Marcel,

Sounds great I am going to search see what I can find.

thanks for the info

Andre






-----Original Message-----
From: piclist-bounces@mit.edu [mailto:piclist-bounces@mit.edu]On Behalf
Of Marcel duchamp
Sent: Tuesday, January 16, 2007 4:05 PM
To: Microcontroller discussion list - Public.
Subject: Re: [PIC]: transducer dead band


Andre,
All you need is a little bit smarter deadband routine. You might find 
that a software filter is a better choice.  If the transducer has a 
small back and forth jitter, then it will look somewhat like a fixed 
reading that has occasional excursions.  The excursions will average out 
over time.

You might try a simple filter where you add, say, 16 readings and then 
divide by 16 (just 4 shifts). This will slowly converge on the right 
answer.  Adjust the filter constants to mask the jitter.  Maybe 256 
samples, maybe only 4 or 8. Experiment a little.

Check the archives for software filtering; there has been lots on this 
topic in the past.



Andre Abelian wrote:
> Hi to all engineers,
> 
> I am working on transducer that outputs 0-5v.
> center position outputs 2,5v so the range is 0 - 2,5"center" - 5v  
> X or Y direction. The output is connected to pic ADC input and 
> the problem I am having is called "dead band" witch means 2,5v
> may jump up or down so what we need to do is mask it that little
> change won't output ant thing.
> graphical way of explaining is:
> 
> 	                 x+ direction 
> 		2,5v |-|--------------| 5v
>                    ^     
>                dead band
> 
> I added the dead band routine that checks if the ADC=2,5v it will output value 0
> the problem with that is right after dead band I am getting 21 because of
> my dead band is 20. what I need is right after dead band start with 1 so what
> I did is "ADC - dead band" it took care of this problem but now I am having problem
> with high side not getting to 1023 because of "- dead band".
> another words right after dead band should start 1 and go all the way up to 1023.
> The way I see it I have to stretch the rest of data in order to get to 1023
> but I am not sure how to do it. Should I use ADC * value ? 
> I am not sure if I could explain clearly?
> any idea will appreciate.
> 
> thanks
> 
> Andre
> 


-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist