Black screen A2000 (rev 4/german)

Het guys,

I’ve just begun unearthing my old Amiga collection that’s been sittning in storage for the last +20 years. And I guess some of it Will need minor/major overhauling.

I’ll start with trying to get the A2000 up and running and move on to the other hardware later

One of the machines is an old A2000 (rev 4) with the CPU-slot 1Mb ram expansion and two floppies, original kbd and mouse.

It also has some sort of joystick/mouse split adapter with a switch and a 1084 RGB->SCART (seeming to lack some color signals, the image is quite desaturated, but that’s for another thread).

While demoing a game of Rodland for the kids I flipped the switch on the joystick/mouse-adapter to enable the second joystick on the mouseport and... BOOM. Sceen turned black

I kinda forgot, after +20 years, that the splitbox wasn’t supposed to be used hot swap-style.

So, I suppose I’ve fried something while flipping the switch. The machine starts, power led is on solid, caps lock works on kbd, ctrl+A+A seems to reset the machine, but floppies aren’t scanned and the screen stays black.

What to do?! I can’t stand to bin the system, I know it’s the sucky german motherboard, but the machine is in excellent condition and I was either going to ATX/Vampire it or sell it, but neither will happen when it’s in this state.

Help a guy out here?

Best regards
Christoffer, Sweden

[MigaTech]

This A2K can be fixed, so please do not bin it! Black screen usually means NO CPU detected. If the keyboard is responding this is a good sign. Check the CAPS LOCK LED is that blinking when powering up? If so check the blinks against the following:

When caps shows 1 Blink this could be ROM checksum failure. This has to be continuous as the system LED does blink once on first power up, but then stays off. If it continues to blink once, then ROM checksum has probably failed.
2 Blinks usually means RAM test failed.
Watchdog timer failure is down to 3 blinks.
Finally, 4 Blinks could be a short exists.

Although the above is usually associated with the A500, the A2K is similar.

Switching the switch on that joystick adapter could be like connecting and/or disconnecting a device from the rear ports, when power is ON. When doing so with the DE9 ports, you can fry the CIA ICs. Although this would display a light green screen rather than black.

(Power should always be OFF when connecting and/or disconnecting any device to the rear, of any Amiga computer.)

Did you removed the evil battery off that motherboard? This can also cause issues specially if the unit has been stored for long periods of time, without power. Check for leakage around that area, it can reach the 68K and cause black screen error code.

Try disconnecting everything and start with just the bare motherboard then work from there. It is a process of elimination, sourcing the offending item and replacing if necessary.

[Daedalus]

Originally Posted by MigaTech

This A2K can be fixed, so please do not bin it! Black screen usually means NO CPU detected.

A CPU isn't "detected" as such in Amiga systems. A non-responsive CPU or a lack of CPU will of course leave the screen in its default black state, but there are many other, more likely, causes that should be ruled out first.

When caps shows 1 Blink this could be ROM checksum failure. This has to be continuous as the system LED does blink once on first power up, but then stays off. If it continues to blink once, then ROM checksum has probably failed.
2 Blinks usually means RAM test failed.
Watchdog timer failure is down to 3 blinks.
Finally, 4 Blinks could be a short exists.

Although the above is usually associated with the A500, the A2K is similar.

I'm sure I've pointed this out before, but anyway, the caps lock LED blink patterns mean precisely nothing about the health of the machine itself. They're purely a result of the keyboard's internal controller self-checks (on any Amiga model), and those failures won't cause a machine to not boot, nor will if offer any help in the case of a non-booting machine. What is of far greater value is the state of the machine's power LED. Does it start dim, then get bright? If it does get bright, then repeatedly pressing the caps lock key will tell you a little bit more - does it keep turning on and off every time, or does it stop responding after maybe 10 or 20 times? If it stops, it shows the keyboard buffer is filling up and not being responded to by the machine.

Switching the switch on that joystick adapter could be like connecting and/or disconnecting a device from the rear ports, when power is ON. When doing so with the DE9 ports, you can fry the CIA ICs. Although this would display a light green screen rather than black.

