Amiga.org
Amiga computer related discussion => Amiga Hardware Issues and discussion => Topic started by: selco on July 06, 2005, 01:33:46 AM
-
Hello all,
I have a big problem...
My C=A4000T worked here for ages as webserver with MySql. No hard- or softwareupgrade for a long times. Everything was fine.
C=A4000T, Cyberstorm060, Toccata, Ariadne, Cybervision64
Then it suddenly died without any reason.
When I power on it now it dies not try to boot. The fans and the harddrives start spinning but that is all.
- The Screen is black.
- The power-LED does not change its brightness.
- There is no access to the disk-drive.
- There is no Early-Startup-Menu.
I checked the Powersupply and +-5V and +-12V are provided.
I replaced the CPU-board (Cyberstor060) with an Cyberstorm 060+PPC and even withn the original 68040/25 CPU-Board but same effect. Nothing happens.
I checked the two fuses F160 and F161 on the motherboard and they seem to be OK.
I checked the Voltage on the Kickstart-Roms (+5V on pin 30) and +-5V and +12V on the RGB-Connector and they are present. I removed the Kickstart-Roms and checked theit contents - its OK.
Mechanically the Roms are properly in theire sockets. The same applies for the CPU-board.
I removed all Cards (Ariadne, Cybervison and so on) and all harddrives but nothing changes. The power-LED still does not change its brightness when I turn on Power.
Does anybody have an idea what could have died here? Does anybody have a tip what other tests I could do?
Are there (readable) schematics available somewhere? Is there a service-instruction available on the net?
There must be a way to awake my Amiga again!
I am really unhappy at the momment!
regards selco, http://selco.da.ru
-
checked the battery?
-
@selco
I know it appears to power up the devices, but have you tried changing the power supply? If in fact it is a C= A4000T that would mean the power supply is now 11 years old...a flakey power supply _could_ potentially be causing you grief.
-
Agree with HammerD. The A4000T has an AT power supply so finding one to test with should be easy. Alternatively, an ATX power supply with an adapter. The symptoms you're describing would make me test with other power supply first before going further.
-
Have you looked at the P8/P9 powerconnectors on the motherboard. Are the contacts on the plug slightly dark or did the plastic of the plug even turn brown near the 5V line ? I have the 3rd powersupply in my A4000T now, as the 5V lines to the motherboard regularely burn away over time.
I admit that I have a lot more stuff in my A4kT that drwas power from the poor 200W PSU (CSPPC, CVisionPPC, 4 HDs, 5 Zorrocards etc.) but I don`t let it run 24/7, so there is a good chance that you PSU quit after 11 years while mine did after just 3.
-
What's the state of the Caps lock light? Is it on or off?
I seem to be cursed with A600s that the Caps lock light stays on for a full hour before cold booting... but they always eventually boot.
You could try just leaving it on for a few hours and see what happens....
(incidentally, I take no responsibility if you do this and your house burns down or something. :) )
-
Hmmm... My guess is it's a problem with the proc card. I had the same problem with my 4000T when I first bought it. It came with a bad CSPPC, and the machine would exibit the same exact symptoms.
Here are some ideas:
- Are you sure the other Proc. card is good? I went through three cards until I found one that had a good proc.
- Have you checked to make sure the proc. card is fully seated? Also, check and make sure the card is level, there is a stand-off on the mobo that can interfere with the card seating properly.
- Might also try some contact cleaner/lube on the connector. The best stuff I have found is the Caig Laboratories ProGold G5 (same stuff I use at work on $100,000 broadcast gear) just make sure it all evaporates before you turn in on again.
Good luck...
-
Hello again,
thanx for your tips so far.
Let me summarize the current investigations:
* Connectors for P8 and P9 look perfect.
* Caps Lock is Off. By pressing Shift Lock you can turn On/Off the Caps Lock LED 12 times.
* Currently I have no Zorro-boards and no Harddisks in the system, just the pain nacked A4000T.
* The PowerLed does never flash, just permanant green.
* The CPU-board seats perfectly in the CPU slot.
* The Kickstart-roms seat perfectly in their slots, the roms itself are OK (I calculated the Checksum over the read-out-images)
I took an Voltmeter and an Oscilloscope and measured all values I am knowing:
* All voltages on the external connectors (Printer, serial, Keyboard, Joystick/mouse, Video) are correct.
* All voltages on the Zorro-slots are correct.
* All clocks (that I could find in my books) are present on the Zorroslots (7MHz, 3.5Mhz, 0.7 Mhz)
* Pressing the reset-button on the case is reflected on two pins of the Zorro-slots as it should be.
* Pins for BUSERR and RESET are in passive state on the Zorroslots
* There is NO clock on the external video-connector!
Does anybody know a description/pinout of the CPU-Slot? Can I measure voltages/clocks there?
Does anybody have a detailed description of the A3640, Cyberstorm 060 MK3 or Cyberstorm060/PPC that I can measure anything on the CPU-board?
Does anybody know/has any repair/service/test instructions for the A4000T?
best regards http://selco.da.ru
-
Hi,
@selco
If you find the failure, let me know the cause!
