home *** CD-ROM | disk | FTP | other *** search
- Path: news.gate.net!not-for-mail
- From: dhaire@gate.net (doug haire)
- Newsgroups: comp.dcom.modems
- Subject: Re: faster than 28.8
- Date: 8 Feb 1996 10:53:48 -0500
- Organization: CyberGate, Inc.
- Message-ID: <4fd6ac$vsm@seminole.gate.net>
- References: <sumner-2001961038000001@sumner.tiac.net> <4ds0fp$4ap4@news-s01.ny.us.ibm.net> <AD29910A96685C7229@asd-stat13-153.dial.xs4all.nl> <bgrubb-2301960739100001@10.0.2.15> <4e3lbi$r3m@brachio.zrz.TU-Berlin.DE> <eric-2601960120540001@sobt.accessorl.n
- <4famg4$1nai@seminole.gate.net> <4faoqp$15ci@navajo.gate.net> <eric-0802960057500001@sobt.accessorl.net>
- NNTP-Posting-Host: seminole.gate.net
- X-Newsreader: TIN [UNIX 1.3 950824BETA PL0]
-
- Eric Shaw (eric@accessorl.net) wrote:
- : In article <4faoqp$15ci@navajo.gate.net>, dhaire@gate.net (doug haire) wrote:
- :
- : >I suggest, Eric, that you detail the environment for your tests. Systems
- : >used, load possibilities, etc. My tests were run between my 486dx/4-100
- : >and my BBS on a 486sx33, both running DOS 5.0 (the BBS also is using
- : >Desqview with only one window open, the PCBoard system) and using DSZ by
- : >Omen Technologies with a command line of 'dsz port 1 pB4096 rz -m'
- : >The connection is over dial-up voice-grade telco lines from my home to my
- : >home (a 20 mile loop to the CO and back over a SLC). the differing
- : >connect speeds for the USR were done by restricting the maximum connect
- : >speed via the &N settings. The Supra simply connected at the rate that it
- : >did (26400/28800) without alteration [sigh].
- :
- : The environments for the tests were identical for both types of modems.
- : The exact same two computers were used, and the same phone cord was strung
- : between the modems, using ATX3D to dial without dialtone (no real phone
- : line used), and ATA to answer on the other modem.
-
- What type machines? What OS? What conditions on each? Would you try the
- tests again using phone lines, not a direct connect?
- The reason that I ask the last question is because a direct cord connect
- is not always the best way to go, you can set up some rather poor
- conditions under stress that way that do not happen in the "real world".
- My tests were made over dialed up connections.
-
- I think your testing process is flawed.
-
- : Perhaps the Couriers have changed since the ones we have (some say "v.34
- : ready" instead of v.34), or they need a real phone line.
-
- My Couriers are so old, they surprised the tech when I gave him the
- serial number and also the part and rev # on the motherboard. The
- Couriers have changed but not like you think. The "brains" (firmware) is
- the same no matter the age of the Courier.
-
- : Is there any reason that the actual near zero delay between the modems
- : could be a problem, or the lack of attenuation? I don't think this could
-
- Lack of attenuation may be what is happening with your tests. When I hook
- a 6 ft phone cord between my 2 Couriers, I only get a 21600/21600
- connection. A longer cord gives a better connect. There is such a thing
- as "too hot" in a phone link. This creates an echo condition which will
- affect transmission.
-
- : be the problem though, because if it was, it shouldn't matter whether the
- : files are compressed, and the Couriers did get about the rate they should
- : on compressed files whenever they didn't exhibit the "halting" problem
- : (where it doesn't send whats in its buffer for several seconds, sometimes
- : over 30-45). I did notice that this halting problem got *MUCH WORSE* when
- : I flashed the modems to 33.6K - it used to just pause for half a second
- : once in awhile.
-
- The 'halting' problem you mention here is on those tests? If so, they
- mean the modem had to resend data because of errors. These errors are not
- reported (unless you do an ati6 after the transfer). And there's where
- your poor results seem to be coming from.
-
- Try using a dialup connection and run your tests again.
-
- --
- "Things are more like they are now than they ever were before."
- [Dwight D. Eisenhower]
-