Russell, at 22:14 on Feb 25 1996, you wrote to Bill Grimsley...

BG> * 25 Feb 08:39:07.03 Processing node 3:711/934 -- Ten Minute Limit
BG> : 25 Feb 08:39:38.31 Sysop: Paul Edwards
BG> # 25 Feb 08:40:45.16 Carrier Freq (Hz)        1959/1959
BG> # 25 Feb 08:40:45.16 Symbol Rate              3429/3429
BG> # 25 Feb 08:40:45.28 Preemphasis (-dB)        8/6
BG> # 25 Feb 08:40:45.28 Recv/Xmit Level (-dB)    35/13
BG> # 25 Feb 08:40:45.31 Roundtrip Delay (msec)   18

BG> * 25 Feb 08:40:59.97 Processing node 3:640/305{at}fidonet -- Jabberwocky
BG> : 25 Feb 08:41:23.81 Sysop: David Drummond
BG> # 25 Feb 08:42:08.78 Carrier Freq (Hz)        1829/1829
BG> # 25 Feb 08:42:08.81 Symbol Rate              3200/3200
BG> # 25 Feb 08:42:08.91 Preemphasis (-dB)        4/4
BG> # 25 Feb 08:42:08.94 Recv/Xmit Level (-dB)    23/12
BG> # 25 Feb 08:42:08.94 Roundtrip Delay (msec)   3

RB> I don't understand why you dont get a better Carrier Freq and sysmbol rate
RB> to David, as we proved that you can download from him at 4000Cps on a 
RB> zipped file using your line & my modem.

That's actually one of the advantages of using modems which support 64
State trellis coding, as it allows a higher bit rate for a given carrier
frequency.  Rockwell chose not to implement this feature though.

RB> The above must prove that your line & exchange are capable of 1959Hz and
RB> 3429 symbol rate yet these are lower to Dave (with a better result in
RB> connect speed and thoughput.

Exactly, and this also goes a long way towards explaining why those modems
with only 16 State coding will connect at lower link rates, such as 26400,
where a USR would show 28800 for the same call.

Regards, Bill

--- Msgedsq/2 3.20