Comment by userbinator | Hacker News Reader

Comment by userbinator | Hacker News Reader

Parent Full thread

Unless some radical change happens to laptop power supplies in the next few months - a change from designs that have remained the same for over a decade - it's very likely this adapter will work with the newest (normal) laptops available when it comes out.

Also, laptops have very flexible power input circuits that will accept a wide range of DC input: they'll start running from ~12V (the voltage of an almost-dead battery), and go as high as the voltage specs on the input components allows - 25V (filter caps) is the usual absolute upper limit. Higher voltage means lower I^2R losses but manufacturers like to leave some safety margin, which is why 20V is somewhat of a de-facto standard given the parts available. The general power distribution circuits in a laptop are like this: there's a main power rail, whose voltage varies with battery voltage when running on battery, and is the DC-IN voltage with the adapter plugged in. The voltages for the CPU, GPU, chipset, etc. are all generated by DC-DC converters supplied from this main rail. There's a set of diodes/MOSFETs that wire-OR together the battery and DC-IN to the main rail, preventing current flow in the wrong direction. It also powers the battery charger.

I've read a lot of laptop schematics, and all of them use very similar power circuitry, often with the same components. The only real concern with device compatibility is in the ID schemes that some manufacturers use, either to "lock out" third party adapters or encode wattage selection. The simplest ones are nothing more than a resistor in the plug, different values for different wattages (IBM/Lenovo is one example). More complex ones use a 1-wire interface to an EEPROM or other storage element (Dell, Apple) (http://www.laptop-junction.com/toast/content/dell-ac-power-a... http://www.righto.com/2013/06/teardown-and-exploration-of-ma... ). Then there's all the others that have no ID at all, just +/- input, and these are the most compatible.

About the Dart itself: the technology used here (VHF SMPS) has been around for a long time in military/aerospace applications, where size/weight is more important than reducing switching losses.

http://www.eevblog.com/forum/crowd-funded-projects/vhf-lapto...

"Unless something radical happens" - well yeh, the USB spec now supports the wattage required to charge laptops so I think that's a pretty radical change which will be coming in the next few months.

I think if you took a look at the timeline between USB spec proposals and deployment times, you wouldn’t be so optimistic.

I would be shocked to see even one laptop ship with USB power before 2015.

Is this a USB 3.1 supply? Because that is where everything is going.

For small mobile device, maybe. However, this project is for laptops, and laptops don't charge themselves over USB, nor will they any time soon.

No. USB 3.1 supports power delivery at 60w or 100W, easily enough for most laptops. Yes, they aren't used at the moment, but it's definitely presumptuous to suggest that laptops won't use USB any time soon when the technology is so nearly in place.

That is simply not true, the USB spec now supports charging of laptops over USB and I would be very surprised if over the next few months it didn't become fairly common in new laptops - it adds an extra USB port to a laptop and is very inexpensive.

Charging via USB adds extra complexity and there is nothing wrong with the existing system, so I don't think it will have as much adoption as you predict. OEMs rely on manufacturers like TI to provide ICs and reference designs, and for the moment it doesn't look like there's a highly-integrated solution for USB-PD.

> and there is nothing wrong with the existing system

Yes, there is: it causes massive electronic waste because of incompatible chargers. Expect the new EU regulation that covers phones etc. (that replaced "voluntary" measures) to be expanded to cover laptops too, or at least for manufacturers to get threatened with changes to th regulation, if manufacturers don't voluntarily switch to USB once suitable chargers start being widely available.

There is nothing electrically wrong, and while I agree that there have been many different connectors used and having one standard would be great, manufacturers do tend to settle on one of relatively few standard connectors (the majority being barrel plugs of various sizes.)

However, the solution is not a grossly overcomplicated standard like USB. It is a 20V barrel plug. Maybe if you want to get really fancy, include a resistor on a third pin for wattage encoding. Standardise on that instead - or pick an existing size - there are enough to choose from already ( http://en.wikipedia.org/wiki/Coaxial_power_connector#Standar... ).

This is a somewhat uninspiring attitude. There was nothing wrong with the PS/2 mouse and keyboard plugs either, nor the 25-pin printer plugs, nor some barrel plug for charging phones. But the problem is there were so many different things that all needed their own plug. Now there's no mouse plug or phone plug or printer plug since they all use USB. Laptops with a standard non-USB plug will still add to the duplication and inconvenience just like those old keyboards did. You won't be able to charge your phone with your standard barrel plug laptop charger, so you'll still have to carry two adapters when you travel.

There was nothing wrong with the PS/2 mouse and keyboard plugs either, nor the 25-pin printer plugs, nor some barrel plug for charging phones.

With the exception of the phone charging example, I would agree. The PS/2 protocol is vastly less complex than USB, and parallel ports are so simple they can be used as GPIOs with ultra-low guaranteed latency. Phones migrated to USB because they already needed the bulk data transfer capability along with power, and USB made for a good fit.

You won't be able to charge your phone with your standard barrel plug laptop charger, so you'll still have to carry two adapters when you travel.

The Dart has both USB and a barrel plug, which is the same as many other aftermarket adapters. (They could've made this more obvious in the marketing material, e.g. by showing USB devices plugged into it and not just laptops most of the time.)

I'll choose a laptop with a big robust barrel plug, and one that can be powered from anything outputting 12-24VDC, over one with a USB socket and requiring complex negotiation protocols to even start consuming power. It's the KISS principle. Standards that try to do everything don't tend to do any one thing particularly well, and introduce unnecessary complexity.

You don't need to implement full USB for USB power delivery. USB's data transmission is not used in it at all.

Detection of power delivery plugs is mechanical for standard A connectors (their male standard-A connectors are 1.3 mm longer so they could be inserted deeper in female power delivery connectors than conventional connectors so metallic shield connects detection pins) and based on detecting resistors and capacitors between ID pin and ground and power pins for other connectors.

For voltage negotiation new simplified bus and protocol was invented; it uses high frequency transmission on the same pin as power.

Traditionally all manufacturers have been using a little bit different connectors on the laptop. Also tho voltages seem to differ a little bit. Hard to see any other real reason for this than to make switching to another brand a little bit harder and to make a few extra $$$ by selling accessories.

So even if adding a new USB port for charging is cheap and easy, the manufacturers might still have reasons not to do it.

Why do you think laptops wont use usb 3 power like everything else will?

Asus Windows 8 laptops use USB to charge.

IIRC USB Power Delivery requires to provide 5V and 12V if you want to provide 20V. According to their FAQ question about compatibility with Microsoft Surface and Chromebook Pixel (https://www.kickstarter.com/projects/215201435/dart-the-worl... ) they don't have 12V support. Probably they are not ready for USB 3.1.

I'd guess 12V is for lower-power devices, where (slightly) cheaper 16V-rated (the next step below 25V) capacitors can be used. The other parts (switcher ICs and MOSFETs) are usually rated for much higher, and the system would be fine with a 20V input if it weren't for the 16V caps.

E.g. the EEE 1000HE uses a 12V adapter, and it has 16V caps in its power circuit. Probably the same for the Surface/Pixel.

Reply on news.ycombinator.com