More Data, this time with my x230 as it has a RJ45,
For some comparison, I have here an X200 (2.4GHz C2D) and a T430s (2.6GHz i5), both with RJ45 and iwn wifi, but via a phone hotspot, so I'll just quote wifi off/on data.
Cable modem idle:11.1W x230 idle 12.5W (acpiconf -i 0)
Downloading 13.1 for RPi WIFI : Cable modem 12.9W x230 16.5W
X200 idle 800MHz, wifi switch off 9.3W, on 10.1W, but not currently using wifi.
T430s idle 1200MHz, wifi switch off 12.6W, on 14.2W, connected.
Switching WIFI off in the modem, connecting cable...
Cable modem idle:11.4W x230 idle 13.5W (acpiconf -i 0) <-- Wut?
When the cable is connected, power consumption of the modem goes up from 10.8 - 10.9 up by half a W.
Er isn't that 0.1W? About all I'd expect for ethernet (idle).
The Laptop goes up by a W also, from 12.5 to 13.5. As soon as DHCP ran, and the interface actually had a link - 11.4W modem, 14.2 Laptop.
Interesting, surprising even. Sometimes need patience with acpiconf data, settling time, asynchronous probes?
Disable wlan0 device, laptop now at 12.2W
Disabled by power switch or software
devctl(8), disable or suspend?
Downloading 13.1 for RPi copper : Cable modem 11.7 laptop 14.5W
So what have we learned?
(Posting this for backup purpose as the machine cold booted at re-docking mid-writing)
Another factor is time: here, ethernet clearly uses far less power than wifi, not least because apart from the odd keepalive packet, it's mostly idle. Wifi OTOH is nearly always busy all day doing nothing much at all ...
OS tuning will win more power savings than either. I have another X200 to replace the above one which has a CPU fan that usually fails on boot, meaning any intensive use shuts down at 100C.
However it's running a so far untuned 12.3-RELEASE, the older one is 9.3-STABLE; the difference at idle is the above 9.3-10.1W versus 12.3-12.7W!