commencement of debug and testing errors found
N.G. 3079


- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 

correct to April 13 2009

Comments earlier written look fine. 

Might be worth pointing out this is version 1.4 (and earlier) 
of the PIC program. I presume when Geoff gets back from  
Easter holidays, he will release version 1.5 which hopefully will  
correct this issue.

It might be worth adding my earlier comments about the stepper motor.  
If you have the polarity of the wires to the stepper motor around the  
"wrong" way, you will always be a second slow. I will reproduce below  
my comments to Geoff as they explain it fairly well...

------

My observations are, that when powering up the clock the very first pulse 
seems to be lost as the hands just tremble and don't move. I am guessing 
that the stepper motor has received two +ve pulses, and thus discards the 
second one. After that it gets -ve, +ve and so on, and moves normally. 
Thus the clock is instantly one second slow, which it never recovers.

Trying to stop the clock when the hands are back at 12, and starting  
again doesn't help, as it receives an even amount of pulses to do  
that, and the problem repeats itself.

My theory is, that to correct this, the only simple way is to reverse  
the polarity of the wires to the stepper motor, as I guess the circuit  
always starts off with a pulse in a certain direction, and if the  
magnet in the motor is the wrong way around for that pulse, it will  
always discard the first pulse.

So, my suggestion to constructors is, if on powering up, you observe  
the first pulse does not move the hands, wait until the second hand is  
back at 12, reset the minute and hour hands, and then reverse the  
wires to the stepper motor (and re-power up the circuit). After that  
it should work consistently.

------

(Of course, this only happens 50% of the time, by chance you may have  
the wires around the right way to start with).

- - - -

end of file - April 13 2009