CIA failures will cause all sorts of issues, including none at all during startup. an Amiga will happily get to the Kickstart screen even with a CIA missing. A lack of a light green screen (which isn't a formal error colour anyway) tells us nothing about the state of the CIAs. However, CIAs have limited involvement in the mouse/joystick ports, and it would be very unlikely to take an entire machine down like this. Typically, CIA I/O failures affecting the ports will stop the relevant peripheral from working, but the rest of the machine to carry on fine.

(Power should always be OFF when connecting and/or disconnecting any device to the rear, of any Amiga computer.)

Indeed that is the standard instruction given, even though the Amiga 2000's controller ports are at the front Nevertheless, switchboxes, and mice and joysticks in general don't cause any problems when hot-swapped. It's devices that are also connected elsewhere (printers, modems, monitors...) that are the real concern.

Did you removed the evil battery off that motherboard? This can also cause issues specially if the unit has been stored for long periods of time, without power. Check for leakage around that area, it can reach the 68K and cause black screen error code.

Yes, this is far more likely to be a concern, though it doesn't really explain the sudden failure on switching the mouse and joystick.

Try disconnecting everything and start with just the bare motherboard then work from there. It is a process of elimination, sourcing the offending item and replacing if necessary.

Indeed, removing any cards, switchboxes and other peripherals is a good place to start.

I would also investigate the power supply. I'm not sure if the original A2000 has the same joystick protecting fuses that the B2000 has, so it's possible that the switch, through some fault, has shorted out the 5V rail and left you with an out of spec or unstable power supply. It could still be causing a partial short if it's still connected, dragging the 5V rail down too low for the machine to start properly.

The machine starts, power led is on solid, caps lock works on kbd, ctrl+A+A seems to reset the machine, but floppies aren’t scanned and the screen stays black.

So when you say the LED is on solid, you mean it stays dim? Does it change brightness at all during a reset? How do you know that the machine is resetting?

[MigaTech]

Originally Posted by Daedalus

I'm sure I've pointed this out before, but anyway, the caps lock LED blink patterns mean precisely nothing about the health of the machine itself.

Yes you have indeed pointed this out before Daedalus, my bad, sorry. I just get ahead of myself sometimes. I am so used to working with early A500s and the later A1200s, that I can get mixed up with the actions of both machines.

Thank you for your adding a very detailed break down of my assistance to the OP. I really appreciate it, as I would not like to offer incorrect information that may lead to more issues, for the OP. Although you was quite fair with me in respect to certain information offered.

Originally Posted by Daedalus

Typically, CIA I/O failures affecting the ports will stop the relevant peripheral from working, but the rest of the machine to carry on fine.

Does this apply to both or just one, at any give time of failure?

[Daedalus]

Originally Posted by MigaTech

Does this apply to both or just one, at any give time of failure?

To both. The I/O pins of the CIA chips are what take the abuse from the ports. If the I/O driver is damaged, that input or output won't physically work, but the rest of the chip will be fine. This is often what happens when a serial or parallel port stops working - the damaged pin simply behaves as if it's constantly at either 1 or 0. In this case, the only function of the controller ports connected to a CIA is pin 6, which is fire 1 or left mouse button. It's very unlikely that anything at the joystick or mouse ports could do damage to a CIA without having external voltages involved, and even more unlikely to take out the whole chip.

However, on the off chance that there was a total failure of the CIA, swapping the two CIAs would allow the system to boot. It's worth trying anyway.

[MigaTech]

Originally Posted by Daedalus

However, on the off chance that there was a total failure of the CIA, swapping the two CIAs would allow the system to boot. It's worth trying anyway.

This was the reason why I mentioned the CIAs as a possibility.

