Wow! Last month's Foolish Things article seems to have struck a chord. After reading about the foolish things various Answer Gang members did, six readers wrote in with their own anecdotes. Let's see what they have to say....
When I was in college, I worked for a small computer store in NYC. This had been a one-man shop (I was #2), and he built Imsai computers for customers who didn't want a kit. He also fixed broken memory boards. One day I came in to find the shop computer broken. He had been testing a memory board (the case of the computer was ALWAYS open with an extender card plugged in). He told me he had just fried the whole computer and it was hopeless.
He had reached for a pair of needle-nosed pliers, and the end of the pliers had become impaled in a steel wool pad that he used for cleaning the tip of his soldering iron. The steel wool pad became airborne and landed in the backplane of the Imsai, setting off a fireworks display.
I pushed him aside and grabbed his logic probe (I passed on the oscilloscope) and started probing the front panel and CPU board. After about an hour I found a dead 7400 on the front panel. Believe it or not, that was the ONLY part that had gotten fried! A $0.25 part replaced and the Imsai was back online.
A friend of mine used to work for one of the companies that made pinball machines (I think Balley). Their first computerized game used a z80 micro. The code was burned onto 2708's (1k x 8 EPROMs). Seems like they had gotten a good deal on these eproms, because they were a little slower than spec. Well the pin game logic checked out just fine in the lab but as soon as one of the machines was put together and buttoned up it wouldn't work. In fact none of the did. So they opened the game up (exposing the CPU board to the light) and it worked fine. Well, it seems they didn't cover the glass windows on the 2708's and when exposed to light those eproms got a little faster and decreased the access time to what the z80 wanted to see. So the fix was to put a light bulb inside the pin game. I wonder about those service calls when the bulb burned out.
A year ago I was building a new system for myself, a PIII 450 (which was replacing my old P166). So, this is the first time I've dealt with a Slot type processor. Anyway, I put the system together, plug everything in, and while the thing is lying on it's side on the floor, I turn it on for the first time. It goes into BIOS POST and then starts booting Linux...in short it works.
So, I power it down, put the cover on the case and set it upright in its final resting place, then turn it on to make sure everything works right...and nothing happens, no beeps, no video init, no BIOS POST, nothing.
So I figure I knocked something loose while I was putting the cover on. I pull it out, lie it on its side, remove the cover, check everything and then power it on (w/o the cover again) and it works. So now I'm confused. I figure now it's running, maybe the cover of the case is doing something, so I put the cover on...and it still runs. Then I set it upright and move it into place...and it locks up. I pull it down again, this time completely confounded, and remove the cover. It starts running again.
I later find out that I didn't seat the processor completely and by standing the case up the processor half fell out of the connector and did a fast system lobotomy. When it was lying on it's side, the processor was seated enough that it worked.
About a week before I had a stroke (that's my excuse anyway), I had to make up a power extension cord for a PC on our office network. The network consisted of about nine PCs connected via coaxial cable.
I was working as a Computer Service Engineer at the time, and one PC had to be moved to a location where the power cord was about two meters too short.
While making up the extension cord, I unwittingly transposed *active* and *earth*, <always triple check!> and then stupidly just plugged the PC into the new power extension cord!
There were cries of shock throughout the office, as the mains current raced around the coaxial cable, from one PC to the next resulting in melted NIC connectors and sparks. A co-worker later described as them as "an angle grinder throwing sparks behind my PC".
We were all shocked at how easy this was to do, and fortunately for me I managed to repair 8/9 NIC cards, and no PCs were damaged due to the isolation at the NIC connector.
[And no animals were harmed in the testing of this cable. :) -Mike.]
Many years ago--about 2years BIPC (Before IBM PC)--I bought a kit Z80 single board called Nascom.
As expected, on completion it didn't work. I traced everything on the board and found two tracks that "didn't make it". Scrape the varnish off the board, and two lengths of hair thin poly-coated wire , and the tracks were fixed.
I found that it only worked for about 30 seconds. The guess was the Z80 -- so I ran it for a week with a tobacco tin into which I popped an ice cube. I imagine that with today's processors, the ice cube would sublime instantly using the coefficient of fusion and the coefficient of vaporization in one go.
Sitting on the fence is not always a good strategy
Look what happened to Humpty Dumpty.
I wouldn't classify this as foolish - just warped but practical. In the very early days of home computing, the first machines on the British market to be cheap and (very basically) usable were produced by Sinclair. Initial models had about 16k of RAM. But you could buy an upgrade to 48k. One snag with the upgrade: the computer ran too hot and began misbehaving.
Someone realised that putting something cool on the right area of the case (I'm talking ice here, not music) would keep the temperature under control. The "something cool" he had to hand was a pint of milk in a cardboard carton straight out of the fridge. He must have written this up somewhere, because it soon became the standard fix for hot Sinclairs!
For economy, the carton went back into the fridge to cool down for the next session. It was a very good idea to mark which carton you were using, of course...!
My foolish thing was misreading the little book that came with my motherboard and inadvertently overclocking the CPU when I was putting the wretched thing together.
I wish I knew precisely what happened. Anyway, that CPU - and probably that motherboard - is headed for the rubbish bin as soon as I can afford a decent replacement for it.
One day I was trying to figure out what this strange scratching sound coming from my computer was, and I eventually narrowed it down to the hard drive (no-name brand Pentuim3 with too much stuff in it) So being the young know-it-all that I was I decided to see if I could fix it. I started by removing the top cover of the hard drive, and found inside that one of the read/write heads was slightly stuffed and the majority of screws were loose. (They just have to use those silly triangle shaped screws so no one can have an easy time, don't they!) From that point I decided to replace the read/write heads and tighten all the screws, and so i did. Eventually it looked sort of normal again, so i closed it up and took it for a test drive.
I was successful, and it is still working with Linux Mandrake today. From 2000 - 2001, only 3 crashes so far. Fingers crossed! I now know that HDDs are made in laboratory conditions but I also know you can fix em and still have them be just as reliable. p.s: Ice is a really good idea except for it melts! Try Gel packs for sports injuries. They are in super-strong double lined plastic. You put em in the freezer for a couple of hours and they stay cold for about six hours. Just throw it on yer motherboard and your heating problems are solved. Sold at chemists.
That's it for this month. The need for cooling seems to be a strong theme. Perhaps I should call this series "The Foolish Ways We Cool Our Computers".
If you would like to tell us about the most foolish thing you've done with your computer, maybe we'll publish it.