At mt former place of employment we were running RBMware ver. 4.31 at my new job we have the latest and greatest ver. My question is I have had a few instances where I see high vibe at gear mesh. When I auto cor the waveform then go to circular plot In the instances I have had high gear mesh The circular plots show a gear looking plot if that makes sense. Every one of them matches the tooth count of the suspect gear. Am I losing my mind or does this have any thing to do with whats happening. Just trying to educate myself as to the value of cicular plotting the waveform. As always thank you for your valued input.
Kind regards,

I have seen that before too, so I would say that is the case. I don't use CSI anymore so I have not seen it in a while.

It might mean something if there are maybe some bad teeth or broken teeth on a gear, but what I think it is doing is simplying taking the waveform and plotting it in a circular form. Yours and the one I have here looks like a gear because of the correlated waveform prior to the circular plot and the correlated peaks in the wave is primarily the gearmesh. In my case here the peaks are at 1/2 gearmesh since that is the predominant frequency in the wave. When gearmesh or a harmonic or a fractional part is present, be it high or low amplitude and is the predominant frequency in the wave, then the circular plot will "represent" the gear but will show only the predominant peak, as in this case I have, the gear actually has 98 teeth but the circle shows only 49 teeth.

If a tooth or teeth are impacting excessively in the waeform, this impact might be or seem to be obivious in a circular plot where the speed is correct. The key to the circular plot is the "speed". If the waveform is limited to 1 rev by the analyst, the basic same signal as what is seen in the circular can be seen in the wave also.
Again, the "key" to the circular is the speed assigned to the wave.

But on the other hand, one might could be mislead into thinking one thing is wrong when in reality it is simply something else in the wave giving the appearance of something in the circular. To me the circular plot of the wave should be used with the same caution as is a Peakvue spectrum, never assume anything until I prove it to myself that what I think I am seeing is in reality what I am seeing or is it not.

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

wavecircular.doc (155 Kb, 54 downloads)

I agree with Ralph--speed is key. I think the real benefit in using the circular plot is for synchronous problems: vane pass, gear mesh, rotor bar, etc. It makes a nice picture to get your point across.

If you are going to count teeth on a waveform, you need to make sure that the sampling rate is high enough to give you a good representation...

Agree with the above posts about speed. The circular plot has in my short experience (1.5 years of vibration) a great way of helping make sense of some synchronis faults. It has helped me with 2 gearmesh problems ( one had worn teeth the other had 5 broken teeth) also love joy coupling problems show up nice as well. A great way to confirm your circle plots is Peak-Vue readings if you are using CSI. The gearmesh freq's really stand out well along with the circular plots from your regular readings. It helped me to convince a seasoned old school thinking boss he'd better change a gearbox soon. That was the turning point of when they started to believe me, when things started to show signs of failure. We took the gearbox apart and it was exactly the gear I said it was.
Another interesting side effect of circular plots with the speed set correctly is that we have found you can pretty much tell how much of the race has spalling on it. We realized this after a few repeat failures on a particular machine. After cutting the bearings apart the inner races matched what the circular plots were showing for damage and how far around the it had gone.
Anyone else found good examples of how they use circular plots?

Just recently "discovered" the circular waveform plots (who needs 'training'?). As I see it, the circular waveform plot lets you see how "repetitive" or cyclical a fault really is. Unless I'm mistaken, all it does is take each revolution of the waveform, folds it into a circle, and then overlays all the rest on top of it. It is absolutely critical to have the running speed right... if it's off even a little, it will make a big difference in how the plot looks. And I've found that velocity plots seem to work best since lower order mechanical faults don't generate much acceleration.

One thing I need to do is take an actual velocity waveform, and compare it to a calculated velocity waveform (from an acceleration waveform) to see how it compares.

It'd be great if they'd make it where you could overlay 2 different waveforms on the same plot. That could really make a 'change' pop out at you.

Hi all,
Sorry about long time for reply. Equipment is an air hearter gearbox at a coal fired plant. Prob freq is 12X rpm and amp is .5 in/sec with impacting showing up in regular waveform. Input pinion has 12 teeth. Auto cor circ plot has 12 "humps" which tells me I have a once per rev "something" going on (speed is set up correctly). I have reccomended remove input pinion and inspect. Thanks for clarifying circ plot usefulness.
Kind regards,

Ron, how many "segments" does the airheater have? If 12, the vibration may be as simple as the segment dividers hitting the seals, caused by a seal set too close. Just a thought.

I think one of the advantages of a circular plot is its ability to show repeatability of the signal at every revolution assuming that non-autocorrelated TWF is used. Of course, as Rusty pointed out, having a precise RPM is paramount for this purpose.

Circular plot also helps in visualization of a problem. In this case the autocorrelated TWF is better to use. If we are dealing with a gear box and GMF is the dominating frequency, then circular plot will resemble a real gear.

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

Hello Guys,
I’ve been using circular plot a lot over the past few years. I especially use it to diagnose gear problem but I assume that it can be use for any synchronous problem. First the speed is critical but the length of the waveform MUST cover at least a few turns (8-10), Otherwise it won’t be possible to confirm the origin of the phenomenon. It wills allows you to see at least 4 correlated rotations of the faulty shaft, which should be enough to confirm the origin. I always use special time waveform acquisition. You will best see the fault in acc. with sufficient definition to get many points per teeth usually 4096 points then you adjust max resolution to get around 8 to 10 revolutions. Circular plotting will help you diagnose gear eccentricity, hunting tooth, broken teeth, transition and pitch line problem, backlash issue, wear. To do so you will need to pay attention to the sensor polarity in regard with its position.
It is very visual and has a great impact on people, which make it a good tool to get your point across. Finally something that a none-trained person can understand! Still you have to be careful, it allows you to see problem that wouldn’t show itself otherwise but it doesn’t mean that it is critical, be conservative. Finally I would recommend reading “gear analysis handbook from James I. Taylor, It really help me with waveform analysis.

Best regard, Marcel