Amiga.org
Amiga computer related discussion => Amiga Hardware Issues and discussion => Topic started by: Frankencelery on March 11, 2017, 02:31:05 PM
-
I've got an Amiga 2000 rev 6.x motherboard that gives an instant green screen on power on. There WAS slight battery leakage, but the damage, if any, appears to be superficial.
I know that green screen means chip ram, but I've had no luck so far. I've installed sockets and replaced all the RAM. Replaced Agnus AND the socket (gotta love desoldering an 84-pin PLCC socket) and also the CPU socket, because that's the one that always gets the battery damage. No luck so far.
Couple other items I forgot to mention. I am working with the bare board on my bench, so no peripherals are attached. Also, I don't know if this is important, but the green is instant as soon as power is applied. There are no levels of gray first.
Any other ideas to try? I'm open to suggestions!
-
Suggest you do a search here for "green screen".
You would not get it if CPU was damaged - at least ROM code to do the tests is being run.
However, the tests are pretty quick and slapdash. A bus error preventing chip RAM from being read may also cause a green screen.
Sometimes is a faulty RAM chip, sometimes faulty Gary or (on an A2000) Buster or Ramsey or even just a short on the slots - old bit of solder, that kind of thing. I've found usually is a short or a corroded socket pin, not a real problem with a chip as such.
Also try swapping the 8520s, if one is bad and causing bus error, might resolve the boot and show the fault to be in an 8520 rather than a RAM chip.
If you do end up swapping out DRAMs (not applicable this case), do fit sockets if not fitted already. :) Also check Rev 6 mods are all in place and haven't been undone.
EDIT: Pic shows very tidy board. Is cap to left of CPU shorting on pin? Probably not, just optical illusion.
-
I haven't tried Gary, but I did swap the 8250s as you suggested.
That cap was not shorting. I checked.
RAM sockets are installed. CPU socket was replaced, Agnus socket was replaced, no corroded pins or anything shorted that I've been able to find yet.
I have not checked for Rev 6 mods. I suppose that's something I can google and verify.
-
I would recommend you run John Hertells diagrom and see what it tells you. If you dont have an eprom burner to make one I can help you out for the cost of the eprom.
http://www.diagrom.com/index.php/download/
-
I would recommend you run John Hertells diagrom and see what it tells you. If you dont have an eprom burner to make one I can help you out for the cost of the eprom.
http://www.diagrom.com/index.php/download/
Fascinating! I had no idea such a thing existed. This might give me just the edge I was looking for. Funny you should mention that today. My company just shut down its legacy business, and among the interesting toys I got to bring home just yesterday was a PROM burner. My old one was never capable of doing Amiga sized ROMs, maybe this one will. Now to see if there's any 27C400s around here....
Thanks for the awesome tip. I'll be sure to come back here and post an update when I have one.
-
Ok...progress made. First, the unrelated stuff:
1. Found out my new-to-me eeTools TopMax programmer DOES work, but I had to drag out an old WinXP machine to make it work. 32-bit and Parallel Port only!
2. Found out that I DO indeed have 3 27C400 EPROMS available. 2 of them are bad, but one programmed with the 0.8 DIAGROM successfully!
Now back to the original subject...my ailing Amiga 2000:
3. DIAGROM boots, gives some flashes (green/red), which is consistent with Chip RAM problems. Connected Null Modem cable. Wow, this guy has done some awesome work! The system appears to detect 512K of RAM (should be 1MB), but it fails to scan any of it. Here's what I get, if anyone can help me decipher this:
Checking memory from $00080400 to $00100000 - Press any key/mousebutton to stop
Checking Address: $00000003 | Bit error shows max $FF errors due to space
8|7|6|5|4|3|2|1| 8|7|6|5|4|3|2|1| 8|7|6|5|4|3|2|1| 8|7|6|5|4|3|2|1|
Bit errors: 0000000000000000 0000000000000000 0000000000000000 0000000000000000
Byte errors: 0 0 0 0 Number of errors: 0
Checked memory: 0KB Usable memory: 0KB NONUsable memory: 0KB
--------------------------------------------------------------------------------
Bad Block start at $FFFFFFFF and ends at $00000003 with a size of 0KB
Shadowmemory detected. Scan stopped. You can ignore the last error if any!
