Comments
Im looking at machine tool spindles that run 2,000 to 4,000 rpm on angular contact ball bearings. I have IRD Odyssey software, and a DataPAC 1500. In the collection type area I have options from 30K to 300K cpm. In the filter type area I have the usual options 100Hz to 2K Hz. What is the Proper set-up, and or is there a general rule that may apply. Thanks Brad

Yea me too, and futher some of the filters are gone after installing ver. 3.0 so now what do I do. I Need to better understand g's Spike Energy. . . . HELP

Jan and Brad,

In setting up the spike energy, you need to set the filter depending on the speed of the machine.
Basically, anything over 1700 you use the 5Khz filter. Over 1000 but less than 1700 use th 2Khz. Over 800 but less than 1000, use the 1 Khz. Over 500 but less than 800, use the 500 Hz. Over 300 but less than 500, use the 200 Hz. Less than 300, use the 100 Hz.

For collection specs, on the magnitude, pick the Standard G's High Freq.
You'll need to create some collection specs for the spectrums. take some of the specs you are using for the velocity readings and copy them. Call them something else that will denote spike energy somehow (I use things like gSe/60K/1600). The thing you need to change within the spike energy specs that are not in your velocity specs, is the signal detection. Spike energy is a Pk-Pk reading.

If you use lets say a 60K or 90K F-max for 1800 rpm general rotating equipment, the demodulation and algorithims of spike energy spectrum will show you the root causes of what is causing the impacting type ultrasonic vibration. therefore, if you see a peak at the 1X rotating speed, then you know that something is causing a very high frequency "impact" or "Rub" once every revolutionl. One thing I have found with spike energy is if you have your resolution fine enough (Never below 800 lines unless the F-max is 6000 or something like that), bearing fault frequencies will hit exactly on the top of the peaks. It's a pretty neat technology.
Of course, you must also understand that are many other things that can cause high frequency ultrosonic vibration (ie: steam leak, hyd. fluid under pressure flowing, pump cavitation, etc.).

I hope I understood your question, and have given you some place to start.
If you have upgraded, and some of your filters disappeared, be sure you have the right data collector as the active collector (tools, set active collector). Odyssey is set up for only those items that the active collector will support is shown in any of the drop downs.

Dave

Dave,
Sorry about the slow reply. I have just set the data collector. Also I have spoken with my Entek rep John G. Between his patient tutoring and your explanation I now have a grasp of this application. Now to implement!

I will be attempting to setup measurements on 24,000rpm spindles. I want to hold running speed levels to .01ips and 1.5g peak but the manufacturer says .118ips! These spindles have Ceramic Hybrid angular contact bearings. I will need to get out the manuals and figure out some envelope filters and set up band alarms.
Do you feel my expectations are too high at .01ips and 1.5g peak ?
Would 1.5g peak be equivalent to 3g peak to peak?

This message has been edited. Last edited by: <Jan Van Z>,

Jan,
First, if you're talking about John G. from Chicago, then you're in good hands. Tell him I said Hi the next time you talk to him.

0.01 IPS is not unreasonable for something turning 24K.
And, yes, 1.5 G,pk is 3 g's Pk-Pk.
What transducer are you using? if the g's you are talking about are spike energy g's, then it will depend on the transducer you are using, mounting method, etc.
If you are talking about regular g's Pk, then mounting method will be about your only repeatability concern.

Dave

Dave,

I am using six IMI 607M13 accelerometers attached to a six channel collection box. These are the flat magnet base type. The spindle is in a machining cell and the array of accelerometers are placed on the spindle while I collect data from outside. This allows data to be collected with just one setup. The accelerometers are numbered and the same sensor is used at each location for each collection interval. The spindle is encased in a cast steel cradle on the end effector of a Tricept robot. Our toolmakers have machined flats on the case above the bearing locations. The signal has to travel through at least 1 inch of steel with one interfacing surface between the spindle housing and the cradle. Future plans are to machine openings in the cradle to allow direct access to the spindle. I am attempting to gather ips, g's peak and gse peak to peak.
I do agree that John G.has been a great resource!

Jan, Brad,and Dave,
I have a couple of questions. First, are you checking the spindles during production and in non-production with full rpm with no load? Secondly, how are you checking multi-head spindles with up to 16 spindles per head? These are the things i am confronted with. Any replies would be appreciated.

Michael,
All of my applications are single spindle tools i.e. CNC's. I measure at full speed no load. Installed is a balanced empty shrink type tool holder. Brad can help you with multiple spindles since he monitors many of these!

Jan Van

Michael, are you talking about gear drive gearboxes ? if so I am currently checking and verifying my own set-up's but maybe I can help. On 5500 rpm gear driven spindles Im a using 120K fmax and a 5kgSE filter to gather gSE Pk-Pk, I may have to go up to 300k yet but for the next two or three months Im just going to watch. I hope this helps and Thanks, Brad email brad.noel@gm.com
in Sunny Spring Hill Tennessee

Yes they are gear driven. My biggest problem is time. At the present we only have about 15-20 hours per week to do collection (idle).I am trying to see what other individuals are doing. One of the spindles has 4 heads with up to 32 spindles each. With this time restraint i do not have time to do a formal anaylisis.

Michael, on some of ours we have a fixed mount transducer between two spindles, that still is just an early warning and thats about all, if you want to pick out the particualar spindle you'll need a mag mounted set-up.