Re: Slow varying inputs

Give it to me. ;-) I'll add it to comedirecord. Once we've got 
everything working as we want we can start thinking of comedi-routines.

My idea for a comedi command is that it is sampling in the background 
and performing averaging. Either with a syncrounus command or with a 
soft timer for simpler cards. The user then can get averaged data 
whenever he/she wants it and then issues a reset so that averaging 
starts over again.

/Bernd

www:    http://www.berndporr.me.uk/
         http://www.linux-usb-daq.co.uk/
Mobile: +44 (0)7840 340069
Work:   +44 (0)141 330 5237
         University of Glasgow
         Department of Electronics & Electrical Engineering
         Room 519, Rankine Building, Oakfield Avenue,
         Glasgow, G12 8LT


Sandy Pond wrote:
> On Thu, 2005-03-03 at 20:38 -0500, Daniel Nilsson wrote:
> 
>>For my application I assume a low-pass filter should work just
>>fine. For example averaging the last 100 samples thus creating a
>>windowing average function might be enough ? Or is there more to it
>>then that ?
> 
> 
> If the averaging give you what you want it's fine.  It appears all
> you're really interested is the low frequency transient.  But if your
> noise is primarily sinusoidal the moving average might need to be very
> long.  The code for a digital filter is not that complex.  I even have
> one that works on integers (all integer math) therefore, very efficient
> for operating on A/D counts before scaling.  As I recall it is only
> about 10 lines of C.  But alas, I'm at home and the code is at work.
> 
> Regards, Sandy
> 
> _______________________________________________
> comedi mailing list
> comedi_at_comedi.org
> https://cvs.comedi.org/cgi-bin/mailman/listinfo/comedi

Received on 2005-03-04Z09:37:12