Well, if you're prepared as a DIYer, to spend that much on a plug-in 'socket converter', you might as well buy a dual gang mains socket faceplate with built-in USB ports[1] in Home Bargains for your £9.99 spend.
I noticed them just yesterday and they appeared to be rather better made than the typical examples reviewed by Big Clive. I was rather impressed by the fact that there were no exposed USB circuit boards or wiring which could be damaged by careless handling during installation into a back-box.
One can only assume that, if such care and attention to these basic safety measures has been applied to the socket, the same level of quality control will have been applied to the smpsu circuitry that's been rendered inaccessible to all bar the most kack handed of DIY enthusiast.
In any case, if there's any doubt about this, it's fairly easy to wire up a chopped off 3 core appliance cord to the socket terminals in order to test with either an energy consumption meter that can read down to one tenth of a watt[2] or else connect a test meter in series with the live wire to measure the AC milliamps current drawn by the USB module, both on and off load. The off-load idling consumption figure is a reasonable indicator of the efficiency of such SMPSUs and their likelihood to blow their safety fuse or, worse still, burst into flames.
[1] If I remember rightly, one port had a 1A rating and the other a 2.1A rating, clearly marked. [2] The cheap (10 quid) plug in energy monitors sold by Maplin provide reasonably accurate sub 1 watt readings compared to my trusty Metrawatt analogue watt meter. However, this leaves you guessing a little as to by how much a zero reading is an under-reading of say a a 0.07 to 0.14 watt load. At least with the analogue meter's mirror backed scale and close scrutiny with a jeweller?s loupe to estimate such low readings, I can interpolate down to just 50mW which is a half pointer's width movement when *unplugging* either of the two Poundland USB wallwarts. Plugging in the wallwart kicks the pointer to about the 5W mark on the 100W scale (calibration marks every two watts) due to the smoothing capacitor inrush current.Obviously if you don't have such an analogue meter (who else does, btw?) and you're planning on fitting more than just one of these dual gang sockets, you can wire the whole lot in parallel to your test cord and divide the resulting plug-in energy monitor reading by the number of sockets you've wired up. This test should assure you that the annual energy cost per socket will amount to just a few pennies a year.
This no load, idle demand testing is of interest in part because of the impact of leaving them permanently plugged in on the annual electricity bill (another way of expressing their carbon footprint in a more meaningful way to most consumers) but also as a guide to their efficiency
- the early smpsu usb wallwarts used to take around 250mW or more idle consumption and a lot of those 5/12v power bricks for external hard drives typically a couple of watts and ditto for those 12v wallwarts supplied with document scanners and similar peripherals.
Some of the earlier high idle efficiency wallwarts cheated by pulsing the chopper circuit that maintained the 350vdc charge on the smoothing cap but the current crop don't appear to be using this 'economy trick' to keep the average idle consumption down, electing instead to just simply be 'damned bloody efficient' at idling. Observation with an analogue meter (watt or current) over a short period of time will reveal such behaviour, a digital energy monitor less so.
The only major concern over efficiency at full load centres on the heating stress on components in a tiny unvented smpsu rather than the slight extra cost of electricity when your phone or whatever *is* actually drawing power to charge or energise it.
If you're going to test this with a 5 or 10 watt rated 5 ohm resistor or two expect to see an efficiency of 80% or better (eg 6.2 to 5.5 watts mains input when driving a 5 ohm load with a 1A rated USB PSU). An improvement from 80 to 90 percent efficiency is hardly going to impact your electricity costs but the halving of wasted energy within the components of a SMPSU will improve its reliability considerably, hence the importance of the full load efficiency rating.
Since there's little you can about this, other than to choose a better design where they've properly allowed for the heat generated by a 3KW jug kettle load current flowing through the 13A plug/socket contacts as well as getting their calculations right with regard to the various temperature gradients between the solid state devices (typical max junction temperature rating of 125 deg C) used within the smpsu and the external environment, you're left with little choice but to rely on the warranty and to trust that the manufacturer knew what he was doing when he designed (and hopefully tested) the product.