david, at 21:22 on Feb 26 1996, you wrote to Bill Grimsley...

db> On Feb 25, 1996 at 10:01, Bill Grimsley of 3:640/305.9 wrote:

BG> # 25 Feb 09:58:48.50 Modulation               V.34
BG> # 25 Feb 09:58:48.53 Carrier Freq (Hz)        1959/1959
BG> # 25 Feb 09:58:48.56 Symbol Rate              3429/3429
BG> # 25 Feb 09:58:48.56 Trellis Code             64S-4D/16S-4D

db> To TML...

BG> # 25 Feb 09:59:46.50 Modulation               V.34
BG> # 25 Feb 09:59:46.53 Carrier Freq (Hz)        1829/1829
BG> # 25 Feb 09:59:46.53 Symbol Rate              3200/3200
BG> # 25 Feb 09:59:46.56 Trellis Code             64S-4D/64S-4D

db> ...and to DD.  Apart from the fact that the USR supports 64-state trellis 
db> coding and the Rockwell-based NetComm only supports 16-state, this is about 
db> the second or third time that I've noticed a higher carrier frequency and 
db> symbol rate on your connects to Paul's TML compared to connecting to DD's 
db> USR.  Could this have something to do with it?

I strongly suspect so, given the less-than-perfect lines somewhere between
here and Wollstonecraft.  The higher the frequency, the worse the line
becomes (see the line probe stats I've posted lately?).

db> Let's stop being silly quoting modem dumps and let's start *analysing* 
db> them.  What *exactly* do these lines tell us, hmmm? 

The USR manual is a little obscure, and only lists the various carrier
freqs for a given protocol, although I assume that the better the lines,
the higher the freq and symbol rate, with the proviso that with 64-State
trellis coding, the highest freq/symbol rate for a given link rate will be
lower than the same connect to a modem which only supports 16-state trellis
coding.

db> I tried to download a copy of the V.34 spec today from the ITU's Web site - 
db> but it's a subscription-only service.  :-(

Arthur Marsh (Aust_Modems) will likely know where to obtain a free copy.

BG> I'm glad you suggested trying that, Paul.  Your lines are FUCKED.

db> :-)

I was half-serious, as the problem does appear to be peculiar to his end of
the link (other STD V.34 calls work perfectly from here, even to NetComms).
:)

db> BTW Paul, yes I did just get a 26.4Kbps connection with you, using the 
db> oft-recommended %L11 option.  Dare I say it - the Spirit Viper is stuffed?  
db> ;-)

It's starting to look like Rockwells just can't handle poor lines as well
as the USR (not surprising, if you read what some NSW country sysops have
to say about their maximum link rates improving dramatically just as soon
as their Rockwells were replaced with USRs).

Of course, Rod would simply call that zealotry...  :)

Regards, Bill

--- Msgedsq/2 3.20