19 Nov
2001
19 Nov
'01
9:25 a.m.
Well, much to my dismay, but not too surprising, my last set of floppy
diskette controller boards has failed and my Intel MDS is severely
crippled. In my days of designing systems with Intel Microcomputer
Development Systems I saw many of the floppy controller boards fail.
That system had a two-board set of multibus cards in order to control up
to four floppy diskettes. Seems kind of funny to call those huge 8"
floppys "diskettes", but that's what they are.
Anyway, of that two board set, I always have seen the same one fail.
One card called the Channel Card which was based on Intel's 3000 series
bipolar microprocessor never failed. It was always the other card, the
Double Density Interface Board. When I "retired" all the MDSs where I
worked, I kept three sets of those boards to keep my MDS going, fully
expecting to have one or maybe even two go "bits up". Alas, last week,
my third and last board set failed. And, yes, it was the interface
card.
The purpose of this post is to see if there might be someone on this
list who would be able to advise me on a strategy for troubleshooting it
(Tony Duell?) Better yet, but I'm not going to hold my breath on this
one, would be someone who has an SBC-202 board-set he might be willing
to part with, sell, trade, etc.
Any ideas or leads would be appreciated. And, Tony, I might have a
space ICE-80 manual for you. Will know next week.
Thanks.
Dave
--
Dave Mabry dmabry(a)mich.com
Dossin Museum Underwater Research Team
NACD #2093
19 Nov
19 Nov
10:18 a.m.
A non-text attachment was scrubbed...
Name: not available
Type: text
Size: 1589 bytes
Desc: not available
Url :
http://www.classiccmp.org/pipermail/cctalk/attachments/20011120/3413e50b/at…
11:52 a.m.
I have that manual (98-422A) and that is the correct one for the
SBC-202. I think Intel just gave the 202 name to that board set when it
was bought for integration into an application.
I am anxious to try to troubleshoot this, but I'm not sure how to go
about it. So far all I have tried to boot the MDS. It does select the
correct drive and load the head. After a pause of about one second the
Interrupt 2 light (on the cpu board, so it is coming from the multibus)
lights solid and the boot code in the MDS system rom returns a
nondescriptive error message "disk error". When it works correctly (two
weeks ago) the interrupt 2 light would pulse (very lightly illuminate)
during read. Now it comes on solid. I believe that Int2 is the signal
from the controller boards signifying end of operation.
FWIW, this seems to be the exact failure mode I've seen in every one
I've had fail. And I have three of them to experiment on, if that
helps. They work ok one minute, then the next they are bad. Nothing in
between. And when I had one good set of boards, the other two sets
definately had the interface board bad. I could make any of the three
channel boards work with my one good interface board. That is until the
last interface board failed.
I just read through the manual tonight, but I still need help with a
direction for shooting it. I was thinking that I should code a simple
routine (in hex since that is all I can think of to load memory in this
machine) to try to read somewhere on the disk and see what error code is
returned. I would guess it will be CRC error.
Note to list: Should we take this private or do you all want to see it?
Tony Duell wrote:
[Intellec FDC Channel board]
--
Dave Mabry dmabry(a)mich.com
Dossin Museum Underwater Research Team
NACD #2093
The purpose of this post is to see if there might
be someone on this
list who would be able to advise me on a strategy for troubleshooting it
(Tony Duell?) Better yet, but I'm not going to hold my breath on this
Oh, what the heck, I've got the manuals to hand. Or at least I've got
some manauls to hand. The most appropriate one is 'Intellec Double
Density Diskette Operating System Hardware Reference Manual', order
number 98-422A. I can't find the name 'SBC-202' anywhere on any of the
manuals, so this might not be 100% applicable.
The interface board described in that manual seems to have 5 separate
sections on it :
Read Data clcock recovery/sync
CRC generator/checker
Write data encoder
Disk drive control/status signals (basically ports from the microcoded
machine on the other board)
Bus control (for DMA transfers, mostly, I think).
My first idea would be to find out which block was malfunctioning. Does
the FDC card appear on the bus at all? Can you select a floppy drive and
move the head around? What goes wrong if you try to read something?
