Amiga 2000 - white screen

[Sanxion]

Hi all

I have come into possesion of an early version of Amiga 2000 with a Revision 4 motherboard. The battery has been replaced and it did work before the power supply failed. However, I suspect that happened many years ago.

I have replaced the power supply, but when I turn the machine on, I am greeted with a white screen that seems to change to a grey one?
According to the error code guide, a white screen represents a processor failure.

The motherboard appears to be in good condition with no obvious damage. There are however two strange things:

1. There are a few pins on the 68000 that seem to be bent. The attached picture shows this.
2. There is a wire soldered to connect two pins on the Kickstart ROM - why would this be done?!

What do you guys suggest I do first?

[MigaTech]

The CPU probably needs to be removed and checked properly. Even though you stated the battery was replaced maybe the traces around the 68000 wasn't checked properly. Check also the condition of the CPU dill socket and for continuity on all traces in the area of the Battery.< Specially if there is signs of leakage on the board.

Re fit the CPU once you have checked the area and make sure all pins are straight clean and correctly in contact with the dill socket.
Try the system again, if it fails to boot, try a different 68000.< If you have one.

The ROM wire is a jumper wire which was required for certain revision 2000 motherboards. Detailed information in the link below:

http://www.oldsoftware.com/kickstart.html

[Sanxion]

You are corrent about the ROM. I swapped it for one that does work and now there is nothing on screen.

Thanks. I will have a look at the 68000 next.

[Sanxion]

I had a quick look at the state of the CPU socket and although I have not tested it with a multimeter, it does seem to be in good condition. I have attached a picture. I also tried a different 68000 and it shows the same few white/grey screens before going black.

(Actually, how do you attach an image to a post other than the original one)

[Sanxion]

I swapped the two 8520A-1 ICs as the light grey screen apparently represents a "CIA" error. However, that did not resolve the issue.

[MigaTech]

Did you check power LED to see if it flashes any error codes? Not sure if the 2000 does the same as 500 with caps error codes?

How is the continuity around the Battery area and CPU dill traces?

[Sanxion]

The power LED comes on and stays on, there is no flickering - as does the caps lock light.

Could you ezplain how to check for continuity in the area you mentioned please?

[Daedalus]

Caps lock flashing error codes mean nothing to the computer and concern only the keyboard controller. And the flashing of the power LED has very little relevance other than telling you that a Guru has happened, and that it's ready to connect to a remote debugger. What's important to note is whether the LED starts off dim for a second, then turns bright. With a white screen (that comes after a second or two), I'd expect the LED to behave correctly. as that screen is quite late in the boot sequence.

Ugh, those bent pins don't look too healthy, looks like they've been bent trying to refit the chip - maybe a previous repair attempt? But if they're making contact, fair enough. To check continuity of those traces means using a meter at known good points of the same circuit on each side of the damaged area, one by one, and making sure each trace is still intact electrically. The schematics will help you here, but bear in mind it can be time consuming.

A white screen *can* mean a CPU issue, but at that point in the boot, the CPU has already done a number of things successfully, and so there could well be other issues to blame. For example, issues with Paula or interrupts can cause the machine to pause indefinitely just before the Kickstart screen would be shown. Have you looked at all the custom chip sockets to check them for tarnishing or corrosion? Do you have another machine to swap chips with for testing?

[Sanxion]

Thank-you for the response.
To date I have swapped the following for known working chips:

1. The 68000
2. The two 8520A-1

The exact colour sequence on start-up seems to be:

Light Grey - Dark Grey - Light Grey - Black

The power light seems to be constantly bright - but I cannot check as I am currently at work

I have corrected the bent pin issue on the 68000.

I did read in another post that a user changed the Agnus. Unfortunately, the Agnus in my machine is the original "thin" version and I don't have a spare...It appears - after doing some research - that my 2000 is nothing more than an Amiga 1000 with a few extras.

A part of me is wondering if this is actually worth the effort as it is the first version of the 2000 which does not seem to have a good reputation...but I then remind myself that when I received the machine, it did come with a SCSI card and drive, an 8MB Zorro card, an extra 1MB card, a modem card. The battery has also been replaced with an original Varta barrel, the Kickstart has been "wired" to make it work...so somebody has put considerable effort into this machine...at one point in time...so perhaps I should try to revive it?!

[Daedalus]

Well, whether it's worth it or not is entirely up to you. It's certainly a more rare machine than the later 2000s, so that in itself is a good reason to revive it. And if its limitations mean it's not that useful to you, I'm sure there are plenty of other people who would like it from a collectors' point of view. It is more difficult (though not impossible) to modify that board to use a Fat Agnus for more chip RAM, but on the other hand, it's still quite useable as it is, especially if you still have those expansions around. The 1MB of fast RAM in particular will make a difference, meaning that some tasks could actually be faster than a later, chip RAM-only 2000.

