• Please review our updated Terms and Rules here

Reviving a D9060

KC9UDX

Space Commander
Joined
Jan 27, 2014
Messages
7,468
Location
Lutenblag
I figured I may as well post my experience with this, because it seems there is very little information otherwise about these things.

I have a D9060, which did not work when I bought it. I bought it from someone who claimed the LEDs didn't light up all the way, and the fan and hard drive sounded too slow.

So, the first thing I found was that it is designed for 240VAC, and was plugged into 125VAC (cord still attached).

Upon connecting it to the proper power source, I found that I could get "73,cbm dos v3.0,00,00,0" at startup. Upon issuing any valid command, the unit would hang. If I issued an invalid command, I would get the appropriate "syntax error" message. So I presumed that communication between the two 6502s on the "DOS board" was broken. (I have since learned that a loss of communication between the "DOS board" and the "SASI board" can cause this)

Finding both CPUs running in valid address space, knowing that the IEEE-488 interface 6502 was working correctly, and having quite a bit of experience with other hardware containing bad 2114 static RAMs, I shot-gunned the problem and replaced all eight 2114s. Now, I could get the drive to successfully report "0,ok,00,00,0" after issuing "i" or "uj" commands. Upon trying to retrieve the cATALOG, lOAD any files, or issue "v", I would get read errors.

I could hear the drive seeking. In fact, upon execution of "v", it would seek quite a few times before failing. Always, read errors, usually in what must be the directory track.

Having checked everything else to my satisfaction, I got gutsy and issued "n:lxx,mw". It sounded like it was seeking properly, and about 15 minutes in, my ?ds$ command returned, blank. In my experience, this is what happens if you turn off a drive in the middle of whatever it's doing. So, I raced to the other room (the controlling computer and D9060 are in different rooms) to find the power LED extinguished, and the activity LED still on, the drive motor and fan still running, but no sign of seeking.

I found that the 5V supply gave up the ghost. I haven't nailed that down yet, but it seems either the center tap from the transformer is open, or, I have two rectifiers with high ESR (but both seem to check good). I'll deal with that later. In the mean time, I disconnected the rectifiers and connected a 5V bench supply to the filter capacitor, to make sure the capacitor and regulator circuits are still functioning properly. I have the D9060 power supply disconnected, and I'm powering the SASI board and the Tandon drive with a small PC-type switching supply.

One thing I had noticed earlier was that the six socketed 82S137 chips on the SASI board get VERY hot during operation. So whilst I had everything apart, I did a quick internet search for 82S137s, and found them to be RAMs (wrong!). So I pulled them, and cleaned the pins and sockets, just because I could. Because I thought they were RAMs, I didn't pay any attention to which socket they were in. That was a big mistake. They are PROMs.

After only 25 iterations I got lucky and got the PROMs back where they belong, or so it seems. If I get an opportunity (I won't...) I'll read these PROMs and see if I can get something that doesn't run so hot. Also, I wonder if this is common, and if so, maybe these SASI boards are prone to fail due to overheated PROMs.

So, now, I am back to square two. Upon power-up, I get "73,cbm dos v3.0,00,00,0". If I issue "i" or "uj", I get "0,ok,00,00,0". If I do anything else, I get "74,drive not ready,00,00,1". I figured at this point I had nothing to lose, so I issued "n:lxx,mw" again. It seemed to be seeking along nicely for the first few minutes. It has been running for an hour and a half with no sign of progress, and I don't hear it seek anymore. But, if it doesn't seek very often, or does so quietly, I won't hear it. I'm in a different room, and there's a large box fan over the top of the D9060 keeping it cool.

I have no idea what a reasonable time is to complete a format. I don't know how I'll know if it finishes, unless it finishes successfully.
 
Last edited:
Well I spoke too soon. It finished whilst I was typing.

But, I'm back to square three.

"74,drive not ready,00,00,1"
 

Attachments

  • KIMG0046.jpg
    KIMG0046.jpg
    97.7 KB · Views: 2
OK, the 'problem' was staring me in the face. The DOS board suffers a quirk similar to the 1540/1541. Beyond that there's absolutely nothing wrong with it.

I wonder how many D9060 and D9090 went in the garbage because of this.
 
So here's the deal.

After issuing a "n" command to format the D9060, it tries to access drive 1 instead of drive 0. Somehow this becomes permanent. I don't know how. Drive 1 of course does not exist.

What this means is any time after formatting, any command results in "74,drive not ready,00,00,1" which of course indicates Track 0, Sector 0, of Drive 1.

All it takes to get it working again, is issue a command referencing drive 0. I used "i0", but a BASIC 4 command of "cAd0" ("catalog d0") works just as well.

View attachment 25564
 

Attachments

  • Untitledy.jpg
    Untitledy.jpg
    11.9 KB · Views: 1
Last edited:
Hello Dave,

Do you think a Flash ROM could be fitted in there?

