Yes, I filed an expense report for a hair dryer, why do you ask?


Back in the late 1990's one of my colleagues (who is now in Office Labs—check it out, they've got some pretty cool stuff) filed an expense report for a hair dryer, and it was accepted. But what valid business purpose would there be for a tester to buy a hair dryer?

At the time, my colleague worked as a tester for Windows power management. One of the things that needed to be tested was whether the motherboard accurately reported thermal stress (translation: overheating) to the operating system and whether the operating system responded appropriately to these reports. And when the project started, the most convenient way to get a motherboard to overheat on cue was to blast it with a hair dryer.

Of course, more official equipment arrived later, but it was definitely a cruel (or clever, depending on your point of view) way to introduce new people to the team: "This is your desk, and here is your hair dryer."

Bonus makeshift testing hardware: During approximately the same era, a joystick tester hooked up the joystick to a series of mechanical linkages, all driven by a record player turntable, in order to stress test the joystick for hours on end. And no, I don't know whether it ran at 33 rpm or 45.

Comments (25)
  1. Adam says:

    I wonder if the Xbox 360 team used this testing method…

  2. Nathan_works says:

    33 or 45 ? I thought everyone ran at 78..

    I’d be really impressed if one of the engineers found a use for an 8-track player in testing ;)

    (and of all the dumb things to keep in the oven…)

  3. Karellen says:

    Nathan > Like she said, it worked fine when she lived alone.

    I’m wondering why she left it switched on when she put it in the oven.

  4. jim says:

    It’s not that unusual. We’ve had the hair dryer out in our office within the last few weeks for an SBC that was suspected of overheating.

  5. Tom says:

    Nathan, you must be too young to remember the last days of records.  33 rpm was for albums, 45 rpm for singles.  78 rpm was on its way out by the 1950s.

    Think of it as follows: 78rpm = DOS, 33 rpm = Windows (9x = original LPs, NT = hi-fi stereo LPs), 45 rpm = Mac.  68.3423421 rpm = Linux.

    People subject their computers to shocking conditions.  Everyday stuff can introduce all sorts of stresses.

  6. Mike says:

    A few weeks ago, I needed to test the microbolometer of a deep infrared camera for spacial defects.  After spending hours trying one thing or another to make a good testing tool for documenting the defect, I gave up, walked to the pet shop down the street, and came back with a $15 reptile lamp.

  7. Sebastian Redl says:

    68.3423421 rpm = Linux.

    You can’t just type in a random number and say that it’s Linux. Linux has two settings, 31.4159265 RPM and 27.1828182 RPM. We’re weird – but with a system. ;-)

  8. Andrew says:

    That’s a good testing hack. Here’s another good example of the lengths to which a good engineer will go when confronted with a novel real world testing example:

    http://woodgears.ca/wood_machines/rotating_machine.html

  9. Brandon Bergren says:

    8-tracks might not have a use whole, but they do contain a few interesting parts.

    I made a working electronic door lock out of the solenoid coil from an 8-track player, a lantern battery, and a barrel key switch once. It wasn’t a particularly challenging hack, but it DID use part of an 8-track player.

  10. Nathan_works says:

    I am too young to remember the change over, but did have a dad with plenty of old 78s and reel-to-reel tape players, etc. I think he actually hooked his reel-to-reel into his box at home and made MP3s from his collection.

    I had the snoopy ‘portable’ record player myself, with the pop-up adapter for different spindle types.

  11. Frode A. says:

    I used to work in the avionics repair industry a while back, and a lot of intermittent errors were troubleshooted with refrigerating aerosol spray cans and hair dryers.

    Do something like this: http://www.facebook.com/photo.php?pid=238905&l=9bb2d&id=642195209 and then point at the general area of the chip you think is broken with a hair dryer, and voila magic happens.

    Andrew: Wow. That is pretty awesome. The Lego one (link at the bottom of the page) is pretty cool too (I’m a sucker for Lego’s)!

  12. computer baker says:

    I once baked a computer in the over at ~150 (the lowest setting on the oven) for about an hour or so, in an attempt to dry it out before stuff started oxidizing. It had been left in the back of a pickup without any covers or the case. The PC worked for a few months, one of the power supply caps decided to blow up, and the CPU fan started making some strange noises.

    sometime later my roomate spilled some beer on his keyboard, decided to wash it out and dry it in the oven. He got impatient and turned up the heat and forgot about it, 5-10 min later we had one melted keyboard. when it cooled everything fused together, it looked sort of like modern art.

