BG> I strongly suspect so, given the less-than-perfect
BG> lines somewhere between here and Wollstonecraft.

DB> No matter what I change, I never get more than 21.6Kbps
DB> with Dave Hatch;  I also get lower-than-expected connections
DB> with 3M. Maybe those guys are all in an area that to each
DB> other is fine, but to them from somewhere else is pot luck
DB> (with DD being lucky, you and me varying in luck)?

Its more likely that the biggest warts are seen when BOTH ends of
a particular session arent that brilliant line wise. The limitations
essentially are combined to produce the end to end result for a session.

We do know from those decent stats you deleted that Pauls line aint
terrific when called from an immaculate like like Russ's. Bills ditto.
And we know the two together produce an obscene graph. Hardly surprising
that that produces a pretty obscene outcome with a modem like the M34F
which doesnt do too well on that rather obscene session.

Thats the only way you can explain why Bill had such problems
even with the V232bis Sporter and the Spirit, when some like
Brenton and Bob never had any real glitches at all.

We also know that the M34F is a pretty poor performer on poor
lines when Steve Anderson got a hell of an improvement by changing
to a Courier, and yet his modem sold to Sydney works fine for who he
sold it to. Wouldnt mind betting the buyer has a rather better line.

BG> The higher the frequency, the worse the line becomes
BG> (see the line probe stats I've posted lately?).

DB> Yes .. the "%L" command has a coded minimum of -11dBm, but
apparently
DB> an undocumented command ("#T") will let you go higher, so maybe Paul
DB> can change it in his M34F to see if it makes any difference.  I think
DB> I tried -9dBm to Dave Hatch, and it didn't work - but it might help Paul.

Depends, the problem that Bill is getting is handshake
failure. You cant necessarily fix the rather obscene session
stuff just by increasing the transmitted signal level.

It seems pretty clear that the M34F does a lousy job in that
case, coz V34 is supposed to be able to handle the poor session
gracefully, get the best it can out of it. The M34F cant even
do the handshaking successfully in that situation tho.

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

DB> Entirely possible - wasn't Paul complaining about poor TV reception
DB> at one stage?  The 'phone lines may be questionable too,

Farkin ell. Nope is the short story. They have nothing to do with
each other. Sydney has some real problems TV wise due to it being hilly.
The phone wiring under the streets has nothing to do with hills at all.

DB> and as you say the USR may be better able to handle that.

Well, it clearly can do a lot better with poor lines,
particularly with one at each end of the session.

DB> NetComm *may* still be able to do *something* with their ROMs, but
DB> of course Rockwell Semiconductor have control of the situation here.

Thats a bit unclear. Certainly the session handshake
detail may well be considerably under Netcons control.
@EOT:

---