02-09-2011 09:11 PM
@falkpl, i think your situation is same as mine... so, what did you do? i was reading about the 'Digital Filter' in the datasheet of
USB 6229, chap 7-32, wherein they mention we can implement a digital filter on the input channel itself...
Did you try this method? if not, what did you do?? please let me know
02-10-2011 05:19 AM - edited 02-10-2011 05:20 AM
02-10-2011 10:18 AM
@ceties... yes i think it has.. its a M-Series card
02-11-2011 10:28 AM
Mostly I fixed the problem at the mechanical side (that is that we reduced the vibration by balancing the axis of rotaion better) but I also handeled it in software, I knew the freq max to be 5k rpm, I measured my 1per rev using the semibuffered perioid (measured both high and low times and filtered the results in software (that is if the period od a single high low was out of range it was added to the period of the next high time (essentially a digital filter). On top of this I would use a running median filter of the calculated rpm over the previous 0.5 seconds of measurments (this filter is chosen for my ap which saw relativly slow changes in rpm values in the order of 500rpm delta per second max and usually a relativly stable signal). This is how I handled the filtering- but if the card has a built in filter it is worth a thy. Again I am assuming that sensor vibration id giving you false signals as was the source of my error when I experienced a similar effect (high freq not measuring correctly).
02-11-2011 10:28 AM
Mostly I fixed the problem at the mechanical side (that is that we reduced the vibration by balancing the axis of rotaion better) but I also handeled it in software, I knew the freq max to be 5k rpm, I measured my 1per rev using the semibuffered perioid (measured both high and low times and filtered the results in software (that is if the period od a single high low was out of range it was added to the period of the next high time (essentially a digital filter). On top of this I would use a running median filter of the calculated rpm over the previous 0.5 seconds of measurments (this filter is chosen for my ap which saw relativly slow changes in rpm values in the order of 500rpm delta per second max and usually a relativly stable signal). This is how I handled the filtering- but if the card has a built in filter it is worth a thy. Again I am assuming that sensor vibration id giving you false signals as was the source of my error when I experienced a similar effect (high freq not measuring correctly).