home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.unix.wizards:3790 comp.unix.misc:3499
- Newsgroups: comp.xenix.sco,comp.unix.wizards,comp.unix.misc
- Path: sparky!uunet!sun-barr!ames!agate!dog.ee.lbl.gov!hellgate.utah.edu!fcom.cc.utah.edu!cs.weber.edu!terry
- From: terry@cs.weber.edu (A Wizard of Earth C)
- Subject: Re: SCO ODT 1.0 (1.0.0y): UUCP hangup bugs; uugetty fix?
- Message-ID: <1992Sep4.063704.17716@fcom.cc.utah.edu>
- Keywords: uucp, uugetty, stty, hupcl, clocal
- Sender: news@fcom.cc.utah.edu
- Organization: Weber State University (Ogden, UT)
- References: <1992Sep03.222723.14461@tcsrtp.uucp>
- Date: Fri, 4 Sep 92 06:37:04 GMT
- Lines: 135
-
- In article <1992Sep03.222723.14461@tcsrtp.uucp> uunet!duke.cs.duke.edu!wolves!tcsrtp!royc writes:
- >
- >1) If I use the modem control device (sat tty1A) so that DCD
- > is listened to and causes hang up
- > a) stty clocal fails to alter behavior of uugetty
- > b) uugetty fns inbound AOK but always talks even with '-r'
- > c) which in turn creates a lock file and stops outbound usage
- > d) I can't 'cu -l tty1A' until DCD is high ...
- > (using DSR for DCD fixes)
-
- The modem control device is definitely the correct one. The problem appears
- to be that you have not set your modem so:
-
- o Respond to DTR
- o Verbose replies
- o Send Result codes
- o No character echo <---
- o Autoanswer
- o Respond to DCD <---
- o RJ11 single line
- o The modem should listen to commands
-
- (Items marked with "<---" are probably incorrect).
-
- The switch settings for a Hayes Smartmodem 300 or 1200 are:
-
- 1 up
- 2 up
- 3 down
- 4 down
- 5 up
- 6 up
- 7 up
- 8 down
-
- An Avatek modem should have switched 6 and 7 down. An additional switch
- (can't remember which one) needs to be down for autoanswer.
- An NCR 1200 should have all switches but 1,2,6, and 9 up.
- A MicroCom 2400 should have switches 1, 5, and 6 up (on the front) and
- switched 5 and 7 down (on the back)..
- A Peachtree 1200 modem should all switches but 3, 4, and 6 down.
- A US Robotics Courier or Microlink should have switches 3,4,8, and 9 down.
- A US Robotics Password should have all switches up.
- A "soft-programmed" modem should have the following commands typed at it:
- AT&C0
- AT&D1
- AT&W
- You should reset to factory defaults before doing this.
-
- >
- >... all on a 9pin connector.
- >
-
- Also, for a 9 pin connector, the cable should force RTS high on the computer
- side by tying it to DTR (since 9 pin connectors are 1 pin short of full
- modem control, which requires 10 if chassis ground takes a pin). Pins 4 and 5
- should be tied together (in the connector hood! No connection to the cable!)
- on the modem side.
-
- The PC's serial port is DTE. A 9 pin DCE port (such as on an Altos) needs
- the following cable:
-
- 2 --- 3
- 3 --- 2
- 5 --- 7
- 6 --- 8
- 4 -+- 20
- 8 -+
-
- This indicates that 4 and 8 on the 9 pin computer side should be tied to pin
- 20 on the 25 pin modem side.
-
- On 25 pin to 25 pin, 6 does not connect through -- instead, it connects to
- pin 8 on the same side. Pin 8 will connect to pin 20 on the other side
- if 2&3 have to be crossed, or pin 8 on the other side if this is not the
- case. 4&5 should be tied together on the computer and modem ends with no
- connection in between (unless you have the device driver writer tell you
- different, in which case, he's probably ignoring a lockup-on-line-loss
- window or is unfamiliar with your modem). Chassis ground should be connected
- on one end only (the one with the best ground) it's purpose is to act as
- a Faraday cage for wire noise (all loose wires in a cable should hook to
- chassis ground on one end only) and connecting both ends may result in a
- ground loop (ie: pin 1-1 is connected and pin 7-7 is connected and the
- power supplies have different ideas of a 0 voltage reference).
-
- If you are using a Bell 103-C DataSet, feel free to connect pins 4 & 5, as
- this is one of the few modems that understands CTS/RTS and how to handle it
- as UNIX perceives it with regards to DTR/DCD transitions. (Built in
- monopoly).
-
- > Basically, uugetty ought to keep its mouth shut
- > until a character whatever happens to the cvontrol
- > lines on a modem control port and 'clocal'/'-clocal'
- >
- > should WORK RIGHT on tty lines ...
-
- Basically, using tty1A instead of tty1a will make getty or uugetty keep
- quiet until the modem asserts DCD.
-
- Your *modem* should ignore data from the computer until DTR is asserted
- and your *modem* should reset as if powered off-then-on-again on an
- on-to-off transition of DTR by the computer.
-
- Your *computer* should use a modem control port (requires a kernel or boot
- monitor patch on Sun -- call Telebit) and have -CLOCAL and HUPCL set in the
- /etc/gettydefs file. Your *computer* should ignore RTS and assert CTS at
- all times if DCD is not high to work with Hayes style modems, but doesn't
- (hence the cable mods).
-
-
- Oh, and *never* use an internal modem and expect it to work; there's no
- way you can modify the cable, and the cheap ones won't have appropriate
- pullup and pulldown resistors to assert/deassert the proper signals prior
- to a connection being made (at which time the line is held active low or
- active high, as appropriate. If you are good at card surgery (like cutting
- the IRQ 2/IRQ 9 vertical retrace interrupt line so your VGA card lets you
- jumper your ethernet card to IRQ 2/IRQ 9), you may want to buy an internal
- modem and solder your own pullup/pulldown resistors.
-
- If you need any more help, buy a comm product (like TERM from Century
- Software, 801-268-3088) and make the vendor support you.
-
-
-
- Terry Lambert
- terry_lambert@gateway.novell.com
- terry@icarus.weber.edu
- ---
- Any opinions in this posting are my own and not those of my present
- or previous employers.
- --
- -------------------------------------------------------------------------------
- terry@icarus.weber.edu
- "I have an 8 user poetic license" - me
- -------------------------------------------------------------------------------
-