Originally Posted by Daedalus

To both. The I/O pins of the CIA chips are what take the abuse from the ports. If the I/O driver is damaged, that input or output won't physically work, but the rest of the chip will be fine. This is often what happens when a serial or parallel port stops working - the damaged pin simply behaves as if it's constantly at either 1 or 0. In this case, the only function of the controller ports connected to a CIA is pin 6, which is fire 1 or left mouse button

Although I have got to admit, I didn't know about this! Thanks Daedalus!

Wow! This is amazing feedback and it makes me feel Very enthusiastic! I’m even starting to recall some of my lost Amiga-skills just reading your posts.

I’ll try pulling all extra hardware and evaluate the power led. AFAIK the led did dim/brighten, that’s how I came to the conclusion that the machine could intercept a ctrl+A+A.

I looked at the battery, and indeed it’s an original, but encased Varta without and apearent signs of leakage(?!) - but I’ll drill down deeper and remove that badboy. Surely I could find suitable CR2032 replacements somewhere? (My A4000 is also in need of a battery swap, it’s begun to look just a bit ”hairy”, but not overrun or spread).

The PSU is marked with Amiga 2500(?), P/N 380708-01, 220w and is an absolute Beast. I’ll check the fuse, sounds like a likely culprit due to a potential short.

I’d like to replace it with a silent ATX. Has anyone tried this?

Stable power is always nice and a silent psu would be nice on the ears.

[Daedalus]

If the LED is starting dim and going bright, then that's something. It does also mean that the reset sequence is completing, which in turn means that the PSU is at least mostly working.

There are battery replacements available that use CR2032 cells, such as this one. If it's encased and hasn't leaked, that's probably not the root cause however. Still, might be worth changing (and the one in the A4000 definitely should be changed as a matter of urgency).

That is indeed a noisy beast of a PSU, and would be ready for refurbishment even if everything was running fine. I'm not sure if a full ATX supply would fit, and even if it did, you might have trouble meeting the minimum load requirements. I would check out a small form factor ATX PSU instead, these will be smaller, quieter and better suited to running low loads like an Amiga. Ian Stedman has some useful ATX adaptors available that might make life a bit easier if you're going down that route.

I’ll have a look at the battery situation asap, especially on the 4000. I have a feeling that machine it’s about to deserve it’s own thread soon enough.

Any way, status update for 2000.

I’ve pulled the RAM expansion, inspected the MB and the PSU - and nothing at least looks out of place, loose or damaged. No bulky caps either, from what I can see. Can’t find a dedicated 5v fuse in the PSU, is there even one? The mains fuse is ofc still intact.

I double checked the LED, and I was mistaken. It doesn’t brighten.

It’s the Caps-key that blinks on ctrl+A+A. And sure enough, after 20-30 keystrokes the buffer fills and the caps key becomes irresponsive. It does however seem to clear the keyboard buffer when ctrl+A+A. Is that a keyboard thing, or does that mean that at least part of the reset cycle runs?

I also noted that the PSU makes a sort of ”electronic / arc-like” noise, apart from the buzzing fan that is. I can’t recall what a fresh PSU would sound like anyway, but would an ATX-adapter and an alternative PSU be a way forward? Is there some intuitive way to test the PSU?

I have a bunch of ”modern” ATX PSU’s that I could perhaps load down with, harddrives, would that be enough for testing purposes? Or does the minimum load have to be drawn from the ATX-connector?

Eliminating a possibly damaged PSU just seems like a reasonable next step. Removing/swapping the CIA’s also seem, perhaps, easy enough. I’m a bit reluctant to start bending and removing those old chips

[MigaTech]

Before you remove any of those ICs try reseating them. Over time they work loose and this can cause sudden failures and/or non booting systems. Just push down evenly on each chip on the motherboard and then try the system again. It might just be the problem, certainly worth a try.

@Daedalus, Your thoughts on the following please.

Black screen code can also mean unspecified ROM error often a result of CIA error, if I am not mistaken. Also if the power LED is dim does that refer to a possible Buster error? The OP stated that the CAP LOCK LED can be depressed more than 10 times, which indicates the CPU is servicing the CIA interrupt requests.

