iBook freezes when on charge

You might try resetting the PMU and see if this helps:

http://support.apple.com/kb/HT1431?viewlocale=en_US

Good luck!

which seems to freeze (and crash) whenever it is attached to the power.

Since you tried a different power adapter the first suspect would be the DC in board.

the other day and when I returned, after a few hours, the case was very hot all over. I switched it on but it would not boot past the blue screen, after a few hours cooling down it worked perfectly again...

Try downloading:
http://www.bresink.de/osx/TemperatureMonitor.html
monitor the CPU temperature and report back.

until I re-attached it to the charger.

So there could be multiple problems here.

The fan comes on during the boot sequence,

That's normal.

but otherwise I don't hear it.

Also normal. I have 6 iBooks here at present and have worked on many more. I never hear the fan running unless it's a U28 problem.

Could there be a problem with the power supply?

possibly

a temperature sensor? system fan?

Both unlikely.
Next I'd look at the attachment of the heat sink. It's possible it's not installed correctly, missing thermal pads or the mounting studs esp on the CPU are broken.
I'd also be careful running the iBook constantly if it's going into a forced shutdown mode that's temp related. That could ultimately kill the CPU.
Richard

The only thing I can think of is that something is overheating as a result of the power supply getting hot

Luke,

all of which would seem to indicate a graphics problem related to heat build-up.

Not necessarily heat unless that temp apt says that.

I switched-on again before it had finished cooling-down and the screen was flickering garbage

So that definately seems like a GPU detachment problem.

What should I do next? Perhaps the heat sink needs re-seating?

Unlikely

Is it possible that the fabled 'shim-on-the-gpu' thing that might work?

Yup. So here's some pics.
This is the position of the GPU on top. That bay right below is where the hard drive goes and that's it's connector lower right:
http://s291.photobucket.com/albums/ll306/spudnuty/?action=view&current=13312GPUt op.jpg
Here's a view with the heat sink in place and a ghost image of the GPU:
http://s291.photobucket.com/albums/ll306/spudnuty/?action=view&current=13312iBoo kGPUheatsink.jpg
The spot where I'd recommend applying pressure from underneath:
http://s291.photobucket.com/albums/ll306/spudnuty/?action=view&current=13312GPUu nderneath.jpg
and where that spot is on the lower shield:
http://s291.photobucket.com/albums/ll306/spudnuty/?action=view&current=133GPUBot tomshield.jpg
So to access that you'd take the lower case off. Go to ifixit for that.
If that solves your problem or makes it better it will run like that for a while. If it gets worse you'd need to reball it. $79 plus shipping one way in the States.
Richard

Richard has given you some excellent advice, and I would recommend that you follow it, as I think he has a handle on what's wrong here. I had my iBook reballed in 2007 following his advice, and that certainly fixed the freezing issues that I had at the time. It has been working without any issues ever since.

About all I would add is to hold off on the new RAM and Leopard until your overheating problem is resolved. You could even contemplate installing a new and bigger hard drive while the iBook is apart.

Good luck!

Hey Luke,

I followed your instructions (and helpful photographs) to locate the GPU 'pressure point' than attached rubber domes to either side of the aluminium base cover so that there is now firm, continuous pressure against that point.

Unfortunately, it has not resolved the problem.
So it might be that there's not enough pressure. I'm working on a 15" power book that shows no display. Briefly it flickered on and there were video artifacts so I thought GPU. I tried pressure from underneath and torquing the board but that didn't work. I found the position of the GPU by looking at another logic board and went to that point which was covered with a large heat sink. Applying down pressure with both thumbs I got the display back and I'm now installing Tiger but this board will have to go in for reballing.

Once again, after a period of use (both with and without the power attached) the screen froze after a period of time. It sounds like it must be heat related but I am at a loss to know what, where? (do you suppose the heat-sink on the GPU needs re-seating?).

Well my guy at Superior wonders why the GPU is under the heat sink at all since it doesn't put out that much heat.

What's more, I switched-on again before it had finished cooling-down and the screen was flickering garbage - all of which would seem to indicate a graphics problem related to heat build-up.

That's what I'm looking at and that does indicate a GPU problem. The bottom might have to come off.
Richard

Luke,
To expand my last post.

The bottom might have to come off.

Here I mean the bottom shield.
That PowerBook I'm working on is sitting here doing updates and after a while the display went crazy with lines and static. This happened after a while when it warmed up. It didn't shut down or freeze so I just put pressure on the top of the heat sink and it came back.
We know the GUI ( graphical user interface ) runs through the GPU and failures there would look like this. Failures in other sections of the GPU would cause freezing maybe a shut down and of course as in the PowerBook a no display situation.
You might try pressure on the top but the bottom case, top case and upper shield would have to come off. Not a trivial job.
If you feel up to it email me. If we can diagnose the problem a reball would be $79.
Richard

Hey Luke,
Good job. That chip referred to on the Corey Arnold site is U28. It's a dual voltage regulator. Here's a pic:
http://tinyurl.com/dezvs4
These are the regular failure points: Pin 1 or Pin 15. Mostly Pin 15 then Pin one. Usually you can tell when it fails if you look at the pins under high magnification it will look corroded right near the board connection. However I have seen one fail that wasn't apparent. It wound up failing a continuity test however.
They both connect to ground. Pin one to the top of that small capacitor (the rectangular thing) you see above the "U28" and 15 to the large end that's directly to the right of "U28" ( You only see the end in the pic )

I have repaired this but it took a special soldering iron tip that I modded into a blade with a flattened tip the length of the legs of the chip.

Richard

Message was edited by: spudnuty

Message was edited by: spudnuty

Hey Luke,

Here's my current dilemma: do I live with it on the basis that it's kind of working and to fiddle any further with the soldering iron is likely result in terminal problems? Or do I have one more go... hold my nerve a little longer, attack a few more pins?
and held the hot iron to the chips legs a little longer this time

Well like I said it's usually pins 1 or 15. Now it's not a good idea to just go attacking pins willy nilly since they are easily messed up. When I've done this I've used a special iron I made with a tiny tip. It measures .009"wide by .066"long. I run it off a dimmer control so it doesn't get too hot. Often just heating it is not enough and I'll pick up the tiniest bead of solder. Now I've used too much and soldered leads together. It took a long time to clean that up. The problem here and with all these soldering problems in iBooks of this era seems to be the transition to lead free solder. That stuff was crap for a while. This affected many things the first that comes to mind is the X Box 360.
So when I do this I verify the problem with my continuity meter and verify the repair. In that picture I supplied I test from the shoulder of (say) pin 15 to the end of that capacitor. I just repaired one that had that problem. The joint looked fine but it failed the continuity test and here again I'm using special tools that I made. The probes are needles attached to the ends of my regular meter probes. So at first there was no continuity between 15 and that cap and after there was.
Now your problem might be that the problem doesn't happen until that junction gets warm and then opens up. Those kinds of problems can drive you nuts!

Maybe I'm being too timid - is this chip particularly heat sensitive?

Yes ultimately any chip is.

Or do I force more pressure on the chip from beneath the case (always feels like a crude solution to me), maybe using some kind of metal strip... is there a danger that I will crack the chip with that kind of pressure?

Absolutely. Perhaps the best solution would be to find someone or someplace that can do the work for you. There was just a post of someone in Japan who found a guy who soldered this chip for the equivalent of $10 American! Go figure.

This has absorbed way too much time (but has been curiously satisfying!).

Just like any learning experience.
I'd start looking if I was you. After all the UK is chock a block full of inventor types and there's got to be many who can do this kind of work. University? at least they could give you leads.
Richard