KANTRONICS KPC-2 CLOCK MODIFICATION
-----------------------------------

I have a Kantronics KPC-2 TNC and noticed that the internal software
clock (accessed by the DAYTIME command) would speed up over a 24 hour
period.  Use of the DAYTWEAK command to slow down the clock to an
acceptable accuracy proved of no use.  Since I planned to use the TNC
continuously and did not wish to be bothered with having to correct 
the 10 to 15 second error each day or have the error accumulate over
a longer period of time, I decided to modify the TNC to solve the 
problem.  I determined that the CPU clock was running too fast and 
thus caused the timing cycles used by the internal software clock to 
be too short.  Using the DAYTWEAK command to slow down the software 
timing intervals had little effect since the CPU oscillator was so 
far off that even setting DAYTWEAK to 15 wouldn't bring the DAYTIME 
clock within acceptable adjustment range.  To correct the problem, 
one must slow down the CPU clock oscillator.  To do this, open the 
KPC-2 case by removing the two front screws and removing the front 
panel and bezel.  You will find a screw on the upper right hand 
corner of the case which holds the 5 volt regulator chip; remove 
this screw.  Pull the circuit board straight out of the front of 
the case.  Holding the board with the front toward you, you will 
see the largest IC chip, the 63B03 CPU, U8, on the board in the 
middle, located somewhat to the left of center.  Just behind the 
right side of the CPU is the 7.3728 MHz crystal for the CPU oscillator 
clock.  Near this crystal should be printed a half-moon shaped area 
for a trimmer capacitor, C33, which Kantronics decided was not 
necessary, so it was omitted.   It is at this location, C33, in 
which I soldered a capacitor to correct the clock speed.  I used a 
47 pf (picofarad) ceramic disc type capacitor with an NP0 rating.  

The NP0 designates a zero temperature coefficient which means there 
should be no change in capacity with any change in temperature.  
Other types may work fine, but accuracy versus temperature may be 
a problem.  Other users may find the 47 pf capacitor too small 
(clock still runs too fast) or too large (clock now runs too slow) 
for their particular TNC.  Use of other values may be necessary in 
these cases.  As long as it is possible to get the software clock 
within adjustment range of the DAYTWEAK command, final corrections 
may be made using the DAYTWEAK command.

I hope this is of help to KPC-2 users if they find this a problem.

I have seen this situation in two KPC-2 TNC's so far and wonder if 
it is a widespread problem.

If anyone has any questions, please don't hesitate to contact me via 
the K3MC BBS in Pittsburgh, Pa.

73, Greg Day, KC8JN - ARRL Assistant Technical Coordinator, Ohio 

Section, Steubenville, Ohio