Since this board contains a fair amount of analogue circuitry (the read
data separator phase-locked loop), it's possible that the only thing
wrong with it is that that the oscillator in said loop has drifted way
off frequency and the loop can no longer lock. But don't start tweaking
things without doing some tests.
Do you have the hardware manuals? My manual contains a reasonably good
theory-of-operation chapter which explains what's going on on that board.
-tony
20 Nov
20 Nov
9:50 a.m.
A non-text attachment was scrubbed...
Name: not available
Type: text
Size: 2719 bytes
Desc: not available
Url :
http://www.classiccmp.org/pipermail/cctalk/attachments/20011120/5e371bf1/at…
25 Nov
25 Nov
9:16 a.m.
Tony and whoever might be interested in this problem...
Fortunately I was able to borrow a scope and I have scoped the signal
DATA SR STB. It starts up the first time the diskette is accessed and
stays on from then on as long as power is applied to the system. It is
a square wave with a period of 1900ns (1.9us). It is +5v for 1100ns and
0v for 800ns.
The trimpot (the only one on the board) will vary this slightly.
Also, that signal does not change that I can see on the scope when the
drive head is loaded and a read initiated.
Before I started the scoping, I put a small program into ram that would
start a read operation and let me see the error bits returned. It was a
0Ah, which means, if I am interpreting it correctly, Address error and
CRC error.
Let me know what you think I should check next, if you still have time
to assist.
Thanks,
Dave
Tony Duell wrote:
--
Dave Mabry dmabry(a)mich.com
Dossin Museum Underwater Research Team
NACD #2093
I have that manual (98-422A) and that is the correct one for the
SBC-202. I think Intel just gave the 202 name to that board set when it
OK, then we're both looking at the same manual.
was bought for integration into an application.
I am anxious to try to troubleshoot this, but I'm not sure how to go
about it. So far all I have tried to boot the MDS. It does select the
correct drive and load the head. After a pause of about one second the
We know the microcode, etc, is working correctly as that's on the other
board. And this would seem to imply that the disk controller is talking
to the Multibus (at least well enough to receive commands) and that the
drive control logic on the interface board is working correctly.
Interrupt 2 light (on the cpu board, so it is
coming from the multibus)
lights solid and the boot code in the MDS system rom returns a
Interrupts are generated by the channel board, really. The interrupt
line driver is on the interface board (A55/b, '125), so that's working
correctly. Yes, the standard INT/ line is number 2.
nondescriptive error message "disk
error". When it works correctly (two
weeks ago) the interrupt 2 light would pulse (very lightly illuminate)
during read. Now it comes on solid. I believe that Int2 is the signal
from the controller boards signifying end of operation.
Well, an interrupt from the controller, anyway.
I just read through the manual tonight, but I
still need help with a
direction for shooting it. I was thinking that I should code a simple
routine (in hex since that is all I can think of to load memory in this
machine) to try to read somewhere on the disk and see what error code is
returned. I would guess it will be CRC error.
I am not so sure. Particularly not if the PLL isn't locking. It might
never find an address mark or something like that.
Read section 2.4 carefully. In particular the last section about the
extra error codes (several bits set together). And see what you get.
If you have a 'scope or logic analyser (and I think you're going to need
one), the first signal I'd look at would be the SR STB signals (5/B1 to
use Intel's notation for finding them on the schematics). From what I
understand, this should be a clock waveform at twice the data rate, from
the crystal oscillator when not actually reading from the disk and locked
to the incoming data stream when reading. Make sure it doesn't disappear
or go off-frequency when trying to read a disk.
Note to list: Should we take this private or do
you all want to see it?
I'm happy for it to remain on-list. It's certainly on-topic, and
doubtless somebody else will find it interesting/useful.
-tony
10:26 a.m.
A non-text attachment was scrubbed...
Name: not available
Type: text
Size: 1696 bytes
Desc: not available
Url :
http://www.classiccmp.org/pipermail/cctalk/attachments/20011126/eaac39c7/at…
8429
days inactive
8435
days old
5 comments
2 participants
participants (2)
-
ard＠p850ug1.demon.co.uk -
dmabry＠mich.com