Results of Laboratory Testing of Advanced Power Strips Introduction

Results of Laboratory Testing of Advanced Power Strips
Lieko Earle and Bethany Sparn, National Renewable Energy Laboratory
ABSTRACT
The ever-increasing inventory of miscellaneous electric loads (MELs) in U.S. homes
results in wasteful electricity consumption in the form of 'vampire' loads. The emerging
technology class of products called the advanced power strip (APS) seeks to mitigate this
problem via auto-switching capabilities, where supply power is automatically shut off when the
end-use appliance is detected to be in an unused state. These products are relatively new and
untested; as such, there is a dearth of quantitative data on how effective they are at energy
abatement and whether they impact the functionalities of the end-use appliances they serve. For
utility efficiency programs to incorporate plug load reduction via APS, concrete numbers are
required to document cost-effectiveness of the strategy, but this cannot be quantified without first
demonstrating that these devices perform their basic functions reliably. To address this need, we
have conducted a focused study to evaluate the technical performance of APS devices when
subjected to a range of home entertainment center and home office usage scenarios. Guided by
the test specification and minimum functional standards developed by the Northeast Energy
Efficiency Partnerships' (NEEP's) APS Working Group, we have tested 20 currently available
APS products in the laboratory to determine whether each works as designed and what, if any,
actions or habit changes are required for the user. We report on the results of this laboratory
investigation.
Introduction
As increasingly stringent residential building energy codes improve thermal envelopes,
and mechanical system efficiencies increase due to advances in technology and standards,
occupants continue to introduce more and more appliances and electronics into the home,
creating a moving target for plug-load energy reduction. From televisions to toasters, cell phone
chargers to pool pumps, plug loads are as diverse as they are ubiquitous, and as a class continue
to expand their fractional share in the electrical energy budget of the modern American home.
With so many appliances throughout the house, it is easy for busy people to leave things like
TVs and computers switched on inadvertently when they are not in use, resulting in substantial
'active' power waste. Moreover, this increase in miscellaneous electric loads (MELs) also
consumes energy in the form of 'vampire' or 'phantom' loads: many devices continue to draw
current as long as they remain plugged into receptacles, even after the appliances are switched
off (Hendron & Eastment 2006). In a typical home with 40 plug loads, vampire loads can
account for nearly 10% of household electricity use (standby.lbl.gov). Plug-load energy waste is
pernicious; each individual device may require only a low level of power to be in standby mode
(and even less in off mode), so there is little incentive for the homeowner to walk around and
switch off power strips, much less unplug each device that is not in use. Effective reduction of
both vampire and active loads will require a more convenient solution.
Recent years have seen growing interest in quantifying plug-load energy consumption via
sub-metering, and early studies have focused largely on creating an inventory of household
MELs and their associated power draw characteristics for on, off, and standby states to identify
©2012 ACEEE Summer Study on Energy Efficiency in Buildings
1-51
savings opportunities (Meier et al. 2008; Porter et al. 2006; Roth et al. 2007). Field test results
indicate that typically occupants do not take the time to unplug unused appliances (Bensch et al.
2010). Given these findings, automation seems an ideal solution to curbing vampire loads
without interfering with how people use their plug loads. Advanced power strips (APS) may
provide the appropriate balance of intelligent control and convenience to meet this need. APS is
an emerging technology that includes products offering a variety of different control
mechanisms, and to date there have been few third-party technical evaluations to assess their
efficacy and cost-effectiveness at automating energy savings without requiring behavioral
changes. The results presented here address whether and how APS devices work to turn off
unused appliances and reduce vampire loads. This information is crucial for efficiency programs
to determine where and how APS can be a cost-effective solution.
Advanced Power Strips
Advanced power strips are designed to facilitate convenient energy savings in the home
office and entertainment center where the concentration of consumer electronics is typically
high. Despite the potential promise of APS, the general public has little knowledge of how they
work, where they should be installed, and what to expect from their operation. The actual savings
achieved will depend heavily on how people use them, as APS products inherently affect the
operation of the devices they control. For example, a user may find it beneficial that his game
