Warwick, at 20:12 on May 04 1996, you wrote to Bill Grimsley...

WE> The only difference I note when dialling with ATX=3 in place, is that the 
WE> engaged signal will not cut off and I have to manually hang up.

BG> I have just noted that the incorrect command syntax of ATX=3 will
BG> actually set X0, which in turn disables ALL service tone detection.

WE> That is probably the reason I have to hang up manually, when getting an 
WE> engaged line?

Absolutely.  As I said before, the correct syntax is ATXn, NOT ATX=n, and
as  X0, X1 and X2 actually disables BUSY detect on the Sportsters, it won't
drop DTR on a BUSY tone.

BG> Start your terminal app, and configure it so that you have CTS/RTS
BG> hardware flow control enabled, XON/XOFF software flow control
BG> disabled, lock your serial port rate to 38400 bps, and enable your
BG> 16550 buffered UART if you have one.

WE> My internal Sportster is configured as above, except that I had the DTE set 
WE> at 115,000 bps,( I have lowered it to 38400 as you suggest). I was of the 
WE> opinion that the higher the rate the better??

Within reason, but the majority of 14400 modems won't even work with serial
ports locked higher than 57600 bps.  28800 modems will, of course.

BG> Next, edit your term's init$ to read ATZ^M (where control-M is the carriage 
BG> return character), then enter the following string to your modem (once 
BG> only, and exactly as written)... AT&F&A3&K3X3&W^M

In your case, The X3 above should actually read X2.  Sorry about that.  :)

(If a 28800 internal, you need to change the above to read AT&F1 as well)

WE> Yep, did that,and when I tried it out, it worked, but when I hung up, and
WE> then fired up Telix again, and looked at ATI4, that pesky
"X4" was there
WE> again. On looking at the manual, I read on page 34, for
"&W" commands, that
WE> &W0  modifies NVRAM 0 template
WE> &W1  modifies NVRAM 1 template
WE> I wonder, did you accidentally leave out either the 0 or 1 on &W  ???   

No, with &W, the 0 is automatically assumed.  Same as ATZ.  If you'd
written the profile to &W1, you'd need to recall it with ATZ1.  If your
ATI4 is still showing X4 for profile 1, you've either failed to write the
command to NVRAM with &W, or the modem is faulty.  There is no other
possibility here.

WE> I then tried redialing, but my original problem of dropout on 2nd ring was
WE> back.

Try using X0, and work upwards until it ceases to function correctly.

BG> Set the modem's DIP switches so that 3, 5, and 8 are ON, all others OFF  
 
WE> Being an internal, I don't have any??

My internal USR 14400s sure did, but I've heard that one or two models
didn't, although I've never personally seen one which didn't have the DIP
switches. 

WE> My settings on ati4 were mostly the same as yours above with the following
WE> exceptions. S10=014  S11=055  S34=000  S36=014  

Obviously a different ROM revision from mine then.  If those are the &F
values, leave them as they are.

WE> The above problem with transfer time seemed to start when I started using 
WE> the above "atx=3". The actual rate of transfer still shows
as approx 
WE> 3100-3200, so that isn't affected.

Didn't you say yours was a 14400 modem?  You should be seeing transfer
rates around 1650-1680 cps, certainly not ~3100, which are 28800 class
speeds.

In fact, if you ARE running a V.34 USR, all bets are off, as the command
set is quite a bit different.  In that case, AT&F1 is all you need for
an init string (plus the correct X setting to solve your other problem).

WE> My biggest problem is that I know enough about computers, modems etc, to 
WE> get myself into trouble :(((.

True, and even current manuals can be extremely cryptic.

WE> I envy you guys who have all the answers. Still, I suppose it isn't 
WE> something one can learn overnight?

You'll find that those of us who do have the answers have also been using
modems for several years.  Even Albert Einstein would have had difficulty
learning the Hayes command set overnight!

WE> Before I forget, I also tried using "atx=2". That made no
difference from
WE> using "atx=3"

No, because "ATX=anything" will set the X register to 0 (zero). 
Once again, drop the "=", and just use ATXn instead.  The syntax
is critical here.

WE> Hope I'm not becoming a nuisance .

No comment.  :)

Post a dump of your ATI7 screen if you can, and we'll try again.

Regards, Bill

--- Msgedsq/2 3.20