3

Here comes a computer at school, running XP.

A classmate actually use a USB Flash Drive which is designed badly that every time you insert it to a USB port there is a high chance of short-circuiting the USB port.

Well, short-circuiting the USB port isn't very big deal. The most it will do is trigger the current overload protection, and Windows will pop up a bubble telling you the USB port has been overloaded (can't find an image of it). Not kidding, I've actually tried this a few times when getting 5V from it.

However, when my classmate insert her badly-designed USB Flash Drive (and of course short-circuiting the port), Windows BSOD and reboot.

My question is: How could this actually causes a BSOD and reboot the computer? Shouldn't current motherboards/PSUs have enough current protection to handle short-circuit events? How could this kind of hardware exceptions trigger a "software exception"? Is that only an extreme rare case happening only to that computer?

Because I frequently need to get 5V from the USB port to power circuits that is possible to exist a wiring error short-circuiting the power source, I need to make sure that the current protection will work on my own computer so that even if there exists an error, it won't crash my computer and make me lose all opened documents and webpages. Till now the most serious situation I've met on my computer is the Windows bubble notice, but after I saw the school's computer BSOD-ing, I am quite worried of whether it is safe to use the USB port to power my hobbyist electronic hardware.

2
  • The USB specification required that devices connect in a low-power mode. That limits them to 0.1Ampere (at +5volt, so max 0.5 Watt). That is not a lot of power. Every device is that allowed to negotiate with the host and ask for permission to draw more power. The host can either grant this, or deny this. If you skip this step and simply draw more then you are violating specs and whatever happens is up to the designer of the motherboard. As far as I see a burned out motherboard would be fully without specs and you are lucky if they added additional protection and warnings.
    – Hennes
    Feb 3, 2013 at 15:25
  • Regarding the bubble on overload: search "power surge on hub port".
    – bwDraco
    Feb 3, 2013 at 20:07

2 Answers 2

2

Anything that happens to hardware could cause a software exception... if your software is buggy. (And/or if your hardware is buggy :). I'd suggest not trying to mix general computer use with powering experimental circuits.

I used to use cheap mains-powered multi-adaptors. They included a 9V-battery-style connector that was very easy to attach crocodile clips to. (Failing that, there were also power jacks that looked a little like headphone jacks).

[That said, many modern programs will perform autosave. Modern productivity apps should do it. Firefox does it every 15 seconds (!). And modern FS's don't wet themselves on power failures. So you might not be risking too much in that respect].

1

I think the problem here lies more with the USB device than anything else. So what if it's BSODing when you short connectors out? I would actually expect that to happen.

1
  • 2
    +1. This question is an instance of "why does this thing break when I do something really unsupported that is commonly considered bad/wrong?" It answers itself. A permanent solution would be to use an AC-to-USB adapter and a potentiometer when at an early enough debugging phase in your project that it could actually short.
    – Zac B
    Feb 3, 2013 at 19:19

Not the answer you're looking for? Browse other questions tagged .