1
\$\begingroup\$

I'm inspecting a USB battery pack w/ included USB fan, and noticed an interesting quirk in its design. The battery pack regularly outputs an expected 5V at 1A for most USB devices, however with the included fan, it enables a boost up to 10v to get a three speed fan configuration at 5v/7v/10v. Since the product is too cheap to have a slave controller for USB in the fan, I assumed it was using some kind of passive signaling to enable the battery pack to output 10v. Further inspection reveals that the battery pack enables a 10v 1A output if the connected device has both data pins tied directly to Vcc (as it does on the included USB fan).

I'm curious if this design of convenience poses a safety risk / flaw. Since 10v out of USB isn't standard without power delivery, what prevents any consumer device that happens to internally tie D+/D- to Vcc from being fried with 10v by this battery pack? Does it simply rely on hoping that most power-only consumer devices will either leave D+/D- floating or tie them to ground?

\$\endgroup\$
6
  • 1
    \$\begingroup\$ Unless it's in the standard, it's bad practice. Though perhaps not as bad as just outright using the USB connector for another purpose. \$\endgroup\$
    – DKNguyen
    Jul 31 '21 at 16:50
  • 2
    \$\begingroup\$ Why would a consumer device tie those pins directly to Vcc? \$\endgroup\$
    – Dave Tweed
    Jul 31 '21 at 17:03
  • 1
    \$\begingroup\$ @DaveTweed: my main concern is that nothing prevents a, perhaps poorly designed, consumer device from electing to tie the pins to Vcc. \$\endgroup\$
    – omnicious8
    Jul 31 '21 at 17:06
  • \$\begingroup\$ USB signaling is only at 3.3V levels - no consumer device which is actually a USB slave would tie those pins to the USB 5V rail. \$\endgroup\$
    – brhans
    Jul 31 '21 at 19:41
  • \$\begingroup\$ @brhans wouldn't this then be a problem on the flipside -- that is, if I plugged the fan into a regular USB host that isn't 5V tolerant on its data bus, as the spec wouldn't require, it would potentially cause damage to the host? \$\endgroup\$
    – omnicious8
    Jul 31 '21 at 20:06
0
\$\begingroup\$

USB devices are allowed but not required to tolerate a short circuit between supply and data pins.

Plugging that fan to any standard device can potetially damage it.

Also plugging some device that the battery pack just happens to think that has supply connected to data can also cause damage.

Or simply if the pack gets damaged and will decide to output 10V always to all devices, that would do damage as well.

So it seems the whole idea is just hazardous and damage is waiting to happen.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.