Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!caen!spool.mu.edu!news.cs.indiana.edu!maytag!xenitec!zswamp!root
From: root@zswamp.uucp (Geoffrey Welsh)
Newsgroups: comp.dcom.modems
Subject: Re: Compucom 9600 baud modems
Message-ID: <145.28486D81@zswamp.uucp>
Date: 1 Jun 91 16:00:44 GMT
Article-I.D.: zswamp.145.28486D81
Organization: Izot's Swamp BBS (FidoNet), Kitchener, Ontario
Lines: 83

In a letter to All, Steve -Social Hacker (cg108dbd@icogsci1.ucsd.edu ) wrote:

--=}>> On 28 May 91 14:05:39 GMT, root@zswamp.uucp (Geoffrey Welsh) said:

GW> - Even if you never connect to another Speedmodem, they're good as a 2400 
GW> bps MNP5 modem.

 >The CompuCom implementation of MNP5 seems very jerky and filled 
 >with (IMHO) overly lengthy pauses.

   Funny, I used to say the same about the HST until I got the hang of 
fine-tuning it.  On the other hand, the HST's docs were always good (and 
they're getting even better), allowing me to do that.

GW> - They claim to use a 16550 UART on their modems, though I don't know if 
GW> it's the reliable National part or one of the flaky substitutes (WD, MX)
GW> - FAX model available

 >From what I was able to tell, they are not using a 16550.  I 
 >have ran
 >several different UART identifiers, and they all came up 
 >with "No
 >16550".  Also, the included MTEZ is supposed to list FIFO in 
 >the
 >connection information if a 16550 is detected, and it 
 >_does_not_.
 >Again, this is said without any hardware inspection or 
 >inside
 >knowledge, but on the software side, I, and all my programs, 
 >were
 >unable to find or use a 16550.  If anyone knows for sure 
 >what is on
 >that card, please post.

   As I reported recently, we were unable to identify positively the 40-pin 
DIP on the board as a UART; I may get around to spending some more time with 
it and trying a 16550.  In the mean time, it is either a 16550 (not so bad, 
really, since it can be replaced), OR it *emulates* a 16450 (yuch!)

 >One aspect I haven't seen discussed is the actual 
 >thoroughput for file transfer.  Having owned both
 >the CompuCom and now a USR HST-V.42-14.4K, I was
 >surprised at the dramatic difference.

 >For starters, text reading is MUCH smoother on the
 >HST.  The CompuCom exhibits those lengthy packet
 >pauses in 9600 mode as well, and they drive me nuts.

   I usually find that disbaling MNP5, forcing smaller packets, and using a 
computer-modem link rate as close as possible to the modem carrier speed (i.e. 
don't do talk 19,200 for a 2400MNP link) helps.

 >As for file xfer, I almost always transfer .ZIP compressed 
 >files, so any external protocol they offer (like CCSP) that
 >boasts compression will give me very little, if any.

   I was very disappointed to discover that CSP was an external program.

 >Numbers:

   Please keep in mind that the HST you benched is a 14,400 bps device with 
V.42bis; back in the days of the 9600 bps HST with MNP5, we did up to 1160 CPS 
with data compression turned OFF (turning it on *slowed* the HST considerably 
when transmitting precompressed files).

   Also, please rmember that your benchmark is only as good as the 
configuration of the modem you're talking to.  When I was trying to test my 
HST, I found that almost no boards (at least for the first while) had theirs 
properly configured.  In particular, hardware handshaking was usually not 
enabled, menaing that they couldn't lock the port speed higher than the 
reported CONNECT speed or they'd lose characters with YMODEM-G.

 >Of course, your mileage may, nay *WILL*, vary, so take it 
 >all with a grain of statistical salt.

   Done! <grin>
 

--  
Geoffrey Welsh - Operator, Izot's Swamp BBS (FidoNet 1:221/171)
root@zswamp.uucp or ..uunet!watmath!xenitec!zswamp!root
602-66 Mooregate Crescent, Kitchener, ON, N2M 5E6 Canada (519)741-9553
"He who claims to know everything can't possibly know much" -me