Chip-designer ARM is known for making low power chips for smartphones and a growing number of tablets and mini-laptops. The goal is to offer enough power to provide decent performance without using so much energy that you run out of battery life after a few hours of use. But it turns out that some of the components of an ARM-based chipset are out of the company’s control — such as the WiFi module.
Speaking at an event this week, ARM president Tudor Brown pointed out that WiFi modules on laptops can draw as much as 8 watts of power — when the amount of power it takes to transmit and receive a wireless signal is actually much lower. The problem is that WiFi modules are horribly inefficient — 98% or more of the energy they use may be wasted.
It’s not entirely clear at the moment what the best way to reduce power consumption from WiFi modules would be, but the goal is to make them more efficient which would result in longer battery life and/or smaller batteries, not to mention potential energy savings which is generally a pretty good thing for the planet.
Now… anyone have any good ideas on how to accomplish this?
OLPC XO-1 had a MeshNetworking ability, and I think that they allowed for the Wireless module to shut down if not being used. Regular WiFi does not do that. See this OLPC spec sheet for more info:
https://laptop.org/en/laptop/hardware/specs.shtml
“Wireless networking: Integrated 802.11b/g (2.4GHz) interface; 802.11s (Mesh) networking supported; dual adjustable, rotating coaxial antennas; supports diversity reception; capable of mesh operation when CPU is powered down; Marvell Libertas 88W8388 controller and 88W8015 radio”
Also see:
https://wiki.laptop.org/go/Mesh_Network_Details
And see this for more about power savings for Wireless that OLPC did:
https://wiki.laptop.org/go/Networking
“Wikipedia’s article on Ad hoc routing protocol notes that: “Transmitting a signal half the distance requires one fourth of the energy and if there is a node in the middle willing spend another fourth of its energy for the second half, data would be transmitted for half of the energy than through a direct transmission.” Marvell’s card is designed to scale to very low power usage. Lower power output allows for less wasted battery. Longer range transmissions are also possible with lower power/bandwidth, allowing a large area to be covered by the mesh”.
The word “scale” is key to understand, power use still might be high at full use, with CPU on. Again, the idea is to scale, and shut down when not being used (vs having on full all the time). Hmmm, Still if high use then maybe it still used too much power and maybe ARM can come up with a better design somewhere that does much better.
Someone needs to call Pixel Qi and see where the DIY kit is that they predicted would be available by the end of Q2 (or so)?
The stated power consumption makes no sense. There’s no way a smartphone battery would last 6-8 hours on WiFi if power consumption were that high.
WiFi also has a low-power standard, and ultra-low power chipsets have been available for years (less than 270 mW):
https://www.broadcom.com/press/release.php?id=919562
That’s a long way from 8W.
@Mace, What he mentioned was the power that is consumed with the module is started and ready for transmission.
I read through a thesis of a undergrad student where he tried to change the way Wifi systems communicate to reduce the energy consumptions.
I do not think there is anyway this can be archived other than changing the protocols used in Wif for tx and rx. Changing the protocol would result in changing the base stations and all the other transmission sections.
Wi-Fi power consumption during transmit/receive is not that high, for most smartphones it probably doesn’t exceed several hundred mW. However, power consumption during startup and connection initialization can easily be >1W. In addition, 802.11 is a very “chatty” protocol, requiring a lot of back-and-forth between the access point and the client in order to maintain the connection. This is where most of the “wasted” energy goes. There’s simply no way to reliably maintain a connection on Wi-Fi without expending an enormous amount of energy (compared to, say, a cellular GSM or CDMA connection) just listening for and responding to beacons and other non-productive traffic over the air.
Another point of contention. The speaker specifically stated Laptop WiFi modules, not those on smart phones.