-
IIRC original A2000 shipped with 512K "pseudo fast" and 512K chip... so for some reason thinks it isn't quite a Rev 6 motherboard.
Gary would be my prime suspect, but like I said, typically isn't a bad chip, is a short somewhere. Agnus too but you seem to have covered that base already.
I had no idea about DiagROM either. I guess I am used to diagnosing by swapping chips out...
Thank you too Acii, and to Mr Hertell for doing the work with DiagROM.
Also, check links to select 1MB Chip Ram / 512K+512K split.
Memory map here, looks more like first 256K isn't accessable for some reason... maybe I am reading this totally wrong.
http://amigadev.elowar.com/read/ADCD_2.1/Hardware_Manual_guide/node00D3.html
ALso J500 may be set wrong, does not just disable clock, also disables some chip RAM.
http://www.titan.co.nz/amigaak/AA060161.htm
-
John and I are very good friends and talk nearly every day. If you PM me a list of results and questions I will be happy to pass them on and review them myself. I am in Japan this month on a job, but have internet access thats quite good. The time change just gets in the way some is all. For now I have passed on your last post with the serial output.
-
If you look at the address the first 512K of chipmem is not detected.. so I would check the output on serial port to see if there are biterrors.
Since the kickstart actually had a greenscreen it is unable to allocate chipmem at first memoryblock (it does it in blocks of 16k if I remember correctly)
while the test rom searches the complete 2MB area for memory.
Post the serial dump here and I will review it.
-
Hi and thank you for your help! Is this what you needed?
Amiga DiagROM V0.8 - 22-Oct-16 - By John (Chucky/The Gang) Hertell
- Parallel Code $ff - Start of ROM, CPU Seems somewhat alive
Checking status of mousebuttons for different startups:
Set all Interrupt enablebits (INTENA $dff09a) to Disabled: Done
Set all Interrupt requestbits (INTREQ $dff09c) to Disabled: Done
Set all DMA enablebits (DMACON $dff096) to Disabled: Done
- Parallel Code $fe - Start of chipmemdetection
Addr $00000400 Write: $F0F0F0F0 Read: $FFFFFFFF FAILED
Addr $00000400 Write: $0F0F0F0F Read: $FFFFFFFF FAILED
Addr $00000400 Write: $00000000 Read: $FFFFFFFF FAILED
Addr $00008400 Write: $F0F0F0F0 Read: $FFFFFFFF FAILED
Addr $00008400 Write: $0F0F0F0F Read: $FFFFFFFF FAILED
Addr $00008400 Write: $00000000 Read: $FFFFFFFF FAILED
Addr $00010400 Write: $F0F0F0F0 Read: $FFFFFFFF FAILED
Addr $00010400 Write: $0F0F0F0F Read: $FFFFFFFF FAILED
Addr $00010400 Write: $00000000 Read: $FFFFFFFF FAILED
Addr $00018400 Write: $F0F0F0F0 Read: $FFFFFFFF FAILED
Addr $00018400 Write: $0F0F0F0F Read: $FFFFFFFF FAILED
Addr $00018400 Write: $00000000 Read: $FFFFFFFF FAILED
Addr $00020400 Write: $F0F0F0F0 Read: $FFFFFFFF FAILED
Addr $00020400 Write: $0F0F0F0F Read: $FFFFFFFF FAILED
Addr $00020400 Write: $00000000 Read: $FFFFFFFF FAILED
Addr $00028400 Write: $F0F0F0F0 Read: $FFFFFFFF FAILED
Addr $00028400 Write: $0F0F0F0F Read: $FFFFFFFF FAILED
Addr $00028400 Write: $00000000 Read: $FFFFFFFF FAILED
Addr $00030400 Write: $F0F0F0F0 Read: $FFFFFFFF FAILED
Addr $00030400 Write: $0F0F0F0F Read: $FFFFFFFF FAILED
Addr $00030400 Write: $00000000 Read: $FFFFFFFF FAILED
