Power plugs and low-hanging fruit

by Rick Jelliffe

Is this the dullest, most trainspotterish blog I've ever written, mo mean a feat: you be the judge if you can make it to the end! I've just come back from a few days in Kuala Lumpur. Lots of new words made many newspaper article inpenetrable: I knew the title "Tun" but I had never seen the title "Datuk", let alone all the "Yangs." But I was impressed that the hotel had universal power sockets installed.


Foreign power plugs are always surprising. Here in Australia, we use 230V at 50Hz with a three pin plug in a wilted T configuration, with flat blades: its is used throughout the South Pacific, and supposedly the Okinawan, Argentinean and Chinese sockets are compatible too. What is surprising is to visit foreign countries where there are multiple kinds of plugs, and multiple kinds of power: the countries which initially went for 100V-120V now have also moved to also provide 220V-230V power for equipment that sucks more current, and these usually have different plugs. In the US, for example, I cannot help but thinking "Why on earth don't they get rid of 110V?" since it is technically poor, complicates things, has obsolete plug forms (the two pin versus the three pin.) "Why don't they just all adopt what the rest of world does, at the electrical level at least? They could take the opportunity to go metric too!"

Wikipedia has a good guide to Domestic AC power plugs and sockets. There are about common 12 kinds national or regional AC connectors (but each many of these have variants that don't prevent plugging in, and many nations have an extra handful of other less common plugs, such as the US NEMA plugs), and IEC defines 13 kinds of connectors for electrical equipment. That makes scores of different plugs in use, with scores more now obsolete or obsolescent.

The technical reason for multiple plugs is to prevent you from plugging in equipment and motors designed for the wrong voltage and frequency: and especially between AC and DC of course; the historical reason is empire and inter-national uncooperativeness; the economic reason is that it would be too expensive to replace them all at once or to have twin connectors on equipment; the design/manufacturing reason is because the differences can be isolated by having detatchable plugs, so there is no cost impact on equipment made with IEC sockets or DC plugs because the major component is the same; the parochial reason is that people get used to things being a certain way and it was a barrier to trade felt to encourage local manufacture or to promote empire-first policies. There are some deeper technical issues too: in Japan I was told they adopt 100V for normal domestic outlets because of earthquakes...if a pole or building falls over they prefer that stray wires should be as low voltage as possible; on the other hand, I have been told that higher voltage is better for power transmission over large distances, so countries like Australia adopted 240V (now 230V). I have been told that mosts modern US homes are wired as 220V now, with only a single side use for the 110V domestic circuits.

The nice thing about the hotel in Malaysia was that they had installed universal plugs: the Malaysians use the big clunky British style 220-250V power (also used in Britain, Brunei, Uganda, Liberia, etc) but the hotel sockets could also fit my smaller but still clunky antipodean style plugs and many others.

What was interesting to me, was that though there are multiple standards for the DC plugs going into appliances (for example, into the laptops), multiple national standards for AC power with plugs, frequency and voltages varying (sometimes within a country!) a perfectly workable solution has emerged.


So in the power plug world, the answer to multiple incompatible standards that is emerging is not based on standardizing on a single exact connection type or electrical type but quite the opposite: the underlying systems are increasingly starting to look more like each other (230V, 50Hz) , the connections to consumables has been standardized (IEC sockets), transformers support electrical plurality (100V-250V, 50 or 60Hz), and sockets are being deployed which accept as many different plugs as can be accommodated as low-hanging fruit.

Ah, I hear you cry, aren't you just advocating N-N which is inefficient? Well, not really: the notebook approach is layered, so that the physical connection becomes an N-1 issue (a universal socket support multiple plugs going to an IEC connector) and then the electrical connection becomes an N-2 issue (the transformer handling multiple voltages). The thing is how to gradually reduce the N-1.

When there are multiple existing technologies deployed, and where a unified standard technology would provide no significant difference in capability that would drive consumers to demand the unified standard by their wallets, the way to achieve a unified standard is by providing a bridge or on-ramp: a technology that neutralizes differences so that users can continue to use their existing systems. That is what happened with XML and character sets: no-one was using Unicode for data interchange; XML provided a bridging mechanism where users could incrementally switch from their locale-based encodings to Unicode systems gradually and incrementally, and now UTF-8 is widely and fairly painlessly used.

Applying this back to power plugs, the solution is not to unify the plug standards, but diversify the sockets specs so that each nation requires a socket that accepts both the national standard and some intended future international standard. If it turns out to be impossible, because countries still want different plugs and sockets for 110 and 220V, then two standards are better than none. If it turns out that the countries with round pins and the countries with blade pins cannot agree or the design/cost problem is too hard, again, two standards is better than none.


Kurt Cagle
2007-06-01 10:46:09
Yaaawwwnnnn .... yup, an exciting post *blink blink*. ;-)

Seriously, I think this is actually one of your better ones, as it points out that oftentimes standardization occurs not necessarily at the edges but from the core out. I see this a lot with syndication formats - historically the syndication space has been a mess, with any number of formats attempting at any given point trying to carry the banner, to the point where the market is now fragmented fairly evenly between RSS 2.0, Atom, RSS 1/RDF, and the various RSS 0.9x formats. However, the syndication feed architecture in general is beginning to expand into the realm of competing (quite effectively) against the SOAP based standards, and as such people are increasingly building applications (GData, anyone?) where the details -- whether flanges or pins should be used in sockets, whether given terms should be used in the transport schema -- are recognized to be different, but the underlying delivery system is standardizing towards a feed based vs SOAP based architecture.

Rick Jelliffe
2007-06-03 01:55:55
Kurt: Yes. But it challenges the XML "syntax-first" mentality (which I certainly have): perhaps the point with consolidating the syntax is that you minimize tooling, while consolidating the semantics minimizes data management.
2007-06-04 09:56:48
The reason the wisdom of crowds works for Linux is because of sharing debuggers. Crowds aren't inherently smarter or more parallelizable until you add a force. The tool is the force.

The tools act as a force on the million monkeys typing. That is clear. As long as they keep communicating and the environment doesn't change, the cost (energy budget) drives them to convergence. The rest is 'fine tuning' smoothing of local effects.

Yes: XML minimizes tooling. OTW, it enables convergence by removing more local effects (curly vs pointy, graph vs tree, myName vs yourName) and adds some simple tests to remove tedium. XML is grease.

BTW: Kurt, by mandate, ebXML/SOAP is being fielded in a very big way these days for real-time alert systems. So the demise of the WS* stack is greatly exaggerated. One hopes the payload specs stay XML plus binary enclosures just in case the REST Rapture does happen.

2007-06-06 14:48:03
Dude, rants about power systems may be appropriate for the general blogosphere, but they're totally out of place on an XML blog. Save it for your personal blog.
Rick Jelliffe
2007-06-06 17:09:18
Isaac: Power points per se are not about XML of course. But discussion on how and where to harmonize alternative standards is a pretty topical XML subject, certainly at ISO and national standards bodies at the moment. If we see the range of approaches, we are better informed about various claims: for example the idea being floated that standards cannot or should not overlap.