Mike Harrison wrote:

>On Fri, 01 Jul 2005 22:46:23 +1200, you wrote:
>
>  
>
>>>Done. There are eight button cells, about 3mm tall and 9.5 in
>>>      
>>>
>>Way to go. Here are some tips from Microchip. The nW parts
>>are what you really want to be using for longest battery life (I've
>>tested the 12F675 at 7nA SLEEP current)
>>
>>================================
>>
>>http://www.microchip.com/stellent/stellent/idcplg?IdcService=SS_GET_PAGE&nod
>>eId=1484
>>
>>Q. How can I make my PIC16/17 run as low power as possible?
>>
>>A. Section Tips Posted: 10/25/1997 10:35:49 AM Discussion:
>>
>>1. Run the clock as slow as you can.
>>    
>>
>If running continuously, yes, but if sitting in sleep most of the time and waking up occasionally,
>faster clocks can sometimes allow less on-time and therefore less current. You need to do tests as
>every application is different and there is no simple answer to which is the optimum rate.
>
>For example, I did some tests on a 16F636, INTRC waking every 30ms and executing a short loop 16
>times : 
>
>31K : 6.4 uA (awake 12ms)
>125K : 17 uA
>250K : 10.1 uA (awake 1.5ms)
>500K : 6.5 uA
>1M : 4.6 uA
>2M : 3.6 uA
>4M : 3.1 uA
>8M : 2.9 uA (awake 46uS)
>
>The above was at 3.6V and INCLUDES the WDT current!
>
>The other neat thing on the nanopower parts is you can change the clock speed as required,
>on-the-fly. e.g. if you need a few tens or hundreds of of microseconds' delay, just kick the clock
>down to 125K or 31K for a few cycles.   
>
>  
>
>>2. Disable the watchdog timer
>>    
>>
>
>Nanopower parts have MUCH better WDT power draw, and a lot of flexibility on WDT interval, both
>faster and slower than the older parts - from memory it's something like 1ms to a minute or so. 
>
>  
>
>>3. Put the part to sleep whenever possible.
>>    
>>
>
>But if using xtals or resonators remember that there is a time (and therefore power) associated with
>wakeup
>
>  
>
>>4. Use MCLR to wake part from sleep instead of the WDT if possible.
>>    
>>
>Pin-change interrupts are also good. 
>
>  
>
>>5. Do not let any inputs float.
>>    
>>
>
>A 'quick test' for this is to run your finger (or a finger-held probe) along the pins while watching
>the current consumption. Mains hum going into a floating pin will produce at least 10s of microamps.
>A more scientific test is to apply a 10-ish khz sinewave swinging within the supply rails via a 1M
>resistor to each pin in turn. A floating input will give a noticeable (>10uA) increase in current
>draw.
>
>One they missed on that list - run at the lowest voltage you can. 
>
Thanks. This goes into my bin of list emails never to delete. All the 
low power tips (and nW modes and techniques) seem daunting, but 
individually they seem achievable. Wish me luck!  :-)

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