jdg@ncrmud.Columbia.NCR.COM (Jim Griggers) (01/27/89)
A while ago I asked about the Packard Bell PB2400PLUS modem. My problem
was the modem was not sensing the remote carrier tone, therefore the
modem would not go on-line. Nothing has changed, but I understand the
operation of the modem a little better.
To explain the problem in greater detail: The Packard Bell modem is
connected to a Wyse 60 terminal on a single residential telephone line.
Modems called were Hayes 1200's (2), NCR 1200's (2), and Microcom
AX/2400C's (4), for a total of eight remote modems. Only about two out
of ten calls would result in the PB2400PLUS going on-line. The other
times it would just sit there until the 30 second timer expired (s7),
then print NO CARRIER. Whenever the called phone was busy, the modem
should have printed BUSY, but only did so about 40% of the time. Several
times when the modem was sent the command ATDT5551234, it would dial the
first digit in tone, then repeat the number in pulse. The analog loop-back
test also acts strangely. If I run AT&T1 which should echo characters typed
back onto the screen, about 50% of the time the modem will go into a mode which
throws garbage onto the screen in a steady stream without anything being typed.
The modem uses an Exar chip set for all the modem operations, and uses
a 8039 processor for decoding the AT commands and retrieving setup
information from an EEPROM. The Exar chip contains call progress sensing
circuits for detecting dial tone, busy signal, and ringing tone.
Apparently, the call progress part is not working properly. If I set up
the modem for no call progress detection; by entering ATX0 to select basic
result codes; everything works correctly (the factory default setting is X4).
I discovered this by remembering a problem I had several years ago with an
NCR 1200 modem which was built by Cermetek. If you selected the call
progress detection with the dip switches, the modem could not dial out in tone
from a pbx system. (This modem was too smart for its own good) What happened
was the modem would dial the first digit (9) and see if it had broken the
dial tone. Since the dial tone was still there, it would back up and
dial the whole number 9xxx-xxxx in pulse. If the extended functions
were selected using an AT command, the modem would dial correctly. I
concluded that the modem had a bug in its firmware.
Except for the call progress detection, I will admit the PB2400
works flawlessly*. It looks like it is well made and even stores more non-
volatile information that the Hayes 2400. Everyone that sent me mail
so far has had no problems with their PB2400s.
*(the loop test still does not work properly, but I don't care as long as
the actual modem operation is okay.)
- Jim Griggers
jdg@ncrcae.Columbia.NCR.COM
P.S. The guy I talked to at Packard Bell's technical support had information
that was pretty much worthless. His only suggestion was to hook the
modem up to a PC running their communications software. He said
he had heard of people having trouble using the modem with a plain old
terminal and I should try a using a PC. I don't even have a PC, but I do
have a 68020 system that runs Unix and hope to use the PB2400 for UUCP.frank@rsoft.UUCP (Frank I. Reiter) (01/30/89)
The problem you describe with the PB200 not recognizing a carrier when
calling out is a faithful reproduction of a bug in the Hayes 2400. If
the X option is set to recognize busy signals then the modem will not
respond to a carrier unless it hears a RING first. Note that if neither
end times out they will connect after 30 or 40 seconds.
--
*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
Frank I. Reiter \ / UUCP: {uunet,ubc-cs}!van-bc!rsoft!frank
Langley, British Columbia / \ BBS: Mind Link @ (604)533-2312
*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*