10 ms·

Not using the proper cable could lead also to fires, for example. We are talking about a significant amount of current. If you don't use a cable rated for that

Not using the proper cable could lead also to fires, for example. We are talking about a significant amount of current. If you don't use a cable rated for that current, the cable or the connector will overheat, and possibly start a fire. And it's not a remote possibility. It happened to me with a fast charging phone power supply and the phone, and the cable was the one provided by the manufacturer! The type-C connector at the phone side was red hot and started to melt, probably caused by a bad connection caused by dirt in the phone connector. Fortunately I was there, and smelled the burnt plastic and disconnected the cable, but what if that happened at night? And it's not something trivial that power supplies and phones can detect, there is not a way to determinate the voltage drop in the cable built into the standard (basically all they needed to do was to add a voltage sensing pin on the connector, to be shorted with VCC at the load side, so the power supply could sense the voltage at the other side and determinate if there too much drop, but as far as I know it doesn't have it).

dragontamer 5y ago
> We are talking about a significant amount of current. Are we? I'm not sure how they deliver more power in this specification. But traditionally, the additional power (from 5W to 20W to 100W) was through additional *voltage*, not current. > And it's not something trivial that power supplies and phones can detect, there is not a way to determinate the voltage drop in the cable built into the standard (basically all they needed to do was to add a voltage sensing pin on the connector, to be shorted with VCC at the load side, so the power supply could sense the voltage at the other side and determinate if there too much drop, but as far as I know it doesn't have it). Surely we just set a current-limit, written into the specification. Then we choose AWG wires / connectors as appropriate to support that current. Voltage can't go up arbitrarily, but voltage can safely be increased to ~48V or so in most applications. After 48V, humans start to get shocked / hurt, so that's probably the reasonable limit. ----- And I'm pretty sure you can sense the voltage drop across a cable. USB-3 delivers 5V by default, and then you send protocol commands to increase your voltage to 12V or whatever. If you detect that the power-supply is only supplying 11V (after the 12V command), then its either a PSU-error or a cable-error. And I'm not sure if it even matters which is which (either way, you're not getting enough power, so your device needs to probably shut off) You can then disconnect / reset your device and maybe stick to 5V default specs.
- dangerlibrary 5y ago
  > Voltage can't go up arbitrarily, but voltage can safely be increased to ~48V or so in most applications. After 48V, humans start to get shocked / hurt, so that's probably the reasonable limit. Car batteries in ICE cars are nominally 12v. Grab both terminals and tell me it doesn't hurt.
  - sjruckle 5y ago
    I've done this. It doesn't hurt. In fact I didn't feel anything besides the battery terminals.
  - StrictDabbler 5y ago
    Ok, done. Many many many times. 12v can't exceed the resistance of human skin so it doesn't matter how much current capacity it has. You can hold onto those terminals all day. It's settled science. If you doubt it, tests of this are all over YouTube. It's just true. Now your tongue, perhaps...
    - amluto 5y ago
      If t you short a car battery, you might not get electrocuted, but the results are unlikely to be pretty.
      - wearywanderer 5y ago
        You can't short a car battery by touching the terminals with your dry unbroken skin, anymore than you can short a AA battery by touching both ends.
      - stordoff 5y ago
        You aren't shorting the battery in this case. There's still a significant resistance (your body, particularly the skin contact resistance) between the terminals.
    - swader999 5y ago
      So my cat doesn't have to worry?
contriban 5y ago
From what I understand, the cable has to advertise the supported power output, it’s not like a raw power outlet that will start a fire if you use a random thin cable. A 5W USB cable will probably never receive 240W, even by mistake.
modeless 5y ago
