Hacker News Reader

Comment by ajays | Hacker News Reader

Tangential: when I saw the iPhone power supply, I was blown away. Here was this little white box, about 1-inch cube, able to convert 120V AC into low-voltage DC. No tranformer; no big capacitors; nothing. Just some solid-state electronics, and bam! Out comes some sweet DC juice.

I wish all DC equipment came with such power supplies. Heck, while I'm at it: here's a great idea. Establish 2-way communication with the solid-state power supply, so that the equipment can tell the power supply what it wants, and the solid-state supply can then give it out. Then you won't need a separate power supply for every piece of equipment!

View on HN

There's actually a tiny transformer and even tinier capacitors and inductors in these phone power supplies. The transformer is super-important, because it isolates the AC line from the output, which is necessary so you don't electrocute yourself on the power supply. AC switching power supplies from tiny cell-phone chargers to big computer supplies are split by the transformer into a primary-side connected to the line, and a secondary side connected to the output. For safety and by law, there can't be any direct electrical connection between the primary and secondary side, so even the voltage feedback control signals need to be isolated with small transformers or optoisolators. And even any circuit board traces between the sides are required to be separated by a couple millimeters.

P.S. Thanks everyone for the comments on my article.

Thanks for taking the time to do such a thorough write up.

Are you sure the iPhone power supply has no transformer? I haven't cracked one open but I'm guessing it has one. Increasing the switching frequency allow a reduction in core size, and with frequencies in the 1 MHz range these days the transformers are pretty small. (Further increases are not recommended due to increased hysteresis loss.)

The 2-way communication is already there, your wish has been fulfilled. Okay, not all DC equipment has it, but that's how USB works. You plug into a USB port you get up to 100 mA free, you can negotiate to increase the limit up to 500 mA. Apple's iPod/iPhone/iPad chargers put nonstandard voltages on the data pins to indicate to the device that they can supply some larger amount of current, like 1A or 2A or whatever. Different voltages on the data pins correspond to different current limits. (So for the chargers it's one-way communication, not two-way like USB, but the two-way communication isn't needed for chargers that only plug into one device.) This isn't the first time Apple messed with the USB power specs, if you had a G4 cube back in the day you'd notice that the speakers would only work if plugged directly into the Mac, which supplied 800 mA from its USB jacks.

Personally, I think it's pretty dumb that USB only supplies up to 500 mA -- that's just 2.5 W. Meanwhile, Firewire could supply up to 45 W, which is why Firewire hard drives don't need to plug into the wall.

"Further increases are not recommended due to increased hysteresis loss."

This is why I am REALLY hoping the materials revolution everyone keeps talking about takes off and we get some decent high frequency, high permeability, low loss, high flux density ferrites out of it (not too much to ask huh?). You can do lots of cool stuff with magnetics, even beyond just power conversion but the materials to do it really well just dont exist.

I said hysteresis loss, but that's just the first barrier. Higher frequencies are also subject to the "skin effect" which requires thinner wires, therefore more wires or ones with lower resistivity (e.g., superconductors). Then you need to get transistors capable of handling the increased switching frequency, and you need to be able to deliver the increased drive current to the transistors.

You don't need superconductors to get around the skin effect. It's actually interesting that you mention superconductors, as the skin depth in a perfect conductor is actually zero. (Anyone know what happens at DC in theory in a perfect conductor? I originally thought it was zero at DC as well, but to me skin depth seems like it becomes undefined.)

Even without superconductors, at 100 MHz the skin depth is still around 2 thousandths of an inch. Litz helps too. High power RF transistors exist. All the things you mention are limiting factors, but I still say the lack of good magnetic materials is the dominant barrier to higher frequency converters.

> Establish 2-way communication

The middle pin on a magsafe power supply is used to send data two ways. It uses the 1-wire protocol and passes along (among other things), the serial number of the charger, the amperage of the charger, the power type (so the computer won't try and charge using an airplane adapter). Negotiation happens between the charger and computer for power and the LED color.

Gladly the industry seems to be (slowly) standardizing on micro-USB plugs, except for laptops. That means nowadays I only have to carry two bricks with me (MicroUSB + Macbook charger).

Personally I'd really love if everyone moved to a common MicroUSB MagSafe form-factor. But apparently there are some patents in the way...