console powers off along with the TV; however, that same user may be inconvenienced if he
must first turn on the TV in order to listen to his stereo receiver.
APS Control Strategies
Advanced power strips feature control options and configurations that differentiate them
from traditional power strips. Many of the currently available products are primarily intended for
use in an office or audio-visual environment. For convenience, we use the labels 'PC' and 'AV' to
refer to the home office and entertainment center, respectively. While many APS devices can be
used with a wide variety of plug loads, the current-sensing and control features are typically
designed to be most effective when used with computer- or TV-related electronics. Table 1
summarizes the general categories of APS most commonly encountered; the technology is
rapidly progressing and this is not an exhaustive list. Note that many devices fall into more than
one category.
Northeast Energy Efficiency Partnerships (NEEP) APS Working Group
To accelerate the inclusion of APS in utility Demand-Side Management (DSM) or
government efficiency and rebate programs, a better understanding of their energy saving
potential is required. Existing studies focus exclusively on market research and occupant
behavior data, so the savings estimates do not account for how well the technology works (e.g.
Koser & Uthe 2011). While these works provide key statistics that inform predictions of how
much energy is available to be saved, they do not address the technical capabilities of specific
APS products or product categories. To close this gap, a rigorous and standardized assessment of
APS performance is needed to speed their adoption in the market place.
©2012 ACEEE Summer Study on Energy Efficiency in Buildings
1-52
Table 1. Categories of APS Technology
Category
Control Strategy
Current-Sensing
Current-sensing power strips monitor the current draw of the controlled outlets to determine when the
connected devices are not turned on. Supply power to the controlled outlets is then shut off. Some
action must be taken to restore power to those outlets. In an AV setting, an infrared (IR) signal is
usually used to 'wake up' the power strip. For PC use, there is typically a button on the power strip that
must be pressed by the user. The algorithm used by the APS to determine when the controlled
appliances are off varies. Some look at peak power use for all the connected devices and shut off the
supply to all outlets when the total power has dropped by a set percentage. Others monitor smaller
groups of outlets and control each group independently.
Infrared (IR)Remote
There are two ways that IR sensing can be used to control a power strip. A simple IR-controlled power
strip uses the IR signal to activate a switch that turns the controlled outlets on and off; in this case the
device is nothing more than a traditional power strip with a remote switch. IR sensing can also be
coupled with a current-sensing power strip. In this case, the IR signal is used to 'wake up' the power
strip and restore power to the controlled outlets. Some IR detectors need to first be paired with the IR
signal from the user's remote; others can detect IR signals from any remote.
Master/Slave
Master/slave power strips use the power state of the 'master' appliance to determine whether the supply
power to 'slaves' (or peripheral devices) should shut off. The master is typically the TV or computer,
which is plugged into a special outlet that does not turn off. The slaves are peripheral devices such as
DVD players or computer monitors. The controlled outlets are powered only when the master
appliance is in use. Most master/slave power strips use current sensing to determine power state of
master device. Some power strips designed for home office environments rely on USB connection to
the computer to determine the power state of the computer, which is described in more detail below.
Motion-Sensing
(OccupancySensing)
Motion-sensing power strips turn off the controlled outlets when no motion has been detected for
some period of time. Unlike many APS's that target standby energy use, these power strips turn off
power to appliances when they are no longer in use (assuming that motion is a good indicator of use).
The motion sensor is typically connected to the power strip via a long cable so that the sensor may be
located in the best place for occupancy detection. The occupancy sensing can include a more complex
suite of sensors than just motion.
Remote Switch
The on/off switches on traditional power strips can be inconvenient to access if they are located behind
entertainment centers or under computer workstations. Remote switches typically use IR or radio
signals to allow users to turn outlets on/off remotely. Some products feature a subset of remotely
controllable outlets on an otherwise conventional power strip.
Timer-Controlled
Power strips with a timer switch can be programmed to shut off supply power during periods of time
when the user knows the appliances will not be needed. They typically have some controlled and some
'always-on' outlets. The primary advantage of the timer-controlled solution is that users do not need to
remember to switch off their power strips. For enhanced automation, this feature can also be
incorporated into current-sensing power strips, where the timer starts counting down once the
appliance is determined by the APS to be in the off state.
USB PowerSensing
This is a variation of the master/slave approach that uses USB rather than current sensing to determine
the power state of the master appliance. Designed for PC use, these power strips rely on a USB sensor
that must be plugged into one of the computer's USB ports. Some plugs are remote; others are
physically tethered to the power strip. Power supply to peripherals is shut off if the computer is
determined to be in an unused state.
In June 2010, the Northeast Energy Efficiency Partnerships (NEEP), New York State
Energy Research and Development Authority (NYSERDA), and Efficiency Vermont (EVT)
jointly sponsored a Consumer Electronics / Plug Load Summit in Albany, New York. The event
brought together key stakeholders in energy-efficient consumer electronics programs to discuss
strategies for optimizing energy savings through application of APS technology. The event was
attended by more than 60 stakeholders, including efficiency program managers and evaluators,
APS manufacturers, and state and federal policy and regulatory interests. A key result of the
meeting was the formation of a NEEP-led APS Working Group, which was tasked to formulate a
robust assessment method for APS devices. The laboratory tests presented in this paper are
©2012 ACEEE Summer Study on Energy Efficiency in Buildings
1-53
guided by the February 2012 draft of the test specification authored by the APS Working Group.
That document is still under development, and the final version will incorporate practical lessons
learned from the work presented here.1 For example, the procedures in the current draft consider
only a limited number of use-case configurations and the test plan is difficult to tailor to different
variations of APS products. Auto-switching technology will continue to evolve, so a
standardized test specification must be broad enough to accommodate future generations in
product design.
Laboratory Testing
The tests were performed in the newly commissioned Automated Home Energy
Management (AHEM) laboratory at the National Renewable Energy Laboratory (NREL).
Table 2 summarizes the basic features of the 20 products that were tested. Manufacturer
and model information is shown in this table, but the device ID numbers assigned in the first
column are used throughout this paper to refer to individual products. The numbers of controlled
and total outlets for each device are also listed. 'Adjustable threshold' refers to whether the
device has a user-adjustable load threshold that triggers the controlled outlets to turn on or off.
Where applicable, it is noted whether the adjustment mechanism is a switch with discrete
settings or a continuous dial. In one case, the threshold is adjusted by training on the intended
end-use appliance. The 'AV' and 'PC' columns indicate which products are designed for use in
which environment; in most cases they are suitable for use in either. The UL standards 1449 and
1363, listed in the last column, correspond respectively to 'Surge Protective Devices' and
'Relocatable Power Taps.'2 The APS Working Group determined that all APS devices should
have both of these UL listings.
Of the technology categories listed in Table 1, motion-sensing, timer-controlled, and
remote switch devices are excluded from our study. The functionality of a motion-sensing or
timer-controlled APS is primarily dependent on the reliability of the motion sensor or timer.
Regardless of whether these categories are effective at mitigating plug load loss, we feel they are
not relevant to our present study, which is focused on the comparison of control algorithms.
Remote switch devices are likewise excluded; they help users turn off their power strips more
conveniently but do not offer any level of automation.
These devices are all currently available for purchase by consumers. The retail cost
ranges from $19 to $70, with a median price of $30.
1
The authors are contributors to the NEEP working group's test specifications effort.
UL defines a 'surge protective device' as a device designed for limiting of transient voltage surges, and a 'relocatable power tap'
as an extension of a grounding AC branch circuit for general use.
2
©2012 ACEEE Summer Study on Energy Efficiency in Buildings
1-54
Table 2. Summary of APS Products Tested
ID
Brand
Model
1
APC
2
Belkin
3
Bits Limited
4
Bits Limited
5
Coleman
Cable Inc.
Power-saving Surge
Protector P8GT
Conserve Smart AV
Surge Protector
Smart Strip Surge
LCG3
Wireless Smart Surge
Protector
SUG7
Smart Strip
FMH6-6MS-SR
6
EcoStrip
USB Ecostrip 2.0
7
Ecotek
8
Ecotek
9
Embertec
10
Embertec
11
Ethereal
12
iGoGreen
Power Smart Tower
13
Monster
Power
14
NuGiant
GreenPower HDP
900G
Energy Saving Smart
Surge
15
Panamax
16
Rocketfish
17
Rocketfish
18
TrickleStar
19
TrickleStar
20
TrickleStar
Standby Saver IR
Version
Standby Saver USB
Version
Emberceptor AV
Series
Emberceptor
Computer Series
Green Power Surge
Protector EGP10
Category
#
Controlled
/All
Outlets
Adjustable
Threshold
Master/Slave
4/8

