03-18-2010 12:32 PM
Darn it! I made a mistake the analysis is not made on the corrected valleys. Can you take a look at the picture and correct the mistake
03-18-2010 12:43 PM
yes, I saw
you were not feeding the multiply node with the unbundled'locations'
got it
03-19-2010 08:27 AM
On my way home, the train get stuck, so I killed some time by using the zero crossing method on the data. As you can see the gain is not very big. But it is somewhat less jitter. I guess jitter are caused by AD resolution. With better AD resolution you will probably get more accurate readings hence less jitter. Better resolution can be achieved by using more of the full AD converter span ( if possible ). Or increasing the AD resolution by changing to an AD converter with more bits. But this is cost/benefit problem. If the costumer is happy just leave everything as it is 😉
03-19-2010 08:41 AM
Spoke to the client yesterday and explained the issue. He was real apologetic, Says he always runs the data through a Butterworth low pass filter before working with it but he sent me raw data.
Ran it through the filter before doing the peak analysis - it looks great
thanx for all ur help
when you are in NY I'll buy you those beers
03-19-2010 12:03 PM
03-19-2010 01:46 PM
with a lowpass filter in the front, you wouldn't even need to decimate and average to reduce the noise.
03-22-2010 08:59 AM