PCB track repair in an A3000

chiark · 31 January 2008, 15:27

Just thought I'd share my experiences of trying to repair my A3000 which has been dead since 1995. I never knew why it died, and as I had a 4000 I didn't worry about it.

Now the 4000 is dead and awaiting pieces to diagnose the fault, I'm busying myself with the 3000.

Lessons I've learnt:
1 - don't buy a "sold as seen" motherboard off ebay. They're junk.
2 - don't try to fix small traces with conductive silver paint.
3 - don't try to fix small traces with conductive PCB pens
4 - the one true way of fixing PCB tracks is with solder and wire. Anything else is just a little bit effete

3000 died. I left it alone. A few years later I decided to fix it and found the battery had leaked, and didn't seem to have done much damage so I removed the battery, cleaned up the mess, tried to get it working by reseating everything, and failed.

Now I'm having a proper go at fixing it.

The tools!

Agnus' socket was cracked (oo er). Paula's legs had a bit of green goo on 'em, and Denise's legs had goo all over 'em.

I started tracing connections on components (namely Paula, Agnus and Denise) and found to my horror that Denise was not really connected to much on pins 33 and 48 through 42. And pin 1 wasn't doing much either.

Damn.

So I decide to desolder the sockets and replace them. Each pin on the DIL socket can be removed individually by teasing out the folded over bit of metal in it, applying heat to the bottom, and gently teasing out.

Paula and Denise live in 48 pin sockets. That's 96 pins.

Agnus is a PLCC, and sits in a carrier. The carrier is soldered onto the board in two squares of connectors, one inside the other. The outer pins of these can be removed using the same technique as for a DIL, however to get the inner pins out relied on me breaking the plastic socket before unsoldering the pins.

Agnus lives in an 84 pin PLCC carrier. That was *such* fun.

Clean up with solder wick and solder sucker.

There's a pin in each socket which is a sod to clean up completely because it's connected to the ground plane so needs a shedload of heat to unsolder. Persevere with this and it'll give in.

Rechecked the connections and found no continuity on the traces.

Using a fibreglass pen and isopropyl alcohol (99%), I removed the PCB coating, exposing the copper traces.

I then thought I'd take a shortcut and try to repair using a PCB repair conductive pen. That worked, but resistance on the trace was around 10 ohms which I think is too much.

So I cleaned that off with isopropyl alcohol and the fibreglass pen again...

And tried again with conductive silver paint. I found the easiest way to apply this was to paint the area rather than a single trace then, using a scraper, remove paint to create individual traces out of the big area.

still 10 ohms on a trace. Not good enough. Cleaned it off again.

So, shaky soldering hand in place, I tinned the tracks.

Then I stripped the thinnest wire I could find, and split the wire bundle into single strands, and tinned a few.

Soldered this onto the PCB and, for good measure, through the PCB hole too leaving enough space for the replacement socket... Traces are now 0.1 ohm

My experience suggests that trying to use these PCB conductive pens or other similar shortcuts just don't work, and are a lot more expensive than doing it the "right" way.

Tonight's job: fixing the last track, refitting sockets. I know some people advocate soldering chips directly onto the board, but I don't want to do that in case denise has died... And I know I'm using dual wipe DIL connectors, but for me they work

thread to be updated with results

musashi5150 · 31 January 2008, 16:52

Best of luck

I hope it turns out alright after all that work!

coze · 31 January 2008, 16:57

great job dude, keep us posted. I once undertook a similar rescue operation, replaced agnus and denise sockets, however I couldn't rescude her

though I didn't know what I was doing back then, and I think I didn't check the traces properly ...

Good luck ! (and pics will be nice too !)

chiark · 31 January 2008, 17:09

Thanks for the encouragement

- yes, I'll put pics up when I get home. It's not terribly pretty, but at least it shows what I've done. At the very least, people can point and laugh, and use 'em as an example of how not to solder

I've got a horrible feeling I'll do all this work and still get nothing. At that point, I think I'd be at the limit of my knowledge. I've posted up on amibay asking for any known good components from an A3000 so I can eliminate things... Who knows, it might just leap into life?

Actually, I think I could test Paula and Denise in the working A500... hmm...

Zetr0 · 31 January 2008, 17:19

Hardware Pr0n!!!!!

chiark · 31 January 2008, 17:40

Would ya like pictures of the implements too, eh? Would ya?

cosmicfrog · 31 January 2008, 21:46

yes yes and yes give us lots of pictures
hope all your hard work pays of with a working machine

chiark · 01 February 2008, 09:56

Well, every trace on the new sockets is good.

