Edison's revenge | Hacker News Reader

Edison's revenge

economist.com

72 pointspclark19 comments

If the miraculous ABB breaker is like any of their other equipment, it'll be quickly deposited in the rubbish bin along with the sanity of most electricians once they start trying to roll any of this out.

I suspect its the principle that matters. If they discovered a design that hits the goal someone else can build on the concept (once patent becomes available, if not before then through licensing) and build it efficiently and properly if ABB is incapable.

How can 50Hz frequency be similar to a frequency of a human heart? I cannot read this rubbish.

Its been a while since I tool biomedical electronics, however suspect the current research is still fairly much the same. Muscles are basically ionic pumps. when electricity moves through the skin this transfer from electric to ionic energy happens. originally sodium chloride (salt water) was used to help with this transfer with some of the first ecg machines. The frequency that current can be highly felt (and and also interact highly with the heart) is between 50 and 60 hz. I remember a discussion in class that happened during a lab where the students were wondering why 1000hz wasn't used instead since the chance of biological interaction would have been much less dangerous. At the time I believe the ad-hoc conclusion that was arrived at was that until switching power supplies were invented the transformers for that frequency would have been too large and expensive comparatively and motors and lighting of the day would have required these kind of transformers to work properly.

Actually, required size of the transformer drops with the frequency. Mains switching power supplies still use regular transformer for isolation which run on comparatively high frequency and thus can be very small.

50/60Hz mains is motivated by compromise between visible flicker from lightbulbs and required frequency for practical synchronous motors with useful rotational speeds (some electrified rail systems use lower AC frequencies because they are better match for motors). Also, significantly higher frequencies will cause even relatively short mains runs to behave like a (RF) transmission line and also cause more significant skin effect.

Agree on that generally for electical equipments size falls off as square of frequency, the main motivation was the light bulb flickering 50/60 times was close to human persistence of vision which allowed the light bulb to appear constant.

Damn, and this is from The Economist. They're supposed to be good with numbers! I guess that's only true if they're prefixed with a $...

I love the Economist's dry sense of humor (italics are mine):

"That could lead both to cheaper power, and to the burial of many of the pylon-borne power lines that disfigure so much of the rich world's countryside. In a deeper, sense, then, perhaps Edison will have the last laugh, after all."

This article is not actually very well written. Definitely the author doesn't understand some of the stuff he's writing about. The section describing the interlink between Japan's different power systems is especially awkward.

I believe this will be used first in datacenters, if not already being tested.

AWS and Google seem likely candidates to test this out.

For a very long time it is possible to buy almost anything data center-worthy in version with 48V DC power supply (although often with significant premium) and most collocation facilities provide 48V DC power. This has nothing to do with efficiency but with 48V being the standard voltage in telco environments, because it's 4 car batteries in series. Phones need DC and until about 50's it wasn't economical to rectify mains AC with enough power to run phone switch and without harmful noise and AC ripple, so phone switches were generally battery powered (today it's typicaly generated by rectifing mains with battery backup). Servers and networking equipment around these switches generally shares these same power circuits. Problem with 48V is that for any significant powers the wires get very thick very quickly.

Many larger server installations noticed that (1) normal AC server SMPS will happily run from 150-300V DC without any modification (reliability is an issue, omitting input rectifier and PFC is good idea) (2) there is almost nothing in today's servers that need any other voltage than 12V. Thus there are various approaches to DC power (for increased efficiency, not telco compatibility) in datacenters:

* DC power supplies with DC voltages around normal mains voltage (IIRC Facebook) * power supplies with inputs for two different voltages (AC mains + sub-100V DC for backup power from batteries) * battery backup inside the server on 12V power bus (Google) * shared 12V DC power supply for multiple servers (almost any blade solution, although I've seen that done with normal 1U servers) * also I vaguely remember some proposal to replace local 12V->something buck converters in modern server with larger input voltage.

You mean low-voltage DC distribution within a datacenter? That's useful but it's a different issue: the object here is long-distance (~1,000 km), high-voltage (~500,000 V) transmission. Economist didn't make this clear enough.

http://www.abb.com/cawp/seitp202/65df338284e41b3dc1257aae004...

https://en.wikipedia.org/wiki/High-voltage_direct_current

I would gladly have read this, except some piece of javascript kept redirecting me to the top of the page. Tried multiple times, but sadly unreadable.

Same here. Click the "print" link in the upper-right area to get a bare version of the article.

http://www.economist.com/blogs/babbage/2013/01/power-transmi...

Obligatory Oatmeal: http://theoatmeal.com/comics/tesla

... which is not really bad... but not really good either. Too exaggerated and populist to be interesting.

Saw the world 'electrocuted' being used incorrectly and got annoyed.

For the record, here's the definition: "Electrocution is death caused by electric shock, either accidental or deliberate" (taken from Wikipedia. Google includes injury). Sure the elephant could have been injured or killed but the article seems to imply that it wasn't.

The correct term, by the way, is 'electric shock'.

Don't know where it implies in the article that the elephant did not died, but it did in fact. Here you can read more: http://en.wikipedia.org/wiki/Topsy_(elephant)

Reply on news.ycombinator.com