Vibration Analysis?

[johnsold]

The math looks reasonable, although I just read through it and did not verify the calculations.

Good luck.

Lynn 

[preston_johnson]

Sounds like you are making some progress. 

I am interested to revisit the data acquisition and signal conditioning.  The bearings and rig are smaller in nature, raising for me the sensitivity and dynamic range of the data acquisition system.  Can you tell us more about the DAQ board in use and the sensors and signal conditioning in use?

For your referenence, here is the documentation for our Sound and Vibration Measurement Suite.  It includes functions such as integration to go from acceleration to velocity.  We also provide an envelope spectrum which is useful for bearing analysis.  These tools alone will save you weeks of time, justifying the $4000 price.  For now, you can review the manual which has tutorials on rotating machinery vibration. 

http://zone.ni.com/reference/en-XX/help/372416A-01/

[JMBacker]

I'm using a 100 mV/g Wilcoxon Research model 785A accelerometer (see attached spec sheet).  I am supplying it with a 24VDC power supply.

I am taking readings with an old DAQ card, DAQCard-AI-16XE-50.  200Ks/s 16-bit Daq board, tied to a scb-68 connector block.

The accelerometer naturally outputs a 12 volt signal at 0g, so to bring it down to a level I can read with an ordinary 0-10 volt channel, I simply run it through a pull-down resistor network.  See the schematic in the second attachment.  I realize this probably isn't the best way to do this, but it was quick, cheap, and easy.  Please let me know if there's something I need to change here.

Thanks for the link and the info.  I'll look through it.

Jeremy

[JMBacker]

The "other" board I was trying to use, but didn't get good readings from, was a USB-6251.  I haven't gone back and tried to troubleshoot this yet.

[johnsold]

You might try putting a ~6 V zener diode in place of the 1.27 kilohm resistor.  That will drop the DC voltage without attenuating the signal.  You will get about twice as much signal with the diode.

The accelerometer is sensitive along one axis only according to the data sheet. It is not completely clear, but I interpret that axis to be through the mounting bolt.

From the pictures (DSC04166.jpg) it appears that the axis is vertical in the thick plate about a diameter out from the bearing.  I would try mounting it directly below the bearing or horizontally on a bearing radial line to see if either of those positions produce better results.

You indicated that the bearing is just sitting in the fixture.  The "loose" coupling may reduce the vibration you can pick up.  The weights you were talking about may help.

Old mechanic's trick:  Hold the handle of a screwdriver to your ear and press the tip of the blade against the vibrating machinery to hear the noise.  This may a be a way of finding a good spot for the transducer.

Lynn 

[JMBacker]

Alright, I'll try the zener diode.

You are correct about the sensing axis of the accelerometer.  I also have a 3-axis accelerometer handy I might try out.

I currently have our machine shop relocating the accelerometer to the center of the bearing, modifying the motor tooling to center the motor on the bearing, and making me some weights to add in 10lbs increments up to 30lbs.

I ran some tests with just loosely placing about 20lbs on top of the motor.  This seemed to help show the bad bearing.  It also made me realize that the motor wasn't centering on the bearing, causing unrepeatable readings.  The mods mentioned above should take care of this.

I've heard of the screwdriver trick before, I'll try it out...

Thanks,

Jeremy

[preston_johnson]

Thanks for the replys

You might consider a vibration designed data acquisition board.  The USB-9234 (24 bit) is designed to work with vibration transducers and has a much better dynamic range than the 16 bit boards.  You can purchase it with the Sound and Vibration Measurement Suite for a $1000 savings.  The Sound and Vibration Measurement Suite also includes an assistant which will let you record and perform zoom FFT (adjustable span and zoom) and zoom order spectrum on the bearing data along with integration to velocity without having to program in LabVIEW.  LabVIEW VIs also include the envelope detection function.  

[varun menon]

hai everyone

I am doing a work for finding the fault on a ball bearing using labview. For sensing the acceleration am using NI USB 9162 connected to TEDS. For measuring the speed of rotation am using a tachometer connected to a software PULSE from Bruel & Kjaer. My system consists of a faulty bearing and a ideal bearing , i have to compare both and find which type of fault is there in the faulty bearing. I have to attach my readings to a waterfall diagram in labview. i need some advice as wel as some help in doing so.

am also attching my VI with this .

VARUN MENON

[Ben_S]

Varun Menon,

Hello! Is your question about putting your data into the Waterfall Graph?

If so check out this Link: Visualizing Data with the LabVIEW Math Plots

Also, there are examples of each 3D Math Plot at C:\Program Files\National Instruments\LabVIEW 2009\examples\Math Plots\3D Math Plots

[Henrik_Volkers]

I haven't looked into the spec of your sensor to see the freq. range, but since you don't need a DC component I suggest to use a 470nF (or 1µF non electrolytic) capacitor to cut off the DC and provide a 5V bias with 100k to 1M Ohm.