Switch
Master/Slave
5/8
Master/Slave
6/10
USB PowerSensing or
Remote Switch
4/6
Master/Slave
3/6
USB PowerSensing
5/6
Remote Switch
6/8
USB PowerSensing
CurrentSensing
Master/Slave
CurrentSensing
CurrentSensing

Dial

Dial
N/A
AV










Remote

USB







6/8


4/7

7/10


Train

1363

4/5
4/8
PC
UL
1449
&
1363
1363




Master/Slave
3/8



Master/Slave
4/6



Master/Slave
4/8



Master/Slave
4/7



Master/Slave
4/7



Master/Slave
4/7



Advanced Power Tap
Master/Slave
2/4



TV Trickle Strip
Master/Slave
3/6



PM8-GAV
7-outlet
RFHTS105
Energy Saving Home
Office Surge
Protector
RPCS7ES
7 Outlet Advanced
Power Strip

Switch

Switch

Switch

Switch
Data Acquisition
The measurement and data-acquisition system used for the majority of the test consists of
Modbus-output WattNodes (WNC-3Y-208-MB) and current transformers (CTM-0360-005,
CTM-0360-015) from Continental Control Systems (CCS), and a CR1000 data-logger from
Campbell Scientific. Instantaneous power and cumulative energy were recorded for each plug
©2012 ACEEE Summer Study on Energy Efficiency in Buildings
1-55
load and for the power strip at a sample rate of 0.25 Hz, which was limited by the time required
for each Modbus query. The approximate time delay of power shut-off and power restore were
also recorded. In addition, the standby power consumption of each APS alone (with no
appliances connected to it) was recorded using an Ohio Semitronic AC Watt Transducer (GW5106E), which has a superior accuracy level to the CCS system, but was inconvenient to employ
for the entire testing program.
The current thresholds for the master/slave and current-sensing power strips were not
explicitly measured but testing was used to determine whether the factory-set threshold levels (or
threshold ranges) were appropriate for the appliances in our test suite.
AV Testing
The home entertainment center used for this testing consists of a Samsung TV, a Marantz
stereo amplifier, a Sony DVD player, and an Xbox 360 game console. Because there is great
variation in the power draw profiles of TVs, and because the TV is often the master appliance,
we tested two different TV units: an LED-backlit LCD display and a plasma display.
For all master/slave power strips in our tests the TV was designated as the master device.
This is the recommendation in the NEEP test specification as well as many of the instruction
manuals for the power strips. This configuration assumes that the entertainment center
components are always used in conjunction with the TV, but this may not always be the case,
and therefore could cause inconveniences for the user. For example, the TV would need to be
turned on before the stereo can be powered.
Each set of tests was started with all appliances initially in the off state. The TV was first
turned on and left on for 30 seconds to reach its nominal on-state power usage level, and then the
peripherals were powered on. After two minutes of all units being on, the TV was turned off and
the response of the APS was recorded. Typically in a master/slave APS, supply power to the
controlled outlets would soon be shut off, turning off the peripherals. After the peripherals turned
off, they were left to remain off for two minutes before the TV was turned back on. The response
of the APS was observed; if working reliably, the supply power to the peripherals would be
restored at this point. These steps were repeated twice for each configuration to test for any
inconsistencies in behavior.
The plasma TV used for testing has an energy-saver feature that shuts itself off if the
HDMI input is inactive; this meant the DVD player had to be connected using the RCA port, so
there was no space for the Xbox 360 in the plasma TV configuration.
PC Testing
The home office equipment in our laboratory consists of a Dell computer, a Dell display
LCD monitor, an Epson printer, and a Realspace desk lamp with a compact fluorescent bulb.
This configuration was tested separately using a laptop and a desktop computer.
For master/slave APS, the computer was designated as the master appliance and the
remaining appliances were plugged into the controlled outlets. Current-sensing power strips give
the consumer the option to plug the computer into a controlled outlet or an always-on outlet. For
our tests the computer was plugged into a controlled outlet because this configuration has the
most energy saving potential as well as the highest likelihood of posing difficulties for the user.
©2012 ACEEE Summer Study on Energy Efficiency in Buildings
1-56
For all of the configurations, the computer and the connected appliances were put through
a sequence of power states. Each test set began with all the appliances off. The computer was
turned on first, preceded by any action required to wake up the power strip. Then the monitor,
printer, and desk lamp were turned on. The power settings on the computer were adjusted for the
tests so that after two minutes of inactivity, the computer would enter sleep mode. The computer
was allowed to remain in sleep mode for two minutes and then was revived with keyboard
activity. The power strips that were not based on master/slave control required the user to press a
button before the supply power to the computer was restored. Any peripheral appliances that did
not come back on automatically were turned back on and the group was allowed to remain on for
another two minutes. The computer was then shut down manually. This sequence was performed
twice for each power strip to ensure that the behavior was consistent.
A limited number of power strips were tested using a laptop with a battery that was only
partially charged. This scenario presents a challenge for the current-sensing and master/slave
power strips: unless the laptop battery is fully charged, the laptop will draw 20~40 W or more for
charging, regardless of whether the computer is on, off, or in standby mode. Although the power
draw for off-and-charging (or standby-and-charging) is lower than for on-and-fully-charged, it is
higher than the threshold used to shut off power to the controlled outlets. This limitation may not
be a major concern, however, because laptops are portable by definition, while the benefits of
using an APS for a computer environment are only realized if the computer is used with
stationary peripherals.
Scoring
Using the NEEP test specification as a guide, functionality and usability scores were
determined for each APS. The functionality score reflects whether the APS performed as
expected, per the manufacturer's user manual. Each APS was subjected to the AV and/or PC
testing sequences as appropriate. If the user manual described additional controls that would not
be triggered during the baseline tests, further tests were conducted as needed to verify other
functionalities. The functionality score does not provide an assessment of the suitability of the
controls employed; it simply indicates whether the controls worked as expected. Higher scores
indicate better performance. The functionality score was determined as follows:


Start with a perfect score of 5.
Subtract 1 point for each instance of incorrect response (e.g. supply power not restored as
desired, standby mode not detected).
Separately, a usability score was determined for each APS based on the following
criteria, according to the NEEP test specification. Usability is a paramount concern for consumer
acceptance and persistence, and in our tests higher scores indicate fewer interactions required. If
the APS demands too many changes in the habits of the users, they will be more likely to replace
it with a standard surge protector. If the power strip requires too much action from the users to
reduce energy consumption, then the users are less likely to go through the required motions
consistently. Because this is potentially a key barrier to broad market adoption, a portion of the
NEEP APS test specification is designed to quantify how much user interaction is required for
desired performance. Following is a summary of the usability scoring procedure:
©2012 ACEEE Summer Study on Energy Efficiency in Buildings
1-57




Start with perfect score of 5.
Subtract 1 point for each additional step required in the initial set-up beyond simply
plugging the appliances into the sockets (e.g. need to adjust threshold, need to pair with
TV remote).
Subtract 1 point for each additional step needed to operate the equipment (e.g. waking the
power strip by pressing a button before the master appliance is turned on).
Subtract 1 point for each further adjustment or set-up changes required to reconfigure
APS when appliances are changed out.
A perfect score in both usability and functionality means that no additional set-up or operation
steps are required and that the device functions as expected in all cases.
Results and Discussion
Results are summarized in Table 3. For each APS the standby and active power use of the
product was measured. The ranges for active power reflect variations that occur even while the
APS is connected to the same suite of appliances. In most cases, supply power was shut off or
restored within 4 seconds of the trigger; exceptions are noted in the table. Individual devices with
point deductions or other features worth noting are discussed below. Several devices earned a
perfect score of 5 for both functionality and usability, and so not all devices are discussed.
Device 1 is a master/slave that can be used as an ordinary power strip if the auto/manual
switch is set to manual. We do not expect that this switch should cause any complications. A
point was deducted from the usability score because the threshold must be adjusted initially.
The threshold adjustment for Device 3 is a dial rather than the more typical high/low
switch. Some difficulties were encountered in adjusting the dial to the appropriate setting, as it
was not clear which way it should be turned to achieve a higher or lower threshold. In PC testing,
once the threshold was found the APS functioned as expected. In AV testing, the APS failed to
work correctly with the LED TV as the master because the load-sensing threshold of this APS
cannot accommodate the low-power states of this TV, in particular when the screen goes black.
Device 4 uses a wireless USB sensor to determine whether the master is in use, making it
more suitable for PC environments. If the appliance does not have a USB port, the APS has no
automation capability and functions as a remote-switch power strip. In our tests, the USB sensor
was successful in detecting the standby state of the laptop but not the desktop computer.
Device 5 has an adjustable dial threshold that required much reconfiguring with every
new master. Finding the correct threshold level was more difficult in the PC test than in the AV
test. We were not able to find a common threshold level that would both turn off the peripherals
when the computer entered standby and that would reconnect supply power when the computer
is revived. It seems the threshold that is used to turn off the peripherals is different from the
threshold required to allow peripherals to turn back on.
Device 6 has a USB plug that is tethered to the strip. It failed to detect when the desktop
computer entered standby mode (i.e. peripherals did not turn off). When tested with the laptop,
the controlled outlets turned off as expected when the computer entered standby.
Device 7 requires initial pairing of the IR sensor with the master appliance remote,
