mike@octel.UUCP (Michael D. Crawford) (08/23/89)
We have a great deal of trouble here with our Sun serial ports either hanging up (meaning rebooting fixes them) or being blown (meaning we need to replace the board). Sun Service says it is from our long serial cables being left unplugged at the far end, or being plugged into powered-off equipment, which causes the unterminated cables to pick up interference and generate lots of interrupts. I do use the Sun perfmeter program to monitor interrupts, and sometimes see the rate jump from a few hundred per second to tens of thousands, which is cured by unplugging the cable at my end; however, this behavior is not always observed before ports go out. My users have not complied to well with my requests to leave their cables plugged into turned on devices, understandable since most of them are used to download into equipment with noisy fans. What I would like to know is if there is some kind of isolator I could make or buy that would overcome this problem, at least some kind of terminator I could plug in the other end of the cables when they are not being used. I would greatly appreciate helpful suggestions of any sort, as Sun does not seem to know anything better to tell me. Thanks in advance, -- Michael David Crawford Consulting for: Oddball Enterprises Octel Communications Corp 694 Nobel Drive 890 Tasman Drive Santa Cruz, CA 95060 Milpitas CA 95035 pyramid!vsi1,!octel!mike CI$ 72377,623
grr@cbmvax.UUCP (George Robbins) (08/24/89)
In article <33@octel.UUCP> mike@octel.UUCP (Michael D. Crawford) writes: > We have a great deal of trouble here with our Sun serial ports either hanging up > (meaning rebooting fixes them) or being blown (meaning we need to replace the > board). > > What I would like to know is if there is some kind of isolator I could make > or buy that would overcome this problem, at least some kind of terminator > I could plug in the other end of the cables when they are not being used. > > I would greatly appreciate helpful suggestions of any sort, as Sun does not > seem to know anything better to tell me. You can play various games with resistors to pull the idle/disconnected lines further away from the threshold and to reduce the "zap" effect when things are plugged/unplugged or switched. The better solution, especially if your cables are real long, is to invest in some "line drivers", which are in effect cheapo modems that convert the RS232 signals into "analog" signals that can be tranmitted over two pairs of wire. Depending on distance you can reuse your serial cabling or just just cheap telephone type twisted pair station cable. An assortment of flavors/prices are available, either from the DP supply catalog (Inmac, Black Box, etc) or modem manufacturers (Gandalf, etc). Just make sure that the one you pick supports asynchronous terminals, and if it is supposed to be "self powered", that the Sun's and terminals put out appropriate voltages on whichever pins the the things draw power from. Do test a pair to make sure their behaviour when the remote end is "powered off" is one of appropriate silence and not random noise. 8-) -- George Robbins - now working for, uucp: {uunet|pyramid|rutgers}!cbmvax!grr but no way officially representing arpa: cbmvax!grr@uunet.uu.net Commodore, Engineering Department fone: 215-431-9255 (only by moonlite)
ingoldsb@ctycal.COM (Terry Ingoldsby) (08/26/89)
In article <33@octel.UUCP>, mike@octel.UUCP (Michael D. Crawford) writes: > We have a great deal of trouble here with our Sun serial ports either hanging up > (meaning rebooting fixes them) or being blown (meaning we need to replace the > board). > > Sun Service says it is from our long serial cables being left unplugged at the > far end, or being plugged into powered-off equipment, which causes the ... > What I would like to know is if there is some kind of isolator I could make > or buy that would overcome this problem, at least some kind of terminator > I could plug in the other end of the cables when they are not being used. Make up a connector with terminating resistors that could be plugged into the end of the cable that is dangling free. The connector should consist of all active lines connected via separate resistors to pin 7 (signal ground). I would try 1K resistors for starters. This should provide a low enough impedance to keep any inputs from picking up noise, and still be high enough resistance so that any outputs won't drive excessive current. I think most RS232 outputs are protected anyway, but better safe than sorry. -- Terry Ingoldsby ctycal!ingoldsb@calgary.UUCP Land Information Systems or The City of Calgary ...{alberta,ubc-cs,utai}!calgary!ctycal!ingoldsb
hedrick@geneva.rutgers.edu (Charles Hedrick) (08/27/89)
The referenced article recommends a terminator that you can plug in when a line is open. That is a clean solution, but it's inconvenient, because users have to remember to plug in the terminator. Some kinds of terminals because unterminated when they are turned off. You may be able to get around this by installing terminating resistors permanently, probably in the machine room. We saw similar problems with open lines on a DEC-20 about 10 years ago. DEC field service whipped up an ECO for us using parts from Radio Shack. (It was thereafter known as "the Radio Shack ECO", and was dutifully installed by field service in all of our newer DEC-20's as they arrived.) I didn't look at the details, but I believe they did pretty much what was described here, except they used larger value resistors. They found a value that provided enough termination to prevent open lines from ringing, but didn't interfere with normal communications when something was plugged in. I don't know what value they used. Presumably somewhere between 10K and 100K. You might experiment.