Fazit: Angehende SF-AutorInnen löten/entkernen/reparieren/kennen ihre Schreibgeräte selbst! Selbstverständlich, oder? Habt ihr Anekdoten besonderer Abstürze zur Hand, und evtl. auch wie ihr sie gelöst habt? Erzählt mal. (^ DAEY - do almost everything yourself )LANCE DIES As if that wasn't enough to get me thinking about heat sinks and cooling, a week or so later Roberta reported problems with Lance, her D845PEBT2 Intel 2.53 GHz Pentium 4 (see the February 2003 column) which was intended to last for years. She kept getting messages from Intel Active Monitor (a program that comes with Intel Motherboards and is one good reason for using Intel motherboards) that one of the fans was too slow. The problem would fix itself before I could look at it, so nothing was done. But the next morning she turned on her system, and it wouldn't come up. She called me and I tried it. We'd get the Intel splash screen, then the screen would go dead black—and nothing. It was obviously running, but nothing was happening. I took it upstairs to the workbench and tried again. There were some funny noises from the power supply, and the CPU chip fan was running very slowly. Very. I dismounted the chip. That wasn't easy. Intel did some wonderful engineering on the mechanical system for attaching the heatsink and chip fan to the motherboard—much more on that later—but I would like to find and beat senseless, or at least skin the knuckles of, the man who designed the actual plastic mechanism that holds the heatsink/fan (HCF) assembly in place. It's not that it's hard to get it on properly: it's that it is a painful experience to get it off. This is I suppose fine for the vast majority of systems. Most PCs don't even get their cases opened, much less have their CPUs replaced; but it's different for BYTE readers, and I sure wish Intel had designed an easier way to uninstall the heat sink. I expect that's uncharitable, because for the most part the system works well. Once the clip mechanism was off the HCF still wouldn't remove. The thermal paste had cooked into something like epoxy. Eventually the heatsink came off, with the Pentium 4 chip still attached, and it took careful work with an Xacto knife to get the chip loose from the heatsink. The chip looked all right, and all the old paste cleaned off with alcohol. SLOW DOWN When Pentium III chips overheat they stop. It's different with Pentium 4: When those get too warm they slow down until they're in no danger, but they attempt to go on working. Clearly Roberta's machine had been trying to boot up. Since in five minutes it hadn't got past the black screen to the Windows XP splash screen, it was going to take a long time, but it was doggedly trying. I used some bottled compressed air to clean out the chip fan, put the Pentium 4 chip back in, added new thermal paste, and held the chip fan and heat sink in place. The result was as before: The system would try to start, but it wasn't getting anywhere. There were also funny noises coming out of the power supply. I replaced that power supply with another just like it, a PC Power and Cooling 400 Watt Silencer, and set the system going again, clamping the chip and heat sink down even though it would be a pain to remove. This time it booted up, but Intel Active Monitor reported that the chip fan was far too slow. Interesting. I shut down and got an adapter so that I could plug the chip fan directly into an outlet from the power supply rather than to the fan socket on the motherboard. This time the chip fan spun much faster. The system came up, but Active Monitor was very unhappy, because it couldn't see any fan, and thus couldn't estimate the fan speed. I plugged the fan power back into the mother board. The fan slowed down, not enough to stop the system, but enough to alarm Intel Active Monitor. Periodically it would speed up, then slow down again, and when it did slow down it looked very slow to me. This wouldn't do. Could it be the motherboard? The CPU temperature sensor (i.e. the CPU itself)? Replacing either was drastic; best first to replace the fan. Meanwhile, Roberta was being noticeably patient about not having her machine. COMPUSA I offered to take her to dinner at Crabby Bob's in Burbank. We often go there before going to a movie but it's also close to the CompUSA store. Previous trips to Fry's revealed that their selection of Pentium 4 heatsink/fan systems is limited to massive things that sound like you have a Cessna in the room. I also talked to Bob Thompson about what to order in case I didn't find a satisfactory unit at Fry's or CompUSA. He recommends Dynatron and on his recommendation I have a Dynatron unit in one of my systems. He also recommended Zalman as making very quiet units. The only problem would be waiting for them to come; Roberta really wanted her system now. Fortunately CompUSA had what we needed. The heat sink and fan are relatively massive, larger than the now non-working one that Intel sent with the 2.53 GHz chip. The CompUSA unit is rated for a 2.5 GHZ Pentium 4, and the only brand name is CompUSA. It cost about ten dollars. I brought it home, installed it and plugged the fan power lead into the motherboard. The system came up normally, and Intel Active Monitor believes the fan is running at about 3 times the speed it saw from the original Intel heatsink/fan. I also noticed that the system didn't have any exhaust fan other than the power supply. I didn't think that had overloaded the chip fan enough to wear it out, but perhaps so; in any event while I was at CompUSA I got an Antec "Blue neon" muffin fan (it has a ring of bright blue LEDs) and set it up as an exhaust fan. The little muffin fan is very quiet and adds no noticeable noise. It also serves as a night light. And that's pretty well the end of the story. Roberta's machine works just fine, none the worse for its overheat experience;
Bearbeitet von yiyippeeyippeeyay, 29 August 2006 - 11:42.