making it suitable for only AV environments. A user would press the remote's power button to
turn on the power strip, then press again to turn on the TV. If the TV and the APS's IR sensor are
in close proximity, it is difficult to point at one device and not the other. This could be
©2012 ACEEE Summer Study on Energy Efficiency in Buildings
1-58
misleading to the user, who may intend to turn off the entire power strip, and instead end up just
turning off the TV. In this case it would not be immediately clear to the user that only the TV
was turned off. These difficulties resulted in a low usability score for this device.
Device 8 has a USB sensor on a tether. Similar to Device 6, the USB failed to recognize
when the desktop computer entered standby mode. A button push by the user is required to
restore power to the peripherals once it has been shut off. The tether to the USB is only 53" long,
which could affect the relative placement of the computer and peripherals.
Table 3. Results of Laboratory Testing
[+/- 0.1 W]
Active
State
Power Use
[W]3
1
0.8
3-7
<4
2
0.3
2-7
3
0.8
4
ID
Off State
Power Use
Time for
Peripherals
to Turn Off
[s]4
Time for
Peripherals
to Turn On
[s]4
Functionality Score
(out of 5)
Usability Score
(out of 5)
AV
PC
AV
PC
<4
5
5
4
4
8-12
4-8
5
5
5
5
3-8
<4 sec
<4
3
5
4
3
0.4
2-7
120-180 (USB)
0 (remote)
< 4 (USB)
0 (remote)
5
4
3 (USB)
2 (remote)
4
5
0.4
3-7
<4
<4
5
4
3
2
6
0.2
1-3
<4
<4
N/A
4
N/A
4
7
0
2-6
4-8
<4
4
N/A
1
N/A
8
0
1-3
4-8
<4
N/A
4
N/A
3
9
0.8
2-6
60
<4
5
N/A
4
N/A
10
0.8
1-4
60
4-8
N/A
5
N/A
5
11
0.6
3-7
180
<4
4
N/A
4
N/A
12
0.5
2-7
60
<4
4
4
2
2
13
1.8
3-8
<4
<4
4
5
5
5
14
0.6
2-7
<4
<4
5
5
5
5
15
2.7
4-8
4-8
<4
3
5
3
4
16
1.0
2-7
<4
<4
5
5
5
5
17
1.0
2-7
<4
<4
5
5
5
5
18
0.5
2-7
4-8
<4
5
5
4
4
19
0.8
3-5
4-8
4-8
5
5
4
3
20
0.9
3-7
<4
<4
5
5
4
4
3
The manufacturer's stated accuracy for the CT's used here is +/- 1% of the measured current down to 10% of full scale. The
current levels measured in our tests were approximately around this lower limit. Our experience has been that the accuracy
remains linear well below what the datasheet specifies, and we are reasonably confident that these power use measurements are
accurate to within a few percent. Any variations in supply voltage is negligible compared to measurement uncertainties in the
current.
4
Resolution is limited by the 0.25 Hz sample rate imposed by the Modbus WattNode equipment that was readily available for
these tests. While a 1 Hz sample rate would have been preferred, our set-up was sufficient to determine whether there is a
significant delay that the user should expect.
©2012 ACEEE Summer Study on Energy Efficiency in Buildings
1-59
Device 9 is an IR-sensing APS that is intended for AV use. Power to the peripherals is
shut off automatically if the APS does not detect a remote control event (i.e. volume change,
channel change) over a specified time period, which can be adjusted by the user to be 1, 2, or 3
hours. A green LED light flashes to warn the user when the time limit is approaching. The
automation works reliably. The timer function can be turned off, in which case the APS will shut
off supply power to eliminate vampire loads once all the electronics are turned off by the user.
Even after all connected devices are turned off by the user, the bright green warning light flashes
before the supply power is shut off.
Device 10 is a master/slave APS designed for PC use. The auto-switching function works
well but there are no always-on receptacles. This could be inconvenient, as the standard suite of
office equipment often includes devices that should stay powered on, such as modems and
routers. This is also the only APS tested that can determine when a laptop computer is off but
charging and turns off the controlled outlets.
Device 11 relies on an IR remote signal for control and so is designed for AV use. We
encountered usability problems with this APS. The controlled outlets turn off when the power
strip detects a reduction in total load of approximately 60%. This prevents the user from
switching directly from a high-power appliance (e.g. Xbox 360) to a low-power appliance (e.g.
LED TV) because when the Xbox 360 is turned off, the supply power is shut off to all outlets.
Device 12 requires the user to train the APS on the desired threshold using the actual
appliances. It must be retrained each time the appliances are changed out. Once supply power is
