B2000CR Rev. 6.2 - Black Screen of Death

I recently got a hold of a really beefed out Amiga 2000 Rev. 6.2 (Vortex Golden Gate with Cyrix, GVP 030/50, Multivision 2000 and other stuff - for free, long story). It has been sitting in an attick for many years and now ended up in my garage.

First I opened it up and saw a massive battery leakage. Removed the battery, cleaned the board and replaced the CPU socket, RTC chip, keyboard connector and associated ferrites and caps. Also had to replace one 74LS245 busdriver (U604 I believe) and its resistor network. Agnus' socket was severely damaged, so I replaced that one too.

On first power up, all I get is a black screen. Floppy does not click and Power-LED comes up with full brightness right from the start - no half-bright LED in between.

I checked the traces of the CPU: all pins from the CPU socket show continuity to their respective pins on the MMU slot. Also checked continuity from the pins of Agnus' socket to the next via/connection along that signal's path to make sure I did not mess up the soldering. Everything seems to be ok here.

I tried swapping all chips with another A2000 I have sitting around here (but with different problems): no changes at all.

Voltages from (3rd-party) PSU are within specs, CPU is getting its 7 MHz clock (checked with a scope).

Now I am pretty much out of ideas. What keeps this thing from booting? Do you have any ideas or suggestions what I can check to identify the culprit?

[DDNI]

Hello, welcome to EAB.

Strip it back to bare bones and try again.
Did you only visually check traces? Try with a continuity checker.
What voltages are you getting at the PSU?
What video output are you using? Have you tried other outputs?

Thank you for your welcome.

The system is stripped down to its bare bones, a case and add-on cards would be quite hindering when I try to debug the board. At the moment, it is just the board, the PSU, a floppy and the keyboard.

Of course I am using a DMM with continuity tester (Gossen-Metrawatt, a pretty reliable one).

Voltages:
5 V - 5.03 V
12 V - 12.21 V
-5 V - -4.97 V
-12 V - -12,32 V

All of them pretty clean, very little noise or ripple.

Currently, the video output is an RGB to Scart cable. No difference with an A-520 or composite though...

[Loedown]

Those A2000 boards get in a right mess from the battery leakage. Look around the board and I mean really look, I even had battery damage in the top ISA connector of one I did. Remove the ROM sockets and CPU socket, replace CPU socket with SIL machined pins, you won't easily find a 64 pin DIL socket anymore and the machined pins are better anyway. Look for open circuit vias. Remove Fat Agnus and check under that for corrosion too. Push all socketted chips into their sockets with firm force and the board supported on a hard surface. Remove keyboard connector and clean.

If it's still no go write back.

The good thing is that this board was mounted in a tower so most of the battery liquid just dripped down without touching the board (there is a huge rust stain on the tower's base).

I already replaced the CPU socket with a new one and also replaced Agnus' socket. Of course I cleaned the board thoroughly, this is actually not the first A2000 board with battery leakage that I had to deal with. The keyboard connector is brand new, so are the caps (C912, C259-261), the ferrites (FB301-303) and the picofuse F1. All other chips are sitting snug and tight in their sockets.

When I replace sockets, I always use the machined type and not that crappy stuff Commodore originally used. Machined DIL64 sockets are actually still available from a big German electronics supplier but they are not cheap - they go for the equivalent of 2.50 AUD each.

So I guess I should still focus on the area around the CPU/KickStart/Memory? How would the machine react if one of the 245 busdrivers between the MMU-Slot and the Zorro-Slots was out? Black screen or just other stuff?

[Loedown]

Originally Posted by RoterFlieger

When I replace sockets, I always use the machined type and not that crappy stuff Commodore originally used. Machined DIL64 sockets are actually still available from a big German electronics supplier but they are not cheap - they go for the equivalent of 2.50 AUD each.

So I guess I should still focus on the area around the CPU/KickStart/Memory? How would the machine react if one of the 245 busdrivers between the MMU-Slot and the Zorro-Slots was out? Black screen or just other stuff?

$2.50 is a small price to pay in my opinion for a socket like that, considering rarity, Beurklin?

Regarding the F245 it would only have an effect on the Zorro slots unless it's shorted or faulty across one or more bus lines.

Check the resnets for +5v continuity, if one of those is missing voltage then lines from the CPU will be floating. If you have a scope then look for data on the data lines to verify that the CPU is trying to read the ROMS. I still think there's an open circuit somewhere but look for obvious stuff like clocks, floating lines, chips resetting etc. Failing that take a couple of nice high res pictures around CPU / ROMS.

So I poked around with my scope on the address and data lines of the 68 k. There is something happening in the address bus (I can see two adresses before the whole thing stops) but nothing on the data bus. During reset, the voltages are around 1V, afterwards there is one L-H transition on some data lines and that's it. /BERR and /HLT are not asserted.

Is it even right that H level on the data lines is 4 V and not 5 V like on the active low signals of the CPU (/DTACK e.g. sits at 5 V)?

I guess I am still very far from a working board... :-(

[Loedown]

Originally Posted by RoterFlieger

So I poked around with my scope on the address and data lines of the 68 k. There is something happening in the address bus (I can see two adresses before the whole thing stops) but nothing on the data bus. During reset, the voltages are around 1V, afterwards there is one L-H transition on some data lines and that's it. /BERR and /HLT are not asserted.

Is it even right that H level on the data lines is 4 V and not 5 V like on the active low signals of the CPU (/DTACK e.g. sits at 5 V)?

I guess I am still very far from a working board... :-(

The board essentially operates on TTL levels so 4V will be considered a logical high.

If you cannot see any data then U103 / U105, check for OE signal and indeed data on the respective inputs to them.

I am happy to announce that my board is working again.

Your suggestion about checking the resnets lead me onto the right track. After checking their resistance to +5V, I noticed that /VPA had only 4.7 Ohms. The culprit was R1000. Possibly due to the battery damage, that bastard, although absolutely ok upon visual inspection, must have lost most of its resistance so that Gary became unable to pull /VPA down.

For a first test, I cut R1000 off, measured the resistance of /VPA again (2.7k this time) and gave it a go. A few seconds later (with my fingers crossed) I finally heard that beatiful "klick-klick-klick" noise from the floppy. It never sounded better... :-D

[multivac]

Originally Posted by RoterFlieger

I am happy to announce that my board is working again.

Your ...

For a first test, I cut R1000 off, measured the resistance of /VPA again (2.7k this time) and gave it a go. A few seconds later (with my fingers crossed) I finally heard that beatiful "klick-klick-klick" noise from the floppy. It never sounded better... :-D

Well done, another Amiga lives again. Please post how you are getting on with the rest of the rebuild.

[Loedown]

Originally Posted by RoterFlieger

I am happy to announce that my board is working again.

Your suggestion about checking the resnets lead me onto the right track. After checking their resistance to +5V, I noticed that /VPA had only 4.7 Ohms. The culprit was R1000. Possibly due to the battery damage, that bastard, although absolutely ok upon visual inspection, must have lost most of its resistance so that Gary became unable to pull /VPA down.

For a first test, I cut R1000 off, measured the resistance of /VPA again (2.7k this time) and gave it a go. A few seconds later (with my fingers crossed) I finally heard that beatiful "klick-klick-klick" noise from the floppy. It never sounded better... :-D

Battery damage can be subtle, pleased you found the fault.