• Please review our updated Terms and Rules here

Amiga 2000 Rev 4.4 Black Screen

lowlytech

Experienced Member
Joined
May 17, 2018
Messages
103
Location
Oklahoma
I have a post over at English Amiga Board, but thought I would post here and see if I can get any leads on what to do. I have a Rev 4.4 Amiga 2000 board that worked when first tested. Battery leakage was minimal, best cosmetic board i have seen. So I cleaned up the leaked area, scratched away the mask, vinegar'ed up the board to neutralize the alkaline, and replaced the capacitors. When I went to power it back up I got nothing, just a black screen.

I have a Rev 4.3 board that was in worse shape and has bodge wires from torn traces and pads, but it worked, so it was going to be a backup board. Now I have found myself using the Rev 4.3 board to help troubleshoot this 4.4 board. Things that I have ruled out...

All custom chips are good (Plus cpu and BIOS, basically all socketed IC's are good)
Power supply is good
Replaced CPU socket
Replaced Gary socket (seemed it was not seated properly from factory, rasied a bit on one side)
Tried KS 2.04 and 1.3 along with DiagROM 1.2.1
Pulled buster completely, no change with DiagROM 1.2.1 installed
Made sure 5V was at all typical places on the board.
Pulled Agnus and made sure pins were making good contact, bent them out a bit to make sure.
Went over the whole board with magnifying glass on top and bottom looking for any cracks, solder joints, etc.
Checked for Clock and Reset and it is okay. Also counter showed the clock CPU pin 15 to be 7.159MHz
With Amiga Keyboard attached, the cap lock light does not flash, it does respond on and off when you press the caps lock, for about 8-10 cycles before locking up the light, the CTRL+Amiga+Amiga does reboot the system
The power LED is solid bright no pulsing or flashing

So what happens is the CPU goes into Halt on boot up. The only way I have found to keep the CPU from halting is when Gary and Paula are BOTH removed. It will still halt if one is left from the two. I know these two IC's are good, so I mimicked the same setup on the working Rev 4.3 board and took some measurements to compare between the two boards. Everything should have been the exact same, except the bad 4.4 board does not have the 1Meg Chip mod done yet. I attached a PDF with my summary of what I found, and it looks like some logic is in different states in some of the 74 logic chips. I tested a few random IC's, but I am happy to check anything if someone can suggest a specific component we need to focus on.

So far only have used the mono composite out for testing til I can get a commodore RGB monitor hopefully next.

The different shade of green is the UV mask I reapplied over where I scraped off the blackened mask. It looks kind of funny with the light reflecting, but it all is good. Basically the only green gunk and damage was just in that one small area from the best I can tell, so I am not sure if this black screen issue is from the battery damage.

Appreciate any help on what maybe to check moving forward. I have been working on this board for about 2 weeks and kind of hit a wall on what to do to resolve the cpu halt.

Thanks
 

Attachments

  • 20221231_202045.jpg
    20221231_202045.jpg
    6.1 MB · Views: 10
  • a2000 troubleshooting.pdf
    556 KB · Views: 3
Last edited:
I went ahead and did the 1Meg chip ram mod to this board and retested all the points I previously did. Surprisingly only one of the tested IC's changed behavior and it still isn't consistent with what the working board did. IC U106 tested different after the 1 MEG mod. Here is what changed.

Pin 1 now shows High and has bus noise, not really a pulse or repeating beep, but static. I don't know what to call this other than bus noise? The working board doesn't have anything on this pin.
Pins 3-4,7-8,13-14,and 17-18 changed from Low with bus noise to High with Pulse (repeated beeps). This still isn't consistent with the working board which just showed Low with no pulse on these pins.

Also kind of surprising to me is the CPU pin 47 did not change after the 1Meg mod. It still shows high when the good board shows pin 47 low. Between the two boards, pin 47 is the only one that is different.

So now both boards are truly identical other than whatever the difference in Rev 4.3 and 4.4 were.
 
Have you tried to see if you get any output on the serial port while using the diagrom? Might help point you in the right direction. Also, if you have a pistorm you can configure it to use the diagrom, and then watch the serial output through an ssh connection to the pi. I would also suggest hooking it up to via the RGB port so you can see what colour screen you are actually getting. Composite out is only black and white.
 
So I cleaned up the leaked area, scratched away the mask, vinegar'ed up the board to neutralize the alkaline, and replaced the capacitors. When I went to power it back up I got nothing, just a black screen.
I don't even see why you took time to test all the chips, the ROM etc. You worked on the leak-affected area and after that, it would no longer work. So apparently you caused some damage there. Check all the caps you replaced (maybe one is faulty or reversed?), check every single trace in the area you worked on for continuity. Scraping away the mask may have damaged a trace. There is absolutely no sense in trying to find the error somewhere else.
 
Timo,

I checked all the caps first thing, I removed them double checked them in a meter that the values were correct, polarity was correct, no shorts, good connections , etc. I even changed out the HE Series low ESR Nichicons I used originally back for standard ESR thinking this may have caused some strange behavior in a circuit, but didn't make any change. The battery leak area seems totally fine. All traces seem to be connected and working.

Eight088,

I hadn't thought to connect up anything to the serial port. I am very interested in doing this. I don't have any Pi units, but could I use like a null modem cable over to my current WIN10 machine and use PuTTY for this? Only thing I am worried about is I am not sure Diagrom would still output any useful data on the serial port with Paula removed? I know if I have Paula installed the CPU immediately goes into HALT state which I figured wouldn't be good if were trying to process anything on the system.

Late last night I did find a Resistor that was way out of spec. It was R215 and it was suppose to be a 1ohm resistor, but it measured around 93ohms. I pulled it out and confirmed it still read almost 10x what it was suppose to be. Hopeful I replaced it with a new resistor, but the board is still behaving the same after the resistor swap :(

Maybe we have a bad resistor pack? I am gonna keep looking for any discrepancies on passive components to rule that out.
 
Yes, you will need paula installed, a null modem cable and a serial port on the win10 machine. Putty will work fine. From memory you will need to configure putty with 9600,8,n,1.
 
Well I have had a bit more time to poke around with this Amiga 2000 board. I am convinced the issue is somehow related to the address line 19. I know this had to do with the 1 meg chip ram mod, right? I have the mod done and verified that J500 is now open, and J101 is on the side closest to the power. I however have a short to 5 volts on A19 on Gary (pin 35) and Agnus (pin 59)I can confirm the working 4.3 board does not have these same faults to the 5v rail. I am hooking one lead of the multimeter to the 5volt power pin (board powered off of course) and the other lead on Gary Pin 35 and Agnus 59 and I show 0.32 ohms. On the working board these pins show OL or open.

So if I was going to start dissecting this A19 issue do you all think U103 thru U106 is a good place to start?

Other thing I keep running into while comparing the boards is there is a huge difference in resistance in certain points, but alot of the points are the same. I tried to look over the schematics but I get lost before to long and can't narrow down the source. For example hooking up test lead to ground and connecting the other one to U103 in resistance mode I get the following between the two boards...

Code:
U103

Pin Good Board       Bad Board

1     86 Ohm             81 Ohm

2     10K                    10K

3     10K                    10K

4     10K                    10K

5     10K                    10K

6     10K                    10K

7     10K                    10K

8     10K                    10K

9     10K                    10K

10   0.10                    0.10

11   16.8Mega           0.63Mega

12   23.5Mega           0.60Mega

13   31.4Mega           0.72Mega

14   21.4Mega           0.80Mega

15   30.7Mega           0.75Mega

16   27.2Mega           0.69Mega

17   25.3Mega           0.67Mega

18   34.1Mega           0.66Mega

19   OL                      OL

20   86                      82

I am sure I got the Open Line on pin 19 cause I have gary and paula removed during these tests, but if I am not mistaken when I used the logic probe, on pin 19 that is the one I get the nasty bus noise static on that I don't get on the working 4.3 Rev board.

Additionally I see these exact pairs of values, for example 16.8 and 0.63 on other pins of the U104, 105, 106 family with what is suppose to be a 20-30 MegaOhm resistance, being under 1 MegaOhm. As well as on some of the resistor packs. Do you think this lack of resistance could come from one of the 74LS logic chips in the U103-106 section?

For giggles the resistance from 5 volts to ground on both boards is pretty close, 86.4 Ohm on the good board, and 80.3 Ohm on the bad.

I have a few other components I tested and compared that had these same megaohm discrepancy pairs, for instance the empty Paula socket showed these same MegaOhm differences( mainly in pins 1 thru 7, 42 thru 48, and a few more), but I won't post them unless someone thinks they could be useful. I think if we can figure out this MegaOhm issue and the A19 short to 5V that may go a long way to storting this board out.

Thanks and forgive the long post.
 
Last edited:
You have a short somewhere. Look where you did work like replacing sockets where you could have accidentally dropped solder or a trace came loose and is touching another signal. It's hard to find mistakes you made yourself. Have someone else check it out if possible.
 
Well I finally found the smoking gun so to speak. I found that 5V was shorted to pin 5 of U602. I convinced myself that the problem was an issue with 74 logic and I desoldered U602 off the board. Well the short was still on that pad with the IC removed. The trace only went one way toward the CPU socket. I took C404 off again so I could see the traces better and my short went away. The cap tested good at 51uF, infact I reinstalled it and it is working now without issue. Boots DiagROM, and KS 2.04 just fine. I sure didn't see where C404 was shorting or touching anything, but when that was removed, that is when we made a breakthrough. Maybe some solder was lodged in the 86 pin Processor slot? I just am not exactly sure, but it seems to be all good from a point of this board booting up. Thank you so much for everyone and their suggestions. It really helps when you have some great feedback from other users to help work out some of these issues, especially the issues where you have spent multiple days trying to figure something out.
 
Back
Top