shut off, the user must push a button on the APS to turn on the power supply, making it more
suitable for a desktop product than for use in an AV environment. The threshold, once set, did
not turn things on/off consistently in our tests.
Device 13 is a non-adjustable master/slave APS. While the TV was warming up to turn
on the controlled outlets cycled; a slightly lower threshold would eliminate this problem.
Device 15 has an adjustable threshold, and switching the setting between high and low
cycles the power on the controlled outlets. This APS was more successful in the PC testing.
When tested with the plasma TV the peripherals turned off when screen dimmed, but then when
screen turned brighter peripherals failed to turn back on. The threshold required readjustment
when the master appliance was changed.
Device 17 earned perfect scores, but the labeling (red on black) may be difficult to read.
Device 18 has a manual threshold adjustment of high/medium/low and so requires the
user to select the setting initially.
Device 19 has only two switched outlets, and the device plugs into the wall socket in
such a way to cover up the second receptacle, which could present an inconvenience.
Device 20 has an adjustable threshold that can be switched between 18W and 36W. The
actual thresholds that triggered the appliances to shut off differed from these nominal numbers,
but other than that confusion, the master/slave function worked reliably.
In general, the master/slave devices behaved predictably with the LED TV in the AV
testing. For APS products with an adjustable current threshold, it does not seem to make a
difference whether it is set to low or high, presumably because the television's off-state power
draw is much lower than the 'low' threshold setting, and the on-state power draw is much higher
than the 'high' setting. For the AV test with the plasma TV the situation is more complicated, as
the screen brightness significantly affects the power state of the TV.
The USB-sensing products have the interesting feature that it does not matter which
receptacle the computer (or other master) is plugged into for power, since the on/off state of the
©2012 ACEEE Summer Study on Energy Efficiency in Buildings
1-60
computer is detected via USB alone. However, using a USB port for power sensing does not
reliably detect standby mode for all types of computers.
Conclusions and Future Work
This work should be interpreted as a snapshot of the state of APS technology, and a first
pass at evaluating APS technologies and creating a means for future evaluations. New APS
products are continually being introduced into the market, making it difficult to do an exhaustive
study. Future versions of the test specification may include weighting to reflect relative
importance of different features.
It is important to note that while the scores are a best effort to present an objective
comparison of the features and functionalities offered by different APS products, they should not
be interpreted as a strict ranking. The control strategies of APS devices vary, and the suitability
of any one product to a particular AV or PC environment is highly dependent on user habits and
desired convenience features. Factors such as price, number of available outlets, and aesthetics
may play a role in the choice of APS employed.
The appropriate APS for any user or appliance suite first depends on the particular energy
savings goal: there is a noticeable difference in design between products that target vampire
loads of off-state appliances and those that are designed as intervention, to turn off appliances
when users forget or neglect to do so. It seems a challenge to design an APS that serves both
purposes equally well. Master/slave APS products target both active and standby loads of slave
devices only (so they may not be ideal in an environment where the master is often left on
inadvertently), while other APS products are effective at reducing only standby loads (but of all
connected appliances.) A notable exception is Device 9, which targets both active and standby of
all connected devices.
As expected, in many cases the power-use profiles of the end-use appliances affect how
effective an APS is at maximizing savings and how much it interferes with the normal usage of
those appliances. The vintage of TVs, computers, and other electronics matter. Interestingly,
