up could very well be:

* 10.726 (BPFI)

* 11.274 (BPFO + 3X)

* … which LOOKS LIKE 0.274 orders sidebands around 11x even though it's not really.

 

To distinguish, go to log scale, identify as many peaks as possible, find the pattern.    Fault pattern will likely show harmonics of the BPFO and BPFI along with multiple 1x sidebands.  Nonsync sidebands around running speed harmonics would typically show up around all running speed harmonics. There area a lot of possible variations, but looking at the whole picture available on log scale makes it possible to give a much better judgement imo.

Sorry for repeating myself. I'm not disagreeing with anyone but thought it's interesting.

Previous Topic | Next Topic
Print
Reply

Quick Navigation:

Easily create a Forum Website with Website Toolbox.