6

I shorted by accident a circuit creating large voltage to my computer, luckily MBA has some sort of shorting protections -- yet this kind of thing should not happen. I would like to have a fuse between my microprocessor and my computer.

What are this kind of protection mechanisms called? USB fuse did not return anything in eBay so I am thinking I am trying to find things with wrong word.

hhh
  • 1,860
  • 8
  • 25
  • 43

3 Answers3

3

Note: this answer concerns marginal but far from comprehensively effective half-measures for protecting devices. It does not address PERSONAL SAFETY in the slightest, because such was not part of the question at the time this was posted.

USB ports often (but not always) have thermally activated self-resetting fuses to protect against excessive current draw. But that is only one type of electrical risk - it does nothing to protected the data lines for example.

I believe there are full USB "isolators" (likely with optical coupling inside) but these are expensive for everyday use, so likely to be found only in industrial or test setups (for example, a programming pod used while developing motor controls).

Using a cheap sacrificial USB hub outboard of the computer may provide some degree of protection, though far from absolute as there is an electrical connection through the circuitry of the hub.

Chris Stratton
  • 33,282
  • 3
  • 43
  • 89
  • No, using a not so cheap USB hub would provide protection. The cheap ones tend to be simple hub chips + a few passives. USB power is routed straight through. – Passerby Apr 22 '13 at 15:00
  • Interesting, I'm yet to see a hub which doesn't have an external power input which can optionally be used. Or are you implying that the jack is fake? These days anything is believable. – Chris Stratton Apr 22 '13 at 15:02
  • Really? Most cheap hubs rarely have a power input. I don't often see Self Powered hubs, let alone for cheap. – Passerby Apr 22 '13 at 15:07
3

You're looking for a resettable fuse, PTC, or thermistor.

They're basically a resistor with low resistance in normal conditions, but as they heat up, the resistance goes up significantly. This causes a massive voltage drop, stopping the short circuit (until it cools down again). Since it has some resistance, some electricity is burned off, heating the PTC a little bit. As you increase the current, it will heat up more until it trips.

talsit
  • 145
  • 2
  • 8
  • I think resetable fuse is the right thing +1 but eBay returns a lot of different things for it -- are the [capacitor-looking things](http://www.ebay.com/sch/i.html?_from=R40&_sacat=0&_nkw=resettable+fuse&_sop=15) correct? – hhh Apr 22 '13 at 14:01
  • 1
    @hhh Yes, the capacitor looking things are one type of resettable PPTC fuse. – Anindo Ghosh Apr 22 '13 at 15:27
  • +1 thank you, I am trying to get [0.01A fast-blow fuses to protect test-personnel](http://electronics.stackexchange.com/a/66418/2387) and 0.5-1A resettable PPTC fuses to protect devices. – hhh Apr 22 '13 at 15:44
  • 2
    Trying to find a fuse which will trip before dangerous current flows through a person is a very unwise idea. Active measures such as ground fault interrupters, *when applied as intended* may *sometimes* do so, but actual melt-the-wire-with-current fuses will not! Fuses generally will not protect logic-level semiconductors either, though they may protect power devices. – Chris Stratton Apr 22 '13 at 16:07
  • @ChrisStratton if you think so, why don't you downvote? You probably understand this thing better than me. I upvoted but I am not an expert in this area ... can you come up with better safety precautions? – hhh Apr 22 '13 at 16:38
  • @hhh I did downvote the answer where that particular suggestion was made, which you linked in your comment. However, **this** is not that answer and does not make that unwise suggestion. Also, your question appears to be about device protection, not safety, if that is not your intent, you should change it so that you get applicable answers. – Chris Stratton Apr 22 '13 at 16:48
  • @ChrisStratton I hope your input [here](http://electronics.stackexchange.com/questions/66538/why-do-0-02a-fuses-not-protect-a-person-why-no-protection-from-shocks). I think it is better to get this point to a thread. It is now a bit hard to read it over many threads. I currently use no fuse but I had a plan to use fuses to protect people/devices alike before your input, now I am uncertain whether fuses help at all -- I would like to understand this. – hhh Apr 22 '13 at 16:56
  • The only advice I would be comfortable giving you online **is not to attempt such projects**. – Chris Stratton Apr 22 '13 at 16:59
0

Newer USB ports have a protection system most of the time. I once burnt down a processor after shorting a USB port, but that was a very old computer. However, it is a good idea to be sure you don't draw too much current.

An extensive article has been written on this. It addresses a lot considerations when it comes to protecting the USB port.

You can always buy a special chip for this, like the LM3525, which is a power switch and over-current protection in one. They also have the LM3526, which is a dual-LM3525.

  • That is such a broad and wrong statement. USB protection circuits vary by manufacturer, brand, and target costs. I wouldn't even consider half to have usb current protection. Most don't even have usb esd protection. – Passerby Apr 22 '13 at 14:58
  • @Passerby If you place flat end of a fork in USB port and it burns, the part was destined for trash bin anyway. – AndrejaKo Apr 22 '13 at 19:04
  • Why would you do that to a poor innocent fork @AndrejaKo Joking aside, not all electronics are baby proof. – Passerby Apr 22 '13 at 20:40
  • Or are not all babies electronics-proof? It's all about priorities :) –  Apr 22 '13 at 20:41
  • @Passerby Not all electronics are baby proof and I agree that not all should be or that it is reasonable for all to be, but USB is explicitly supposed to be one of the baby-proof things. There's this common quote: – AndrejaKo Apr 22 '13 at 21:31
  • `A USB transceiver is required to withstand a continuous short circuit of D+ and/or D- to VBUS, GND, other data line, or the cable shield at the connector, for a minimum of 24 hours without degradation. It is recommended that transceivers be designed to withstand such short-circuits indefinitely. The device must not be damaged under the short-circuit condition when transmitting 50% of the time and receiving 50% of the time (in all supported speeds). The transmit phase consists of a symmetrical signal that toggles between drive HIGH and drive LOW.`cont.. – AndrejaKo Apr 22 '13 at 21:31
  • `This requirement must be met for max value of VBUS (5.25 V).` Therefore I can conclude that those devices that can't sustain required levels are in themselves dangerous and should be avoided. I do admit that in practice they can be commonly found which is a situation needing rectification. – AndrejaKo Apr 22 '13 at 21:33
  • 1
    @AndrejaKo Might want to point out that that section applies to high impedance data lines, and not to direct VBUS and Ground Shorts. In fact, USB actually makes use of shorting a data line to ground or Vbus (or through voltage dividers) for power supplies. It's not talking about power shorting. – Passerby Apr 22 '13 at 22:50