A4000 sudden death

[McTedson]

First post here, so hello everybody.
Like topic title says, my 4000 decided to die on me...

This is a 68040/25 A3640 3.1, rev. B mainboard, had some damage from the battery, but was cleaned. A3640 had its capacitors changed before leaking.

Happened while playing a game, image just froze, and it would not respond do C-A-A. Power off and back on results in...nothing, just a black image. No harddrive spin, nothing. Power LED is on.
Things done so far:
Changed almost all capacitors on the mainboard (still need to replace some but need to get them). Power supply voltages are within range. Power section on the mainboard checks out ok, i have all the voltages.
Got my logic probe connected. Checked reset line _KBRESET from Fat Gary to U141 (7407) pin 1, and found it stuck on low. Removed 7407 to check if there's a pulse coming from Gary upon reset - there is. Put 7407 back, _KBRESET from Gary again stuck low ?! Removed U140 (74F32), but left U141 on the board. Now _KBRESET is behaving correctly, and all signals on 7407 match the schematic. Changed U140, _KBRESET again stuck low ?!.
Ok, so removed 7407 and checked signals on 74F32, behaves like in schematic. Checked all resistors on pull-ups (R120, 121, 122) , all ok. Checked the R125 - OK.
Manually pulsing _RESET to high does leave _RESET on high, and Alice starts working - so it seems - as it starts heating up, and i can see activity on the CHIPRAM, also there is some sort of image on TV - dark grey. Also with A3640 hooked on the mainboard, the CPU starts heating up and there's some activity on the 543's.

Basically with either of the IC's on the mainboard, signals are ok. Both IC's on mainboard, reset line gets stuck low. At this point i kinda ran out of ideeas. I don't see any reason why the input pin on the 7407 would be pulled low. Oh and I did substitute the 7407 with another one and it behaved the same.

Does anyone have any ideeas ?

[Tension]

Turn it off and on again.

[save2600]

@Tension.. LOL!  I think he's a little more experienced with electronics than that  

Wow, sounds like you're pretty handy with a logic probe! But did you try the simplest stuff first? Like removing and cleaning the pins where the 3640 interfaces the mobo? Horrible design IMO as they're too small. But since you're getting some kind of a picture... is it real dark with wavy lines? If so, recheck the video output. Do you have access to an RTG video card or just the AGA RGB output? How does it look when you hook it through the composite out? Alice might be bad if you've checked all the other components near her. If you don't want to go through the hassle of removing a SMD chip such as she, mobo's for the A4000 come up on ePay from time to time and not too expensive.

In any event and pardon my memory if this is not feasible, but I'd crazy glue a tiny heatsink (like from a PeeCee's video card) on her next time around...

Oh and Amigean stated what I overlooked... reseating ALL removable chips, or if they look corroded, cleaning their pins with light sandpaper or fiberglass brush (I guess they are no longer sold, but I have a large stash of replacements for mine). Only other bit of advice I could offer is to be sure your power supply caps are all good. GREAT idea to replace them anyway as they're known to go bad on any computer this age or older. Other components that fail in the power supply are bridge rectifiers and diodes - both of which can easily be tested with a DMM. Lastly, I'd check for cold solder joints everywhere (mobo and power supply especially), but especially where the power supply connector goes into the mobo.

[amigean]

Hi,

You seem to be an electronics "black belt"so I couldn´t hope to give you any meaningful advice at the very low level stuff, however, I am wondering whether you have tried the more down to earth advice usually dispensed on such occasions:

1) remove all socketed chips (does the A4000 have any?), daughterboard, processor cards, expansions and any peripherals, clean all connectors with isopropyl alcohol, look for wear&tear and reseat them. Check memory (simply remove simms bar one at a time).

2) check traces around the battery, acid leakage damage could conceivably progressively deteriorate over time. Clean area with appropriate chemical solution (some advise water, others vinegar, I would say check with someone who really knows first).

3) The either one IC work / both ICs won´t work deal makes me think it is related to power supply. Power supply damage also squares with the manner in which the problem occured to begin with (while the computer was on). You said you tried out voltages at the "power section" of the motherboard - not sure what you meant. Get an Atx to A4000 connector (amigamaniac in australia does them and so does amigakit I think) and try a new one.