Do you think changing setting on J300 might have any effect on this system? He would have to change it when using a modern ATX anyways, right?

I also thought that maybe the OP could try a different video output, although if there are no drive clicks, would it be worth bothering?

[Daedalus]

Originally Posted by voltagectrl

I’ve pulled the RAM expansion, inspected the MB and the PSU - and nothing at least looks out of place, loose or damaged. No bulky caps either, from what I can see. Can’t find a dedicated 5v fuse in the PSU, is there even one? The mains fuse is ofc still intact.

In general, no, these PSUs don't have output fuses. The B2000 has a fuse on the motherboard that explicitly protects the joystick ports, but I'm not sure about the A2000. Anyway, failure of that fuse won't stop the machine from booting.

I double checked the LED, and I was mistaken. It doesn’t brighten.

It’s the Caps-key that blinks on ctrl+A+A. And sure enough, after 20-30 keystrokes the buffer fills and the caps key becomes irresponsive.

Hmmmm, that tells us that the CPU probably isn't running, so quite a fundamental issue.

It does however seem to clear the keyboard buffer when ctrl+A+A. Is that a keyboard thing, or does that mean that at least part of the reset cycle runs?

That's just the keyboard's internal circuitry doing all that - the motherboard isn't completing its reset sequence for whatever reason.

I also noted that the PSU makes a sort of ”electronic / arc-like” noise, apart from the buzzing fan that is. I can’t recall what a fresh PSU would sound like anyway,

There shouldn't be any arcing noise, so that's a possibly issue right there. Do you have access to a multimeter?

but would an ATX-adapter and an alternative PSU be a way forward? Is there some intuitive way to test the PSU?

Yeah, a small ATX-style PSU might be a good option - full-sized ones will struggle to fit, but slimline ones will provide more than enough power for any Amiga, and will be quiet and efficient. Really, the best way of testing the PSU is with an oscilloscope, but a multimeter is also quite useful for checking the output voltages.

I have a bunch of ”modern” ATX PSU’s that I could perhaps load down with, harddrives, would that be enough for testing purposes? Or does the minimum load have to be drawn from the ATX-connector?

Usually a minimum load is required, yes. This can be in the form of a couple of drives, or a large wattage resistor. PSUs with a lower power rating will, generally speaking, have a lower minimum load requirement, so that's another reason to look for a small form factor supply of maybe 150-200W.

Eliminating a possibly damaged PSU just seems like a reasonable next step. Removing/swapping the CIA’s also seem, perhaps, easy enough. I’m a bit reluctant to start bending and removing those old chips

Indeed, it's worth trying, just in case. One CIA is responsible for the ROM overlay during boot on the older-style Amigas (since your A2000 is based on the A1000 I would guess that that's the case on your machine). But if you have a multimeter, check the 5V output of the PSU first. If you don't... get one They're so very useful for anyone with a retro computing hobby, and aren't very expensive these days.

Originally Posted by MigaTech

Black screen code can also mean unspecified ROM error often a result of CIA error, if I am not mistaken.

Yep, but really a black screen isn't a code for anything. It just means no other colour has been applied. That can be because of any number of things, including:

- Bad/no CPU
- Bad/no ROM
- Bad/no Buster
- Bad/no Denise
- Bad/no Gary
- Bad/no Video Hybrid
- Bad/no bus buffer chips
- Bad/no connections to any of these (sockets, vias etc.)

Also if the power LED is dim does that refer to a possible Buster error?

Possibly, along with anything else from the list above, and more.

The OP stated that the CAP LOCK LED can be depressed more than 10 times, which indicates the CPU is servicing the CIA interrupt requests.

Well, it is still halting after 20-30 keypresses. If the interrupts are being serviced, it won't halt. If they're not, it will halt when the buffer's full. I'm not sure exactly how many keypresses the buffer can hold, but bear in mind that the caps lock key is special in that it has a toggle action and therefore only produces one keypress for every two mechanical actions. This might be the difference (20 keypresses = 10 characters), or the 10 characters test that is often quoted could just be an estimate, I don't know. anyway, the interrupts haven't even been set up at that stage since the power LED isn't coming up.

