Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Such a thing would have to be really smart about advertising power-delivery ability for each port without exceeding a total commitment of 1,800 watts (US 120 volts x 15 amps), but not actually needing an inefficient 1,800-watt AC/DC transformer, and then somehow still managing to be able to charge at least a couple high-power laptops simultaneously. All this to satisfy the reasonable desire to replace six bulky 15-watt transformers with one simple device.

Contrast your typical $10 AC power strip, which lets you plug in 10,000 watts of hair dryers at once, because it's understood that a circuit breaker will (hopefully) blow if you do. It doesn't need to be smart.

I think the basic problem is whether USB-PD allows a source to commit to delivering 100 watts to a port that asked for 20V/5A, but then notice that the port only needs 2 watts right now (because the device finished charging), so that it can reclaim 98 watts in its total power budget. So far I haven't seen a power supply that does this well.



Not sure what's the exact problem you talk about.

If a USB-PD over commits, then the devices it tries to power will obviously get less power than advertised, or the internal over-delivering protections trigger and the source shuts down. The user will (maybe) notice and can do something about it.

I'm pretty sure everybody understands that you can't just plug 6 high power consumers in one USB device just because it has 6 ports and it supports USB-PD, just like people don't plug 6 kettles in one extension cord.


People absolutely do plug tens of thousands of watts' worth of appliances into a single 15-amp home circuit that can supply only 1,800 watts at once. And as long as they don't exceed 1,800 watts of momentary usage by too much or for too long, it works. AC power doesn't have a "commitment" in any sense but a physical plug shape. It just has circuit breakers to prevent fires.

A good example of the problem is a hypothetical USB-PD power strip that has 5 ports. The Amazon listing says it can deliver 120 watts. A person buys it, thinking they can use it for their 100-watt laptop and four other small things that each need barely 5 watts, 1 x 100 + 4 x 5 = 120, so the math adds up. They're still working from home post-pandemic, so they leave their laptop plugged in all the time on their desk. The laptop charges for maybe 45 minutes each day, but it still tells the power supply that it needs 20V x 5A = 100W constantly, and each of the small devices asks for the minimum of 5V x 3A = 15W.

For actual USB-PD supplies I've bought in this approximate situation, the laptop won't charge because the supply can't promise 100W. You hook up your Kill-A-Watt and find that the supply is actually drawing only 20 watts. The power supply is keeping all its promises -- 15W to each of four small devices, meaning it can't agree to supply 100W to the fifth device. So it instead says it can do maybe 12V x 3A for the laptop, which says forget about it and refuses to charge. Yeah, the laptop sucks for lying that it needs 100W 24/7, but what am I going to do? Get a new laptop? Or just put the USB-PD supply on a shelf and return to a regular old AC power strip with five individual USB transformers plugged into it? That's what I'll do, because that's what works.

In today's world, the "If" in your comment happens at the AC circuit-breaker level. In the USB-PD world, it would happen at the level of this little multi-port power supply. Your average consumer in a hurry doesn't want to be bothered by this kind of detail.


It seems there are more alternatives. Use the USB-PD supply for the laptop and another one for you 5 devices.

It's a bad laptop design. An iPad or Android phone/tablet can fast charge with a good supply or slow charge with a cheap one. The laptop should do the same, accept whatever the supply can deliver.

Also, what's the problem with the user using 5 individual small USB chargers versus a big one? The efficiency of all USB chargers is about the same, so it doesn't matter how you split them up. In fact, a small USB charger working at maximum output is more efficient than a large one working at 20% load.

You are correct that ideally the USB-PD supply should monitor the actual usage and route power accordingly, but I'm not sure it's such a problem in practice.


what's the problem with the user using 5 individual small USB chargers versus a big one?

:) This thread started with iamchp asking about "single multi-port USB-C PD converters."

The laptop should [...] accept whatever the supply can deliver.

And both sides should renegotiate when available/needed power changes. If a laptop could say "I could use 240W but I can deal with 18W," a supply said "I can give you only 18W, but I'll let you know when I can give you 240W," and the laptop later said "you know what? Now I really need only 36W," then I think USB-PD would be good. My experience is the negotiation happens once, possibly days earlier, when the device is first plugged in.

(By the way, it's not always the right product decision to include boosting circuitry from lower-voltage supplies. While a MacBook will charge from a 5-volt supply, you'll be sad when it dies during a critical work presentation even though it's charging, because the presentation needed more than 15 watts. It's a defensible product decision to require a minimum wattage, and thus effectively a minimum voltage, or refuse to charge.)


It would be pretty neat to have some kind of refresh interval, where both devices list their current capabilities and have it renegotiate to some optimal level.

As for the refusing to charge, my ThinkPad on Windows will alert me when the dock or power source isn't putting out enough power for it to maintain the current power usage trend.


Wouldn't that kind of overcommit be disastrous though?

If a device was granted a 100W budget, but only needs 2W now, maybe it suddenly needs to burst up to the full 100W? If the hub reclaimed those other 98W, now it can't deliver what it promised, so something fails to work, or gets powered off.


Depends on your definition of disaster.

In the hair-dryer case, all six actually do turn on -- but only for a few seconds. The circuit breaker heats up and trips before the wires in the wall get hot enough to ignite the wood. Which sounds scary, but that's part of the design that enables normal spiky power usage of refrigerator compressors starting up, running a garbage disposal for a few seconds, etc., without constantly tripping a home's breakers.

If USB-PD has some way for a single supply to "overbook" its commitments, then it could model this system. Otherwise, its competition will be people plugging lots of individual USB power supplies to AC power strips, which is bulky and wasteful, but extremely functional.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: