Powerbook G4 black screen hang/freeze

[mac-man6]

Hello I have a http://www.everymac.com/systems/apple/powerbook_g4/stats/powerbook_g4_1.67_15_hr.html . She's a very nice machine but the screen goes black 2 minutes after boot sometimes 20 minutes. Sometimes it doesn't freeze at all. It has a fresh install and I can't reproduce the problem, it just happens at random times.

Any ideas? I would like to hold onto this machine and keep it in top condition. Have you heard of this before, is it worth buying a new battery for this troubled Powerbook?

Symptoms:

Black screen, fans normal speed, unresponsive. Computer does not seem hot and the computer has crashed a cold start.

 

[phreakout]

What version of OS X are you running, 10.4.x (Tiger) or 10.5.x (Leopard)? Sometimes a blank display with no system response could be either a software issue (possibly a driver or extension, preference file) or hardware (the graphics chip ATI Mobility Radeon 9700, video data ribbon cable, back-light/inverter cable or Ambient Light Sensor). The entire line of PowerBook G4s, although despite brilliant and innovative design, were plagued with problems since day one. Read here under the quality and display issues.

One thing you can try first is try deleting (drag to the Trash) the preference file for the display settings (under your Home user account, not the root account area) in the Preferences folder and empty the trash bin, then restart your PowerBook. You might need to manually readjust the settings for the display to your liking (in the System Preferences).

Another suggestion is to try deleting all the cache files, using an app called Onyx (free). After you clean house, you'll want to restart twice. The first restart will re-assemble all the cache files; the second time will boot faster than the first restart. This might fix the problem.

If you can, see if there are settings you can change for the Ambient Light Sensors in the keyboard area. If you can shut it off or make the light dimmer to a cooler temperature, it will prevent the sensors from tripping early and causing the whole system to cycle between awake and sleep mode.

Finally, connect the PowerBook to an external display, be it a monitor or video projector if you'd like, and see if the symptom happens again for both displays. If so, it won't necessarily mean a hardware failure, but it will allow us to know that the hardware is working and that either a cable is loose or shorted or that the problem is with the LCD display.

Give these suggestions a try and report back with your findings.

73s de Phreakout. :rambo:

PS: I forgot to mention. If you can, with a second display hooked up, check to see if any software updates have come down the pipe relating to any display fixes, ALS problems, etc. Maybe those will fix your PBG4's ailment as well.

 

[mac-man6]

It's running 10.4 and the last time I had it running it was fully updated. I normally have it hooked up to a TV as a media player. It's hung without monitors attached and with a TV attached via s-video. The computer is hung/frozen because I can't restart it using shortcuts, typing the volume keys doesn't provide the clicking noise.

I know that the internal PRAM battery is dead, it always thinks it's in the past (I don't think it's 1956 like classic macs, but something like that. Also the main battery is dead. I've looked and even most online suppliers are sold out of that battery. It's unique to that model and if it was in stock it would be $80 or more.

I'll try the other tips you mentioned. My thoughts I'm willing to reapply heat paste and maybe reflow some joints but I don't want to invest too much time and money if it's not going to work. I was thinking of getting a main battery too but I don't want the purchase to be a waste if the problems continue. I was thinking the battery may help because even though the PRAM battery is dead, the main battery will keep all the PRAM settings and date.

 

[BGoins12]

The screen going black sounds like a bad video connection or failing ambient light sensors. A bad video connection can cause enough voltage fluctuation to lock it solid. I had a ThinkPad A31 with a bad video cable, and the screen would go black and lock the computer up.

Usually a GPU separating from the logic board causes screen artifacts or kernel panics. That, and the fact the GPU separation is rare in these particular models. It does happen, but not as often as the iBook G3/G4 and the earlier G4 Powerbooks. It really only happens to the ones that were constantly carried around and flexed or picked up by a corner.

 

[mac-man6]

So I had some time to try somethings.

Updates are up-to-date except iTunes.

