dead A4000

[whaka]

hello,

recently, i got an A4000, and started to restore it.
i'll try to make the story short, but after replacing capacitors, troubles began...

i first got a very garbled video for a short time, via a scandoubler pluged in video slot as i don't have any 15Khz monitor where i am.
video back to normal again after moving the scandoubler card/daughterboard, so... i though simply a bad contact...

i installed 3.2 kickstart and workbench on CF and from the cd rom sucessfully, and let it run for few hours, without any issues. switch it off, go to sleep.

then, i did the PIO2 mod, and when i switched it on, a get the garbled video again... it booted successfully on the workbench, so the mod himself is a succes. but very shortly after, it crashed and reboot, got a guru, rebooted again... then, dead end. only black screen, no sign of life of the cpu (caps lock switched on/off for 6 times, then dead.)
i tested the CS Mk1 in it on a 3000 mobo, and it work fine, so this is excluded.

then... i heard weird noise comming from the motherboard, a very low "sparking noise" from time to time... well, i finally spotted it... and here it was :





yes, a nice carbon tracking phenomena. not to mention the new capacitor was not the culprit. it's just the pcb himself, when for any reason, the insulator become conductor...
i mounted tantalum capacitors, so, can't see what really happened until i touched it and feel how hot he was.

this capacitor is for the video DAC supply, so, i know why i got garbled video now. but it won't really explain why the motherboard still refuse to boot.
maybe except if a dead video DAC can prevent boot.

can someone confirm me if the supply of the DAC go near to 0V (probably what happned during pcb "carbon tracking thing" can it be killed ?
may sound weird as a question, but it may be a bad biasing story as lisa and some others peripherial related to him were still pushing their current in it during this time.
but before desoldering it, maybe there's others things i can check ? i have an oscilloscope and a logic analyzer.

so, all ideas are welcome
thanks.

edit : i checked amiga pcb explorer, and there's no tracks passing under the "sparking hole" it might just be the ground plane. so, certainly not the reason of non booting.

[A4000Bear]

I think your statement 'I mounted tantalum capacitors' may provide the answer to what happened. When tantalum capacitors fail, they usually do so catastrophically. Usually they go dead short and can get extremely hot, sometimes to the point of bursting into flames or exploding. It is quite likely yours had become so hot it scorched and carbonised your PCB. It is possible you now have a conductive path between the positive terminal of the capacitor through the burnt PCB into the ground layer in the middle of the PCB. If you still have the capacitor, you can see if it was the cause by checking to see if it is shorted.

Tantalum capacitors have been known to fail for no apparent reason, however they are very sensitive to reverse polarity or over voltage, as shown here when one is fed a higher voltage than what it is rated at : [ Show youtube player ]

[cpiac64]

tantalium caps have positive signed
electrolitics caps have negative signeed

[whaka]

hello,
i said, and confirm again, the new capacitor was not the culprit. i checked it after desoldering, he was not shorted at all. and his casing was clearly not the source of the sparking.
and when i say new, he was. i only buy capacitors on farnell/element14 or mouser.

i didn't mounted it backward, and not underrated it. it might be hard to believe, but sometimes, pcb can fail too

edit : here is the capacitor


as you can see, his case is not totally drilled by a spark source. if he was the sparks source, i seriously think he'll go into big smoke mode before the pcb had time to burn in that way.

[cpiac64]

add a 10uF radial for view if the board smoke

[whaka]

i cleaned all the area of carbon residue, so with or without capacitor, there's no reason for the pcb to continue burning.

i did something i didn't think to do before, checking the 5V at the capacitor pad. i think i now why it doesn't boot...

it's really noisy in this area and the average voltage is 4.76V, it's low. that can explain why.
i don't know what generate such noise, maybe the DAC, as it's his supply rail.
maybe my first idea was good, and the DAC is maybe well dead from underpowering.

[Mick]

Is it possible the pad has undergone a repair? directly underneath the pad is either 5V or ground plane but I'm not sure which. Have you tested for shorts? +AVID has a small sectioned off power plane in that area, the power enters it through L500 and R465 so test those.

