Hi all.

I am building a transmitter/receiver combination both of which use a 
PIC12F675 as the primary controller. Both are battery powered (2xAA 
lithium cells, i.e. mostly operate at about 3.5V for 90% of the cell's 
life, and then falls off rapidly. Devices are programmed to shut down 
below 2.5V). So, I essentially operate at between 2.5 and 3.6V.

The transmitter sleeps for a period (one sleep instruction with WDT 
enabled to wake device, and 1:128 prescalar), wakes, reads some sensors 
(2 digital inputs), monitors the system volts (one A/D reading), 
transmits its state (about 32ms), then goes back to sleep.

The receiver uses an algorithm to basically synchronize itself to the 
transmitter's 'period' by using multiple short sleep periods to sleep 
most of the time, then wakes up, enables the receiver, receives the 
transmission, then goes back to sleep.

In building this system I have come to realise that the actual sleep 
time as described by the datasheet is very different to what I expected. 
The datasheet indicates a 'nominal' WDT wakeup happens after about 18ms. 
With 1:128 prescalar the datasheet indicates about 2.3 seconds. My 
calculations (using a stop-watch) are that 10 transmit cycles take 29 
seconds, or 2.9 seconds per transmit cycle.

Now, I expected variations in the sleep times, and the receiver's 
algorithm is able to adjust itself to pretty much any sleep period, 
within reason, but, for the moment, I have set limits on a 'valid' 
transmission being received between 1.5 and 4 seconds.

Now that it all seems to be working quite nicely, I am trying to tune 
the power consumption by reducing the 'on' time of the receiver as much 
as possible. In order to do so I want to 'fine-tune' the 'window' in 
which a valid transmission can be expected. I want to make the 'valid' 
window less than half the transmit period so that I can't possibly get 
two transmissions in one window....

Thus, the question....

what sort of variance can I expect from the WDT wake-up from sleep 
period in the PIC12F675?

Currently I am seeing 25% longer than 'spec'.
The transmitter will be mounted in an unheated building in Canada and 
must operate year-round (temp range from -40 to +40 I guess - hence the 
lithium batteries....).  The receiver will be in a 'climate controlled' 
house, thus limited temp variances between +15 and +35 I guess.

What can I expect as the range of sleep times? where can I find such 
data? Does the sleep time increase with temperature? Does it decrease 
with Vcc?

Thanks for any input.

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