> Not using the proper cable could lead also to fires, for example Only if the cable or device is defective or damaged. Which is true of every type of charging cable or connector.
- tom_mellior 5y ago
  > defective or damaged. Which is true of every type of charging cable or connector. Yes, that's my experience with USB-C as well. (Seriously, my Dell laptop and Dell dock connected by a Dell cable can't keep up a reliable connection. People are saying that the spec is fine, but a spec that's this hard to implement well is maybe not fine.)
zlsa 5y ago
It should be impossible for any set of non-damaged, non-defective, spec-compliant set of USB-C cables and peripherals to cause fires. The specifications are very carefully designed to prevent any combination of cables and devices from causing damage.
- koube 5y ago
  The Nintendo Switch can be bricked if you use the wrong cable, although my understanding is that it's not built to the USB-C spec: https://www.reddit.com/r/NintendoSwitch/comments/87vmud/the_switch_is_not_usbc_compliant_and_overdraws/ https://www.reddit.com/r/NintendoSwitch/comments/87vmud/the_... Either way, this is an incredibly common device that can be damaged by incompatible hardware where it's difficult to determine compatibility without bricking the device.
  - fouric 5y ago
    > it's difficult to determine compatibility without bricking the device This is the problem! This is what everybody talking about the specs doesn't get - it's difficult to (1) tell which spec a device claims to support and (2) verify that it actually supports it (see: a lot of cheap devices on Amazon) and also (3) in real life many devices do not support the specs (i.e. this isn't a theoretical problem).
    - vokep 5y ago
      Unless I'm mistaken, there is a license to call a port a USB port, though goods are commonly sold without that. In theory the solution is simple: only buy products that have the actual USB logo and ensure certification, and the USB-IF should retain the right to significantly fine any manufacturer advertising being to spec and using the logo who is in fact not. In practice, people aren't going to stop buying nintendo switches, unfortunately. Still, its not entirely necessary to support all specs or have that be entirely clear, what does need to be clear is if you can expect safety specs to be followed. If I plug in my nintendo switch to a charger thinking it'll charge at lightspeed but it takes hours, oh well. If I plug it in and it destroys the device then that should be pretty much an unforgivable problem. Personally I'd be happy to abstain from buying such a device due to that, though also personally I did buy one, expecting no issues, and only found out months later when seeing it mentioned somewhere online. Ideally research is done on every product to ensure things like this aren't the case, but again in practice, that doesn't always happen. I'm not sure what the fix is, other than outright making USB not so universal by requiring a license for any and all vendors using the design, if the license is cheap enough, maybe that could work? I don't know the legalities so much but maybe a free license could be required for any implementation, which would only have the spec of "make sure it doesn't brick charger or device, and make sure it doesn't catch on fire"
unparagoned 5y ago
When I last worked with USB, there were all sorts of resistors over the pins telling you about its capabilities. I'm pretty sure usb-c has that as well. Your device should check the spec of the cable and inly draw the amount if power its rated for
- freeone3000 5y ago
  There's a microcontroller in the cable negotiating the power transfer rate, data transfer rate, and other capabities.
- x0 5y ago
  Is a resistor really enough to determine if a 5W cable gets 240W?
  - spookthesunset 5y ago
    Current is often measured by putting a small value resistor of a precisely known value and measuring its voltage drop. I = V/R
dTal 5y ago
I have also had my phone's USB-C port spontaneously self-destruct, effectively ruining the phone. As you say, burning smells and melted plastic, and I shudder to think if I hadn't been there to power it off and yank the battery. And just this past weekend, a close friend of mine had their Galaxy S6 drain completely to 0 in a minute or two, and then claim water damage in the boot screen (which tests resistance across the port). The phone wasn't wet. Both times, the phone wasn't even plugged in. I am beginning to look very askance at USB-C ports.

