Slashdot Mirror
BBC Writer Tries PC Repair, Finds Poor Software
twitter writes "BBC author Paul Rubens tried out amateur computer repair and wrote about it. All of the software was for Windows, and he finds what most of us do: "Most of the problems I've been called to look at have been caused by viruses and spyware, some by strange software [conflicts], and only one by faulty hardware." He then flames the whole world of computer repairmen as 'a bunch of unqualified amateurs.'"
Perhaps if they were paid more than your typical McDonalds employee they'd be a bit better than said fast-food dispensers.
I had a discussion with a friend about people in the IT world (I'm not one of them). One works for a large corporation's IT department. He was unaware of SSH, VNC, but was concerned that the wireless router they use in their house doesn't use MAC filtering or WEP. He also has a piece of software phoning home to Toshiba constantly, uses an unencrypted IM client, and gives you a blank stare when you talk about SSH tunnels to a squid proxy.
In this case the unpaid amateur knows 100x what the well paid IT professional does.
As long as the market continues to allow that sort of crap to go on the computing world will continue to suffer as a whole.
Do you have any reason to believe that the power supply that comes with a $40 case is more likely to fail than a more expensive one?
Yes. My sample included 10 computers with hardware failures, about a half of them with cheapo PSUs. 3 of the failures were due to a PSU blowing up, all 3 being of the the cheapo variety.
One of the PSUs took the motherboard with it, and another took a hard drive full of data, and, by some strange fluke, a DIMM socket and a DIMM (the rest of the motherboard still works). This is not real statistics, mind you, but it gives you an idea.
==
A typical failure mechanism is like this. Cheap PSUs can't make it through a brown-out. The H-bridge transistors have to pass higher current to compensate for the lower input voltage, and start to overheat. A good PSU would use heftier transistors, and the controller would shut it down if the voltage dips too low.
In a lousy cheap PSU one of the FETs reaches the point where the silicon starts to melt, and it becomes permanently conductive. Then, the controller switches the bridge, causing a through current that melts the other FET in the same side of the bridge. One of the FETs then vapourizes with a loud bang, leaving a visible crack in the plastic case of the FET. During all this the current through the transformer gets switched chaotically, causing spikes in the secondary windings, and killing the cheap underrated regulators in the secondary circuits, which then pass the spikes to your expensive components. Something like that.
Another problem with poorly built PSUs is that the irregularities in the input sometimes make it to the output, causing crashes and hang-ups. If you want to build a stable system, start with a high quality PSU. It doesn't have to be expensive. Just pick one from a company whose name you (or your friends) recognize.