4) Of course, another alternative is that one of the custom chips itself decided to give way. I don´t  have a clue how you could check that though, sorry.

Good luck and keep us posted on progress - let us have more info on whatever else you´ve tried, it might stimulate further ideas.

[McTedson]

Hehe Tension Only did that for a few thousand times

Thanks for the replies guys.

I don't have access to any video board , i do however have an Indivision 4000, but it does match the RGB output , as in no image (monitor does sync though). I thought that Alice might be holding my reset line low so i did cut the reset trace that goes to it, but there's no change in the way reset goes, so it's not that.
The battery area I have cleaned at the time when i replaced the battery. I have used some vinegar to neutralize the white stuff, then i have removed everything that could be removed from the mainboard and washed it and dried it it. I have also removed all chips in the RTC area, cleaned the board underneath them and repaired a few corroded traces,then soldered everything back.
The only socketed chip on the board is the Buster which i did remove to see if there was any corrosion underneath or broken pins, but everything looks ok. Oh and the roms, but nothing there.

Buying a new mainboard might solve the problem, but erm, would like to fix this one, bit short on the dosh these days. Definetely going to put some more cooling into this thing if I ever get it back online...some of the PAL chips get *really* hot, i was considering replacing the with GAL's to make them run cooler - but that would be some other time project, if ever
There's no problem for me to remove the chips , was thinking into maybe getting sockets and socketing at least the big chips (except LISA, the indivision might have a problem fitting over a socketed one i think :roflmao:)
@amigean: I'm sorry for not being clear, i meant the power section as in power connector , the diodes on +12v and - 12v, and the 7805 used for -5V (U198 in the schematic). The voltages are there and in range, but  with no scope i cannot verify the quality of the rails.

I will try to gather the capacitors to shotgun the ones in the power supply. I don't have scope to check for ripple unfortunately, but it's a good ideea. Worst case scenario will order that ATX adapter.

Could any of the 4000 owners tell me what their amiga shows with no cpu board and no ram plugged ? Screen should be black ? Any colour ?


Thank you for the replies guys, keep'em coming

[tone007]

Could any of the 4000 owners tell me what their amiga shows with no cpu board and no ram plugged ? Screen should be black ? Any colour ?

Black with no CPU board as far as I've seen.

Might be hard to tell if your '040 has a heatsink, but is the CPU heating up at all when the system is on for a couple of minutes?

edit: never mind, I re-read your first post.

[tone007]

No harddrive spin

That doesn't sound right to me, hard drive isn't even spinning up?  Sounds like a dead drive or lack of 12v. (Unless of course it's a SCSI drive that's waiting for the interface to spin it up.)

[McTedson]

No, as long as reset is stuck on low, CPU doesn't heat up. Hardrive doesn't spin due to the same problem, reset is going on IDE pin 1 keeping HDD in reset

[Argus]

I had a dead (still dead) A4000 board with similar symptoms (black screen to dark grey then nothing).  I had all the caps and most of the custom chips swapped/replaced.  Turns out after all that I examined the pin contacts of the Bridgette chip and found one of the chip legs had broken free of the pin trace so you might check there.  My attempt to reattach the pin leg resulted in a broken pin so I desoldered the chip completely and have been looking for some time for a replacement Bridgette.

[Argus]

As for screen colors:

Black: no cpu
Green: no/bad chip ram
Yellow: expansion memory problem
Red: ROM error
Blue: Custom chip (Denise/Alice/Paula)
Purple: Custom chip/rom error
White/light grey: Hardware fault somewhere stopping final boot

[Castellen]

been looking for some time for a replacement Bridgette.

I have spares of these if you need any, though they're generally untested.  I can also fit the part to the board for you if that helps.

http://amiga.serveftp.net


@McTedson
If any of the various reset lines are being held low, I'd certainly suggest focusing more in that area.  If _KBRST is permanently low, there could be a problem with U150.  If _IORST is being held low, it could be one of the CIAs.  You could quickly swap U300, U350 positions to check them, they're the same part. The individual reset lines are formed from U140 and U141, though it's very unlikely for either of these two gates to fail.  But it's a good starting point in the circuit for tracing reset line problems.