Addr $00038400 Write: $F0F0F0F0 Read: $FFFFFFFF FAILED
Addr $00038400 Write: $0F0F0F0F Read: $FFFFFFFF FAILED
Addr $00038400 Write: $00000000 Read: $FFFFFFFF FAILED
Addr $00040400 Write: $F0F0F0F0 Read: $FFFFFFFF FAILED
Addr $00040400 Write: $0F0F0F0F Read: $FFFFFFFF FAILED
Addr $00040400 Write: $00000000 Read: $FFFFFFFF FAILED
Addr $00048400 Write: $F0F0F0F0 Read: $FFFFFFFF FAILED
Addr $00048400 Write: $0F0F0F0F Read: $FFFFFFFF FAILED
Addr $00048400 Write: $00000000 Read: $FFFFFFFF FAILED
Addr $00050400 Write: $F0F0F0F0 Read: $FFFFFFFF FAILED
Addr $00050400 Write: $0F0F0F0F Read: $FFFFFFFF FAILED
Addr $00050400 Write: $00000000 Read: $FFFFFFFF FAILED
Addr $00058400 Write: $F0F0F0F0 Read: $FFFFFFFF FAILED
Addr $00058400 Write: $0F0F0F0F Read: $FFFFFFFF FAILED
Addr $00058400 Write: $00000000 Read: $FFFFFFFF FAILED
Addr $00060400 Write: $F0F0F0F0 Read: $FFFFFFFF FAILED
Addr $00060400 Write: $0F0F0F0F Read: $FFFFFFFF FAILED
Addr $00060400 Write: $00000000 Read: $FFFFFFFF FAILED
Addr $00068400 Write: $F0F0F0F0 Read: $FFFFFFFF FAILED
Addr $00068400 Write: $0F0F0F0F Read: $FFFFFFFF FAILED
Addr $00068400 Write: $00000000 Read: $FFFFFFFF FAILED
Addr $00070400 Write: $F0F0F0F0 Read: $FFFFFFFF FAILED
Addr $00070400 Write: $0F0F0F0F Read: $FFFFFFFF FAILED
Addr $00070400 Write: $00000000 Read: $FFFFFFFF FAILED
Addr $00078400 Write: $F0F0F0F0 Read: $FFFFFFFF FAILED
Addr $00078400 Write: $0F0F0F0F Read: $FFFFFFFF FAILED
Addr $00078400 Write: $00000000 Read: $FFFFFFFF FAILED
Addr $00100400 OK Number of 32K blocks found: $10
Startaddr: $00080400 Endaddr: $00100000
Using $000EE24E as start of workmem
- Parallel Code $fd - Start of motherboard fastmemdetection
- Parallel Code $fc - Memorydetection done
- Parallel Code $fb - Starting to use detected memory
Detecting if we have a working raster: DETECTED
Detected Chipmem: 512KB
Detected Motherboard Fastmem: 15360KB
Basememory address: $000EE24E
- Parallel Code $fa - Detected memory in use, we now have a stack etc
Set Start of copper (COP1LCH $dff080): Done
Starting Copper (COPJMP1 $dff088): Done
Set all DMA enablebits (DMACON $dff096) to Enabled: Done
Set Beam Conter control register to 32 (PAL) (BEAMCON0 $dff1dc): Done
Set POTGO to all OUTPUT ($FF00) (POTGO $dff034): Done
- Parallel Code $f9 - Starting up screen, text echoed to serialport
Amiga DiagROM V0.8 - 22-Oct-16 - By John (Chucky/The Gang) Hertell
DiagROM V0.8 - 22-Oct-16
-
Here also is the dump from the System Info menu option, which may be what you were wanting in the first place!
Information of this machine:
Dump of all readable Custom Chipset HW Registers:
BLTDDAT ($dff000): $FFFF DMACONR ($dff002): $4380 VPOSR ($dff004): $B000
VHPOSR ($dff006): $BD9D DSKDATR ($dff008): $FFFF JOY0DAT ($dff00a): $1404
POT0DAT ($dff012): $E0E0 POT1DAT ($dff014): $E0E0 POTINP ($dff016): $5500
SERDATR ($dff018): $0B0D DSKBYTR ($dff01a): $9000 INTENAR ($dff01c): $0000
INTREQR ($dff01e): $0028 DENISEID ($dff07c): $FFFF HHPOSR ($dff1dc): $0064
-
I would recommend you run John Hertells diagrom and see what it tells you. If you dont have an eprom burner to make one I can help you out for the cost of the eprom.
http://www.diagrom.com/index.php/download/
Interesting.
