pulling 5v from pc usb port

[Mp3]

Let's say i take a pair of diodes and cut a usb cable and add the diodes to the Vcc and ground, and then clip stuff to the diodes to be powered by the 5v (or 4.5v, whatever it will be after the diodes)

Is there any reason I shouldn't do this, if i be sure the wires don't short? any risk to the pc or other devices on usb bus if i plug it in backwards?

[narkeleptk]

Yes it will work as long as your device does not exceed the current output of your usb. You don't need the diodes. Just add a connector of some type at the end.

[bobbydazzler]

It will work I don't understand why you need the diodes though?  Also if you use a usb3 connector the wire will be slightly beefier and able to supply more current, I believe 900mA is usb3 spec.

[fordem]

You may find that USB-C won't provide power that way - the USB-C spec allows for power to be delivered at three different voltages, and some devices will not deliver power when they cannot determine what voltage is required.

[tooki]

You may find that USB-C won't provide power that way - the USB-C spec allows for power to be delivered at three different voltages, and some devices will not deliver power when they cannot determine what voltage is required.

But it has to deliver 5V in order for a device to power up enough to negotiate what voltage and current it does want. IIRC a device is allowed to draw up to 100mA without negotiation.

Also, it’s 4 voltages: 5V, 9V, 15V, and 20V. The newest revision adds 28V, 36V, and 48V, as well as support for variable voltage from 15-48V in 0.1V steps! See link in next post.

[tooki]

It will work I don't understand why you need the diodes though?  Also if you use a usb3 connector the wire will be slightly beefier and able to supply more current, I believe 900mA is usb3 spec.

1. USB-PD (power delivery) doesn’t require USB-C. The higher voltages and currents can be negotiated over traditional USB connectors, too. It’s just not very commonly seen.
2. It can be way more than that: the current real-world maximum is 20V at 5A, but the most recent USB-PD standard allows up to 48V at 5A!! cf https://www.eenewspower.com/news/usb-pd-boosts-usb-c-power-delivery-240w-48v#

[tooki]

Let's say i take a pair of diodes and cut a usb cable and add the diodes to the Vcc and ground, and then clip stuff to the diodes to be powered by the 5v (or 4.5v, whatever it will be after the diodes)

Is there any reason I shouldn't do this, if i be sure the wires don't short? any risk to the pc or other devices on usb bus if i plug it in backwards?

You can do this, but I would only do it for very small loads. Without negotiating, any load above 100mA theoretically risks the USB controller shutting it off, or even cutting power to all the USB ports. (In practice, up to 500mA is unlikely to be a problem. But more can be.)

Unless you have a very specific need for a device to be powered from the computer power supply, which it doesn’t sound like you do, I would use a separate power supply like a (high quality) USB charger or even better, a lab power supply.

[David Hess]

The limit is suppose to be 100 milliamps without negotiation, and there are also limits on the amount of capacitance without current limiting which can be used.

But in practice a lot of devices, including Fenix flashlights, ignore the specifications and draw what they want from any USB port without negotiation.  I was pleased to find that my new USB charger for Milwaukee power tool batteries properly enumerates power from a USB port; that is the exception now.