[Dylan16807]

There is an electrical circuit that suppresses the voltage spike when you suddenly unplug the cable, to suppress arcing. This improves the immediate physical safety at high power levels, and improves the amount of wear that happens. No physical changes.

[ProllyInfamous]

My lay understanding is that USB-C PowerDelivery isn't even initiated until comms have established the supported wattage? ...or perhaps some very low 5W USB-A-like amount. On sudden disconnect, I presume you're talking about a debouncer (RC) circuit?

----

The concern I have is less about initial arcing (i.e. intentional [dis]connections), and more about long-term sustained powerdraw (I have seen soooooo many melted neutral terminals on 120V receptacles) on a loose connection. Connections become loose for a variety of reasons (including but not limited to bad installation), particularly on thermal throttlers (e.g. small wires, corrosion, cycling).

Does low voltage world have the same 80% derating as insidewireman-land (NEC/AHJ)? i.e. does a 240W PD USB-C allow continues 240W delivery (by protocol/standard/regulator), or is it neutered to 180W for "long-term loads" == 3+hr runtime (e.g. a computer display), with only ≥181W-peaking allowed..?

I just cannot see how such a small connector/cable can deliver sustained 240W, in the realworld that I've lived in.

[Dylan16807]

> My lay understanding is that USB-C PowerDelivery isn't even initiated until comms have established the supported wattage? ...or perhaps some very low 5W USB-A-like amount. On sudden disconnect, I presume you're talking about a debouncer (RC) circuit?

Correct that this is only a worry about disconnects.

> The concern I have is less about initial arcing (i.e. intentional [dis]connections), and more about long-term sustained powerdraw

I think devices usually monitor voltage to make sure there isn't too much loss, and you're probably not going to get enough loose pins at the same time to see dramatic issues.

It's a valid concern, but it's a concern you'd see on almost any type of plug, isn't it?

> Does low voltage world have the same 80% derating as insidewireman-land (NEC/AHJ)? i.e. does a 240W PD USB-C allow continues 240W delivery (by protocol/standard/regulator), or is it neutered to 180W for "long-term loads" == 3+hr runtime (e.g. a computer display), with only ≥181W-peaking allowed..?

They're not worried about heating that takes more than 3 hours, so that specific kind of derating isn't part of the spec.

The 3 or 5 amp limit is designed around continuous load.

> I just cannot see how such a small connector/cable can deliver sustained 240W, in the realworld that I've lived in.

Well for sustained current we're worried about the amps, right? You get the same resistance and heat in the plug regardless of voltage.

Before USB C, we were putting 3 amps over a single pin each way in a USB Micro connector. Now with USB C we're putting 5 amps over 4 pins each way, with the new pins almost as big as the old pins.

[ProllyInfamous]

>you're probably not going to get enough loose pins at the same time to see dramatic issues ... it's a valid concern, but it's a concern you'd see on almost any type of plug, isn't it?

nVidia_12VHPWR_sweating_bullets_.gif

(if unfamiliar, the 12VHPWR is the fire hazard found on some modern GPUs)

----

In the trade-offs of amps verse volts, there are tradeoffs to be made. Yes, I agree that amperage is the primary generator of heat... but is voltage not the primary degenerator of insulations/gaps (particular one so user-interfacing). In a perfect world...

kids_phone_cord.frayed

----

Thanks for the great discussion. I'm learning/adapting. This oaf breaks.things.lots

[Dylan16807]

I'm aware of the nvidia thing. But these particular pins have much less individual leeway, and they're a few mm apart in a pretty tight shell so you can't get the same kind of crooked install.

More voltage has more dangerous aspects, but 48 isn't all that high and in a steady state it's not causing problems.

[ProllyInfamous]

>I'm aware of the nvidia thing.

For anybody unaware: a product has been built specifically to avoid a "rated" connector from melting down brand new perfectly installed GPUs.

[•] <https://www.tomshardware.com/pc-components/gpus/load-balanci...>