Looks like a 2000 needs a 27C4000? http://www.mouser.com/Search/Refine.aspx?Keyword=27C400
Where do you find yours?
-
Interesting.
Looks like a 2000 needs a 27C4000? http://www.mouser.com/Search/Refine.aspx?Keyword=27C400
Where do you find yours?
27c400 and I get them from China off ebay. Here is the ones I use and also the same seller.
http://www.ebay.com/itm/10pcs-IC-AM27C400-150-IC-DIP-40-512K-x-8bit-256K-x-16-bit-COMS-EPROM-AM27C400-/191274549477?hash=item2c88da00e5:g:PoYAAMXQKLdRy-xJ
As for the A2000 serial dump I can see the first 512k is totally dead.. so I'd guess power to memory bufferchips or something. Look around for bad traces near them. Check U451 as the fist thing and I expect its bad too, it handles _WE and _RAS1 and RAS0, CASL och CASU. Thats the first thing I would replace. Its an LSF244 and all memory select signals go through it. Here id data sheet: http://www.mouser.com/ds/2/149/fairchild%20semiconductor_74f244-608438.pdf
-
Yeah, Mouser is the only place that seems to have those 74F244s. I ordered some of those, and the 245s. Gonna be a few days to get those.
I'm glad you mentioned that you were getting good results from an Ebay seller in China. I bought some Commodore CIAs a couple months back and not only were they counterfeit, but they were all defective. Made me nervous about buying from China sellers. I do need some more 27C400s as well.
I didn't want to muddy the waters with too much information, but I thought I might share some more information about this board. I bought it as a spare, a couple years ago and it has never worked. I bought it because it was Rev6 and looked to be in good condition with no battery damage.
-
Hopefully, this is all it is. The fact you can get the diagrom working and are getting a green screen is actully a good thing. It means the KS rom is at least executing its first command.
-
OH!
Acill, you are the MAN!
Tonight, I decided to unsolder U540 and U541 and put in sockets. My replacement chips are supposed to be here on Friday.
Then I looked closer at U541....do you see it?
-
The trace going to Pin 8 was making no connection. This is _RAS0, which I can only assume stands for Row Address Strobe 0. Memory Access...of course!
The pic I posted was AFTER I picked at it with an X-acto knife. There was just the tiniest bit of green fuzz there at the pin that was totally obscuring the broken trace.
Repaired that trace, and....TA DA!!!!
-
Great news!
-
I love that white screen!
-
The trace going to Pin 8 was making no connection. This is _RAS0, which I can only assume stands for Row Address Strobe 0. Memory Access...of course!
The pic I posted was AFTER I picked at it with an X-acto knife. There was just the tiniest bit of green fuzz there at the pin that was totally obscuring the broken trace.
Repaired that trace, and....TA DA!!!!
GREAT JOB!!! Even from Japan, I was able to get you in the right location. Damn, I am good! hahha.. No, but in all seriousness, you did all the work and great job finding that trace and fixing it!
Now get some pure lemon concentrate and kill off all the rest of that leaking battery crap, get some GOOD board cleaner and flood the crap put of it, blow it down and inspect it thoroughly for any more weak traces.
-
I haven't had any time to get it back together yet, but I hope to this weekend.
This all started when I bought another A2000 a couple weeks ago. The seller gave the ever popular "it worked when I used it X years ago, but now I have no way of testing it", and I fell for it because I thought he was a stand up guy. Anyway, it was an early model with a rev 4.4 motherboard, and it had pretty severe battery damage, which he also failed to mention. That one has a yellow screen. But I figured if I was going to get this thing going, it would be a better bet to fix my spare rev 6 board, and put that one in the machine.
So now I might try my hand at fixing the old 4.4 board. If I need help (which I probably will) I will start a new thread.
Thanks again for your help everyone!
-
Interesting.
