From: ohland@charter.net   
      
         THE SCSI (SMALL COMPUTER SYSTEMS INTERFACE) IS NOT A FIXED   
         DISK CONTROLLER.  IT IS A CHANNEL ATTACHMENT THAT MEETS   
         INDUSTRY STANDARDS AND IS CAPABLE OF CONTROLLING ANY DEVICE   
         WHICH MEETS THE SCSI SIGNAL PROTOCOL.  IN TROUBLESHOOTING   
         PROBLEMS, WHEN THE SCSI DEVICE(S) WILL NOT CONFIGURE OR COME   
         "READY", THERE ARE A NUMBER OF POSSIBLE CAUSES.  THIS TIP   
         IS TO HELP PROVIDE A BASIC UNDERSTANDING OF SCSI OPERATION   
         AND TO AID IN PROBLEM DETERMINATION.   
      
         FOR A SCSI DEVICE TO COME "READY" AND BE CONFIGURED, THE   
         FOLLOWING EVENTS MUST OCCUR:   
      
         A. UPON POWER-ON-RESET, SCSI POST ISSUES AN "INQUIRY" COMMAND   
            TO EACH SCSI ADDRESS TO POLL ALL DEVICES PRESENT.   
      
         B. ALL SCSI DEVICES ATTACHED TO THE SCSI BUS MUST ANSWER WITH   
            THEIR INQUIRY DATA.   
      
         C. A "TEST UNIT READY" COMMAND IS ISSUED TO ALL DEVICES   
            WHICH RESPONDED TO THE "INQUIRY" COMMAND.   
      
         D. IF A HARD DISK DRIVE ANSWERS "NOT READY," A "START UNIT"   
            COMMAND IS THEN ISSUED TO SEQUENTIALLY START THE DRIVE   
            (MOTOR START).  THE DEVICES MUST THEN RUN "SELF TEST" AND   
            REPORT THEIR STATUS.   
      
         E. ONCE THE DRIVE IS READY, THE "READ DEVICE CAPACITY" COMMAND   
            IS ISSUED TO DETERMINE THE DEVICE SIZE (320MB, 1GB, ETC.).   
      
         F. ALL DEVICES RESPONDING ARE COMPARED WITH THE CONFIGURATION   
            DATA.   
         G. PS/2 POST BEEPS TWICE AND DISPLAYS ERRORS IF THERE ANY ARE   
            NEW, MISSING, OR DEFECTIVE SCSI DEVICES.   
      
         H. FOLLOWING A SUCCESSFUL POST (ONE BEEP), THE SYSTEM WILL   
             ATTEMPT TO LOAD THE OPERATING SYSTEM .   
      
         IF SCSI POST OR CONFIGURATION ERRORS OCCUR, THE FOLLOWING ITEMS   
         SHOULD BE CHECKED:   
      
          A. DIAGNOSTICS SHOULD BE RUN TO TEST THE EQUIPMENT.  IF A   
             FAILURE IS INDICATED, USE THE ERROR CODE EXTENSION TO   
             ISOLATE THE CAUSE.  COMPLETE SCSI DEVICE ERROR CODES ARE 12   
             CHARACTERS LONG.   
                         *****NOTE THE FOLLOWING*****   
             THE FIRST EIGHT CHARACTERS OF THE SCSI ERROR CODE ARE DEVICE   
             INFORMATION, NOT ERROR INFORMATION.  THE ERROR INFORMATION   
             IS IN THE  NEXT 4 CHARACTERS (ERROR EXTENSION BYTES).   
             FAILURE TO USE THE ENTIRE ERROR CODE MAY RESULT IN INCORRECT   
             FRU PARTS REPLACEMENT.   
      
          B. EACH DEVICE MUST BE SET FOR A UNIQUE SCSI ID VIA JUMPERS,   
             SWITCHES OR REMOTELY WITH A CABLE CONNECTED SWITCH.  RUN   
             "SET AND VIEW SCSI CONFIGURATION," FROM THE CONFIGURATION   
      
            MENU, TO CONFIGURE NEW OR POWERED-OFF DEVICES.   
      
         C. ARE ALL DEVICES RECEIVING POWER?   
      
            1. IS THE "POWER GOOD" (GREEN) LIGHT ON THE SYSTEM UNIT   
               TURNED ON?   
      
               "POWER GOOD" IS MORE COMPLEX THAN A "POWER-ON" LIGHT.   
               IT INDICATES THE POWER SUPPLY IS DELIVERING ALL THE   
               CORRECT DC OUTPUT VOLTAGES WITHIN SPECIFICATIONS.  (+5   
               VDC FOR LOGIC AND +12 VDC FOR MOTOR, THOUGH SOME SCSI   
               FIXED DISKS MAY USE +5 VDC FOR THE MOTOR ALSO.)  ANY   
               MISSING OR INCORRECT VOLTAGE WILL CAUSE THE SYSTEM TO   
               SHUT DOWN (SELF-PROTECT).   
      
