Dead Amiga 4000 motherboard; what next

[PulsatingQuasar]

I bought a dead Amiga 4000 motherboard to fix. It was missing the ramsey so I soldered that back on.

If I remove the chip RAM I get a green screen. So I know the processor board is working and the ROM is also doing it's thing. If I stick the RAM back in the screen stays black. Well, not completely. If I turn the brightness and contrast way up on my monitor I can see there is a signal. I also get the white short line on the screen from powering up the 68040 so I know the CPU board is working.

No flashing lights and I can use Ctrl-Amiga-Amiga to which the Amiga responds with the power light going dim and coming back up. No diskdrive click.

The Buster, Gary and both CIA's stay cold.

The freaky thing is that I got it working once after spraying contact cleaner on the SIMM, ROM socket, power connector and CPU connector and after applying some pressure on the CPU board connection. But I haven't gotten it to work since then. So I think this might just have been a freak occurence.

How can I meassure either of the CIA chips is causing the Amiga not to boot?

It's either that or there is a PCB problem somewhere on the mainbord. There is no battery damage.

It's really odd that it worked once.

[x303]

If I recall right: 1 cia chip is for the diskdrive and the other one for the keyboard. Why not switch them over ?

Or take a look here.

[color=ff0000]x303[/color] :-D  :-D  :-D

[PulsatingQuasar]

Because they are soldered down.

There must some kind of measurement I can do to find out if it's the cause of the not booting.

[arnljot]

I assume that you have played with the jumpers, and that was not in play when it freak booted.

I'm thinking of the INT/EXT jumper for the CPU Clock source. It comes in play between 030 and 040/060 cpu cards.

I do not suspect the CIA chips since you have had it boot once and it was through making sure things are propperly seated and all contacts were clean. And like you say, the CIAs are soldered on.

So reconsider and find a new gut feeling. Which chips seemed to benefit most from your tlc.

Also, anything that can be taken out of the card should also be verified on a working motherboard.

Look over all caps also, are they dry or busted?

And finally, if the CIAs are busted. Look for two 22pin sockes, find your solder iron and find two new CIA agents and socket them :-)

Sounds like the card can be saved since you've had it alive once

[PulsatingQuasar]

The jumpers are in the correct location. But I also tested with other 040 and 030 boards.

All capacitors have been replaced.

The only thing I can take off is the Super Buster 11 which I bought new from Vesalia because the board had a rev 7 buster. So I don't think it's that.

Where do you actually get 2 new CIA chips these days?
EDIT: Found them at Vesalia. I'll order a couple of them.

What is interesting is that the board has flexed a lot around the CPU connector. The connector has actually bended somewhat.

I could stick the CPU board on the motherboard and use a multimeter to test the connections one by one but if that is not it, it might be a broken trace or via. That's going to be hard to find.

[countzero]

did you check the capacitors ?

[da9000]

@PulsatingQuasar:

First of all it sounds like you know what you're doing (see below about diagnosis). A good thing when starting with such complicated things.

Second thing: very strange that Ramsey was missing. Was it desoldered? (you mentioned soldering it). To me it's a sign that someone messed with the board before (they thought it was dead? knew it was dead?)

You mentioned replacing all the right stuff and the caps and also getting it to work at least once. VERY good sign.

Also your description about the "gray screen" (video signal present), Ctrl-A-A resets, white "flash" on power on, no ChipRAM=green scren, are "spot-on" for diagnosis. Good job. I'd agree with you that the CPU is working properly EVEN if you later describe the socket being bent and flexed (explain the flexed part again?).

Now, if it's not the socket, it's most likely one of the CIAs or some other part. I'd blame the CIA first, and the reason why is that it not only controls the stuff mentioned above, but if I recall correctly it's also the part that will map the Kickstart (I assume you've swapped the Kickstarts with known working ones, as well as all other socketed chips) which of course is necessary for the booting to occur. Therefore, if one (I don't remember which one controls the memory mapping, so you'll have to do some research on that and target that one first) of the CIAs is toast, the machine will power on, the CPU will be online, but no booting (aka. "almost black screen").