Do you think changing setting on J300 might have any effect on this system? He would have to change it when using a modern ATX anyways, right?

I'm not sure what J300 is, or if it's even present on the German A2000s. Do you mean the tick generation? If so, that is a possibility.

I also thought that maybe the OP could try a different video output, although if there are no drive clicks, would it be worth bothering?

Nope, if the power LED isn't coming up, the startup is halting way before any video hardware is initialised.

[MigaTech]

Originally Posted by Daedalus

Yep, but really a black screen isn't a code for anything. It just means no other colour has been applied.

Funny you should state this, as I have never thought about the Black referring as a non official Amiga error colour code. From now on I am going to adopt this reference!

Originally Posted by Daedalus

Well, it is still halting after 20-30 keypresses. If the interrupts are being serviced, it won't halt. If they're not, it will halt when the buffer's full. I'm not sure exactly how many keypresses the buffer can hold, but bear in mind that the caps lock key is special in that it has a toggle action and therefore only produces one keypress for every two mechanical actions. This might be the difference (20 keypresses = 10 characters), or the 10 characters test that is often quoted could just be an estimate, I don't know. anyway, the interrupts haven't even been set up at that stage since the power LED isn't coming up.

Yet again you offer another non obvious yet obvious when stated, point! I wonder how many other people have not thought about caps lock toggle action, for key press calculations?

Originally Posted by Daedalus

I'm not sure what J300 is, or if it's even present on the German A2000s. Do you mean the tick generation? If so, that is a possibility.

Yes J300 is Tick generation and I am sure OP has to change setting for use with ATX PSU. Also PIN 14 on the current PSU is one to check.

Originally Posted by Daedalus

Nope, if the power LED isn't coming up, the startup is halting way before any video hardware is initialised.

Thought as much.

What about resistors do you think any of them might be down?

[Daedalus]

Originally Posted by MigaTech

Yes J300 is Tick generation and I am sure OP has to change setting for use with ATX PSU. Also PIN 14 on the current PSU is one to check.

Yep, though without some sort of logic probe or oscilloscope to hand, it's very difficult to check a tick signal - it's too fast for a multimeter to register. Setting it to internal should eliminate any issues with the PSU's own tick signal, but again, I wonder about the presence of the jumper on the A2000, which is a totally different design to the much more common B2000. I also have a feeling that there should at least be some boot attempt without it - power LED should come up, since the signal has to be checked by the system to see if it's valid. I don't know the ins and outs of it though, so that's just speculation.

What about resistors do you think any of them might be down?

Yep, there could be some issue with pull-ups, though typically that would introduce bus errors and give green or yellow screens. I can't imagine so many resistors all failing simultaneously to completely disable the bus in one go. That's way down the list of things to look at in my mind anyway. My approach would be:

- Check the PSU
- Check the reset lines
- Check the clocks
And if all of them are ok
- Check the ROM overlay signal from the CIA.

Ok, I’ve got multimeters and osciloscopes (albeight, also vintage electronics) and can manage reasonable soldering if required.

So, let’s get cracking here...

I’ll rip out the PSU tonight and try to figure out if the voltages are stable. If I can get my scope up and running I’ll also evaluate pin 14/TICK.

As far as replacement PSU’s go, It would seem that the Rev A / 4.0 motherboard lacks the J300-switch. but the adapters from http://www.ianstedman.co.uk/Amiga/de...x_adaptor.html feature a TICK-generator, so that should solve the TICK-issue.

I’ll also reseat/nudge all chips.

How would I go about checking the reset lines?

