Path: utzoo!dptcdc!jarvis.csri.toronto.edu!mailrus!cornell!rochester!rutgers!att!ihuxy!vg55611 From: vg55611@ihuxy.ATT.COM (Gopal) Newsgroups: comp.sys.ibm.pc Subject: Re: Hardware Weirdness with add-on boards Message-ID: <2929@ihuxy.ATT.COM> Date: 16 Apr 89 18:12:05 GMT References: <6809@bsu-cs.bsu.edu> Reply-To: vg55611@ihuxy.UUCP (55611-Gopal,V.P.) Organization: AT&T Bell Laboratories - Naperville, Illinois Lines: 95 In article <6809@bsu-cs.bsu.edu> cfchiesa@bsu-cs.bsu.edu (Christopher Chiesa) writes: {stuff deleted} >Both "PCs" (Please allow me to call them that, even though they're clones?) >have been configured with third-party boards plugged into the internal expan- >sion slots; both appear to be clones of either an "AT" or "XT" (I don't know >what the distinction is), with a hard drive and a single floppy drive. One >is trying to use a "Practical Peripherals" internal modem, the other is trying >to use a couple of boards that together are supposed to perform interactive >video for instructional use -- the intent is to overlay PC graphics and >laserdisk video, and have the PC able to tell the laser disk what frame to >play, freeze, display, etc. etc. etc., "on the fly" under software control. >The only other thing these two systems have in common is that they DON'T >WORK, and that the problem appears to be due to something else "stealing" the >COM2 port signals! >The first system's trouble manifests in the strange fact that you can run >a terminal program and send commands to the modem, and have it obey them -- >proving that data is getting FROM the PC circuitry TO the modem -- but that >NOTHING -- modem response codes, modem internal-data-echo, remote host data >-- makes it BACK TO the PC circuitry FROM the modem, for display. At first >I thought the modem was simply flaky, but then I noticed another board in >a nearby slot, driving two "serial ports." It seemed unlikely that the PC >could properly drive THOSE serial ports and still drive the modem on COM1 >or COM2 as it was doing. I powered down, removed that serial-port card, >booted back up, and VOILA! Perfect performance from the modem. Somehow, >that other card "steals" or "masks out" whatever signals are trying to >send data in from the modem. Unfortunately, the serial-port card has also >the responsibility for a PARALLEL port, to which the system's PRINTER is >attached, so must remain in the system. My questions on this are obvious: >is there any reasonable way to let these cards coexist in the system and >BOTH work? No use is made of the two "jamming" serial ports, but even >when dip switches are set to "disable" both of them, the modem signals >are still blanked. >The second system, the interactive-video thing, actually works fine EXCEPT >that the PC fails to command the Laser Disk properly. This problem is less >well-defined, since there's no real proof yet that the Laser Disk player >actually would respond to commands even if they GOT to it. But on the assump- >tion that if commands GOT there, they'd command the player, the problem appears >to be that the commands just aren't getting out. On this one, I actually >phoned the manufacturer (in Colorado; I am in New York (forget the Indiana >username here...)) and verified that the driver software was in the right >place on the disk, CONFIG.SYS had the proper line in it, and the DIP switches >on the Laser Disk player were set properly. The circuit boards involved in >this system, again, have "COM2" involvement, and that day's session with the >manufacturer ended with his surmise that "... maybe there's a serial COM2 >port built into the main board on that PC clone, and it is STEALING INTERRUPT >THREE..." {stuff deleted} A few things about COM ports which will help you figure things out: 1. COM ports reside at (sets of) physical addresses. Only one COM port can be set to one range of addresses. The range of addresses are well defined for COM1 and COM2 but not for COM3,COM4 etc.. 2. Usually add-on serial port boards can be configured so that they are at a specific range of port addresses. You must ensure that there is no conflict of port addresses. Remember that only COM1 and COM2 have well defined addresses. 3. Many PCs come with built-in serial ports. Check your hardware manual to see if there is (are) any. Also, you should know what port it is set up to be (usually this is COM1 or COM2) so that you know the range of port addresses that it uses. 4. If you have add-on serial ports, ensure that they are at different port addresses from the built-in ones. 5. COM ports also generate interrupts when a character comes in. The interrupt number generated for COM1 and COM2 are well defined. Just like the port addresses, usually the interrupt number has to be set on the add-on board when you want it to be COM3, COM4 etc.. Usually the interrupt numbers used are the same as the ones for COM1 and COM2 - this means that the software that you are running has to be smart enough to figure out which COM port generated the interrupt. Many pieces of software are not that smart. 6. Check to see if the software that you run allows you to set up COM3, COM4 etc. and whether you can tell the software what port addresses and what interrupt number you will be using. 7. Unless you need more than 2 serial ports, the easiest thing to do is to have only COM1 and COM2 in the system. Hope that helps. Venu P. Gopal UUCP: att!ihuxy!vg55611 Internet: vg55611@ihuxy.att.com BITNET: com%"vg55611@ihuxy.att.com" or com%"vg55611%ihuxy@research.att.com" Silence those silent letters and save the world 500 million keystrokes a day.