            2. IS THE POWER CABLE CORRECTLY CONNECTED (NOT INVERTED)?   
      
               ARE ANY CONTACTS PUSHED BACK INTO THE CONNECTOR?  CHECK   
               ALL CABLE CONNECTIONS.  ARE ANY FEMALE CONTACTS EXPANDED   
      
               (DEFORMED), THEREFORE MAKING INTERMITTENT CONTACT?  IF SO,   
               REFORM THEM USING THE TIP OF A SMALL SCREWDRIVER, ETC.   
      
         D. ALL CONNECTIONS AND CABLING MUST BE CORRECT.   
            1. IS THE SCSI ADAPTER FULLY SEATED?   
      
               TENSION ON THE EXTERNAL SCSI INTERFACE CABLE MAY CAUSE   
               THE SCSI ADAPTER TO PARTIALLY UNPLUG (UNSEAT) FROM THE   
               SYSTEM BOARD.  (SYSTEMS JAMMED AGAINST THE WALL ARE   
               SUBJECT TO THIS.)   
      
            2. ARE THE CABLES FULLY SEATED?   
      
          E. IT IS CRITICAL THAT THE SCSI BUS IS PROPERLY TERMINATED.   
      
             1. THE PHYSICAL ENDS OF THE INTERNAL AND EXTERNAL SCSI BUS   
                MUST BE TERMINATED AT THE LAST DEVICE.  DEVICES CONNECTED   
                TO THE MIDDLE OF THE BUS MUST NOT BE TERMINATED.   
      
                    __________________                 ____________   
                    |                 |                |           |   
             ___TERMINATOR         __----________      |    ___TERMINATOR   
            |INTERNAL|            |             |      |   |EXTERNAL|   
            |DEVICE  |            |SCSI ADAPTER |______|   |DEVICE  |   
            |________|            |_____________|          |________|   
      
             2. TERMINATOR TYPES INCLUDE:   
      
                - SOCKETED DIP (DUAL IN-LINE PACKAGE) ON THE ADAPTER.   
                - CONNECTOR PLUG ON ADAPTER'S INTERNAL OR EXTERNAL   
                  CONNECTOR OR CABLE   
                - TWO OR MORE SOCKETED SIPS (SINGLE IN-LINE PACKAGE)   
                  ON A DEVICE.   
                - IN-LINE TERMINATOR ATTACHED BETWEEN THE DEVICE AND   
                  CABLE.   
                - ENABLED VIA A JUMPER   
             3. THE POLARITY OF TERMINATORS MUST BE CORRECT.   
      
                - PIN 1 ON DIPS AND SIP ARE USUALLY INDICATED BY A DOT   
                  OR ARROW.   
                - PIN 1 ON SOCKETS ARE USUALLY INDICATED BY A 1, DOT,   
                  OR SQUARE SOLDER PAD.   
      
          F. READ ERRORS MAY BE CAUSED BY A BAD OR INTERRUPTED WRITE IF   
             THE DRIVE WAS BUMPED OR POWER FAILED DURING THE WRITE   
             OPERATION (SECTOR INCOMPLETELY WRITTEN).   
      
             1. IF ONLY THE DATA IS CORRUPTED, REWRITING THE DATA WILL   
                CORRECT THE ERROR.   
      
             2. IF THE ERROR OCCURS IN THE DIRECTORY OR THE OPERATING   
                SYSTEM'S FILE STRUCTURE, IT MAY BE NECESSARY TO PERFORM   
                A HIGH LEVEL FORMAT (FORMAT C: ), OR EVEN RECREATE THE   
                PARTITION WITH FDISK.   
      
             3. IF THE SECTOR ID IS CORRUPTED (ID OR SECTOR NOT FOUND   
                ERRORS), AN ADVANCED DIAGNOSTIC FORMAT SHOULD BE ATTEMPTE   
                PRIOR TO REPLACEMENT OF THE FIXED DISK.  THE ADVANCED   
                DIAGNOSTIC FORMAT WILL REWRITE ALL SECTOR ADDRESS AND   
                DATA FIELDS.  WHEN THE FORMAT COMPLETES, THE DIAGNOSTICS   
                SHOULD BE RERUN.  IF THE DIAGNOSTICS THEN RUN ERROR-FREE,   
                IT IS UNLIKELY A HARDWARE FAILURE HAS OCCURRED.   
      
          G. IF FAILURES RETURN AFTER A PARTICULAR SOFTWARE APPLICATION   
             HAS BEEN LOADED, SOFTWARE MAY BE CORRUPTING THE FIXED DISK.   
             ASSISTANCE SHOULD BE SOUGHT FROM THE APPROPRIATE SOFTWARE   
             SOURCE.   
      
         NOTE:   
         FAILURES MAY OCCUR DURING SYSTEM OPERATION, YET DIAGNOSTICS   
      
   [continued in next message]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)