How much power do recent and current tabletPCs draw during operation? As part of my external battery testing, I found it necessary to do some power consumption measurements on some of the tabletPCs that I happen to have on hand. The results, shown in the table below, were interesting.
In order to make these measurements, I used a precision meter to measure voltage, current, wattage and power factor through the AC adapter for each tablet. Prior to measuring, I made sure that each internal battery was fully charged, so that power effects due to battery charging were negligible. Once the tablet was turned on, I waited until the desktop appeared and the background processes settled down; that is, with the processor(s) idling and only showing the occasional activity spike. I call this the Stable Desktop. This measurement represents the low end of power consumption during tablet operation.
Many users report that playing videos is a stressful activity for a tablet, occasionally accompanied by high case temperatures and shortened battery operating times. Virus scans have also been identified in similar terms. I therefore decided to make measurements during both operations since they likely define the high end of power consumption during typical use. I streamed a specific video, played with Windows Media Player and observed the power fluctuations over a period of time. This is the Play Video entry. I played the same video on all five tablets. The quality of playback varied considerably. The Q550 was probably the worst of the lot; its playback was very choppy and pretty much unsatisfactory for viewing. Next was the Slate 500; its playback was less choppy, but still not very satisfactory for viewing. The TC1100 showed very little chop; for an old tablet, it performed pretty well. The EP121 also had very occasional light chop but was satisfactory. The XT2 performed best; it showed no observable chop and was easily the most viewable presentation of the video. Interestingly, these results also track the power consumption observed. Coincidence? I don’t think so!
Finally, I started a virus scan and repeated the timed observation; this is the Virus Scan entry. All five tablets used Microsoft Security Essentials. Note that the virus scan was generally less demanding than the video (except for the EP121)
All of the tablets were running Windows 7, but two used Ultimate, two used Professional and one used Home. All of the tablets were set to Balanced power management mode. Each tablet had a unique software load, so the number of processes running in the background varied. Four of the five had Office 2010 installed.
Considering all the variables involved, the measurements shown here should be regarded as approximate, at best. Still, my observation is that current (Intel-based) hardware and software appear to demand about 20 W or more of power to deliver what I would consider to be a “satisfactory” computing experience. Less than that, and I felt like the tablet was lagging or sluggish. As the power diminished, the experience worsened. Sometimes, you just can’t beat the physics!
PS: Keep this table in mind; we’re going to refer back to it when we talk about the external batteries!
@steve where you measuring voltage and current at the AC side of the power brick or at the DC side? – I imagine different bricks have different efficiencies and some waste more energy as heat than others.
On a slightly different topic, is there any published data on how much more efficient tablets with battery pens are 5%, 10%, more, less? (ntrig has both battery and non battery pens, the old TC1000 had a finepoint battery pen compared to the TC1100′s wacom pen).
As regards video playback, it would be interesting to know how efficient (current-draw wise) hardware decoders are compared to software decoding.
Hi, James: All good questions, but some are difficult to answer conclusively!
“…where you measuring…at the AC side…or at the DC side? – I imagine different bricks have different efficiencies…”
…I measured on the AC side, using the remarkable Kill A Watt EZ, a device that boasts “typical” accuracies in the 0.2% to 0.5% range. Regarding modern AC adapter efficiencies, they are much more consistent than one might expect and typically hover in the 90% to 95% range. However, you are right that there are a lot of variables to take into account (as I also say) and that’s why I say the numbers should only be taken as “approximate”…
“…is there any published data on how much more efficient tablets with battery pens are 5%, 10%, more, less?..”
…To the best of my knowledge, no. The battery-operated pen eliminates the need for the excitation pulses that the digitizer produced to powered the battery-less pens, but as to what quantity of power-savings that represented, I couldn’t even guess. Apparently, it was enough to make N-trig baseline it for their new generation of hardware…!
“…it would be interesting to know how efficient (current-draw wise) hardware decoders are compared to software decoding…”
…Interesting, but I’m not sure if it would be of practical use to the user. Software codecs exist because formats change and evolve so rapidly. Hardware decoders would probably work better, but would definitely have a limited useful life (and that might be OK, given the half-life of computer hardware in general…!)
Have you made any study with respect to charging current requirement for various tablets / slates /ipads?