a4000 only unstable when in case

[bvh2k]

Amiga model: 4000D CR
680x0 CPU: 68060 A3660 (with a fan and heatsink)
RAM Configuration: 16mb via motherboard 256 via ZZ
Kickstart, AmigaOS Versions: 47.102, 47.3
ZZ9000 Firmware version: 1.10
Other Zorro cards: FastATA 4000 MK VI
450watt ATX power supply

Hi my amiga 4000 is rock solid when laid out on a test bench with the logic board sitting on top of the cardboard box my laptop came in. It boots every time and is super stable to use.However when installed in its case I get all kinds instability problems from the moment its switch on. The most common problem is I get flashing gray screen (kind of like its boot looping or power cycling) or a just a blank screen.

If it does successfully boot into workbench, the system acts very strange. for example several workbench windows can be opened and moved but suddenly they can no longer be closed dispute the gadget seeming to responding to clicks. Some times programs like iBrowser won't open with an error saying MUI isn't installed, when it it is. A reboot later and it opens fine but then I can't close the message saying its the demo version. but the requester can still be moved and the close button acts like its been clicked. one time when leaving the system idling I've had it randomly crash out of workbench then be stuck at the flashing gray screen.

I've tried powering on the system with just the minimum amount needed to boot it. It seems to get most unstable when the drives (hdd, floppy, cd-rom) are added to the mix. However no given combo of hardware seems to result in a constantly repeatable outcome resulting in a definitive answer.

Given that the system works fine outside the case I wonder if there is a earthing issue with the case? The plastic insulating sheet under the logic board looks intact. There is some ware on the case stand offs and a small ware marks on the logic board from where the screws make contact.There is some ware on the case from from sliding the floppy drive sled in and out. I've tried putting a anti-static bag between the case and the drive sled as a insulating layer but this doesn't seem to have any effect.

Any help in what to troubleshoot next would be very much welcome.

[van_dammesque]

- Shorting as you say.
- Heat? (CPU and/or PSU)
- Cable orientation - that is when in the case the cable may flex into a certain position so that they make intermittent contact.
- Strain on the PCB caused by cable or mounting. This could be causing traces to slightly not conduct. Or expose a dry solder joint somewhere.

Try simply resting the board in the case with plastic for insulation and no screws and the cover removed. Then tighten the screws one by one and after each screw is slightly biting go and test for stability. Hopefully this will pin point at the least the area of the miggy that is faulty.

Good luck.

[A4000Bear]

Given that the instability occurs immediately after powering up, I'll say that overheating is probably not the cause.


I think your first port of call should be the CPU slot connector, these are notorious for bad contacts, especially if the CPU card is not firmly fixed into position. The original nylon standoff clips are usually not enough. I suggest carefully examining the contacts on both connectors to make sure they are not corroded or otherwise damaged, then give them a thorough clean. Replace the nylon standoffs with threaded standoffs of the same size, and attach the board firmly with screws. You mention you have the CR version of the A4000. This has an 030 processor on the motherboard. Try running it without the CPU card (after setting the appropriate jumpers) and see if the problems go away.



Another potential source of problems is the 5V terminal on the power connector. This frequently overheats due to bad contact, which makes the contact even worse. Check the voltage on the 5V supply on the motherboard and compare it to the 5V on an unused drive power connector. If the motherboard voltage is significantly lower, you will need to clean and re tension the 5V pin on the power connector. The 5V supply on the motherboard should be no lower than about 4.8V.


It is possible the motherboard may be slightly flexed when installed in the case, possibly causing the bad contact problems mentioned above.

[bvh2k]

Originally Posted by bvh2k

If it does successfully boot into workbench, the system acts very strange. for example several workbench windows can be opened and moved but suddenly they can no longer be closed dispute the gadget seeming to responding to clicks.

I worked out this seems to be a result of booting with the ZZ9000 and its auto boot firmware 1.10 but only after the second boot when its onboard ram is directed by the amiga. I don't get this problem when the amiga is out side the case and the ZZ is installed.

Originally Posted by A4000Bear

Given that the instability occurs immediately after powering up, I'll say that overheating is probably not the cause.

Agreed, I've always had the lid off the case while testing and the fun running on the cpu.

