In a message dated 08-16-95, Bill Grimsley said to Michael Raiteri:

MR> True however I'm not sure phone lines are the problem.  If the stats I 
MR> posted bring this out then fair enough. 

BG>That's EXACTLY what your stats show.  Your lines are terrible.

Fair enough.

MR> BTW which ones do I look for and what are acceptable ranges.

BG>The S/N ratio should be 38db or better for a start, and yours isn't.

Thanks thats what I wanted to know.

BG> You have shitty lines, and the best modem in the world is 
BG> not going to change that fact.  It's bad luck, but you're 
BG> basically stuck with it.

Agreed however from my understanding of modems isn't error correction built in
to allow reasonably high connect rates on noisy lines.

MR> True but it still alter the fact my Dynalink will not 
MR> negotiate _any_ sort 
MR> of connection with at least two BBS' and ISP' using an init string of 
MR> AT&F1&HI&R2X1 or including &K0.  It just squeals
during negotiation but 
MR> does not connect. 

BG>It could just as easily be the remote modem at fault.

The thought had crossed my mind however I would have thought my ISP would have
had more than one modem and its my ISP that just refuses to negotiate at
anything other than a set 14400.  I wish they would reply to my email as to what
types of modems they are using. BG> 
MR> I can then  hang up and use an init string of
AT&F1&k0X1S32=12S33=32 and 
MR> get a 14400 connect with those same BBS' and ISP 
MR> _every_ time. Why won't my 
MR> Dynalink negotiate down without having to be set at 
MR> 14400 (before you ask I
MR> have checked and they _all_ have 28800 modems but I 
MR> don't know what sort). 
MR> Shitty lines should not cause that.

BG>Sure, but I've already told you to try disabling the 3429 symbol rate and the
BG> V.42 detect phase, both of which are workarounds for 
BG>problems inherent in others modems, not necessarily yours.

I've also tried that S33=32 to disable the 3429 symbol rate as well as both
S15=128 to disable V42 and S15=136 as mentioned by Tom Przeor.  It didn't make
any difference.

MR> BTW on each of these calls I was using Dynalinks 
MR> recommended init string of AT&F1&H1&R2X1.

BG>As usual, I suspect that Askey are talking crap.

I don't doubt that.

BG>Disabling the V.42 detect phase does NOT prevent the modems 
BG>from connecting with V.42, it just stops them from 
BG>negotiating it themselves.  The whole point is to FORCE 
BG>V.42 without the possiblitly of the modems objecting to it.

Thanks for clearing that up.  Unfortunately it didn't help :(

MR> Is there any manual/publication/file which sets out how to compare the 
MR> stats produced by ATI6 ATI11 and ATY11 with 
MR> perfect/ideal ones and what to 
MR> do.

BG>No, but linking into the USR_Modems echo will help, I'd imagine.

I'll do that, it'll give me enough reading for the weekend.

MR> I've thought about it lately with this one anyway. There is one thing you 
MR> may be able to add.  On one particular BBS I connect 
MR> then about 20 seconds 
MR> later it drops carrier.  ATI6 displays the reason for the disconnect as 
MR> GSTN  cleardown. I can't find anything on GSTN 
MR> cleardown.  Can you shed any 
MR> light on it?

BG>  GSTN (General Switch Telephone Network) Clear Down:  

BG>     The connection was non-ARQ and DTR was dropped from one side 
BG>     of the connection, or the DISC frame was corrupted due to noise.

Again thanks.  It had Dynalink totally stumped.  None of their tech team had
ever heard of it.  Pretty bloody poor on their part.  Maybe you should get a job
there to improve things a _lot_.

BG>Sounds to me like the original connect was made without error correction, 
BG>something which disabling the V.42 detect phase is supposed 
BG>to prevent.

Hmmm &m4 would allow that as well. I've since changed to &M5. 

lastly could you confirm that S15=128 does disable V42 detect phase and not
disable V42 entirely.

The modem online help for the S registers has S15=128 - disable V42

many thanks.

Cheers
Mike
Internet: mraiteri{at}dyson.brisnet.org.au
___
 * MR/2 2.25 NR * Help stamp out, eliminate, and abolish redundancy!

--- Squish/386 v1.11