Amiga.org
Amiga computer related discussion => Amiga Hardware Issues and discussion => Topic started by: melange on February 19, 2007, 03:08:40 PM
-
Hi folks, have just landed myself an A2630 Rev 9, with a 25mhz 030 and 25 Mhz 68882. The card came without the 2MB Zip RAm and also without the 68882, I just happened to have the ram and fpu lying around.
Problem is she don't work. One thing I did notice after looking here...
http://www.amiga-hardware.com/download_photos/a2630_1_big.jpg
Is that the Oscilator on my card is only 16mhz, whilst the OSC on the card in that pic is 25mhz.
I found a 25.175 mhz Crystal, but still no go.
How exact does the timing have to be? The one in the above picture is 25.000mhz.
Anyone know?
-
I have used my 2630 with a 24 and 25 MHz osc and it worked fine with booth, perhaps your board is broken?
-
Thanks for the fast reply, yes, I think you might be right. I'd like to unbreak it however. :-D
-
Yeah your crystal should be just fine, as should the 16Mhz one though (Although I guess you could get memory refresh problems)
Have you checked if all the jumpers are correct?
there is one for memory size, what OS your booting and some others.
Any missing chips or anything?
Have you tried removing the 68000 off your A2000 board? (Only needed in german A2000's but you can do it to any A2000)
-
@Oli_hd
Nah, I haven't tried removing the 68000 as it's not a German 4 Layer board. Have checked all jumpers carefully. It has a rather odd set of ROMs, with "Amiga Technology Aust" 390282-07 & 390283-07. Nothing like the pics I'm seeing online.
Anybody got the ROM code that's compatible with 3.1?? I can blow a new set if I can get my hands on the code. I have a programmer handy and the correct EPROMS.
-
Glad I wrote that down there, coz I just rubbed my finger on the sticker as I attempted to lift it off and the ink came off the sticker. Home made ROMS for sure me thinks :lol:
-
@Oli_hd
Come to think of it, I don't get it. I know that on the A500 for instance the 28.xxx OSC feeds the agnus which in turn provides the 68000 with the 7.1x mhz signal.
How could a 16mhz OSC give the 25mhz processor what it needs?
-
How could a 16mhz OSC give the 25mhz processor what it needs?
Simple, it would run at 16Mhz, the whole board would clock at the same speed (if any of the PAL's actually use the system clock, I have the service manual and user manual so I could have a look) and you can underclock a CPU the same wasy as you can overclock one.
It wont give you 25Mhz though.
-
Cheers,
I'm starting to think that the processor might be stuffed.
I have a 68020 @ 25mhz. If I was to put that on the board for a quick test would I fry it?
EDIT: Ignore that, I see that the are not at all compatible now :-)
-
Hi Melange,
I have an A2630 that also will not work, I have replaced the ROMs and no change, the really wierd part is that if I remove the 68000 from the MB the system still works perfectly but @ 7MHZ :-? I am sending the board to Tahoe to test for me, maybe its my MB that has the problem.
The A2630 is fitted with 4MB of RAM but that will not work at all no matter how I configure the system/card :cry:
-
hehe, strange problem you have there.
I wonder where the CPU is getting that 7Mhz signal from... any wires running from inside the shield to the edge connector? What does the crystal say on it? (7Mhz?)
-
HI dbrads,
Wow that's really strange! I have read the roms off my card and unified the file, and found that the bytes were not swapped, as I believe they are supposed to be for a motorola processor. I remade them with the bytes swapped but still no go. I feel very strongly that the processor is dead, but of course it's just a hunch. Looking out for a replacement processor ATM, to see if that's it.
Failing that I can't see what more could be done. The card is absolutley loaded to the hilt with PAL chips, which I think would probably be near impossible to replace now. I could replace caps and the garden variety logic chips, but it could be a whole lot of effort for nothing.
Thanks for your reply, let me know if you get it going, would be interesting to find out what's wrong.
Cheers :-)
-
The PALs are a good starting point. Just remove and reseat them one by one. I once had an A2630 with the same behaviour. I turned out that it was a contact problem with one ore more PAL.
After the described procedure the card ran fine.
-
@Lemmink
Thanks, I will give that a try. I'm very concerned about the ROMS still however, as I believe they were programmed incorrectly. I have to wonder if the version in use on the card is even correct for it. Would love to find a source of binaries so I could try some other versions.
The manual states that the card can take 64K Roms expandable to 128K. The board has 256K Roms on it, which I know to be compatible, however in past experience it's been necessary to repeat the code (or copy it to) the later half of the chip 0x4000 to 0x8000 as well. depending on how the board was designed it may only read the data from the latter half of the chip and ignore the beginning half.
Whomever programmed my ROMS did not do the above, nor did they perform a byteswap before programming the chip
-
ROMs:
http://www.hyakushiki.net/junk/a2630lo.bin
http://www.hyakushiki.net/junk/a2630hi.bin
try holding both mouse buttons when you power it on and see if you get the 68000/68030 screen
-
Wow, thanks everyone so very much. I now have enough to play with.
Still convinced its the processor at the moment, and have ordered a stack of 68030's off ebay. Can't be sure if they will be the correct speed though. Even if they are slower they should be compatible providing I match the speed with the crystal. If they are faster than 25mhz, then I should be able to drive them slower with the current 25mhz crystal. Is that a fair assumption, or have I got it all wrong?
-
melange wrote:
Wow, thanks everyone so very much. I now have enough to play with.
Still convinced its the processor at the moment, and have ordered a stack of 68030's off ebay. Can't be sure if they will be the correct speed though. Even if they are slower they should be compatible providing I match the speed with the crystal. If they are faster than 25mhz, then I should be able to drive them slower with the current 25mhz crystal. Is that a fair assumption, or have I got it all wrong?
You seem to have it right as far as I can tell. However, I really doubt it is the processor that's dead.
I remember back in 1992 I worked at the advertising department of Dumoulin, electronics retailers, and I used to do the newspaper ads with Professional Page on a 2000 equipped with a 2630 that failed one day and had to replace it with an RCS Fusion 40, made here in Montreal BTW.
We gave the 2630 to Software House and they fixed and sold it back. I later learned that the problem was damaged traces inside the PCB caused by the heating of the CPU and that the 2630 often suffered this type of problem.
Commodore probably tried to save money on the PCBs and got low quality ones.
-
@eslapion
If your correct then I have a real task on my hands. Lots of tracing can take a long time to decipher, but the schematics are there so I might see what I can find. I notice that a earth trace has been added between two of the logic chips on the board so perhaps your correct. I will have a look and see if I can find anything obvious. Cheers :-)
At least the processors are cheap so will make good spares for things in the future.
-
Damn - you're out of luck. For _years_ after I owned one (or ever touched an Amiga) I had images of the 2 ROMs for A2630 (as well as the jumper info etc.) on my HD. Gone now though :-/
I inherited a 2630 - put it in my 2000 (or was it 1500 ?) which had KS2.0 ... didn't work - but as I was doing bits of work for a company registered as a C= developer, I persuaded C= UK to send me a pair of ROMs, which a mate then blew for another mate in posession of an old 2630 .
Haaah, happy days :-)
-
Thanks slazzer, I have 2 or 3 different versions to work with now. Tried them all and no difference. I have pulled every socketed chip from the board and reseated it. Tried a host of different things. Now I'm looking for broken traces.