[whaka]

no, the pad was looking really good before, and the capacitor i replaced was the original.
with a very small leakage traces.

i'll check what you suggest, but i don't really know how to test an inductor.
all i can say, L500 is not cut, for sure. but if the coil shorted, i don't know how to see that.

edit : R465 was not factory mounted on my mobo.

[Mick]

I think you need a special meter but I'd probably measure what the 5V reads going into them and what's coming out on the opposite side. It sounds like there's an intermittent short from what you've said though.

[Mick]

How long did you use it for before recapping it? you said it was fine before?

[Mick]

My only other guess would be like A4000Bear says but maybe the tantalum was simply faulty and in shorting out has got really hot on the bottom as well as damaged other components? in terms of reliability I've read that Tantalums will either fail early or last decades but when they do fail they leave a mess.

edit: you said you tested the tantalum after removing it but if it was intermittent issue with it?

[whaka]

voltage is the same before and after L500, 4,68V with a multimeter, and 4,65 on lowest peak with oscilloscope, definitely too low.

i checked voltage on lisa, and it's far better, and on some various points, i get 4,95V.
the drop is really localized in the DAC area.

with original capacitors, i only powered it once when i received it, and played few minutes with it. just to check if floppy drive worked and booted a WB floppy.
so, not very long.

[whaka]

when i said it was hot, it was, but i could let my finger on for few seconds.
so... i'll said around 75/80°C not more. except the pcb and the capacitor, nothing else get burned.

i checked the capacitor again, and he's not shorted at all.
personnaly, i don't blame it because there's no apparent reason to do it.

if he was the problem source, considering sparking was occuring... if there was enough power for sparks to travel from the capacitor case to the pcb, the capacitor would have been the first to seriously blow up...

[whaka]

sorry, i was wrong, i get 4,95V before L500.
so maybe her internal resistance raised too much due to the abnormal power consumption when pcb was burning.

edit : well, i'm lost now... i have a spare motherboard with no custom chips on it, but all the rest is present.
i checked voltage on it for curiosity, and i got 4,5V after L500 on it (and R465 is present)...

and yes, i can check her resistance with a simple multimeter, on the spare, it's around 5 ohm, and 3 on the mobo i work on.
that's weird, because the datasheet of the DAC specify 4,75V as lower possible voltage.
i'm really lost

[RoniWorlds]

I keep my fingers crossed that we can fix it. Each Amiga saved is a plus one to the skill.

[Mick]

I'd replace the scorched tantalum just to be safe, you're probably better off doing what cpiac64 said temporarily just to make sure the PCB doesn't have any damage.

[whaka]

of course i'll replace it, actually the mobo is without the capacitor.
the capacitor case get burned, even if i don't suspect it in first, and despite the fact he's not shorted, i don't trust it anymore for long term service.

the mobo should start without it, it's not very mandatory for that anyway

[Mick]

So there's no activity at all? even outside of video?

[whaka]

no, most of time just a black screen and no answer from cpu.
this afternoon, after letting sit the mobo disconnected for, let's say near 24hr, the first boot gave me a blinking power led, then i heard the floppy drive clicked once, then dead.

unfortunately, i didn't pluged the scandoubler, so, i can't tell what happened on screen.
so, very randomly, i can get a kind of system wake.

i have all the stuff for troubleshoot, but i don't know where to start, as i have no knowledge on how amiga boot. i checked clocks/video synchro, they're all there and clean.
that's all i can say.

[whaka]

ok, this time i think i got why it doesn't boot... the power good signal fall to around 3.80V. that's THE why.

when psu is unplugged, i got 5V on it, and fall once plugged.
so, it might still be the motherboard, or the psu itself providing a weak power good and fall once loaded.

i'm going to try an atx psu, just to see.

edit : well, no. even if i directly plug a 5V rail in the power good from an atx psu , it doesn't boot.
but anyway, i think i got something to follow. i don't really know how this power good had to behave on amiga, but if it's like on pc, he had to stay at 5V and not falling.