Brian:
Some progress has been made in that I am now
able to commnicate serially through a COM port using a simple
QB45 routine running in a DOS box under ME at 9600 and
inconsistently at 19200. 73, Brian
Really! Well take a scope and watch the incoming data at both speeds and
make sure that the start bit directly follows the preceding stop bit, no
extra time allowed. Bet you find your not quite doing what you thing your
doing. If you get errors when there is ZERO extra time then you have a
problem sitting in the bush waiting to catch you.
Some years ago I was paid to disassemble some functions of QB 4.5 and find
out what was going on in the code, unless you are running a very fast
Pentium you will find there is a little bomb waiting for you when incoming
data is going at machine rate, ie no extra wasted time between stop and
start bits.
It can look good but then when you have invested hundreds of hours of
programming you might find a nasty waiting to take you on. My standard
test here, because of the 18 millisecond interrupt in a PC, is to make the
machine faultlessly handle coms for several weeks without
failure. Remember MS makes no statements ever about the maximum return
time from the 18 ms interrupt, it can take up to several hundred
milliseconds and if your running at machine rate only a buffer of 16K or
larger can keep the game sane.
Good Luck.
PS under VB6 we now use minimum buffer size of 32K and my partner Jim has
just been working with a new beta release of MS VB and tells me he sees the
machine wants to assign up to 256K to the buffer for each com port.
PPS, are you active on 144 MHz with a big signal by chance?
Larry
VA3LK
|
