🚀 go-pugleaf

This is a repost of an article I posted recently to demon.tech.modems. I 
am reposting with a wider Newsgroups: scope, because it generated little 
interest in d.t.modems, but fits well with a thread currently running in 
d.i.s.pc. Since it is not platform specific (at present) I have also 
included demon.ip.support with a view to removing the d.i.s.pc cross-
posting later on if the platform independance remains. 

NOTE that my local number is on the Mercury ROMP, so I am guaranteed an 
Ascend modem when I dial.

   -----8<-------8<-------8<-------8<-------8<-------8<-------8<-----

For the last few weeks I've been happily using a USR v34 Courier, but 
unfortunately it wasn't mine and has now gone back to its owner :-( I 
have returned to my old Zoom v32.

In my "dialer" file there were two init strings for the Zoom commented 
out, and I couldn't remember which I had been using, so I picked one and 
off I went.

I managed to connect several times (although less reliably than usual), 
but having received the motd everything died.

After a few abortive attempts (ie no successes) it occured to me to try 
the other init string; I connected first time and everything worked okay 
(and has done ever since). This suggests to me that at least some of the 
problems connecting to the Ascends are related to modem setup - both the 
init strings previously worked okay with the Couriers. 

The init strings were as follows:

    FAILED: AT&F0E1N0Q0X3Y0&C0&D2&G2&J0&K3&R1&S1&T4S37=11\r
    WORKED: AT&F0W1&C1&D2%C2\r

Both are from Modem.txt, I believe.

I do not have access to the Zoom manual at present, so can't determine 
the exact differences (I suspect that the former is largely setting 
things to their default values anyway), which is why I'm passing it over 
to this group.

I can carry out further tests if required, including taking any logs that 
KA9Q (DOS) will take.

In demon.ip.support.pc it has been noted that some people fail to connect 
to the Ascends in this way when dialing with the DOS software, but 
succeed when using Turnpike, which could quite possibly be because 
Turnpike generates its own Init strings which quite possibly differ from 
those in the users' DOS Dialer scripts.

   -----8<-------8<-------8<-------8<-------8<-------8<-------8<-----

I have now found what I think is the Zoom manual, so that the differences 
between the two INIT strings can be determined.

AT&F (set factory defaults) expands to:
 AT B1E1F0L1M1N1Q0V1W0X4Y0 &C0&D0&G0&K3&P&Q5&S0&T4&Y0 %C3%E0 \A1\G0\L0

AT&F0 is equivalent to AT&F. We can remove any redundant parameter 
setting:

Therefore, AT&F0E1N0Q0X3Y0&C0&D2&G2&J0&K3&R1&S1&T4S37=11 (Failed) =
 AT &F N0X3 &D2&G2&J0&R1&S1 S37=11

Similarly, AT&F0W1&C1&D2%C2 (Worked) =
 AT &F W1 &C1&D2 %C2

&D2 is common to both, so can't be an issue, but only means "hang up if 
DTR dropped" anyway.

Of the remaining commands:
 (the following from the FAILED init string)
 N0S37=11  - Force 14k4 V.32bis connect
 X3        - Affects which modem responses are displayed
 &G2       - Enable 1800Hz guard tone
 &J0       - Not defined in manual
 &R1       - Not defined in manual
 &S1       - DSR active after answer tone detected, inactive after 
             carrier loss
 (the following from the WORKING init string)
 W1        - Affects which modem speed/protocol parameters are reported
 &C1       - RLSD (DCD) follows remote carrier
 %C2       - Enable V.42bis compression (note: default is V.42bis *and* 
             MNP5 both enabled)

The two un-defined parameters are interesting (&J and &R), as is the fact 
that the failing INIT string appears to accept the default of V.42bis 
*and* MNP5 enabled.

I will do some more experiments this evening to try and determine which 
parameters are valid, but I repeat that both strings work with the USRs. 
I have also just checked and can confirm that both are indeed from 
MODEM.TXT. I have included the detail of my "workings" here to prompt 
others to experiment, although I would suggest that in most cases the 
level of detail included here need not be posted (although any 
conclusions should be). 

Mark.
-- 
int MeaningOfLife (void) { return 42; }   // mcr@holly.demon.co.uk
function MeaningOfLife :integer; begin MeaningOfLife := 6 * 9 end;