automatic triggering with two devices

[charliedawkes]

Hi,

I've got an acquiring program that logs data from two boards, one PCI-6052E and one PCI-6115, the 6115 can have up to 4 channels running at 1MHz or above and the 6052E can have 16 channels about 10kHz. At the moment it can do this for over 10 minutes which is great but I would like to add some automatic triggering to start the data logging, but I'm not sure where to start. It is very important that none of the starting data is lost, is it possible to even save some of the data just before the event starts? Is it possible to use some kind of moving average technique so it is comparing the noise to the data and using that to calculate when the event has started.

I don't know if this will help you to understand the situation but I am using this on taking readings (say vibrations of the work piece) on cutting machines, say drilling a piece of metal, so here I would want the readings to start the instant the drill bit touches the metal.

Thanks

Charlie

(I am using labVIEW 7.1)

[tbd]

Hi Charlie,

      Sounds like an interesting project.  It sounds like you'll be reading an analog-input buffer in a loop and doing the vibration analysis on each waveform read.  You can save some number of these waveforms as you go, so when analysis says "it's TIME", you'll have the pre-start history you need.  The process of Reading "removes" samples from the input buffer - that's important with continuous acquisitions, because there's a finite amount of PC memory - and the input buffer needs to be configured big enough to accommodate data acquired between Reads.  The higher the DAQ frequency, the bigger the input-buffer, and the longer it takes to do real-time processing for a given time-period.

It looks like you're off to a good start, why not analyze the data from each read, and set a flag when you "see" vibrations(?)  Only log after the flag is set.  You could buffer each read-waveform (by wiring it to a shift register) - then the previous read will be available when "it's TIME" to log...

Cheers!

Message Edited by tbd on 01-27-2008 04:34 PM

[charliedawkes]

Sorry I never said thanks sooner, but I had to put this project on hold, I think I've come up with a solution I just have to fiddle around with the numbers to get the optimum values, Thanks again for your help!