And in fear of stating the obvious, check that _FAIL (U150 pin 56 and power connector pin 1) is at 5V.  From memory, U150 causes the system to be held in reset state if the power supply holds _FAIL low.  Worth checking if you haven't already done so.

[McTedson]

Hello Castellen,

I am trying to read the schematic here, and from what I can understand, things go a bit like this:
_KBRESET from Gary is fed into U141 at pins 1 (probably to be mixed with EHLT) and straight to pin 3. Has a R120 pullup, so idle state should be high. If the signal is propagating correctly, all reset lines should be high.
Now when Gary pulses low the KBRESET, it should propagate through to all reset lines.
FPURST, CPURST got to A3640.
RESET goes to A3640, Alice and IDE pin 1.
_IORST goes to pretty much all other big chips
_ERST goes to A3640 and daughterboard (i think).

A3640 stuff:
If I understand this correctly, CPURST goes into U201, comes out as 040RSTI , which among other things goes to 68040 in RSTI. RSTO from 68040 goes to U104 pin 5 and gets output on pin 6 which is the _RESET line. This I suspect is so the CPU can execute the RESET instruction and have reset control over hardware. On My3640 RSTO is high. So the cpu might also mess with the _RESET line.

And that's why my logic fails here, if I remove U141 i see the reset pulse coming from Gary on pin 1. Put it back, pin 1 is stuck on low. Remove U140, i can see reset pulse coming from gary on pin 1 on U141, and propagating nicely to the U140 pins. Put U140 back everything freezes on low.
Why and what can mess with the input pins from U141 :madashell:

_FAIL is logical high here measures the same as 5V, so it's not that....but maybe Gary thinks it's low ? How does Gary hold the system in reset ? It's not obvious to me from the schematic.

[McTedson]

I had a dead (still dead) A4000 board with similar symptoms (black screen to dark grey then nothing).  I had all the caps and most of the custom chips swapped/replaced.  Turns out after all that I examined the pin contacts of the Bridgette chip and found one of the chip legs had broken free of the pin trace so you might check there.  My attempt to reattach the pin leg resulted in a broken pin so I desoldered the chip completely and have been looking for some time for a replacement Bridgette.

I have checked the pins on the Bridgette and they look ok, i did a bit of touchup on some pins  that did look a bit flakey, but no change

[Castellen]

The reset logic can be complex in that any number of system devices can cause a reset.  The outputs of U141 are open collector, so anything else sharing the _EHLT _CPURST or _RESET lines is capable of causing a system reset, including the CPU board.

I'd expect U140 to have a totem pole output, so therefore the lines _FPURST _IORST and _ERST are only driven from U140 outputs.  That's why these lines don't have pull up resistors, while the lines on the U141 outputs do.

Not sure exactly what drives the U150 internal _KBRST logic, no way of easily knowing what goes on inside the gate array.  I'd imagine _FAIL has some influence on the _KBRST line.  Aside from that, U150 is mostly doing address decoding.

If you've measured _FAIL at 5V, it's probably OK.

Quite possible there's something dodgy happening on the CPU board, as it's capable of generating system resets.  The best suggestion is to try a known working A3640 or whatever to isolate the fault to the CPU board or main board.  Else you could be chasing your tail for days.

[spirantho]

I'm not as good with a schematic as you, but I'd strongly suggest trying a known good A3640 or even a 4000/030's 68030 board. Usually malfunctioning chips will show a diagnostic colour on the screen when the 4000 does its selfchecks. Unless the Kickstart is bad, of course, but that's a lot less likely than the CPU board being bad. If the monitor is getting a sync then the display chips should be ok, otherwise you'd get "no signal", and that also implies you're getting a reasonable +5V line.

Where are you in the world? Maybe someone can loan you a 68030 board? I have one spare but it's at my parents' house 150 miles away....