I've not seen this twist yet. I have seen the problems first hand tho, and
have tried nearly every trick that I could think of to resolve it. Many
times I do see a problem where the dial sequence is completed then the modem
comes back with a dial tone. This is quite common here even among systems
that do work reliably.
My thought is that either the PM-3 or some particular modems don't handle
the 28800 negotiation exactly to standards. On some of the problem systems,
I can establish a 14400 negotiation to the PM3 just fine and not a 28800,
but can get 28800 negotiation to analog modems. Admittedly, a good deal of
these modems that are a problem are a bit proprietary, either combo
sound/modem cards or built by a manufacturer expressly for their machines.
>
>> I personally think it's a 95 problem, as occasionally the problem can be
>> cured by either creating a 2nd dial-up connection and connecting with that,
>> or by deleting the modem driver and reinstalling it.
>
>How often does this resolve the conflict? Has it reappeared on
>instances that have worked?
I can't give much history on this cure as this problem is extremely new.
This usually fixes the problem I'd estimate about 65% of the time. So far,
every time this has fixed the problem, I've never seen it reappear.
The really odd thing, through hyperterm I can talk directly to the port and
about 80% of the time get connections just fine, but if I try to talk to the
modem per 95's driver, either in DUN or hyperterm it fails. That's why I
think it's something flakey with either 95 or the modem driver, or both.
Also, it seems that the 2nd release of 95 is somewhat more prone to these
problems, IMO. I wonder if there is some way that 95 could mis-interpret the
result code being sent from the modem and see a CONNECT as something else?
Or is there something with the PM-3 that could cause the client modem to
return an odd result code? On one system that comes to mind I got a CONNECT
LCMP (I believe) messge in hyperterm on a 14400 negotiation, which I've
never seen before.
>
>Sincerely,
>
>Joshua Norber
>VisualNet Internet Services
>
Thanks for your input Joshua.
Craig Brown
Blueriver Networking Services