Originally Posted by A4000Bear

I think your first port of call should be the CPU slot connector, these are notorious for bad contacts, especially if the CPU card is not firmly fixed into position. The original nylon standoff clips are usually not enough. I suggest carefully examining the contacts on both connectors to make sure they are not corroded or otherwise damaged, then give them a thorough clean. Replace the nylon standoffs with threaded standoffs of the same size, and attach the board firmly with screws. You mention you have the CR version of the A4000. This has an 030 processor on the motherboard. Try running it without the CPU card (after setting the appropriate jumpers) and see if the problems go away.

I've just been using 3d printed stand off for the CPU card as it didn't come with anything. I'll have to investigate some other stand offs that are more secure. The copper pins on both the a3660 and the motherboard look fine but it does look like the pins between the motherboard cpu socket could do with a clean. I've attached a photo. I'll try and clean it over the weekend.

Using the onboard 030 doesn't not seem to have any impact as I still get the same out come.

Originally Posted by A4000Bear

Another potential source of problems is the 5V terminal on the power connector. This frequently overheats due to bad contact, which makes the contact even worse. Check the voltage on the 5V supply on the motherboard and compare it to the 5V on an unused drive power connector. If the motherboard voltage is significantly lower, you will need to clean and re tension the 5V pin on the power connector. The 5V supply on the motherboard should be no lower than about 4.8V.

the 5v on a Molex was 5.2v and the motherboard power socket is spot on 5v.

I've attached some photos of the worst stand off and ware marks on the case a long with the mounting points on the motherbord. One can clearly see copper showing. Should I cover this with some solder to cover the exposed copper? I can also some rust on the edge of the metal around the expiation slot window and vent holes.

[r.cade]

I'm going to take a wild guess and say your case has become magnetized and may need demagnetized.

Either that or there is some component on the board that has failed in a way that it is more susceptible to being close to a strong ground and is malfunctioning. That would be a difficult manual process to find. It happens.

EDIT: If you have already found it's only with the ZZ9000 card installed, then you've found your culprit! It could be added stress of mounting it in the card cage is warping the card or touching or too close to your house ground and doesn't like that.

[grelbfarlk]

I'm going to guess you just have a cold solder joint somewhere. One time I had a randomly cracked SMT resistor on the bottom of the board which was impossible to spot visually.
How about some clear hi-res photos of the motherboard in full, say 4 pictures of the 4 quadrants.
Are you using the same power supply when it is outside and inside the case?

[bvh2k]

Originally Posted by r.cade

EDIT: If you have already found it's only with the ZZ9000 card installed, then you've found your culprit! It could be added stress of mounting it in the card cage is warping the card or touching or too close to your house ground and doesn't like that.

Form memory the other issues still remain even when the ZZ9000 isn't install but the amiga is in the case. But I'll do more testing over the weekend to confirm this.

[grelbfarlk]

I know it's just an illusion but these pins really don't look like they're soldered down. I'm assuming when you are testing it outside of the case the busboard isn't installed?

[Damion]

Wow! So much awesomeness in this thread (from some truly awesome and talented Amiga gurus!).

Not much I could possibly add to all of this excellent advice -- only to mention the board flexing somewhat as it's tightened down (case or board slightly warped), exposing a weakness elsewhere (CPU slot, poor solder contact) sounds probable. I always shim these with flat washers if needed, as you definitely want the board mounted stress-free. Also, regarding checking the PSU plug, and also flip the board and inspect the solder joints for that connector.

[bvh2k]

Originally Posted by grelbfarlk

How about some clear hi-res photos of the motherboard in full, say 4 pictures of the 4 quadrants.
Are you using the same power supply when it is outside and inside the case?

I'm trying but I'm finding it hard to photograph with out getting a lot of shine off the board from lights. keep you posted now that its my day off.

Yes same ATX 450watt power supply used in both cases.

[bvh2k]

Originally Posted by grelbfarlk

I know it's just an illusion but these pins really don't look like they're soldered down. I'm assuming when you are testing it outside of the case the busboard isn't installed?

Your right it is an illusion

The zorro daughterboard was installed when the amiga was outside of the case. I used a PS1 memory card to provide support for the zz9000.

I can have the system fully kited out (floppy, gotek, cdrom, zip drive, zz9000, a3660 and fast ata) and its rock solid just sitting on the desk.

Also I just cleaned the board and its shocking how it can look fine then you take a macro and you can see the Cotton Swab left little bit of cotton on every thing

[idrougge]

I would also guess that it’s a cold solder joint or a microscopic rupture in the card that only shows up when the card is flexed from being screwed into the case.

Finding out such ruptures may well be impossible, but you could try to re-heat solder joints throughout the board.