[ECOS] ADC polling vs. ADC generated interrupt

[Peter Huszar]

Hi Uwe,

First of all I would like to stress, that I am just starting to find out the capabilities 
of  the LPC2000, and I am asking questions in the hope of getting a better 
understanding.

There are two reasons why I posted the question:
1) If I understand correctly, the burst mode is really a continuous mode, 
    where conversion is performed continuously. I have a preconception
    of using ADC taking a couple of samples in about once a second for 
    checking some temperature data. From this I would assume that some 
    tens of thousands of samples will be collected continuously before a read
    is made that actually requires the data. Setting the timer to 1/10 sec, turning
    the burst mode off, using the hardware start conversion signal, would result
    in max 80 reads per second. This would imply to me less time spent 
    servicing interrupts, less power, less noise etc.. Even if the ADC would be run  
       on full speed, the above method should not slow it down considerably just 
    because of the different use of the interrupt. Of course, at full speed, serving 
    8 interrupt might cost more than 1 interrupt plus polling 8 registers.

2) If this mode is acceptable, maybe a uniform approach could be considered
    for the lpc21xx/22xx/24xx driver. 

Please correct me if I am wrong in any of my assumptions. I would also welcome 
your comments since you have practical experience with the ADC.

Please let me know if this idea is not worth pursuing. I would not like to waste
precious time on some futile attempt as a result of my lack of understanding
of the intricate hardware details.

Regards,


Peter




Original Message ----
From: Uwe Kindler <uwe_kindler@web.de>
To: Peter Last NameHuszar <duci750@yahoo.com>
Cc: ecos-discuss@sourceware.org
Sent: Saturday, December 27, 2008 5:10:00 PM
Subject: Re: [ECOS] ADC polling vs. ADC generated interrupt

Hi Peter,

why would your idea of using timer1 without interrupts and let the ADC generate interrupts after each conversion be more efficient? Please give me some more hints why this method would be more efficient.

Thank you, Uwe



      


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss