D&N MEM-CM9 problems

[nama]

Ok, I’m throwing this one out to you guys,

So I’m trying to get my D&N board up and working with my C2 and MPI drive.
Please see my setup for this machine below in my forum signature.

When I got this D&N board, It was missing most of the RAM, and was setup for 8” drives.
You can see here how I got it up and running and working with 24K of RAM:

http://www.neoncluster.com/projects-osi/osi-memc9.html

Now that I know my MPI drive is working (as is my C4 system) I wanted to get my C2 system up and running so that I can use HexDOS.

the D&N manual can be found here:
http://www.osiweb.org/manuals/DN_MEM_CM9.pdf

The first thing I did was to setup the D&N board for 5.25” drives. This involved changing around a few jumpers and swapping out a resistor for one with a different value. Then I noticed that the board was setup for double sided drives. Again, a bit of jumpering I was able to take it back to single sided operation. Finally the manual says that I need to adjust the timing using some onboard variable resistors. I used my logic analyser as a crude scope for this process:

adjustment.jpg (81.45 KiB) Viewed 12300 times

I can confirm that READ works as I can boot with no real issue into OS65D. However WRITE seems to be an problem. I have tried making a duplicate copy of the OS65D and it always fails. I have also tried using Mike’s new util to write data stored in a buffer directly to track 0 to create a bootable floppy. That also didn’t work. Finally I wrote 00’s to the entire track 0 so I could get a better idea what was happening. I then dumped the track 0 content back out using the disk dump tool. This is what I got:

22 00 08 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 80 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
40 00 00 00 00 00 00 00 00 00 00 00 00 81 80 80
80 80 80 80 80 80 80 80 80 80 80 80 80 80 80 80
80 80 80 80 80 80 80 80 80 80 80 80 80 80 80 80
80 80 80 80 80 80 80 80 80 80 80 80 80 80 80 80
80 80 80 80 80 80 80 40 40 40 40 40 40 40 40 40
44 40 40 40 40 40 40 40 40 40 22 20 20 20 10 10
10 10 10 10 10 10 10 10 10 10 10 10 10 18 10 10
10 10 10 10 10 10 10 10 10 10 10 10 12 10 10 10
10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10
08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08
08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08
08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08
08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08
08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08
08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08
08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08
08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08
08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08
08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08
08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08
08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08
08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08
04 04 04 04 04 04 04 01 01 01 01 01 01 01 01 01
01 01 01 01 01 01 01 01 01 01 03 01 01 01 01 01
01 01 01 01 01 01 01 01 01 01 01 01 01 01 01 01
01 01 01 01 01 01 01 00 00 00 00 00 00 00 00 00
40 00 00 00 10 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 80 80 80 80 80 80 80
80 80 80 80 80 80 80 80 80 80 80 80 80 80 80 80
80 80 80 80 80 80 80 80 80 80 80 80 80 80 80 80
80 80 80 80 80 80 80 80 80 80 80 80 80 80 80 80
80 80 80 80 80 80 80 80 80 80 80 80 20 20 20 20
20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20
20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20
20 20 20 20 20 A0 20 20 20 20 20 20 20 20 20 20
20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20
20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20
20 20 10 10 10 10 10 10 10 10 10 10 10 10 10 08
08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08
08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08
08 08 08 08 08 08 08 88 04 04 04 04 04 04 04 04
04 04 04 04 04 04 04 04 04 04 04 04 24 02 02 02
02 02 02 02 02 02 02 02 02 02 02 02 01 01 01 81
01 01 01 01 01 01 01 01 01 01 01 01 01 01 01 01
01 01 01 01 01 01 01 01 01 01 01 01 01 01 41 01
01 01 01 01 01 01 01 01 01 01 01 01 01 01 01 01
01 01 01 01 01 01 01 01 01 01 01 01 01 01 01 01
01 01 01 01 01 01 01 01 11 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
...etc

N.B. the first 3 bytes are the header info.
Obviously something not quite right there, seems to fluctuate in and out of writing 00’s.

As nearly all the IC’s in the disk section of the D&N board are socketed, I next tried to shotgun replace some parts. Below you can see what parts I’ve replaced (in green) and what parts I haven’t because I don’t have spares in my parts bin. the problem still persists.

Disk section.jpg (150.74 KiB) Viewed 12300 times

Finally I hooked up the analyser to the WRITE line, to try to get an idea as to what was happening when the 00’s are written. Again the analyser is probably not the best tool for this. What I saw was a little bit of fluctuation in the signal, but as I’m no pro, I really don’t know what I’m looking for, or even know if the reading from the analyser is accurate enough to draw any conclusions.
Here is a screen shots. It's magnified on a small area so you can see two negative pulses. I believe the pulses are supposed to be 400ns but you can see that the negative pulse, and the gap between pulses seems to fluctuate.

write 00's.jpg (52.11 KiB) Viewed 12300 times

Now, its time to just throw wild ideas out there. I don’t know why I think this, but maybe I have some kind of bad capacitor that is causing my issues. The only capacitors I can see in the WRITE circuit are C8 and C9. Does anyone think it may be a good idea to replace these?

Anyway, just wondering if anyone has any ideas as to what I should try next.