I have an A3000T with the same symptoms. I have already tested everything that isn't soldered on the board (e.g. all custom chips, the processor board resp. the onboard processor, the PSU etc.) without effort. The LED doesn't gets bright :-(
I have also changed the 28.xxx MHz oscillator and checked the RESET signals.
I'm currently disassembling an A3000 motherboard to get all IC's that are soldered on the board and will start to change every single IC on the board (but I have an easier job than you because the A3000T isn't soldered in SMD). My current favorite failure source is one of the chips required to access the chip ram.
I think I have enough to do for the next weeks/months, it is quite hard to solder these multilayer boards, an A500 or A2000 is so much easier to solder...
Noster
-
Have you triedbooting without the harddrives connected ? Dead drives can prevent a computer from getting passed the first self tests.
-
Hi
Interesting that. I have an A4000T here with the exact same problem, which I had put down to processor burnt out. But if you have swapped those over then maybe there is hope for us all. You say no access to the floppy drive.. I always thought that to be strange. Probably cus the machine is not getting past the initial check procedure. What you have is power activity but no computer activity. I put this down to a processor lock out. Maybe its another chip problem and could be the ROM. Dunno.
If all was fine one minute and not the next and the computer hadn`t suffered any trauma ( ie moving and dropping it ) it really points to an electrical failure. The trouble is finding it.
One last thing.... I assume that your other processor cards work .
No help, but hey, your not alone.
scuzz
-
Hello!
Interesting, I as well have an A4000T with a CS060 with the *exact* same problems.
We must find a solution to this!
Can you please guide me how to measure the voltage and clock signals so I can see if those are correct?
I need a good step by step guidance :)
-
I've got A4000T that behaves strangely: it only works when placed horisontaly, maybe yours would work like that too?
-
I've no real experience of dealing with knackered Amigas... but I've been fixing a lot of Sinclair Spectrums of late so I guess it's the same principles.
What I've done on the Spectrum is make my own diagnosis ROM which tells me what works and what doesn't on the motherboard. I don't see why you shouldn't be able to do the same thing on an Amiga.
The first thing I would do is burn a simple ROM with something like (pardon my asm, it's been a while):
START MOVE.L #$0000,D0
LOOP MOVE.L ($DFF180),D0
INC.L D0
JMP LOOP
Now from memory, and it's been a long time so I've probably got the syntax/numbers wrong, but this would just increment the background colour, resulting in multicoloured strips all over the screen - if the system is working. If this didn't happen, I'd know either the display was dead chip or the CPU was dead or not getting a clock signal. From there you can build more features into the ROM to test various parts of the system.
Again, I'm probably way off on the details, but I think the principle should be sound.
-
Please tell me what hardware I need to do this!
-
I imagine any old EPROM and EPROM burner should do the trick. A simple 8K EPROM like a 2764 would work, but you may need to wire it up slightly oddly as the Amiga is expecting a 256KB or 512KB ROM, not an 8K one. I'm not sure what the pin-outs on the Amiga ROM chips are - they might be the same in which case you shouldn't need to rewire too much.
I wouldn't recommend it unless you at least understand basic logic circuits and about chip select lines and stuff... but then you're unlikely to own an EPROM burner if you don't. :)
Maybe someone else has actually done this...? I'm only working from a theory (though I think a good one).
-
great idea, though there is some sort of check already builtin ROM, those coloured screens like green - chip ram etc..
-
My problem is solved!
A friend of mine had a look at it last weekend and found out that the soldering on the pins on the 060 socket had some cracks. After re-soldering them the machine now works flawless! I have had in one for 48h without a crash (heavily modified wb + miami dx + amirc)
-
The first thing I would do is burn a simple ROM with something like (pardon my asm, it's been a while):
START MOVE.L #$0000,D0
LOOP MOVE.L ($DFF180),D0
INC.L D0
JMP LOOP
Now from memory, and it's been a long time so I've probably got the syntax/numbers wrong, but this would just increment the background colour, resulting in multicoloured strips all over the screen - if the system is working.
Too bad this is not enough. When the Amiga system is powered up the colour burst DMA is turned off, thus you won't see anything at all. So to see anything at all, the colour burst DMA must be explicitly turned on first.
Fixed code:
start: lea $dff000,a6
move.w #$7FFF,$9A(a6) ; All ints off!
move.w #$01FF,$96(a6) ; All DMA off!
move.w #$0200,$100(a6) ; Colour burst on!
moveq #0,d0
loop: move.w d0,$180(a6)
addq.w #1,d0
bra.s loop
Further, at this point the ROM appears in place of the chip memory until the ROM overlay bit is turned of from CIA... So in case you must check chip ram, you need to turn off the overlay first.
-
@ZyBeR
whot! this is "weird" seen that you have tryed another 2 CPU board!!!
Have you checked to have the right standoff for your CPU card?
A4000T were bundled with too long standoff . . .no problem with a 3640
but with a CS or CSPPC, this could be a big problem.
There are some useful pictures here in the Gallery.
Thanks to X-ray :-)
http://www.amiga.org/gallery/index.php?n=873