I was looking at some data today and ran across this spectrum. I am going to post two spectrums and waveforms that are off of the same motor but at different times. I have never seen a spectrum look like this and was hoping someone could explain it to me.

Thanks,
Ronnie

Weird_Spectrum.doc (40 Kb, 101 downloads)

It looks like there was some type of spectral display problem or maybe a conversion problem in the instrument.
The waveforms are the raw data and they look too much alike for there to be something actually wrong with the machine.

Ronnie, where do you work in Mobile?

Vibeguy,
Primarily I work in KC Paper mill but I also do routes for other customers that I have worked for the last few years. I am out of the mill about 4 or 5 days a month. Our office is inside KC.

Ronnie

Ronnie,

Looks to me like the fixed vertical scale was enabled to a negative 2 times 10 to 18 power.

The one below is one I forced to look like yours.

Only my opinion and I may be wrong.

Ralph,
I agree, it does look like that but did you notice I have a frequency marked at 1993 Hz that was up side down. Nothing was done on our part to make this happen. It was a random event and just thought it was weird. Thanks for the feed back.

Ronnie

Yeah! Sometimes strange things happen with software. Seems like the little guy directing trafic inside the computer gets so confused sometimes he just throws up both hands and says "go on do do what ever you want".

Ronnie,

Looks like a good job for Ken Piety (MasterTrend father) to figure out!

Walt

quote:

Looks like a good job for Ken Piety (MasterTrend father) to figure out!

Walt,

Is Ken still working? I was thinking he sold his part of CSI and retired, back several years ago. Been a while since I talked to him. He is a great guy, what with the charity work and the children he works with. He really helped me a lot with shortcut keys and stuff back when RBM first came out and I was in a transition from DOS MT.

Ken Piety just came to work for Azima Inc, the company I work for. I got an email earlier this week that he came on board this week. I guess we are pretty lucky, huh?

Ronnie

quote:

Ken Piety just came to work for Azima Inc, the company I work for. I got an email earlier this week that he came on board this week. I guess we are pretty lucky, huh?

Yeah! You are.

I think CSI paid Jim Corley royalities for Master Trend as it was his - am I wrong?

quote:

I think CSI paid Jim Corley royalities for Master Trend as it was his - am I wrong?

They may have, Sam. I really do not know that part of the history.

I do know Ken was the main most man in charge of and doing most all the programming back in the 90's in "plotdata".

I remember one time when I was there in Knoxville, I was complaining about losing all my "F keys" from DOS MT to windows MT and RBM plotdata because I had been using a keyboard with the function keys on the left side of the keyboard. He put some hot keys in for me while I was there that week. One I particular remember was the "A" key for set mark, which, I think, was "F3" in DOS MT.

I remember the "F1" key (top left key on my keyboard) was "accept" in DOS MT plotdata and the reason given me for doing away with it was MicroSoft used it for "help" function, I think, but I see now "F1" in RBM is for "ANNOTATE PLOT", hmmm! times change, huh?

I hated the mouse back in those early days of windows. Still don't like it much. I am having to use it left handed now. Repeative use with my right has caused some bad pain in my shoulder.

One other program change he did, which I remember quite well, was adding a message to "Autoplot" telling that the program was searching for a plot within the time frame requested. The program gave the impression that the computer was locked up and I had from time to time, given up and shutdown in the middle of a search, not knowing it was still searching.

Bottom line of what I am saying is, Ken is an extremely smart man, a good man and helps people out and never really gets upset about suggestions, no matter how basic they may be.

Sorry for getting into such a non-related reply to the topic of the thread, but when Ken's name was mentioned, it brought back some old memories.

Ralph,
FYI, seeing how I started this thread I wanted you to know all of your input is appreciated. Sometimes a question can be asked and many good points branch off from it. I am sure that within reason we should not change a topic but I was glad to hear what you had to say about Ken especially now that we work for the same company.
And by the way, I sent Ken the spectrum and waveform and Ken said,
"You clearly have bad spectral data in the first set of data that you sent to
me. The waveform looks reasonable. When I say bad, I mean that there was a
memory glitch in the analyzer, or a communication error during transmission,
or a corrupted data record in the data base, etc.

You cannot have negative values in spectral data. You could have zero values
or more likely very small positive values close to zero, but never negative
values. The spectral calculation is a RMS-type calculation. So no matter
what type real-value waveform you hand to the algorithm, the result has to
be greater than or equal to zero. You can ignore this data for analysis;
you may want to report it to CSI as a possible software bug"
Just thought I would pass that along if anyone was interested.

Thanks again everyone for all input.

Ronnie