CSI Wizard defaults to the ENABLED Special Time Waveform (STW) collection option. It means that when a spectrum is collected, say, with Fmax=65xRPM, the STW defaults typically to Fmax=80xRPM with same LOR and gets saved. Both parameters also could be modified. The TWF, which was collected for the spectrum calculation, is not saved.

Do you guys find it useful for a survey program or it is more prudent/benefitial to keep STW DISABLED in this case?

Thanks,
Dave

Dave,

Try "special time waveform" in the find feature here and you will find many discussions about this topic.

I'm sure there are some new ideas, but if you are in a hurry, we have discussed this before.

Good Luck,

Dave,

One of the reasons the STW is switched on with the wizard is to get a TW in units of G's acceleration. The route data is collected using analog integration as it has better dynamic range. If you just stored the last TW of 6 averages, you would have a velocity TW. The STW takes one last block of data in acceleration to store so your WF will be in G's.

Hope this helped,

Bill

Bill,

Me and, as far as I know, most of the people use digital integration which makes it to collect the raw signal in acceleration units. And of course the last TWF gets saved.

The main question I have a problem with is a higher Fmax and sampling rate for STW that CSI apprently is suggesting. As a result, the spectrum is calculated based on other TWF setting which user does not see. Both higher Fmax and sampling rate is rather benefitial for analysis in my view, although take more memory space, which is not a big problem these days.

Dave

Dave,

The factory default for 2120s with new firmware and 2130s is analog integration. Analog has better dynmaic range than digital, so it is preferred. Teaching RBMware/MHM software for many years, I know how hard it was to explain this and how to overcome it by setting a STW. The wizard automated the process behind the scenes, which made it easier, but confused many users.

STW is a nice tool for those doing in depth TW analysis. The calculator makes it easy to make sure you get what you want in time/shaft revolutions. Dr. Robinson was recommending 10-20 shaft revolutions in the TW for good autocorrelation results.

A simple way that I look at the STW is I can change the "fmax" of the waveform to what I feel I need to help capture a short "burst" of energy that might not be captured in the waveform from, say, a route's 1500 Hz Fmax waveform. My "normal" setup for a 1500 Hz spectrum is a STW of 3000 Hz. Basically the same data is there that would be in a "route" waveform plus any thing that might be happening in a hgher frequency area not seen by the lower Fmax. This alerts me to collect more data if I see something happening in the STW. Of course I have to increase the sweep to a higher level to get better resoultion in the waveform. Seems like 4096 does fairly well in a 3000 Hz Fmax setup for STW. This setup "change" depends also on the speed of the point being monitoring.

I wish I could find one particular waveform plot, but I can't seem to find it, I had where I saw one peak with a huge G value and upon rechecking the point with different settings, I found the bearing was loosening up on the shaft. The spectrum showed no clue to this but the STW did.

Of course this is only my opinion and I could be totally wrong.

Ralph,

I had the same opinion as far as additional info you can get once STW is set with different collection parameters. But I am talking of not in-depth analysis but of a survey type of settings. Although I agree, it still makes sense to collect TWF with 3000HZ and higher sampling rate instead of 1500HZ.

Bill,

You are saying is that CSI recommends and defaults to analog integration. If so, then TWF is integrated in the box and stored in velocity. It also apprently means that when we want to switch TWF to acceleration units in the RBMware, the software differentiates the signal into g's.

If so, we never see the raw acceleration signal in favor of better dynamic range.

Is that correct?

Dave

This message has been edited. Last edited by: Dave_man,

quote:

Originally posted by Dave_man:

Bill,

You are saying is that CSI recommends and defaults to analog integration. If so, then TWF is integrated in the box and stored in velocity. It also apprently means that when we want to switch TWF to acceleration units in the RBMware, the software differentiates the signal into g's.

If so, we never see the raw acceleration signal in favor of better dynamic range.

Is that correct?

Dave

Yes and No... The TWs are integrated for the spectrum, except for the one it stores (STW) which is independent of integration method and just looks for units chosen in the STW setup. So in effect, if you use 4 averages for your FFT, the analyzer will take the first 4 time blocks integrate to velocity and FFT them for averaging, then take a 5th block of time with the STW settings and store it, so that will be in G's.

Confused? I am now..

quote:

Ralph,

I had the same opinion as far as additional info you can get once STW is set with different collection parameters. But I am talking of not in-depth analysis but of a survey type of settings. Although I agree, it still makes sense to collect TWF with 3000HZ and higher sampling rate instead of 1500HZ.

This is where I collect my waveform,on the routine route. The 1500 Hz spectrum is saved with the 3000 Hz 4096 (or whatever I choose) sweep waveform. This way I get a "free" in-depth at the same time I collect my route data. True the waveform is not the from which the spectrum was derived but it is taken a few millseconds after and contains the same data in addition to the extra data, of course if there is something drastic going on and the waveform is way out in left field from the one from which the spectrum was taken, then the machine is in big time trouble anyway.

I find "trending" the pk-pk value a must!!


I normally will acquire a velocity waveform if I am thinking the issue is alignment,imbalance, etc.

For medium to high speed rotating assembly bearing monitoring, I gotta have the ACC stw.