I reseated the RAM so the 1gb dimm is in slot 0 and the 512mb is in slot 1.

The backlit keyboard function is off.

Deleted display preferences at Home > Library > Preferences > com.apple.systempreferences.plist

I tried to run ONYX but it froze before it finished.

I'm repairing permissions but it froze.

So it still freezes. I installed Temperature monitor and got some specs are these too high?

Battery temp 30°c/86°f

Power supply bottom side 55.5°c/131.9°f

Processor bottom side 60.2°c/140.4°f

Processor/Controller Bottomside 47.8°c/118.0°f

Smart Disk 33°c/91°f

Trackpad 33°c/91°f

 

[phreakout]

Okay, a dead PRAM battery shouldn't be that much of a problem. Besides the obvious date & time issues associated with a dead battery, the other would be corruption of the PMU (Power Management Unit); a little computer within the computer that holds hardware settings in a cache. When the battery is dead, it may have some stray currents or under-volts that trigger the PMU, causing it to be corrupted. Kind of like throwing off a bean counter and making them start all over again. So I'd say that a dead battery is to consider, but not critical at this time.

If you do plan on replacing the thermal compound on everything, go right ahead. Use Arctic Silver type, since we trust it; you can get some online or Radio Shack now carries them for about $10 or $11 U.S. Dollars. While you are working under the hood, you can check to see that the CPU and chipset feel and look secure, not pulled away from the logic board due to overheating.

Looking at the temperature specs, the only one I'm concerned with is the processor bottom side @ about 140 degrees F. Can anyone else out there confirm that this is a normal F temp? It seems a tad bit high, although if it were about 160 to 180 or more, then I'd be really concerned. All the other temps look normal.

A bad video data cable (ribbon) going from the logic board to the LCD screen can and will eventually wear or short out, so replacing that might help. But the only way to tell it definitely needs to be done is if the display cuts out while you tilt the LCD full on open and nearly closed. The back-lighting will still be on, but the video signal will disappear and reappear again, once things settle down. A bad cable could send any short circuit signals backwards and screw things up, but I doubt that.

A bad inverter cable going from the bottom half of the PowerBook to the LCD screen can go bad, but when it does, causes the back-lighting to flicker while you open or close the display; similar to the video data cable mentioned above. The other symptom is no back-lighting but you can see the normal desktop, when you shine a flashlight into the front of the screen.

Software is up to date, RAM reseated, but Onyx crashed. Darn (somewhat). A permissions scan is less critical than if you were to spring clean the cache files, which the chipset and/or GPU require for fast storage of temporary info. The permissions simply tell the user, OS and others what privileges you are granted to each file and directory; whether or not you can read, write, modify or erase anything. You can try executing the daily, weekly, monthly scripts in the maintenance category and then move on to the cleaning category, where you can flush out all the cache files. Try that.

Do you have the install discs that came with that model PowerBook? If you do, there may be a disk containing the Apple Hardware Test (AHT). I would recommend booting that and running the tests to rule anything else out. Just a suggestion.

73s de Phreakout. :rambo:

 

[mac-man6]

I have access to ASD disk (version 2.6.3) and the computer passed the test. I thought I would try looping 25 times but it crashed during testing. I'm starting to think it might be hardware problem and very likely heat.

So that limits the possible solutions but before I open up the case are there any suggestions?

 

[BGoins12]

Open it up and fix the GPU's heatsink problem....it's an inherent design flaw that Apple ignored by glopping thermal compound on the GPU so the heatsink made contact with it...over time the compound will eventually spread outward and will no longer make contact with the GPU, thus overheating it and shutting down the computer. Luckily, my compound was still making contact and I wasn't having any problems except the fact that it was running a bit hot (55* C).

I just took my recently acquired 15" 1.67 DLSD/HR apart to clean it out and replace the OE compound with Arctic Silver, and sure enough, the heatsink doesn't even make contact with the GPU. I measured the airgap and found it to be 0.8mm (really Apple? Really?), in which there are numerous copper shims on eBay in that size. I ordered me one and I am going to fix mine. Just remember, it's a bit of a job to pull it all apart.

