Fairlight-CMI

Dear Group..

I'm at restaurating a CMI III that i managed to aquire but have a strange problem, that i hope the whole knowledge of this combined group can help me getting this wonderful instrument back to life..

What is at hand:

I have the following cards:

WFM-32 Slot 2

CMI31 Rev 2A3 x 8 Slots 8-15

CMI32 Rev 2 Slot 16

CMI41 Rev 1 Slot 17 (with new EPROM and jumper LK5 on 1&2)

CMI28 Rev 4 Slot 18

Q356 Rev 2A Slot 21

Q133 R6 Slot 23 (with new EPROMS)

Q209 R6 Slot 24

Q219 R3 Slot 25

QFC9 Rev 7 Slot 26

Q777 Rev 2 (empty EPROM Socket) Slot 27

Hi Daniel.

I just checked your info against my system.
It has the same boot ROM but a different hardware setup.

I find the following:
What you see confirms that both core CPUs (from the 2xx card) are running.
If I stop either from the front panel, I do NOT get the display cleared, or any message written at all, after a reset.

Re. the boot sequence:
The next step in a good system would be the display of available hardware.
Only thereafter would the message appear that will ask you to press space etc.
That means your system should be stuck somewhere at the test of available hardware, maybe because some of the additional CPUs or other hardware do not run as expected.

Here are all the components that are listed on the boot screen (and therefore definitely contacted in some way, before booting will continue):

BOOT ROM V17.00 - (C) FAIRLIGHT ESP 1992
/
--- WAVEFORM SUPERVISOR HARDWARE ---
KMON20 ROM REVISION KMON20 ROM TYPE
68450 DMA CONTROLLER WAVEFORM BUS
68881 FPU COPROCESSOR P1/P2 BUS
NCR 5380 WS SCSI ROUTER
NCR 53C94 TURBO SCSI HARDWARE PROTECTION
PRIVATE RAM
WAVEFORM MEMORY
-- SYNC CARD HARDWARE ---
NOT DETECTED
--- CHANNEL CARDS ---
...
--- OS9 SYSTEM AND P1/P2 BUS ---
GRAPHICS CARD
WS/TURBO SCSI DRIVER
PRIMARY FLOPPY DRIVER

HIT SPACE BAR FOR COMMAND MENU
/ (rotating wheel, shortly...)
BOOTED FROM TURBO SCSI PORT

FAIRLIGHT OS9 VERSION 6.46:FKW 23RD MARCH 1993
...

(I've left out the results of the tests to save me the typing/formatting, but can put up a photo or video if definitely needed.)

The booting will then proceed to another screen, where the SCSI HDU is displayed and the actual system finally loaded.

You might try reseating all your cards, and then, trying to boot your system with pulled waveform RAM, channel cards, waveform supervisor, or floppy/scsi controller/s - to try to isolate one card that *might* currently block its proceeding to the above hardware overview message.

I don't know however what role the waveform supervisor CPU plays already at this early point. It may be that it takes over a lot quite early in the CMI III - or it may be that it's initialized only during the actual system boot up. You should be able to find that info (and also a better description of the boot process) in the service manual and proceed with fault isolation accordingly.

Another idea: Could you enter the monitor via the NMI button on the front panel? See service manual for details, too.

Good luck & Kind regards, Joerg

Dear Group,

Thanks to JB Emond i managed to get ahold of a Flash-Floppy-Replacement to get some baby-steps further with getting my CMI back to live..

Now it seems i have a problem with the 'Quasar' parts..

The Diagnostics point to missing interrupts and parity errors on the cpu-ram

Dones anyone have and idea on what is wrong here? Interrupts are being handled by the Q133 i assume, and the party errors point to Q356.. the diagnostics i have only work for Q256, but i think party errors should not arise even if you only check the first part of the Q356 memory

Does anyone have a good image of the newer CMI III Diagnostics? The images that i found are faulty and have a lot of empty files on it, that results in E7 Errors once you start such a .CM executable..

I am eternally grateful if anyone can issue some hints or maybe have some known good cards i could borrow for a few days to compare against mine.. located in switzerland would be perfect, but i'm happy with all i can get..

thanks for your kind support
Daniel