Aucbvax.6280 fa.info-terms utcsrgv!utzoo!decvax!ucbvax!info-terms Wed Feb 24 16:13:41 1982 Lear Sieglar ADM-3A quirk >From decvax!duke!unc!mcnc!taylor@Berkeley Wed Feb 24 16:10:24 1982 This is one of those things which nobody documents, but can cause all grades of problems in some isolated circumstances. We have been running Lear Sieglar ADM-3A terminals for years with relatively few problems. Recently, however, we installed a Timeplex stat mux at one of our sites (NCCU) which had a cluster of ADM-3A's which were previously accessing services with phone couplers. All 3A's were working fine. When the 3-A's were connected to the stat mux, which, incidently, was attached at the host to a Develcon Dataswitch, some, not all, of the terminals would, immediately after receiving the 'Request:(cr)(bel)' prompt from the host, start sending a large (~100) number of space characters! Needless to say, the dataswitch considered this to be garbage, and the terminal was never able to connect. To make a 4 month story relatively short, we, and Timeplex, changed EVERYTHING. We replaced muxes at both ends. We replaced dataswitch modules. We checked the terminal on other services. Everything worked beautifully alone, with lots of other equipment in the network working also, but there was something in the combination that was not working. The first big break came Monday afternoon when I changed the phone coupler which i had been using with my ADM-3A for a direct connect modem. When I called the dataswitch, I got the same results that I had gotten through the mux. Same story when I called one UNIX system, but fine results with a different system with a different logon banner. The crucial parameter turned out to be having a bel (cntl-g) in the logon banner. UNDOCUMENTED, BUT TRUE - Lear Sieglar confirmed that if a control-G is sent when pin 22 on the EIA interface has a voltage (plus or minus) on it, this is interpreted by the ADM-3A as an internal test code for which it will respond be sending out a number of spaces. The cables used in this case (to the mux) included pin 22 (ring indicator). The timeplex stat mux and the direct- connect modem had pin 22 active, as they should. since a phone coupler does no know about auto-answer, pin 22 was no active so the terminal worked fine. Why the ADM is like this, i don't know, but I sure wish they had documented it. (It seems that the cntl-G had to be part of an initial connect sequence - maybe it even had to be while ring indicator was high. Also, it seemed to not affect some ADM-3A's at all. Very Strange.) If you have found similar quirks with other terminals, please tell us about them. Maybe this will help you avoid all the hours I spent on this one. Steve Taylor ----------------------------------------------------------------- gopher://quux.org/ conversion by John Goerzen of http://communication.ucsd.edu/A-News/ This Usenet Oldnews Archive article may be copied and distributed freely, provided: 1. There is no money collected for the text(s) of the articles. 2. The following notice remains appended to each copy: The Usenet Oldnews Archive: Compilation Copyright (C) 1981, 1996 Bruce Jones, Henry Spencer, David Wiseman.