JT> From the commands and setup I have fed
 JT> the modem, I believe that EC is enabled.

 RS> Yes, and the evidence from the result when you disable EC does
 RS> indicate that its normally achieve. Corse you cant be absolutely
 RS> certain on that evidence alone that its actually being achieved
 RS> on the PARTICULAR dud calls. I have seen situations when repeated
 RS> calls to the same modem on the same line at each end occasionally
 RS> results in a PARTICULAR call not actually achieving EC at all.

 You may be right, although while testing a while back, purposely disabling
EC results in continuse errors during the link.  Even while in menus, I get
noise with the stuff that comes down, on the calls in question, there is no
noise, just pauses every so often.

 RS> With some modems that stands out like dogs balls on the front panel
 RS> display. 

 Lucky me.  I don't have that luxury.

 JT> What my spastic modem does behind my back is beyond my control.

 RS> Sure, but I was discussing whether its line errors that produce the
 RS> errors at the Zmodem level. Dont believe that one. Something else is.

 I wouldn't be the least bit suprised if it is buggy firmware not being able
to deal with many retrains and re-sends.  I know my modems MNP implementation
(EC and compression) is not really compatible with many modems.  I've disabled
everything to do with MNP now, and my connections have become a LOT more
predictable.
 I am happy to deal with the occasional non-EC links.  The alternative is
NO connection at all.
 
 JT> And yes I'm sure I have EC turned on, I've tried it turned
 JT> off once, sheez it was bad. IMO, EC *improves* the situation,
 JT> but does not always completely eliminate errors.
 
 RS> Yes, if the line is so grossly bad that the
 RS> modems retrain, you can THEN see Zmodem errors.
 
 Yep.

 RS> And the line must be that grossly bad if it drops out at all.

 Yep.  Sigh :-(
 
 JT> Out of the 10 odd lines going to the same place, one is by far the
 JT> worst. I remember asking ages ago if he had any problem reports by other
 JT> users, but there were none.  Most likely it is a compatibility issue.

 RS> Maybe. You can certainly get some weird combinations with modems. At
 RS> one time with lots of us calling here roughly daily, just one person
 RS> ever saw a significant level of problem with a particular modem, the
 RS> others had no problems at all, with the SAME modem. Turned out when we
 RS> measured his line with a modem that can do that, his line is very
 RS> measurably worse. Combined with Pauls which is nothing to write home
 RS> about, and a his modem with a quirk that made it exhibit a significant
 RS> number of dud calls with him. 
 RS> Its even possible that the flow control in the transmitting
 RS> modem has some deficiencys in the retrain situation.
 
 I think I've only had a problem once with Paul's modems, it was with only
one modem, and it did not connect the first time I polled, the next time,
(within five minutes) it worked.  On the other hand, when my other host had
his spirit online, 3-6 of those bad calls per month was the norm.
 Where I work now, I get to see warranty-returned spirit modems (many
different models) all the time.  I have no problem with telling anyone that
they are nothing short of complete rubbish.

 JT> Maybe, I think it is just my modem that is intolerant of bad lines.

 RS> If line errors are getting past the modems EC you have a problem
 RS> somewhere.
 
 JT> I wish I knew where,
 
 RS> Most likely an obscene line if it drops out at all. It shouldnt,
 RS> just fall back speed wise to a speed that isnt too bad error wise.
 
 Like I said, the bad link is usually repeatable with a particular line of
the BBS that I call, I still have to get around to asking what type of modem
he has online there.

 JT> Given a choice of slow, or dropped line, I'd take slow any day.

 RS> True. It may well be that that particular pair of modems combined with
 RS> your and the BBS modems phone line cant actually successfully
 RS> negotiate an appropriate set of speed etc conditions to make the
 RS> session atleast usable. And the Zmodem errors are actually just
 RS> timeouts seen when they are mindlessly retraining trying to negotiate
 RS> the session params, and that quite a bit of the time they just give up
 RS> and the line drops right out. 

 I remember seeing the data come down fairly rapidly with the errors appearing
in between, there were no timeouts at that stage in the download.

 JT> With slow, the data still gets across whatever way.

 RS> And you arent paying for the time, you are for fresh connects.

 True, true...
 
 JT> Either way, fuck all I can do about it except not call that line.

 RS> Well, a better modem at your end may well have more success calling
 RS> that particular modem/line if everyone else does better. Corse you
 RS> would want to try that and see rather than just buy and pray.
 
 Pity no one offers a try-before-you-buy deal.  No bloody wonder.

 JT> May very well be, I'll ask again on reports on that particular
 JT> line, last time I asked, he said there were no problems.

 RS> You should also see a marked difference between summer and
 RS> winter nights with an HR joint, particularly heat wave nights.
 
 I haven't noticed differences between seasons, not taken note to be honest.

 JT> It must be just bad line noise, on certain systems,
 JT> i CAN'T get over 24,000 odd, yet I've called interstate
 JT> at 28.8 without even a hint of a glitch.

 RS> Yeah, certainly the line is down on performance, BUT the whole point
 RS> of V34 is that it should be able to get a good solid session even with
 RS> lines that are down considerably performance wise. Albeit with the
 RS> thruput down a little because the line isnt anything special at all.

 Yep, on one of my calls, it initially started at 28.8, (with the pauses) and
then after the problems stopped, it dropped to 26.whatever.  Come to think of
it, it *may* be that the "trouble" modem might not be able to correctly
negotiate speed changes with my modem.

 JT> The only thing that is stopping me, is the adaptive answering
 JT> works ok most of the time, and I use it.  Even if the data is
 JT> reliable on a new modem, if the fax don't work, it's no good to me.

 RS> Yeah, life wasnt meant to be easy |-)

 (Grumble)  Oh why not!?  :-)

John Tserkezis, Sydney, Oz. Fidonet: 3:712/610  Internet: jt{at}suburbia.com.au

... An error?  Impossible!  My modem is error correcting.
---