Thanks for all the invaluable help!
// christoffer

[Daedalus]

Ah, having a scope makes things much more interesting! The reset lines are active low, and are held low briefly as the machine starts before going high. This should let the CPU run, but if the reset line is held low, the CPU cannot start. That can be caused by a failure in the reset circuitry itself, or in one of the chips controlling the reset lines (there are typically 2 or 3 separate reset lines in an Amiga that feed into each other: CPU reset (goes to CPU and custom chips), keyboard reset (comes from keyboard controller and is buffered to the CPU reset), and peripheral reset (also buffered from the KBReset line). I don't have the German A2000 schematics so I can't tell you all that much about the specifics, or even if it uses the same scheme, but most Amigas have reset circuits that are based on a delay timer on boot, and some also use a voltage detection circuit to hold the reset line low when the power supply is out of spec. This makes sure the PSU is stable before letting the machine boot, so PSU regulation or noise issues could cause the reset line to be held low indefinitely.

The Tick signal should be a 50Hz square wave (assuming European 50Hz mains power), and of course if the scope is up to it, the Amiga's clock signals could also be checked to ensure they're present. The B2000 schematics can be used as a vague guide regarding pinouts for the CPU and other chips, so I would have a look at those. Just don't depend too heavily on them as there are significant differences from your machine.

[Jope]

You can just add a wire between the TICK pin on the power connector (or Pin19 of the CIA that does the printer port) and the VSYNC pin of the Agnus.

edit: looking at the schematics and the motherboard pic on BBOAH, the A2000A has a tick jumper as well. It is J34. You should cut the trace that connects pins 1 and 2 of this jumper, solder a piece of pin strip on there and place your short circuit block on pins 2-3 to get vsync tick.

Originally Posted by Jope

You can just add a wire between the TICK pin on the power connector (or Pin19 of the CIA that does the printer port) and the VSYNC pin of the Agnus.

edit: looking at the schematics and the motherboard pic on BBOAH, the A2000A has a tick jumper as well. It is J34. You should cut the trace that connects pins 1 and 2 of this jumper, solder a piece of pin strip on there and place your short circuit block on pins 2-3 to get vsync tick.

That’s very usable information and the first time I’ve heard anyone mention a TICK jumper/solution specific to the A2000A. I might go down the ian-atx-adapter route that includes a tick generator and a picoPSU as a final solution, but for diagnostics this would be a good solution.

I haven’t had the time to fetch and hook up my scope from the music studio yet, and I’m a bit afraid that it might be suffering from similar issues as the Amiga (old caps and whatnot). Short OT: I’m considering getting a small and portable scope for these kind of simple analytics. Do you have an opinion on for example the OpenScope?

I have however measured the PSU connector with the following results:

Pin 1: 5,2V (yellow)
Pin 2: 5,2V (yellow)
Pin 3: 5,2V (yellow)
Pin 4: 5,2V (yellow)
Pin 5: Ground (blue)
Pin 6: Ground (blue)
Pin 7: Ground (blue)
Pin 8: Ground (blue)
Pin 9: 12,54V (orange)
Pin 10: -
Pin 11: -11,96V (red)
Pin 12: 11,27V (green) - isn’t this supposed to be +5V ?
Pin 13: -4,95V (white)
Pin 14: TICK unconfirmed at this time

All measurements were done without any load on the PSU. I’ve understod that this might be an issue? If so I’ll hook up some hardware and redo the test.

From what I can see, the negative pins seem to be a bit lacking, but perhaps within reasonable margins?

That +5V USER pin though seems to be way out of spec at 11,27V - what is even a ”5V USER” line and how does it differ from the regular 5V lines? I can’t seem to find any information on it.

Could I perhaps just cut it and use power from one of the other 5V lines, or from a separate power source just to confirm the culprit?

The search continues!

[MigaTech]

Originally Posted by voltagectrl

That +5V USER pin though seems to be way out of spec at 11,27V - what is even a ”5V USER” line and how does it differ from the regular 5V lines? I can’t seem to find any information on it.

Could I perhaps just cut it and use power from one of the other 5V lines, or from a separate power source just to confirm the culprit?

Pin 12 is plus 5v and is used to supply the joystick and mouse ports. You could use another 5v supply feed, only make sure ratings are similar to Pin 12.