Same.
Frigging.
Error.

LED lights up, blinks 9 times, then machine resets.

To say I'm a little crestfallen would be an understatement.

chiark · 01 February 2008, 10:26

hmm, hang on, what's this?

Pass or fail the RAM to the Screen (LED Blinks 9short 1long )
http://www.fixya.com/support/t42548-green_screen

I've got a RAM problem

At least I know what the problem is!

coze · 01 February 2008, 12:23

you got any fast or chip ram installed ? 1mb chip ram must be soldered.

chiark · 01 February 2008, 12:29

It's all stripped out apart from the stuff on board, which is sorta hinting that the stuff on board is the problem

Pics coming soon, just need to resize 'em

r.cade · 01 February 2008, 17:01

That's great that you are digging into it. I've got an A3000 board that shuts down the power supply as soon as you plug in the board. It looks to me from my crappy VOM that there is a dead short between +5 and GND on the board somewhere that I can't find. All TTL chips show continuity of 0 between bottom left and upper right pins. Already ruled out the 1488/1489 which are usually the cause of this (plugging in modem cable while powered on).

It was working in the evening, dead in the morning.

chiark · 01 February 2008, 17:14

hmm, that is weird. Mine just refused to boot one day in 1995... This is all being prompted by the 4000 dying last week just after I'd backed up the drive using WinUAE

Pics now attached as you can see

.

31 January 2008, 15:27	#1
chiark Needs a life Join Date: Jan 2008 Location: England Posts: 1,707	PCB track repair in an A3000 Just thought I'd share my experiences of trying to repair my A3000 which has been dead since 1995. I never knew why it died, and as I had a 4000 I didn't worry about it. Now the 4000 is dead and awaiting pieces to diagnose the fault, I'm busying myself with the 3000. Lessons I've learnt: 1 - don't buy a "sold as seen" motherboard off ebay. They're junk. 2 - don't try to fix small traces with conductive silver paint. 3 - don't try to fix small traces with conductive PCB pens 4 - the one true way of fixing PCB tracks is with solder and wire. Anything else is just a little bit effete 3000 died. I left it alone. A few years later I decided to fix it and found the battery had leaked, and didn't seem to have done much damage so I removed the battery, cleaned up the mess, tried to get it working by reseating everything, and failed. Now I'm having a proper go at fixing it. The tools! Agnus' socket was cracked (oo er). Paula's legs had a bit of green goo on 'em, and Denise's legs had goo all over 'em. I started tracing connections on components (namely Paula, Agnus and Denise) and found to my horror that Denise was not really connected to much on pins 33 and 48 through 42. And pin 1 wasn't doing much either. Damn. So I decide to desolder the sockets and replace them. Each pin on the DIL socket can be removed individually by teasing out the folded over bit of metal in it, applying heat to the bottom, and gently teasing out. Paula and Denise live in 48 pin sockets. That's 96 pins. Agnus is a PLCC, and sits in a carrier. The carrier is soldered onto the board in two squares of connectors, one inside the other. The outer pins of these can be removed using the same technique as for a DIL, however to get the inner pins out relied on me breaking the plastic socket before unsoldering the pins. Agnus lives in an 84 pin PLCC carrier. That was such fun. Clean up with solder wick and solder sucker. There's a pin in each socket which is a sod to clean up completely because it's connected to the ground plane so needs a shedload of heat to unsolder. Persevere with this and it'll give in. Rechecked the connections and found no continuity on the traces. Using a fibreglass pen and isopropyl alcohol (99%), I removed the PCB coating, exposing the copper traces. I then thought I'd take a shortcut and try to repair using a PCB repair conductive pen. That worked, but resistance on the trace was around 10 ohms which I think is too much. So I cleaned that off with isopropyl alcohol and the fibreglass pen again... And tried again with conductive silver paint. I found the easiest way to apply this was to paint the area rather than a single trace then, using a scraper, remove paint to create individual traces out of the big area. still 10 ohms on a trace. Not good enough. Cleaned it off again. So, shaky soldering hand in place, I tinned the tracks. Then I stripped the thinnest wire I could find, and split the wire bundle into single strands, and tinned a few. Soldered this onto the PCB and, for good measure, through the PCB hole too leaving enough space for the replacement socket... Traces are now 0.1 ohm My experience suggests that trying to use these PCB conductive pens or other similar shortcuts just don't work, and are a lot more expensive than doing it the "right" way. Tonight's job: fixing the last track, refitting sockets. I know some people advocate soldering chips directly onto the board, but I don't want to do that in case denise has died... And I know I'm using dual wipe DIL connectors, but for me they work thread to be updated with results Attached Thumbnails Last edited by chiark; 01 February 2008 at 14:49.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
A3000 Repair Needed	merlin1968	support.Hardware	3	03 June 2013 12:56
HomeMade PCB	coze	support.Hardware	11	13 March 2011 00:55
Wrinkles in PCB	PowerPie5000	support.Hardware	10	11 December 2010 15:21
Looking for PCB wizzard	kolla	request.Other	5	11 August 2010 12:12
FS: FULLY LOADED A3000, a bare bone A3000 and lots of Amiga Stuff!	emc_errol	MarketPlace	0	30 January 2009 22:33

