RB> 99% of connectivity problems are solved with the ATS27=32 or ATS27=48

PE> 32 disables LAPM?  I want LAPM.  I might be happy
PE> to disable the detection phase though, since I
PE> don't know what that is, so I'll try ATS27=48.

PE> As of 1996-02-09T23:26 I am now running with ATS27=48.
PE> Could Rod + David B report any changes.

Ah, I thought for a moment you had beaten the USR to death wiv a very
large stick and put something else on. I got an MNP4 connect this am.

Now to document the other tests we did at around 13:30 DST, for the others.

Tried S34=2, 'disable advanced, proprietary V32bis modulation'.
Wouldnt have been be a bit surprising if thats causing the problem.

This looked promising on 5 calls, mainly coz although it did show a
retrain on all calls, all did work fine. However a further set of 10
calls in this mode gave ONE failure. So its hard to say that this is
useful, may just be statistics and its no better than the defaults.

Tested S34=128 too, disabling V32terbo, mainly coz its got a very
similar constellation pattern to V32bis and its possible that that
was giving the USR brain fade when called by a Supra V32bis for some
reason. And because V32terbo isnt supported by the Sportster and it
doesnt give the retrain indication at the Supra end.

This looks very much better, no display of any retrains on the Supra
at all. One call did fail in 15, but that was with a different symptom,
not the usual continuous receive and transmit errors that the bad
default calls normally get, just no successful connect at all. So,
since this is showing no retrains, and the Sportster doesnt either, and
it doesnt support V32terbo, this seems like worth testing for a week.
@EOT:

---