Here's the one I bought, I think they are the only ones in the US on eBay that sell them cheap. I would measure the airgap before ordering though, they may all be different. All you have to do is attach the heatsink assembly to the logic board. with the mounting screws around the CPU and GPU, then you can measure the gap.

http://www.ebay.com/itm/320747909222

 

[mac-man6]

Thanks Bgoins12. How did you measure to .8mm? Do you have a micrometer? Or do you have a low-tech solution to measure the gap?

 

[BGoins12]

I used feeler gauges.

 

[BGoins12]

Not to hijack your topic, but WHOO! I got my shim today, got it all back together and ran a few benchmarks in a row. Not a SINGLE TIME did the CPU or GPU temp get over 45*C. That is absolutely insane considering before I put the shim in, the GPU ran at 55-60*, and the CPU hovered at 50-55. I'm typing on it right now, and the CPU is 39.5, and the GPU is 41.2*C, much cooler than my 12" PowerBook and my MDD tower.

 

[mac-man6]

No worries, I'm happy. Now I know there's hope. I went out and bought some feeler gauges, I like adding some diversity to my tool box.

Anyway, I was meaning to ask. Did you have plans on tearing down the machine and how much time did it take?

 

[BGoins12]

Yeah, I always tear down any laptop I get to put Arctic Silver on the CPU/GPU to make them run cooler.

Ifixit.com is the site I used to pull mine apart... and since yours is the same as mine, I bookmarked the link -


Follow that, and be sure to sort the screws either by size or by how you pulled them out. I used a old egg carton to sort mine out.

And be careful removing the logic board... the DVI port gets caught on the case and pulling too hard could possibly snap or crack the board. If you've been in laptops before, it'll be cake.

 

[mac-man6]

I've disassembled the logic board. I'm thinking if I need shims or not. How did measure the air gap between the processor and heatsink? Did you measure the buildup or thermal paste or the actual clearance?

Here's two photos of my heatsinks.

http://imgur.com/a/AXcED

 

[BGoins12]

I didn't need a shim on the CPU. As seen in your pictures, the CPU is making contact with the heatsink, while the GPU isn't... it's evident because the compound isn't flattened like it is on the CPU side. The compound itself is the only thing that makes the GPU come in contact with the heatsink. Clean all the goop off, and take the heatsink out of the computer. Mount the heatsink to the board using the screws that go around the CPU/GPU. Then measure the gap between the GPU die and the heatsink. If it's 0.8mm (which it most likely is), then this is proof enough that everyone with a PowerBook G4 that uses this heatsink setup needs to shim the GPU.

 

[mac-man6]

If the space is 0.08cm and the shim is 0.08cm won't it be too big when paste is applied? For example would it be better to have 0.07cm and then have 0.005cm of paste between each side of the shim? Or is the shim put in contact with the heatsink and thermal paste only applied to one side?

Heatsink -> shim -> thermal paste -> GPU?

OR

Heatsink -> thermal paste -> shim -> thermal paste -> GPU?

 

[BGoins12]

Nah, that paste is going to "smoosh" out so thin it won't even affect it. I did a test fit before reassembling mine, and it fit perfectly. The shim didn't slide around, nor was it too big.

 

[mac-man6]

Update news, I have applied the shim and the computer hasn't when using the finder. It has crashed twice using the ASD disk. I'm leaving it running Rember to test the RAM and going to try stress testing the video after dinner.

Some of the problems I found. Attaching the modem cable (I didn't), I have lost a screw and getting the board into place was a pain.

 

[mac-man6]

Update part 2.

It passes both tests using the ASD disk. During some of the test the mouse can become unresponsive for a 30-60 seconds so I thought it crashed. It passes when booted in to OF and OS partitions.

It crashes when left playing iTunes visualizer for longer than 4-8 minutes.