Richard Cownie (tich@pobox.com) on 9/18/09 wrote:
---------------------------
>Color me skeptical. So you screw your motherboard to a
>rigid piece of metal. Then you turn it on, and it gets
>hot, and it bends. Worse than that, it doesn't get
>uniformly hot, but it has hotspots around the cpu and
>chipset (which are probably also where most of the
>small weak traces go).

If the chassis is a rigid piece of metal, then you don't have any problems. The problem is that $600 laptops tend to be made out of plastic, and without metal support structures. Or they're exceptionally robust because they're 1.5" thick, which makes them undesirable as portable machines. Here's an example of thinness + cheapness resulting in uselessness:

http://www.gizmodo.com.au/2009/05/msi_x340_review_the_unemployed_mans_macbook_air-2/

Your point about hotspots is correct. If you don't adequately cool a chip, there can be excessive stress between the chip-to-motherboard bonds, and they will crack and break. IIRC, the Xbox 360 RROD fiasco was caused by this type of problem. I would think this only strengthens the need for a rigid chassis in the first place.


>But then the anecdotal evidence makes me skeptical to
>start with, because my wife's MacBook Pro seems to be in
>for repairs every 4 months or so - a flaky graphics chip,
>a busted screen, a dead disk, a totally bricked
>motherboard. She beats on it really hard, using it many
>hours a day doing 3D graphics and video editing. But
>it's supposed to do all that stuff without breaking down,
>and it just doesn't.

My Macbook Pro has been off for maybe 250 hours since August of 2006. I beat on it daily, and it's in perfect condition. I've compressed dozens of hours of high-bitrate, 720p h.264 video, which has consumed hundreds of hours of CPU time. Aside from workstations, I've never had a more robust or reliable system. The only thing that causes me grief is the 3GB physical memory limitation. Admittedly, the cable for the magnetic power connector does not have stress-relief ribbing at the connector-end. Instead, it appears to be reinforced with a slip-on ring. I think this is a design flaw, or perhaps a design choice for aesthetic rather than robustness reasons.


>Of course the $1500 laptops are better. They just don't
>seem *that* much better. And when I was shopping in Jan,
>the $700 HP Pavilion and the $400 Acer seemed pretty
>solidly built, while the $500 Toshiba seemed extremely
>flimsy. My impression is that you can find cheap
>machines that are pretty solid, as well as ones that
>are flimsy.

For anyone price-constrained, I wouldn't recommend a Macbook Pro. Although I believe that there is a great deal of value in Apple laptops, I understand that there is also a price premium due to brand. The superior resale value of Apple systems should be noted, as well.


>>http://www.techspot.com/news/23245-lenovo-and-apple-offer-best-laptop-quality-says-rescuecom.html
>
>It's not at all clear from that report how big the quality
>gap is. What percentage of machines need repair ? If
>the answer is 20% for HP, and 5% for Apple, that would be
>impressive. But if it's 15% for HP, and 12% for Apple,
>then you'd hardly notice the difference.

I found the second link from a page which reported the fraction of laptops from various vendors which were brought in for repair. Their data showed that 20-24% of laptops were brought-in for repair, with Apple actually on the higher-end of that range. The nature of the repair work was not disclosed. They then observed the rate of successful repair and customer satisfaction for each vendor, and found that Apple scored 30% higher than anyone else. Their conclusion was that the laptops were all failing at the same rate, but Apple was most successful in repairing the systems, and leaving customers happy.