Phil

[nama]

Just a quick follow up.
So obviously there is some king of repeating pattern running through the bytes being written to disk, so I decided to analyse the binary of these bytes.

00000000 = 00
10000000 = 80
01000000 = 40
00100000 = 20
00010000 = 10
00001000 = 08
00000100 = 04
00000010 = 02
00000001 = 01
00000000 = 00

Although this isn't always perfectly accurate, and occasional spurious bytes appear, the general rule is that there is a single bit thats cascades through the binary data that is being written to disk.

As I've done multiple dumps, another possible clue is that the first part of the data always seems fine, but then the cascading bit appears and screws things up. I have yet to see a dump where the first part of the data is screwed up. Not sure how important that is or not.

Anyway, if anyone has any insights, I'm all ears.

Phil

[nama]

Last update for tonight:

Now that I have discovered what seems to be a transient bit that seems to propagate through the data being written to the floppy, I thought this would be pretty easy to spot when probing the WRITE line with the analyser, so tonight I tried again.

I sent 00's to the floppy as before, and looked at what kind of signal was being sent through the WRITE line. I set the analyser at the highest frequency in the hopes that that would give me a more accurate measurement. I really couldn't see anything that looked like a cascading bit through the signal, Nothing obvious at all.

The only thing that looked a little odd was the 400ns pulse which fluctuated anywhere between 380ns to 420ns. Strangely enough they seemed (although I could be mistaken) to be in pairs. If there was a short pulse, then it was usually followed by a long pulse.

Tomorrow I'm going to try my second MPI drive and see if that exhibits the same symptoms when dumping the data back from the floppy.

Phil

[Jeff]

I notice, as you did that each byte written has no more than 1 bit set. I'm thinking timing.... maybe the "thing" that does the writing isn't waiting for all 8 bits before it writes. Do you have a schematic?

[nama]

Hi Jeff,
Thanks for your thoughts.
Yes, the schematics are towards the back of the manuals which can be found here:
http://www.osiweb.org/manuals/DN_MEM_CM9.pdf

Actually, very occasionally there is a 2 bit set such as 81 or 11 etc. And as dumps go, the one I posted above is cleaner than most. Usually I get a few more of these spurious 2 bit bytes, but not a lot.

Also, I just tried out my second MPI drive and the D&N board exhibits the same issues.

Phil

[nama]

I have now changed out nearly every IC in the disk section of the D&N board. The only IC I haven't changed is the 74LS94 au U74 because it's the only one on the entire board that is soldered with no socket...and still I have the write issue!!!

replaced IC's.jpg (80.64 KiB) Viewed 12279 times

[nama]

Oh well, now the system had died.

I was reseating the 74LS85's at U12 and U13 and suddenly the system wouldn't start, and just showed a screen of garbage. I tried again and was able to get it to reset just once, but that was it, can't reset anymore. Removing the D&N board allows the system to boot, but placing it back in the system won't allow reset.

One step forward, two steps back.

[nama]

UPDATE: D&N board now works (not the disk write part thought) and the C2 can now reset. The issue was the 74LS27 at U5. In fact the only IC I didn't have in my parts bin. Typical.

Now I also seem to have a small issue with the 502 board with a RAM in and around the $00FB and $0200 addresses (first two 2114 RAM) . I've swapped the RAM out but the problem persists.

Can someone confirm the correct operation of the RAM CS lines?
This is what I see with my logic probe:

- When I'm scanning the addresses $0000 to $03FF the CS0 line pulses high and low.
- When I scan through $0400 to $07FF the CS1 line pulses high (not high and low like CS0), and CS0 still continues to pulse high and low.
- When I scan through $0800 to $11FF the CS2 line pulses high, CS1 stops pulsing (just goes high), but CS0 continues to pulse high and low.

I would have thought CS0 would have stopped pulsing starting from address $0400.

Phil

[dave]

Hi Phil,

You are correct, only one /CS should be active(low) for any given access. CS0 should be high for anything above $03FF. That suggests a faulty 74lS138 at U26.

I'm not sure what to make of the write issue. It almost seems as if the clock bit for a previously written track (at a very slightly slower rate) is somehow getting on to your read signal. I would suggest scoping the read signal as it propagates through the signal conditioning and digitizing circuit, but you'd be looking for spurious wiggles on the raw data, spurious peaks on the differentiated signal and the digitized signal, which can be difficult to spot. However, following the write signal through the controller (looks like it's perfect) and through the drive write circuit should be enough to check if the write signal is intact.

Are you using a black floppy, or a previously formatted one? Any chance the disk could have been written before by a machine not quite in alignment (or could your machine be slightly out of alignment?

Good luck,

Dave

[nama]

Hi Dave,
Thanks for the reply.

I also thought the RAM problem might be an issue with U26. I must have previously put a socket at U26 and put a new 74lS138 there, although I can't remember doing that, so it must have been a while ago. I removed the 74lS138 and dropped in another one, but the problem remains.

Yes, the Write section has me stumped. I don't think my old 20mhz scope is up for the challenge, maybe I need to put this fix to one side until I can afford a nice digital scope.
I've been testing on nice new floppies because I was thinking exactly the same thing. I didn't want previously written data to contaminate my readings in any way.

Cheers

Phil