> Gladly the industry seems to be (slowly) standardizing on micro-USB plugs, except for laptops

Largely because the EU pretty much threatened the cell phone manufacturers with heavy regulation unless they came up with some sort of standard. Since they "have" to do micro-USB for the EU, they might as well do it everywhere.

China required this in 2007. EU didn't move on this until 2009.

Unfortunately power is still a problem. While there is part of the standard for power, Apple went the other way so high power chargers have to support Apple or open standards but not both. Todays tablets have very high capacity batteries (and they keep getting bigger) so various manufacturers have been going proprietary again in order to increase power from the charger (Samsung Galaxy Tab 10.1 is an example).

Here is a post showing someone trying to get a standards following USB charger:

https://plus.google.com/117091380454742934025/posts/5suTrJrp...

I've got one of these: http://www.avalontest.com/e-store/media/prod_full/hp_6236b.j...

It's terribly useful.

What you do is read the voltage you want from the device, turn the knob to the voltage you want and plug it in!

It was made in 1977. It works like the day it was bought i.e perfectly. I don't imagine a cheapy switch mode supply making it to 35 years old.

With regards to the "sweet DC juice" - it'll be full of noise if it's a switch mode supply. They're horrible devices that cause all sorts of local RF interference, noise and transients on the mains. Sometimes more than chucking an inductive load like a hoover on the line.

The only positive things is that they are small and cheap. They are definitely not good.

Also a switch mode supply does have a transformer! Just a small one.

Hey, not to shoot you down but there is definitely a transformer in the iphone supply. Second, building switching supplies that can supply an arbitrary output voltage/current isn't practical or economical. Third, don't you want a power supply for every piece of equipment? You don't really want to have to unplug one thing to plug another thing in do you?

I agree with you it probably isn't economical, at least not yet. But the reason for wanting it is for the power supplies to be arbitrarily interchangeable.

One of the big advantages of the move towards micro-USB charging cables for phones, for example, is that in my house we now have several adapters spread around the house, and both my wife and I can now charge our phones at any one of them, despite having different brands.

I can see the advantage, but I still would argue that just having some DC power standard like USB will always be more economical and practical than trying to build "do it all" supplies to work with all gear despite the absence of a standard.

I'm not sure what problem variable voltage is supposed to solve.

The power supply that really blew my mind was this even smaller one for iPhone:

http://storeimages.apple.com/2106/as-images.apple.com/is/ima...

Is it just me or is that actually bigger than the US one?

http://www.empetel.es/WebRoot/StoreLES/Shops/17205045/4CB1/B...

http://images.apple.com/support/usbadapter/exchangeprogram/i...

The UK one is also pretty incredible:

http://www.handtec.co.uk/shopimages/products/normal/Apple_Ma...

I like how the UK one directs the cord in a direction that avoids unnecessary bending (compared to the europlug-one)

It's small in volume but including the rigid piece of the usb connector, it can become quite a lever when the cord is pulled up or down. I wonder whether that causes mechanical problems. Edit: For comparison, the slightly larger iPad power supply supports itself against up/down motion by resting on the outside of the plug.

This is a great idea. Unfortunately, I think I remember one of my electronics lecturers at uni telling us that generally communication/feedback with power supplies is dangerous because you introduce the possibility of unexpected power surges propagating in the other direction. It's apparently very hard/expensive to have both communication and protection against this. I'd love for someone to correct me, especially as this might not apply to the newer designs.

An optocoupler will isolate the communications themselves, but perhaps he was referring to the fact that the overvoltage 'crowbar' circuit would need to be software controlled. Overvoltage protection shuts down the power supply when the output voltage rises too high - say if the FET gets stuck on.

OTOH one could put additional protection downstream in the load itself, and the PS would just need a well-known fixed overvoltage. If the "5V" PS started outputting 15V (but still under the PS max), the device would just shut down. However, don't expect manufacturers to do this on their own - as the article says, $1 is "expensive".

> Just some solid-state electronics

Just so you are aware, the Palm V which came out in 1999 had just such a power supply. A bit bigger than 1 inch, but it's a switching power supply and doesn't get warm (i.e. it's efficient).

So did the Newton (1993 per Wikipedia). I remember it as my first-ever encounter with a (small) switching power supply — it seemed unreasonably light as “wall-warts” went.

Reply on news.ycombinator.com