Thanks to KE7X, KR2J/5, WD8AUB, GI0NWG, AD1C, W3LPL, G3VHB, AE6E,
and N2BCC for resonding to my question regarding the occurrences of
Checksum errors on the CT serial network, specifically as to whether
going to hardware flow control would result in better results.
The respondents had varying suggestions, ranging from "it won't
matter in the least" (1), to "it must be RFI so try using fiberoptic
cable" (3), to "we use a full cable and still see occasional
checksum erros" (3), to "we use a full cable and have no problems
anymore" (2).
As for the results at N3RS, we installed a commercial, shielded fully
populated null modem cable between the computers, and ran some checks
before the WAE contest. Please note that WE DID NOT TRANSMIT DURING
THESE TESTS, SO THERE COULD BE NO RFI INFLUENCES. It took only a
very few minutes of simulated contest before we encountered our first
checksum error, so we can pretty much state with certainty that the
checksums are NOT RFI derived. It is also worthy of note that both
rigs on the network were Icoms, and as K1EA puts it, "Icoms sit there
and babble their frequency info continuously." This really puts a
load on the network. There is a command line switch (-NF, I believe)
that suppresses this babble from being sent on the network. Using
-NF should decrease traffic and hopefully increase reliability.
During the WAE, we had to be very diligent in our observation of the
"synchronism" between the two computers. If they became out of
synch, we had to do a SAVELOG on the running computer and pass that
to the other computer, with the necessary changes needing to be made
in the SETUP screens (computer number, rig type, network
configuration, etc.). This process took about 2-3 minutes on each
occurrence, but it would have been nice to not have to worry about
it. This issue is particularly significant for serial number
contests, and extremely significant for the WAE where you have a QTC
file which has to be kept up to date as well.
So, at N3RS we still are subject to checksum errors. Maybe for CT
v10 Ken will install an error free protocol, or perhaps a third party
will come up with an Ethernet (or equal) approach which will
accomplish the same thing.
73 - Dave N3RD
n3rd@ix.netcom.com
http://www.QTH.com/topten
--
Submissions: ct-user@ve7tcp.ampr.org
Administrative requests: ct-user-REQUEST@ve7tcp.ampr.org
WWW: http://www.ve7tcp.ampr.org/Software/ct
Questions: owner-ct-user@ve7tcp.ampr.org
|