🚀 go-pugleaf

 In article <6880dbrewood@nest.demon.co.uk>
  dbrewood@nest.demon.co.uk (Daron Brewood) said:-

>I'm finding I'm getting lost data at that rate 115200, so have droped down
>to using 134 (for 57600) which seems to work fine, but I still get NOS
>saying 134 is an invalid baud rate. The CPX is of course set to match.

    Yes! I always used 115200, and I thought had no problem, but I just 2
days ago switched to 57600 to test it... and I have a much improved CPS
rate while polling news! AAAArg! So we have transmission errors on Falcons
at 115200...

    Exactly at 115K I have articles comming (in batch mode with 16
buffers) in say 0 to 2 seconds each... then a big pause, an article
comming in say 10 or 15 seconds... for this article the CPS could be 150
CPS. Then a burst of articles at 2000/3000 CPS, a new pause, etc... at
session end the average CPS was say... 1000 CPS.

    At 57K all articles come in burst... At average 2000/3000 CPS only
depending on the servers overloading! To say the truth, my access node for
compuserve in Lille has changed exactly 2 days ago, so I'm not sure this
has no effect...

    Someone has an idea, how to test the serial port, to see if there are
errors at 115K? I have only one computer... perhaps testing local echo of
my modem? Send him big strings and compare with the echo? but perhaps the
modem adds a delay, or... dont know?

    I never had errors while zmodeming with local BBSes, but with all
recent modems having V42 error correcting, perhaps error are masqued?

    And in NOS, where can we find how much errors occured? I have seen a
retransmission counter (in "tcp status" I think) but it's not zero even at
57600 bauds...
__

	/* Bruno Kozlowski 100710.1360@compuserve.com */
	/* -==- A happy Falcon owner from france -==- */