From Tim Clarke
   If it's an actual read error on the IML boot record and it couldn't recover (it tries 20 different 5-sector records for a Type-3) then the area at the "end" of the drive's capacity is suspect. Backup operational partitions and do an LLFormat from Advanced Diagnostics without clearing the "grown defect list". Then Restore System Partition and then verify IML is O.K. before reloading operational partitions. 

   You are here because the system detects an Initial Machine Load (IML) problem or an Initial Program Load (IPL) problem. You can correct most of these errors without replacing any FRUs. IML/IPL problems are associated with the following: 
    An I999XXXX error is displayed. 
    Computer powers-on without POST error, but operating system does not work. 
    The system starts up in BASIC. 

IML and IPL errors can be caused by any of the following conditions: 
    Invalid selectable drive startup sequence (default drive not in startup sequence) 
    Operating system not found 
    IML code not found 
    Related hardware failure (SCSI adapter, hard disk, SCSI cable) 
    Unrelated hardware failure (misleading error due to problem on SCSI Bus). 

Checking Startup Sequence and Hard Disk Drive
1. Start computer from RefDisk and select Set Features from Main Menu. 
2. Select Set Startup Sequence and see if default HD (usually ID6) is in startup 
    sequence. If you get an error, follow instructions on the screen, then return here. 
3. If the default drive is in startup sequence, exit from screen and Main Menu, then
    go to step 5. If default drive is not in startup sequence, follow instructions on the
    screen, then continue with step 4. 
4. Remove RefDisk, power-off computer, then power-on computer. If the "F1" (Insert
    Diskette) icon appears on screen, this means no operating system was found. Go 
    to step 5. If the "Not OK" icon appears, this means that the system cannot access
   the System Partition. Try updating the system programs.
5. Either install an operating system on a device that is in the startup sequence, or 
     change the startup sequence to include a device that has an operating system, then 
     power-on the computer, check for the normal power-up sequence and then run the 
     advanced diagnostic tests. If the problem still exists, suspect the SCSI controller,
    then the hard disk drive.

NOTES:
Password Jumper: 
   t may be necessary to move the power-on password jumper (J16-8595 or J10 - 8590) to enable an IML to the reference diskette in drive A. Moving this jumper forces the stage 1 POST routine to branch to the diskette, rather than attempt to load stage 2 from the Fixed disk system partition. (A reference diskette must be present in the a: drive)
     The Stage 2 POST code on the reference diskette is identical to the code loaded onto the system partition, Moving this jumper sets all default values, except ASCII terminal mode. It should be further noted, that moving the Power-On password jumper is effective for only 1 Power-On cycle because stage 1 POST tracks (records) the positon of the jumper each time it is run.

ASCII Terminal Mode: 
   If ASCII terminal mode is set (96-8N1, Visible in information panell), the battery voltage must be removed from the planar to reset it. (Slide stiff paper under the battery clip for approx 10 to 15 minutes).
   ASCII terminal mode will be automatically set by stage 1 POST if there is no primary video adapter detected (XGA adapter is missing from the system).

I999xxxx Error Codes

Usually Fatal I999xxx Errors
The following errors normally require processor FRU replacement:-
I9990401 - Unauthorised access.
I9990402 - Missing ROM IBM Copyright notice
I9990403 - Non IBM IML routine
I9990600 - IML Overlay Logistics Error
I9990607 - IML Overlay Checksum Error
I9990609 - IML Overlay ID Error