Matt

Well it looks like the SASI board has a bit-slice CPU on it. The Proms are 1Kx4. Wikipeda suggests the CPU is 16-bit so 6 x 4bit proms does not make sense, unless there is some microcode, but however they are used they need to be fast. I doubt that any Flash ROM is fast enough.

I was going to say "no chance" but in fact when I look fast PROMs are available. For example:-

http://uk.rs-online.com/web/p/eprom-memory-chips/7380097/

are 16-bit Flash ROMs with a 45ns access time. Pity they are the wrong form factor, but a small daughter board could perhaps be made...
.. so I don't think you will find drop in replacements, but you will be able to do something
 
Yep, making an adaptor board should be pretty straightforward.

I thought the same thing about 6 x 4bit...
 
Would it be possible to get a good photo of the board so I can see the chips on it. All the pictures, and "schematics" I have seen are so fuzzy as to be useless...
.. in fact all the pics of the controller board seem to be the same as the one on wikipedea which is so small you can't even count the chips never mind read the numbers.

In fact as SASI is an 8-bit bus, and there is only one large chip on the board, (the over lay file from the manual is OK but I can't find a matching parts list) I just wonder if it isn't an 8-bit micro...
 
Hello Dave,

I got a chance to take some pictures last night. Unfortunately my phone does not take good pictures. But here are some pictures of the DOS board and the SASI board.

In fact as SASI is an 8-bit bus, and there is only one large chip on the board, (the over lay file from the manual is OK but I can't find a matching parts list) I just wonder if it isn't an 8-bit micro...

That's my impression too, but I haven't had an opportunity to do any research.
 

Attachments

  • IMAG1372.jpg
    IMAG1372.jpg
    97.6 KB · Views: 1
  • IMAG1373.jpg
    IMAG1373.jpg
    93.2 KB · Views: 1
  • IMAG1374.jpg
    IMAG1374.jpg
    103.3 KB · Views: 1
  • IMAG1375.jpg
    IMAG1375.jpg
    103.3 KB · Views: 1
  • IMAG1376.jpg
    IMAG1376.jpg
    100 KB · Views: 1
IMAG1378.jpgAlso, I found the problem with the power supply. The centre tap pin of the power connector at the power supply board was badly burnt. I don't know how I missed that in the first place.

After many attempts to clean the pin and the wiper, it still has too much resistance. So, I bypassed it. Everything works as it should now.
 

Attachments

  • IMAG1380.jpg
    IMAG1380.jpg
    96.6 KB · Views: 1
I don't think it was a RISC processor from AMD in 1978... I think the number is 2910D, at least that's what I thought it said. Unfortunately the pictures got squashed when I uploaded them here.

I checked the original photo. The chip reads AM2910DC 8232DM (C)1978 AMD

https://en.wikipedia.org/wiki/AMD_Am2900

It's 12 bit. That makes sense. :)

I don't know why I never heard of these before. Or, maybe I did and I forgot. Either way, the 2910 isn't a uP, so the whole board is more complex than I thought.
 
It does except I can't see a 4-bit (or any bit) ALU on there and you would normally use the 2910 with a 3 x 2901 ALU's or even 74LS181 both of which come in 0.6" packages and the 2910 is the only 0.6" chip on the board....
 
It's 12 bit. That makes sense. :)

The 2910 has a 12 bit wide addressing capability. It is a microprogram memory controller usually used with the 2901 bit slice, but not necessarily. For simple sequencer designs that do not require arithmetic, it can do the job by itself. It can keep track of a stack (nine address deep) to support subroutines, perform instructions like RESET (jump to zero), and other control functions. Here is the blurb from the spec sheet:

The 2910A is an address controller that will sequence through a series ofmicroinstructions contained in a ROM or PROM. There are four sources of addresses from whichthe IA2910A may select. The first is a direct data bus (DIN) input, which is straightforward. Thesecond is an internal address register. This register contains data loaded during a previousmicroinstruction. The third source is a program counter register, which generally contains theprevious address incremented by one. The final source of addresses is a nine-word deep stackregister that gives the capability to execute nested subroutine linkages.
 
I kind of know that, but because the manual

http://www.zimmers.net/anonftp/pub/cbm/schematics/drives/old/9090/01.gif

claims its a "Bit Slice Processor from the AMD2900" series I assumed it was a real CPU. Perhaps not, but very useful for implementing the state machines needed to interface to SASI or the hard drive. The XEBEC boards that do the same job generally use a Z80 CPU but I guess that was forbidden at Commodore.
 
I would assume the selected design was at least $2 cheaper to manufacture then using e.g. a Z80. You know the importance of cutting costs even on high profile, expensive gear.
 
I'm not sure Commodore made the board. The DOS board seems to have been contracted out, and the SASI board just doesn't look like something Commodore would do.
 
I found with the D9060 swapping out the Tandon for an ST-225 improved performance (as in the Tandon drive did not work)... you may have a marginal drive.
Bill
 
I don't know... I've been working it pretty hard the past two days and it seems to be doing well (so far).
 
Back
Top