turbo balancing

[hommer]

hello i am looking to set a new system for balancing my turbos at the minute i am using 2 IMI industrail accelerometers and a optical tacho sensor ,i have them going to a sound card in my computer,i was woundering if their may be a more stable ,or sensative way. i heard somthing about a DAQ interface can anyone help?

[DAD]

To give you guidance we need some more information.  How many samples per revolution do you need and what speed do you need to spin the turbine at to balance it?

Let's say you want one sample per degree and you're spinning it at 10,000 rpm.  That's 166 revolutions per second x 360 degrees = 60,000 samples per second.

The sound card probably has a maximum sample rate of 44kHz so that won't work.  But this is just an example, it depends on your answers to the above questions.

[hommer]

thankyou for your reply,my setup will spin the turbine at a max of 10,000 rpm. I normaly start at about 2000 and work up two 10,000 to check final balance but hope in fucture to build rig to go to a higher rpm thanks

[DAD]

How many samples per revolution?

[hommer]

hello i am new to this but i am looking for somthing that will work in the range of 2000 to 10,000 rpm not sure if this is possable if not what what do you recomend

[DAD]

There are a couple of different ways that you could capture the data you're after.

If you can spin the turbine at a constant speed, then you can just set a sample rate in the DAQ hardware that gives you the number of samples per revolution that you need (or something slightly greater).  You still haven't specified if how many samples/degree (or degrees/sample) you want.

You'll need some kind of reference so you can relate the accelerometer data back to an absolute position.  This can be a tachometer pulse of some kind.  You could use this to trigger the data acquisition or just acquire it as another channel.

If the spin rate is not constant, then you might be better off using an encoder.

The other thing to consider is how the two axis are acquired, most NI hardware is multiplexing the inputs to a single A/D so the two axis wouldn't be sampled simultaneously.  This could make the analysis a bit more difficult as you might have to take into consideration the delay between an x and y sample (again, depends on the samples/degree you want).  The alternative is going with simultaneous sampling hardware, but that will have a higher price tag because there's a separate A/D on each channel.

[hommer]

i normaly use one accelerometer in single plane balance ,and i can run the turbine at a steady speed. i mark referance point on the turbine wheel for the optical sensor i don,t know how many samples in need as i am new to this type,can you recomend a sample rate?thanks  

[hommer]

hello i am looking to update my turbo balancing rig it currently runs on a sound card base system, can anyone recomend a DAQ and software for this type of applacation thanks

[markwhiteley]

Hi Hommer,

You specified you were using some IMI sensors and an optical tacho, could you possibly link us to a product page for these devices or tell us specifically which accelerometers and tacho's they are (i.e. their model name)

As DAD mentioned we need to understand exactly how many times you would want to sample per each revolution of your turbo, which means would you want it to read twice every time revolution to take a reading at every 180 degrees, or would you want it to take 360 samples so that it took one sample per degree of the revolution?

A good way to understand this further would be to understand exactly what you are looking to find from your signal or what you are looking to read back, could you let us know what you are doing here?

A basic idea of modular system that you could use to take these readings would be the following set up:

http://ohm.ni.com/advisors/compactdaq/pages/common/intro.xhtml?configid=CD3380030

In this set up you would have your 9234 module to take in your aceelerometer signal (as this has all of its signal conditioning built into the module), and the 9421 to take in your digital signal. As I mentioned before, if we knew better exactly what sensors you were using we could tailor this system a little better to your requirements.

[preston_johnson]

You will need to power the accelerometers with ICP power or NI's name of IEPE. Do you have the power for the access and the laser tach?

NI offers daq, th NI 9232 module for our USB sleeve that provides power and samples at 102400 samples per second. Further our sound and vibration measurement suite for LabVIEW has a ready to run example for single plane balancing including trial weight process with influence coefficients, polar graphs, etc.

If you want a non LabVIEW application consider eZ-Balance and the 650u USB device. Www.ni.com/iotech