And now for the sweet part: hackery time: You don't have to remove the CIAs if you don't want to. What I did is take a PLCC socket which is the right size for the CIA, and flip it upside down, and "crown" the CIA. Then bend up the pins of the socket, and place the NEW CIA inside of the pins. Solder pins to CIA. Now you have a "complete crown". Last bit: cut the VCC pin on the old CIA (the one on the motherboard), so that there's a gap and therefore even side-ways pressure will not allow contact with the chip's portion of the pin. This will effectively disable the onboard CIA. Now place the "crown" on top of the CIA and boot. Note: you might have to file away some of the plastic of the PLCC socket so that the crown will sit and not pop-up. Hope these mini-instructions made sense. If I have some time I'll post photos.

Good luck mate and let us know how it goes.

[PulsatingQuasar]

Great tip about the PLCC socket.

The ramsey was indeed soldered off by the previous owner. I bought one from Vesalia and soldered it back on.

I also fixed jumper J500 and some traces next to the ADV chip since the previous owner took that off too.

I measured all the pins between the motherboard and a CPU board and they all make a connection so the CPU connector is not the problem.

I'll buy a bunch of those CIA's from Vesalia. Since they are a common cause for problems it never hurts to have a few spares around.

Next weekend a new try to fix this motherboard. :-)

[x303]

The ramsey was indeed soldered off by the previous owner

and

since the previous owner took that off too.

Who do we have to beat up for all of this pointless abuse ? :pissed:

Let us know if you got the mainboard working.

[color=ff0000]x303[/color] :-D  :-D  :-D

[PulsatingQuasar]

Small update.

It doesn't look like the CIA chips are the problem. The problem still persists.

What I did next was to fit an 68030 accelerator because it is smaller. I then was able to feel the temperature of the Ramsey chip which within minutes gets warm/hot. That is not OK because on a working A4000 motherboard the Ramsey barely gets warm.

So I was afraid it might have been broken. So I cut it off and this time I soldered on a socket and put a new Ramsey in the socket but again the temperature gets warm/hot within minutes.

The only thing I can think of now is there is a short cirquit on the mainbord that kills the ramsey. I don't think I will continue to fix this board unless it's something obvious I can fix first.


Other Amiga 4000 mainbord problem:
I have another Amiga 4000 mainbord here which works but it doesn't see any fast ram from the mainbord or from a Fastlane Z3 expansion. Should I replace the Ramsey first and then some other chip like the Buster?

[recidivist]

 What is most likely cause of  4000 motherboard not seeing,i.e. zorro cards not working?

[CLS2086]

You may have capacitor leak that cut a track...
Contact amigacenter.com  :rtfm:

[PulsatingQuasar]

The motherboard that didn't see the fast RAM on the motherboard had 4 tracks that didn't connect anymore to U891 from the via's. I fixed that and it now sees the memory again. This board is fully working now.

The other one is still dead. I also looked around on a German forum and asked Amiga Center about a possible cause. It is clear it is a short cirquit that heats up the Ramsey but no obvious cause like a resistor that doesn't work anymore.

There is possibility that the Bridgette or one of the other RAM related chips is causing the short cirquit.

[darksun9210]

wow, you certainly sound like you know what you're doing. last thing i can say is i agree with CLS2086, and look at amigacenter to help out...

[MastaTabs]

there are propably only two possibilities.

1. Shortcut against ground or vcc. Might be caused by a borked ttl chip or resistor.

2. another driver driving on the same net the same time ramsey is active.

At first I'd check all data and address lines from/to ramsey for short curcuits against vcc/gnd. If thats not the cause then have a look at the schematics and check for possible ttls or gals with connections to ramsey. At first possibly only with connections to ramsey as other parts don't seem to get hot.