Power consumption

Post Reply
Site Admin
Posts: 1
Joined: Thu Oct 01, 2015 7:42 am

Power consumption

Post by jacky » Sat Nov 28, 2015 1:28 pm

I am investigating the chance of powering devices from battery and/or solar panels. So first I made some power measurements on the hardware I have available. Here is the information I got so far.

Power Management:
I expect the power management of the processor/Linux itself to help out a lot but the difference in power consumption between no CPU load and full CPU load is like 15%. I was hoping for more. Maybe the power management is not implemented properly.

Stupidity diode:
A lot of devices have a diode in series with the DC power supply, resulting in a ~0.8V drop for the step-down converter. When the device uses a 9V supply the total power consumption can be reduced by ~9% just by shorting that diode.

The Ethernet port also consumes a lot of power. Just by unplugging the Ethernet cable I was able to reduce the power by 10%. Completely turning off the interface reduces the power even much further, at least on some devices.

Many devices have a lot of LEDs which are usually turned on. Yes, LEDs do not consume a lot of power, but when you don't need them, remove them or turn them off in the software. I measured on one device that the LEDs are supplied with 3.3V which is typical and consume ~4.1mA. When 4 LEDs are turned on, this results in a total power consumption of ~50mW. When the hole device only needs 500mW, this is a big portion!

Wifi output power:
Reducing the Wifi output power has no measurable effect on the standby power consumption as the device should not be transmitting that much. As we expect the devices to be not transmitting most of the time, there is basically no reason to turn down the output power. It might even increase the total amount of power consumption as a given transmission may take longer as the data rate is reduced.

Here are a few measurements I did on different hardware, all running Type 1 firmware.

TL-WA701ND v2.2:
Standby power consumption in default Type 1 configuration with Ethernet plugged in and no traffic: 0.97W (100%)
Additionally shorting the protection diode: 0.88W (-0.09W/-9.3%)
Additionally turned off Ethernet: 0,62W (-0.26W/-26.8%)
Additionally turned off LEDs: 0,58W (3x 4.1mA@3.3V/-0,04W/-4.1%)

TL-WR841ND v9.3:
Standby power consumption in default Type 1 configuration with Ethernet plugged in and no traffic: 1.03W (100%)
Additionally shorting the protection diode: 0.94W (-0.09W/-8.7%)
Additionally turned off Ethernet: no effect!
Additionally unplugged the Ethernet cable: 0.74W (-0.2W/-19.4%)
Additionally turned off LEDs: 0.69W (4x 4.1mA@3.3V/-0.05W/-6.8%)

So it is possible to bring a mobile Arcman station down to 600mW with barely any effort. A 16Ah 3.7V battery can be bought as cheap as 20€ which would supply about 59Wh. This would lead to a total run time of the device of 98 hours or 4 days!
Yes, the batteries from China often do not deliver what they are supposed to and the average power consumption will rise a bit when the device is used. However, even when assuming that the real life time of that configuration would only be 2 days, put in two of the batteries and be happy ;)

Post Reply