31 January 2008, 16:52	#2
musashi5150 move.w #$4489,$dff07e Join Date: Sep 2005 Location: Norfolk, UK Age: 42 Posts: 2,351	Best of luck I hope it turns out alright after all that work!

31 January 2008, 16:57	#3
coze hastala vista winny vista Join Date: Feb 2006 Location: mt fuji Age: 46 Posts: 1,335	great job dude, keep us posted. I once undertook a similar rescue operation, replaced agnus and denise sockets, however I couldn't rescude her though I didn't know what I was doing back then, and I think I didn't check the traces properly ... Good luck ! (and pics will be nice too !)

31 January 2008, 17:09	#4
chiark Needs a life Join Date: Jan 2008 Location: England Posts: 1,707	Thanks for the encouragement - yes, I'll put pics up when I get home. It's not terribly pretty, but at least it shows what I've done. At the very least, people can point and laugh, and use 'em as an example of how not to solder I've got a horrible feeling I'll do all this work and still get nothing. At that point, I think I'd be at the limit of my knowledge. I've posted up on amibay asking for any known good components from an A3000 so I can eliminate things... Who knows, it might just leap into life? Actually, I think I could test Paula and Denise in the working A500... hmm...

31 January 2008, 17:19	#5
Zetr0 Ya' like it Retr0? Join Date: Jul 2005 Location: United Kingdom Age: 49 Posts: 9,768	Hardware Pr0n!!!!!

31 January 2008, 17:40	#6
chiark Needs a life Join Date: Jan 2008 Location: England Posts: 1,707	Would ya like pictures of the implements too, eh? Would ya?

31 January 2008, 21:46	#7
cosmicfrog The 1 who ribbits Join Date: Apr 2006 Location: leek, Staffs, UK Age: 56 Posts: 3,557	yes yes and yes give us lots of pictures hope all your hard work pays of with a working machine

01 February 2008, 09:56	#8
chiark Needs a life Join Date: Jan 2008 Location: England Posts: 1,707	Well, every trace on the new sockets is good. Same. Frigging. Error. LED lights up, blinks 9 times, then machine resets. To say I'm a little crestfallen would be an understatement.

01 February 2008, 10:26	#9
chiark Needs a life Join Date: Jan 2008 Location: England Posts: 1,707	hmm, hang on, what's this? Pass or fail the RAM to the Screen (LED Blinks 9short 1long ) http://www.fixya.com/support/t42548-green_screen I've got a RAM problem At least I know what the problem is!

01 February 2008, 12:23	#10
coze hastala vista winny vista Join Date: Feb 2006 Location: mt fuji Age: 46 Posts: 1,335	you got any fast or chip ram installed ? 1mb chip ram must be soldered.

01 February 2008, 12:29	#11
chiark Needs a life Join Date: Jan 2008 Location: England Posts: 1,707	It's all stripped out apart from the stuff on board, which is sorta hinting that the stuff on board is the problem Pics coming soon, just need to resize 'em

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)

01 February 2008, 17:01	#12
r.cade Registered User Join Date: Aug 2006 Location: Augusta, Georgia, USA Posts: 548	That's great that you are digging into it. I've got an A3000 board that shuts down the power supply as soon as you plug in the board. It looks to me from my crappy VOM that there is a dead short between +5 and GND on the board somewhere that I can't find. All TTL chips show continuity of 0 between bottom left and upper right pins. Already ruled out the 1488/1489 which are usually the cause of this (plugging in modem cable while powered on). It was working in the evening, dead in the morning.

01 February 2008, 17:14	#13
chiark Needs a life Join Date: Jan 2008 Location: England Posts: 1,707	hmm, that is weird. Mine just refused to boot one day in 1995... This is all being prompted by the 4000 dying last week just after I'd backed up the drive using WinUAE Pics now attached as you can see .