I am comparing shock pulse measurements with the SPM bearing checker BC100 and the Pruftechnik Vibscanner trying to decide what instrument is better for this method. Doing tests with both, I found the following differences:

1.SPM talks about frequencies detection around 32KHz and Pruftechnik around 36KHz.
2.Pruftechnik asks for RPM, shaft diameter and a dBa value that quantifies the external influence. SPM just asks me the RPM and diameter.
3.Vibscanner doesn’t recommend to collect data with hand held probe (stinger) and BC100 is a hand held instrument.

I took both instruments, configured the same parameters 1460 RPM, 55 mm, they both “calculated” 18dBi (I didn’t set any dBa) and I took readings on an easy to hear failed bearing and in a new bearing getting the following results:

Failed bearing:
SPc SPp Status
Vibscaner -4 6 OK
BC100 17 29 ALARM

New bearing:
SPc SPp Status
Vibscaner -4 4 OK
BC100 9 9 OK
If I would set the dBa in the Vibscanner I would get a right status but it means that I need always to have a reading when I know that the bearing is ok. In the other hand, some times I get false alarms with the BC100.

My questions are:
1-Why don’t I have to configure a dBa in the SPM instrument and I have to do it in Pruftechnik instrument?
2-Can I measure shock pulse with a stringer?
3-What is the frequency range of bearing natural frequencies for shock pulse detection?
4-What other instruments perform shock pulse method?

Why would you want to work in sound measurements using an A weighted filter for a microphone to monitor a bearing where HP & LP filters are employed collecting acceleraton data?

Sam,
Shock pulse method is a little bit different than the method that you probably use. It measures the energy of stress waves that result from metal-to-metal contact in a short duration in dB. Here an article describing this method.

detecting_lube_problems_shock_pulse.pdf (767 Kb, 54 downloads)

Why purchase an instrument that only lets you look for lubrication problems and detect bearing defects that may never turn into actual faults, or that flag a bearing as "bad" when it has perhaps years of service left?

I think your money will be better spent buying a good vibration analyzer (not all of them are expensive). You will be able to detect and diagnose a much larger range of problems.

I have never understood how such "stress waves" can be reliably measured with any kind of repeatability... seems your measurement would be extremely dependent on the coupling of the transducer to the surface.

Shockpulse methods all have the problem of using the transducer (accelerometer compression type apprx 60pC/g) resonance as a detector, you get hi sensitivity and lo linearity. If the resonance is 32 or 36KHz makes no direct difference. In some cases you get a bearing fault that do not excite the transducer resonance and you get no signal. In some cases you have some other signal from a old greasing system or so that generate signal in the sensitive range, then you get signal all the time. Those cases are not that common but they do happen. So I prefere a system where you sum a filtered broadband signal and that you then can analyse the FFT both regular and demodulation to be able to fix the cause of the bearing fail and to make sure it´s a faulty bearing I am looking at and not the shaft cover touching the shaft or the coupling totally worn out. You could otherwise use the yellow handle analyzer or a screwdriver, that I do anyway. Olov

I don't use shock pulse but a modified version of it. I don't need to know when micro fractures develope. I prefere these defects to develope and become visible. I use CSI's VHFD(variable High frequency Detection) parameter. The analysis parameter is set from 300Hz to 5000Hz and returns a G's rms value that can be trended. for 3600 rpm machines, alarms are set at alarm 1 @ 4 g's and alarm 2 @ 7 g's. These values will typically give you 3 month's of run time before the bearing starts to enter the 4th stages of failure. There are many variables that determine the rate of failure, but the exponential curve is present for every failure.

When we get alarms we collect extra spectrum and waveform data in G's from 300Hz to 5000Hz. We look for bearing frequencies in the spectrum and waveform. I know there are other techniques but this method allows us to do a manual peakvue or demod. If you understand the fundamentals of waveform dynamics, you become a better analyst.

There are lots of techniques, they just need to be properly applied and undestood.

Thank you very much for your replys! Just some comments: I was using the BC100 before getting the Vibscanner, that’s why I have both. I know demodulation is my best options but it has these disadvantages:
1- It is time consuming (I am monitoring more around 250 machines) to set alarms up and to analyze.
2- The collecting time is also higher.

Do you normally collect data for demodulation with a stringer? Or magnet??
Don’t tell me that I should screw the sensor!! It would also increase my collecting data time.

An accelerometer measures g's or acceleration and a microphone measures dB. Statements for Oli and Rusty were good.

Are you trying to integrate or transpose to dB?

I monitor ~1200 machines and would not have time for a method to send me in a poor direction.

CE, in 26 years of working as an engineer and vibration analyst, I have found there are some things we just can't change. Foremost, is that contrary to popular marketing "hype" you can NOT have it all. I have found the following to always be true (and it is posted in many offices of people I've worked with as a 'first principal'):

I can do it fast, cheap, and right..... so which two do you want?

CE


1-Why don’t I have to configure a dBa in the SPM instrument and I have to do it in Pruftechnik instrument?

The SPM configures it for you. That is why you tell it the rpm and shaft diameter.

If you want to know more try the spm website, in the UK it can be found at spminstrument.co.uk

I use a leonova infinity with both shock pulse and vibration loaded. I have found that shock pulse picks up a defect before vibration does, so it is possible to track a developing fault for longer to give you time to decide when you need to take action.

The A weighted filter is to match the human ear in sound measurements. Why would you want to 'roll-off' and drop-out machinery vibration. Why use dB or sound measurement?

Sam,
You are mixed up in thinking that A-weighted sound level in dB has anthing to do with the SPM dB. A trip to the SPM web site should clearify what they are talking about. The US Navy and others have used dB for vibration units for long time. Different strokes for different folks!

Walt

The guy was saying he was using dBa and calibrating, not just dB as I understood it. I'm aware of the Navy's use back ~30 years ago. But wasn't that just an overall level program? Reach a level and change it out not really trying to obtain maximum life?

9/21/07
Caution - Vendor response
Hello CE
I read your response and wanted to comment in a non commercial manner. I will just start to comment on the technologies involved.
Bearings are high frequency generators of compression waves. Whether the transducers are set up for 32 or 36 KHz doesn't matter. What is actually happening is the technology is only allowing the peizo crystal in the transducer to vibrate at its resonant frequency and not its linear frequency. The result being a very strong signal off the transducer from a component(bearing) that could have a very weak signal in a large mass of rotating parts. Suffice to say SPM invented the Shock Pulse Method in the early seventies and our transducer is constructed differently from any one else's. SPM then developed the empirical formulas from years of test stand research allowing us to interpret these compression waves with the shaft diameter and rpm(these two parameters define a dBi value) and provide an answer as to the operating condition of the bearing.
It is always better to take readings with fixed point pickup whether it is shock pulse or vibration analysis but the practical answer is probe monitoring with shock pulse can be effective if the readings are collected at the same location and it is on the bearing housing as close to the load zone as possible. We only require fixed point pickup when the rpm is below 100 rpm.
I have attached a non commercial set of slides that go into more detail on the technology involved and I will leave you with that and thank you.
Lou Morando
SPM Instrument Inc
Marlborough, CT

basicDefinition.ppt (2,108 Kb, 58 downloads)