An experimental Setup for measuring Power Consumption in WLANs
The TKN group investigates, develops and tests power save strategies for
RF interfaces. In particular, the recent research deals with reduction
of the energy needs of IEEE 802.11 as well as HIPERLAN WLANs.
For the sake of realistic power consumption parameters to be used in
simulations, power consumption measurements of a WLAN network interface
cards were necessary. Furthermore, we were in need of an experimental setup
to verify our developed power save mechanisms.
In our measurements we used Aironet's PC4800 PCMCIA air interfaces which
implement a DSSS physical layer and a medium access control protocol
according to the IEEE 802.11(b) specification. An IEEE 802.11 network interface
card can basically be in 4 power relevant working modes when in action:
We measured the power consumption in each of these modes while varying
the following parameters:
Furthermore, we measured the averaged power consumption for receiving and
sending. For that purpose we assumed two mobile terminals whereby one of
them sends continuously frames of a fixed size and the other one receives
these frames. In addition to the above mentioned parameters we varied
transmission rate (1, 2, 5.5 and 11Mbit/s) and
RF output power (1, 5, 10, 20 and 50 mW).
and recorded throughput and packet error rate . By doing so, we were able
to compute the average energy needs to successfully transmit/receive one
bit of information. The results are used to verify the outcome of simulations
which use the measured, mode-dependent power consumption parameters of
the above described network setup.
(5, 10 and 15 meters; there were walls between transmitter and receiver
for 10 and 15m)
packet size (64 ... 2312 bytes).
Our experimental setup consists of two Laptops. Each of them are equipped
with an Aironet PC4800 PCMCIA card. We chose Linux (kernel version 2.2.13)
as operating system because of its public available source code. This was
necessary to be able to modify and extend kernel and drivers for measurement
purposes. The wireless cards were configured to work in ad hoc mode. The
hardware setup for measuring power consumption is shown in Figure 1. Probes
as well as the external power supply for the card are mounted via a PCMCIA
extender card (see Figure 2).
Figure 1: Measurement Setup
Figure 2: PCMCIA extender card
Power consumption results were obtained by measuring the voltage which
dropped across the Aironet PC4800 card. In order to obtain the current,
a small resistor of 1 Ohm was placed in series with the PC4800 card.
We measured the voltage drop across the resistor and computed the current
by means of the resistance and voltage values. We used an external power
supply in order to ensure the quality of our measurements (avoiding variances
in voltage delivered by the notebook) and to keep the voltage across the
PCMCIA card in the specified voltage range (4.75-5.25V).
In Figure 3 and 4 we show exemplary results of the measurements. A technical
report which provides more insights of the measurement setup and which
contains all of the results will be published soon (check here).
Figure 3: Energy per successfully transmitted bit of information; RF
output power 1mW
Figure 4: Energy per successfully transmitted bit of information; RF output
Last modified Nov. 9th, 2000 by Jean-Pierre