Separate and graph frequencies using FFT?

[KY Wildcat]

Hello all,

I've been given the task of writing code for a gear tester. The gear is measured by quantifying the total error, eccentricity error, and tooth to tooth error. I'm trying to deconstruct a quadrature encoder signal into these components. The total error is easy, I can simply use the mask and limit express VI to test that. The problem lies in separating the eccentricity and tooth to tooth components while testing and displaying each individually. Is there some built-in VI I'm overlooking? I can somewhat get it to recognize the individual frequencies (one is a once per rev, and the other is roughly a sine wave with a frequency equal to the number of teeth on the gear).

Thanks,

David Joseph

[madengr]

There is a VI to measure harmonic component amplitudes however it does not look like your signals will be harmonically related so it may not work.  Since the FFT is an array, how about just using the array sort VI, that would give you the maximum bins and the corresponding indices, from which you know the frequencies.

Message Edited by madengr on 04-05-2007 04:43 PM

[Conseils]

Perhaps what you need to do is an order based analysis, there is a LabVIEW toolkit for this or you could bash your way through the math and implement your own.

If it's an order based analysis then watch out for any orders that might be introduced by the exciting system.

I recall that to cancel out eccentricity errors it's usual to measure simultaneously at multiple points, one could readily see how this might be used to determine eccentricity error.

I imagine from what you have written that its a standard form of encoder that you are using to check the gears. I wonder how you intend to calibrate the system, perhaps you are proposing to use a master, but unless the sensor is precise, you may have to compensate for errors in the sensor.

I hope I am on the right track here.......

[KY Wildcat]

I am using a linear encoder on a spring stage. The test gear is fixed to this stage and then it is driven by a master gear. The motor has an encoder also. Both are quadrature outputs that are input directly into a 6602 card.What I would like to do is plot displacement of the stage for every degree of rotation. I'll also need to plot three different graphs: total displacement, displacement caused by test gear center being off center, and the displacement caused by the teeth.

So my problem lies in not only measuring each error, but also graphing them.

Message Edited by KY Wildcat on 04-09-2007 09:36 AM

[Conseils]

The following post should point you in the right direction.

http://forums.ni.com/ni/board/message?board.id=170&message.id=212334

Here Kevin Price gives good advice on using counters.

[Szdiablo]

Hi  Conseils !!!

I am looking for a GPIB's driver of the Solartron 1250 for a labview applications...I saw in an old forum that you have already work on it.Can you tell me some informations about that???

Thank you

Ps: I know that it's the wrong place to ask that but it's the only place where I can talk to Conseils 

[Ebele_Obijiofor]

Hello,

A search on ni.com/idnet shows that we have drivers available for the 1260 and 1270. While none of these are specific to your instrument, they may be similar enough that you can use it with little or no modifications. You can also try contacting the Solartron to see if they have any drivers for the 1250. Hope this helps.

Regards,

Message Edited by _Belle on 04-17-2007 12:47 PM

[Conseils]

Well I almost remember the post you are referring to.

I think we were talking specifically serial port communication, as the IEEE 488 drivers existed. So the post you have should get you off the ground.

I would recommend them as serial is slow and these days rare on P.C's

Good luck with your project.