  13. NT says:

    At WebTV, a blob of silly putty was a pretty standard issue piece of test equipment.  The engineering and test teams bought the stuff in massive quantities, because when you have 6 remote controlled set-top boxes per person and two people per cubicle, you need a good way to block IR signals on all the devices you’re not aiming at, and masking tape doesn’t make a very good seal.

  14. EricLippert says:

    I was once chatting with a fellow passenger on a shuttle trip. Asked him what he did for a living. Hardware tester. I asked what sort of physical testing they did, and he said that on that particular afternoon he was spending a few hours seeing what happens when you put Microsoft keyboards into vats of liquid nitrogen.  

    I want that job.

  15. steveg says:

    The record player in my house when I was growing up had 16 as well as 33, 45 + 78. It was useful for slowing The Chipmunks records down to normal speed.

  16. We periodically blast our boards with heat guns during testing. Our old heat gun looked like an all-metal hair dryer. We recently bought a new one, but that one looks like a drill (much more focused heat).

    We also use cans of freeze spray. In a pinch, we will use canned air upside down.

    However, we usually only use these techniques when testing a specific component of the board. For board and system level testing, we have a temperature chamber.  

  17. Office Labs – interesting.  Out of curiosity, were they responsible for that annoying little helper paper clip?  <g>

  18. Cheong says:

    Adam: Are you suggesting "wrap in towel" is the "standard working condition" of Xbox360? :P

    The display card I’m using works at high "reported temperature". The reported temperature is 97 – 100 degree celsius even if it is idle, and I just cleared all the dust on the cooling fan. Let me wonder if it’s really that hot or the sensor is broken.

  19. Don Dumitru says:

    When I was working on the original MSN Network, we did dial-up testing. We discovered that the Brand X modems were awful – if you put them in the freezer for a while, they were able to connect reliably.  But when they warmed up, the number of failures climbed massively.

    –Don

  20. Ken Hagan says:

    "We periodically blast our boards with heat guns during testing."

    Me too, but a heat gun is almost too powerful for the job. A hair dryer would actually be more suitable.

  21. CDarklock says:

    were they responsible for that annoying little helper paper clip?

    Actually, that was one of the most visible takeways from Microsoft Bob.

    I’m fascinated by Microsoft Bob. I’m sure there are eighty million useful lessons to be learned from it, but everyone just turned their backs and pretended it didn’t happen. I’d love to get together with the team that built it and pick their brains.

  22. Worf says:

    At WebTV, a blob of silly putty was a pretty standard issue piece of test equipment.  The engineering and test teams bought the stuff in massive quantities, because when you have 6 remote controlled set-top boxes per person and two people per cubicle, you need a good way to block IR signals on all the devices you’re not aiming at, and masking tape doesn’t make a very good seal.

    The Windows Mobile folks (and OEMs) also use tons of silly putty. For one test, it’s required to keep one’s sanity.

    The test is called Mean Time To Failure (MTTF), but everyone knows it as the "Hopper" test – a program called Hopper basically sends random events to Windows Mobile (opening menus, clicking buttons, sending keystrokes, etc). The goal is basically survive random bashing for 24 hours (it often results in crashing due to memory leaks or oddball race conditions).

    The problem is, the preconditions for the test is a clean boot, and hardware as delivered. And it must be 10 runs that survive 24 hours. So most people run 10 devices simultaneously. And since it’s default configuration, you get all the bleeps, bloops, beeps and buzzes happening quite quickly. You can’t disable the speaker…

    Thus, the silly putty (I think there’s actually an official Windows Mobile Silly Putty – the egg shell has Windows Mobile on it) – put it over the speaker, and keep your sanity.

    (Note: Since the system still has to be running at the end of the period, it doesn’t necessarily have to be running well – it can be degraded and still pass. Thankful is the time when the audio driver crashes after an hour of testing and the silly putty is nowhere to be found).

  23. "Me too, but a heat gun is almost too powerful for the job. A hair dryer would actually be more suitable."

    Not for us.  Our products have an operating range of -40C to +85C.  (Which means we frequently test from -50C to +90C.)

  24. Yuhong Bao says:

    "At the time, my colleague worked as a tester for Windows power management."

    Yep, ACPI, APM, and so on.

  25. Igor Levicki says:

    >I wonder if the Xbox 360 team used this testing method…<<

    Owned :-)

Comments are closed.