On Sat, 28 Jan 2006 18:18:11 +0200, "cquirke (MVP Windows shell/user)"
put finger to keyboard and composed:

On Sat, 28 Jan 2006 08:23:20 +1100, Franc Zabkar
On Fri, 27 Jan 2006 04:57:36 +0200, "cquirke (MVP Windows shell/user)"
On Sun, 22 Jan 2006 19:42:26 GMT, (Olive) wrote:

typical 41k, 44k, 49k connects.

If your connect speeds are really that inconsistent, then I would
suspect a phone line issue.

What may affect our mileage here is telco exchange equipment (which
has improved - the same 41k-49k V.90 modems now routinely do 50k and
even 52k) and ISP peering, which affects hops to overseas servers
(i.e. nearly all servers). Also, the way folks wire up the phone gear
in the house can create line quality issues, e.g. line splitters that
have other phone handsets "in front of" the modem.

Don't be fooled by the initial connect speed. The client and server
modems will speedshift downwards and upwards as line conditions
change. That's just one of the things that the modem's last call
diagnostic report will tell you. You can also listen in on a dialup
session via the modem's speaker by adding L3M2 to your modem's Extra
Settings in your DUN connectoid. Short blips indicate speedshifts, eg
from 46667bps to 45333bps, and a ~10 sec period of caterwauling during
which no data are transferred indicates a retrain. To minimise your
annoyance you could plug a headset into your modem's speaker socket,
if it has one.

See these abbreviated diagnostic logs:
http://groups.google.com/group/alt.c...e=source&hl=en

One final important point. The previous discussion assumes that the OP
has an external serial (RS232) modem. It may be that she has an
internal soft or controllerless modem, in which case it will have a
virtual COM port, with an emulated UART, and will therefore be immune
to overrun errors. It will also be insensitive to port rate settings
and FIFO buffer trigger points. This is because the concept of port
rate only makes sense for a real, single-wire interface.

Depending how far back you go (and remember, this is the poster who is
quoting 386-era Windows 3.yuk articles at us g ) he may have an
internal modem with true internal UART interface with a bufferless
UART. True, that would be most likely V.32bis, i.e. 14400 bps!

The other worry is that he may have an elderly PCI system (say, a
486DX2-66 or Pentium-133) with a modern "soft" modem, and the
processing overhead of these "soft" modems can give problems that look
a bit like serial port overruns... may even present as such.

The term, "softmodem", is somewhat misunderstood.

In fact there are three types of internal modem, "soft",
controllerless, and "hard" (controller based). Softmodems have a DAA
(telephone line interface), controllerless modems have a DAA and DSP
(digital signal processor), and "hard" modems have a DAA, DSP, and
controller.

Among other things, a modem's controller handles AT command parsing,
UART emulation, data compression and error correction. These functions
do not impact significantly on the host CPU. OTOH, the functions of a
DSP are highly CPU intensive, so a softmodem (which emulates the DSP
in software) may impact noticeably on CPU performance.

Examples of softmodem chipsets are PCtel HSP, Motorola SM56,
Smartlink, and Conexant HSF. Controllerless examples include Conexant
HCF, Intel HaM, Lucent Win Modem, and USR Winmodem.

Note that the term "Winmodem" is a USR trademark and refers to their
line of controllerless modems.

- Franc Zabkar
--
Please remove one 'i' from my address when replying by email.