The Kickstart mod appears to have been common enough, and not limited to the A2000 either. Some ROM parts simply needed a couple of address lines swapped to work as the Amiga expects them to.

As for troubleshooting, I would suggest swapping Paula for a known good chip to see if it helps, and maybe the other chips too. But, as suggested, there is always the possibility of damage to traces on the board or the IC sockets themselves.

[Sanxion]

Agreed, I will replace the other ICs.

This may be a strange question but I will ask none-the-less. If I inserted a processor card via the Zorro expansion slots, would the machine attempt to "boot" from that first, or does it only look for an additional processor after performing the complete boot sequence first?

[dalek]

If you have an eprom burner try diagrom: http://www.diagrom.com/

Definitely try and reseat your Agnus chip with a proper chip extractor. Clean the socket contacts with isopropyl alcohol and make sure after reseating the Agnus each pin is connected to the socket (check with a multimeter - painful yes but a flakey socket is a common problem).

Also if you are trying the RGB output, try the composite output as well just to check it's not a display problem.

Failing that grab the schematics and trace every pin from the 68k and make sure they are connected where they need to be. If you have access to an oscilloscope check your crystal is OK - it probably is as you are getting from black to dark gray to white).

I would bet that if you can't boot from the onboard 68k you won't be able to from the CPU slot either, as the CPU slot is electrically connected to the same traces as the onboard 68k (I'm guessing the CPU slot accelerator has to disable CLK or Enable to keep the onboard CPU off the bus? EDIT: did some reading, looks like the expansion needs to put the onboard 68k in a halt state.)

[Daedalus]

Bear in mind that those early A2000s had a DIP Agnus, not the PLCC type of the later A2000s, 500s, 3000s etc.

The CPU handover is a little more sophisticated actually - the 68K series was designed for multi-CPU use, though the Amiga doesn't use that other than for using separate FPUs. Basically the expansion CPU will request the bus from the other and then never release it, so the original CPU idles forever. CPU expansions go in the dedicated CPU slot, not the Zorro slots. And, while the early A2000s have that same mechanical slot, there are some minor differences so it's worth checking if a given CPU card is compatible with both types.

However, a white screen does point towards the CPU actually working and getting part way through the boot procedure before halting. Chucky's DiagROMs are an excellent suggestion! If you don't have an EPROM burner, he'll burn you some for a small fee.

[Sanxion]

I am reading about Diagrom right now.

Please note, I think the last screen before the black appears is light grey and not white. The title of this post is incorrect.

[Daedalus]

Yep, that's still just a stage of Kickstart's boot sequence. Unless there are some other issues going on with the video circuits, those colours are generated by Kickstart, and that means the CPU is running.

[MigaTech]

Originally Posted by Daedalus

Yep, that's still just a stage of Kickstart's boot sequence. Unless there are some other issues going on with the video circuits, those colours are generated by Kickstart, and that means the CPU is running.

This is what I thought, from those colours he is seeing this system has a very good chance of booting again.

@Daedalus, The caps lights you don't think could help? I thought they informed of rom / ram checks and watchdog timer issues? The latter could be a possibility.< What is your take on this? To be fair I wasn't entirely sure if the Amiga 2000 has caps error codes. So your advising here would be more appropriate anyways.
What you stated about the power LED is also correct which is yet another reason to why I asked him to check. I figured whatever he reported back would of told us more information.

@Sanxion, What Daedalus advised you on how to test the continuity is absolutely correct, follow all the instructions he has provided then report your progress. From what you are stating this computer has a very good chance of working again.

As for your whether it is worth it or not statement, should go with saying, it's an AMIGA < course it is worth the trouble to repair. Keep with it, you shall get there in the end.

[Sanxion]

@MigaTech - I spent some time yesrerday checking the continuity of the tracks around the 68000 and they all 'seem' to be fine. I am no expert at this and could well be placing the probes in the wrong position. Any guidance reagrding this would be much appreciated.

I have also ordered a DiagROM from Chucky as that does seem to be the easiest method of detecting the error here...

[dalek]

I didn't read it above but have you tried reseating all the other socketed DIP chips on the board?

[Sanxion]

I have just reseated the processor, kickstart and two CIA ICs.

[dalek]

Definitely try reseating Gary as well - even DIP sockets can suffer corrosion/oxidisation and lose contact after many years.