dragontamer 5y ago

> We are talking about a significant amount of current. Are we? I'm not sure how they deliver more power in this specification. But traditionally, the additional power (from 5W to 20W to 100W) was through additional *voltage*, not current. > And it's not something trivial that power supplies and phones can detect, there is not a way to determinate the voltage drop in the cable built into the standard (basically all they needed to do was to add a voltage sensing pin on the connector, to be shorted with VCC at the load side, so the power supply could sense the voltage at the other side and determinate if there too much drop, but as far as I know it doesn't have it). Surely we just set a current-limit, written into the specification. Then we choose AWG wires / connectors as appropriate to support that current. Voltage can't go up arbitrarily, but voltage can safely be increased to ~48V or so in most applications. After 48V, humans start to get shocked / hurt, so that's probably the reasonable limit. ----- And I'm pretty sure you can sense the voltage drop across a cable. USB-3 delivers 5V by default, and then you send protocol commands to increase your voltage to 12V or whatever. If you detect that the power-supply is only supplying 11V (after the 12V command), then its either a PSU-error or a cable-error. And I'm not sure if it even matters which is which (either way, you're not getting enough power, so your device needs to probably shut off) You can then disconnect / reset your device and maybe stick to 5V default specs.

dangerlibrary 5y ago
> Voltage can't go up arbitrarily, but voltage can safely be increased to ~48V or so in most applications. After 48V, humans start to get shocked / hurt, so that's probably the reasonable limit. Car batteries in ICE cars are nominally 12v. Grab both terminals and tell me it doesn't hurt.
- sjruckle 5y ago
  I've done this. It doesn't hurt. In fact I didn't feel anything besides the battery terminals.
- StrictDabbler 5y ago
  Ok, done. Many many many times. 12v can't exceed the resistance of human skin so it doesn't matter how much current capacity it has. You can hold onto those terminals all day. It's settled science. If you doubt it, tests of this are all over YouTube. It's just true. Now your tongue, perhaps...
  - amluto 5y ago
    If t you short a car battery, you might not get electrocuted, but the results are unlikely to be pretty.
    - wearywanderer 5y ago
      You can't short a car battery by touching the terminals with your dry unbroken skin, anymore than you can short a AA battery by touching both ends.
    - stordoff 5y ago
      You aren't shorting the battery in this case. There's still a significant resistance (your body, particularly the skin contact resistance) between the terminals.
  - swader999 5y ago
    So my cat doesn't have to worry?

contriban 5y ago

From what I understand, the cable has to advertise the supported power output, it’s not like a raw power outlet that will start a fire if you use a random thin cable. A 5W USB cable will probably never receive 240W, even by mistake.

modeless 5y ago

> Not using the proper cable could lead also to fires, for example Only if the cable or device is defective or damaged. Which is true of every type of charging cable or connector.

tom_mellior 5y ago
> defective or damaged. Which is true of every type of charging cable or connector. Yes, that's my experience with USB-C as well. (Seriously, my Dell laptop and Dell dock connected by a Dell cable can't keep up a reliable connection. People are saying that the spec is fine, but a spec that's this hard to implement well is maybe not fine.)

zlsa 5y ago

It should be impossible for any set of non-damaged, non-defective, spec-compliant set of USB-C cables and peripherals to cause fires. The specifications are very carefully designed to prevent any combination of cables and devices from causing damage.

koube 5y ago
The Nintendo Switch can be bricked if you use the wrong cable, although my understanding is that it's not built to the USB-C spec: https://www.reddit.com/r/NintendoSwitch/comments/87vmud/the_switch_is_not_usbc_compliant_and_overdraws/ https://www.reddit.com/r/NintendoSwitch/comments/87vmud/the_... Either way, this is an incredibly common device that can be damaged by incompatible hardware where it's difficult to determine compatibility without bricking the device.
- fouric 5y ago
  > it's difficult to determine compatibility without bricking the device This is the problem! This is what everybody talking about the specs doesn't get - it's difficult to (1) tell which spec a device claims to support and (2) verify that it actually supports it (see: a lot of cheap devices on Amazon) and also (3) in real life many devices do not support the specs (i.e. this isn't a theoretical problem).
  - vokep 5y ago
    Unless I'm mistaken, there is a license to call a port a USB port, though goods are commonly sold without that. In theory the solution is simple: only buy products that have the actual USB logo and ensure certification, and the USB-IF should retain the right to significantly fine any manufacturer advertising being to spec and using the logo who is in fact not. In practice, people aren't going to stop buying nintendo switches, unfortunately. Still, its not entirely necessary to support all specs or have that be entirely clear, what does need to be clear is if you can expect safety specs to be followed. If I plug in my nintendo switch to a charger thinking it'll charge at lightspeed but it takes hours, oh well. If I plug it in and it destroys the device then that should be pretty much an unforgivable problem. Personally I'd be happy to abstain from buying such a device due to that, though also personally I did buy one, expecting no issues, and only found out months later when seeing it mentioned somewhere online. Ideally research is done on every product to ensure things like this aren't the case, but again in practice, that doesn't always happen. I'm not sure what the fix is, other than outright making USB not so universal by requiring a license for any and all vendors using the design, if the license is cheap enough, maybe that could work? I don't know the legalities so much but maybe a free license could be required for any implementation, which would only have the spec of "make sure it doesn't brick charger or device, and make sure it doesn't catch on fire"

unparagoned 5y ago

When I last worked with USB, there were all sorts of resistors over the pins telling you about its capabilities. I'm pretty sure usb-c has that as well. Your device should check the spec of the cable and inly draw the amount if power its rated for

freeone3000 5y ago
There's a microcontroller in the cable negotiating the power transfer rate, data transfer rate, and other capabities.
x0 5y ago
Is a resistor really enough to determine if a 5W cable gets 240W?
- spookthesunset 5y ago
  Current is often measured by putting a small value resistor of a precisely known value and measuring its voltage drop. I = V/R

dTal 5y ago

I have also had my phone's USB-C port spontaneously self-destruct, effectively ruining the phone. As you say, burning smells and melted plastic, and I shudder to think if I hadn't been there to power it off and yank the battery. And just this past weekend, a close friend of mine had their Galaxy S6 drain completely to 0 in a minute or two, and then claim water damage in the boot screen (which tests resistance across the port). The phone wasn't wet. Both times, the phone wasn't even plugged in. I am beginning to look very askance at USB-C ports.