[nrp]

It’s designed around 48V and 5A. Higher voltages introduce additional risks around arcing and also get you into different safety regulations and requirements. Above 5A also starts to make cable and connector design difficult for a small connector like USB-C.

[zbrozek]

Last time I looked at the comms protocol, it seemed possible to request up to 10 amps. I'm not sure there's a physical spec for a matching cable, but there's also no way for a cable to communicate its current limitations.

[nfriedly]

> there's also no way for a cable to communicate its current limitations

I don't think that's quite right. USB-C cables that support >60W (>3A @ 20V) must have an e-mark, which is a small chip embedded in the cable that identifies it as being able to support specific levels of power, such as 5A @ 20V for 100W or 5A @ 48V for 240W.

Spec-compliant chargers won't supply more than 60W unless the cable has the appropriate e-mark.

[zbrozek]

Good to know, thanks! I was interested in the comms side and not the cable side since I was considering building a widget that didn't use USB-C as the physical layer but did use PD for power negotiation.

[lloeki]

Nothing prevents anyone to use the USB-PD protocol over a different cable and plug (and maybe some do but we don't know it) rated for 10A. It's just so much more practical both for users and when designing to just use off the shelf USB-C (which AIUI has its design rated for 5A) as the physical part.

[zbrozek]

Thanks! That's exactly why I was looking into the comms side but didn't spend more than a couple moments on the cable side. I was considering building something that had board-to-board connected modules and used PD to negotiate the power between boards.

[nrp]

USB-C cables other than basic low speed 3A ones are required to have an embedded e-marker chip that enumerate the capabilities of the cable.