Looks like a 2000 needs a 27C4000? http://www.mouser.com/Search/Refine.aspx?Keyword=27C400 (http://www.mouser.com/Search/Refine.aspx?Keyword=27C400)
Where do you find yours?
27c400 and I get them from China off ebay. Here is the ones I use and also the same seller.
http://www.ebay.com/itm/10pcs-IC-AM27C400-150-IC-DIP-40-512K-x-8bit-256K-x-16-bit-COMS-EPROM-AM27C400-/191274549477?hash=item2c88da00e5:g:PoYAAMXQKLdRy-xJ (http://www.ebay.com/itm/10pcs-IC-AM27C400-150-IC-DIP-40-512K-x-8bit-256K-x-16-bit-COMS-EPROM-AM27C400-/191274549477?hash=item2c88da00e5:g:PoYAAMXQKLdRy-xJ)
As for the A2000 serial dump I can see the first 512k is totally dead.. so I'd guess power to memory bufferchips or something. Look around for bad traces near them. Check U451 as the fist thing and I expect its bad too, it handles _WE and _RAS1 and RAS0, CASL och CASU. Thats the first thing I would replace. Its an LSF244 and all memory select signals go through it. Here id data sheet: http://www.mouser.com/ds/2/149/fairchild%20semiconductor_74f244-608438.pdf (http://www.mouser.com/ds/2/149/fairchild%20semiconductor_74f244-608438.pdf)
A2000 REV 4.1 here. History: Battery leak, battery removed, was given to me faulty, I repaired lots of old leak damage, tested all socketed chips in a working REV 6.2, changed CPU, ROM and Agnus sockets, still grey screen, yesterday found missing contuinity between CPU pin 64 and RP900 pin 10, repaired that today and got green screen now.
Since there are no no schematics for A2000 REV 4.1 (only REV 6.1 that has different number and type of RAM chips) and https://www.amigapcb.org/ is still missing any A2000 models, I googled Amiga 2000 green screen for ideas and came to this topic.
After reading Acills good tips (BTW Japan jet lag took its toll ;), since it is U541 not U451 and it is 74F244 and not LSF244) I was encouraged by Frankencelerys success on finding U541 pin 8 track cut. After extensive measuring I concluded my U541 tracks are ok. Did find out U540 pin 15 is routed different on REV 4.1 than REV 6.2. Read on about that.
But since U540 is closer to old battery place, it is more likely to get leakage and after measuring last four lines I found it. U540 pin 8 track was cut on its first via ring tracing that track to Agnus pin 46. That place is located battery length on right side of CPU pin 1. Very notorious place for leakage damage with all those non masked vias.
Thank you for both of you for making my repair easier. Will repair that track soon and maybe notice there are still more faults. I just hope I don't have take on RAM chips since on REV 4.1 there are 32 of them compared to better REV 6.2 with only 8 of them.
How is your older REV 4.4 doing now Frankencelery?
Here are my notes to help some A2000 green screeners to start with:
U540 74F244
1 GND
2 DRA0 U101 AGNUS 43
3 RP500 9
4 DRA1 U101 AGNUS 44
5 RP501 2
6 DRA2 U101 AGNUS 45
7 RP501 6
8 DRA3 U101 AGNUS 46
10 GND
11 DRA7 U101 AGNUS 50
12 RP500 7
13 DRA6 U101 AGNUS 49
14 RP500 6
15 DRA5 U101 AGNUS 48
16 RP500 4
17 DRA4 U101 AGNUS 47
18 RP500 2
19 GND
20 +5 U500 21
U541 74F244
1 GND
2 DRA8 U101 AGNUS 51
3 RP502 10
4 _CASL U101 AGNUS 54
5 (SPARE)
6 _CASU U101 AGNUS 55
7 RP502 6
8 _RAS0 U101 AGNUS 57
9 (SPARE)
10 GND
11 GND
12 RP502 2
13 _RAS1 U101 AGNUS 56
14 RP502 4
15 REV 6.2 GND / REV 4.1 _WE U101 AGNUS 21
16 RP501 10
17 _WE U101 AGNUS 21
18 RP501 4
19 GND
20 +5 (FOR EXAMPLE U500 21)