some newer appliances have built-in energy saving features that can interfere with the
assumptions inherent in APS design.
On the whole, the APS devices tested here worked as intended. Other than the USB
power-sensing products, there were few APS products that demonstrated functional problems.
The results from the usability testing were more varied, with some products requiring many
interactions and some requiring none. More research is needed to determine the true effect of
user interaction on persistence, but there are currently a variety of APS options that live up to the
functional claims of their manufacturers.
In general, USB power sensing does not appear to be a robust solution to detecting the
on/off state of computers. There are many USB power-sensing APS products on the market and
our tests revealed basic problems with this type of technology. First, some computers have USB
ports that are powered even when the computer is off, and it is not straightforward for the user to
know if any of his computer's USB ports is the 'always-on' variety. Second, some computers do
not turn off the power to USB ports when the computer is in standby mode, so the peripheral
outlets would not be turned off even though the computer is not in use. Lastly, many USBsensing devices have the USB plugs tethered to the power strip; this physical layout can be
inconvenient or unsightly.
©2012 ACEEE Summer Study on Energy Efficiency in Buildings
1-61
Plans for further testing in our laboratory are underway, including expanding the variety
of master devices (different computers, different television units, and using the stereo receiver as
master), and measurement of actual threshold levels. A possible configuration for the AV
environment is to designate the stereo receiver as the master and plug the TV, DVD player, game
console, and any other accessory electronics into the controlled outlets. This arrangement would
allow users to listen to music without having to turn on the TV. It would also help prevent the
black screen problems (i.e. TV power level dips below the shut-off threshold when the screen is
black even though the TV is in use) that were seen when playing DVDs in some of our tests. The
limitations to this approach are that not all households have stereo amplifiers as part of their
home entertainment systems (TVs are much more common), and that it may be more likely that
people would leave their receiver on after turning off the TV (thereby failing to trigger the
threshold that would shut off standby power to peripherals).
Because we found that APS reliability is strongly dependent on end-use devices, a
broader dataset of quantitative results is required to provide a robust estimate of potential energy
savings achievable through widespread residential adoption of APS technology.
APS may not be the long-term solution to reducing plug-load energy consumption. As
with other household systems, the baseline energy efficiencies and control algorithms built into
individual MELs devices will continue to improve. However, the trajectory of plug loads
diversity versus efficiency standards is a difficult one to predict, and in the meantime, APS,
which by design targets a variety of consumer end uses simultaneously, can provide a key
inexpensive stopgap solution if they are reliable and easy to use.
References
Bensch, Ingo; Pigg, Scott; Koski, Karen; Belshe, Rana. 2010. Electricity Savings Opportunities
for Home Electronics and Other Plug-In Devices in Minnesota Homes. Madison, Wis.:
Energy Center of Wisconsin
Hendron, Robert; Eastment, Mark. 2006. Development of an Energy-Savings Calculation
Methodology for Residential Miscellaneous Electric Loads. Washington, D.C.: ACEEE
Summer Study on Energy Efficiency in Buildings
Koser, Kelly; Uthe, Carl. 2011. Advanced Power Strip Research Report. Albany, N.Y.:
NYSERDA
Lawrence Berkeley National Laboratory. n.d. . Standby Power. Retrieved from standby.lbl.gov
Meier, Alan; Nordman, Bruce; Busch, John; Payne, Christopher; Brown, Richard; Homan,
Gregory; Sanchez, Maria; Webber, Carrie. 2008. Low-Power Mode Energy Use in
California Homes. CEC–500–2008-035. California Energy Commission, PIER Buildings
End–Use Energy Efficiency Program
O'Neil, Nicholas; Braman, Matt; Hardy, Gregg. 2010. Out of Control: Barriers to Smart Power
Strip Implementation. Washington, D.C.: ACEEE Summer Study on Energy Efficiency in
Buildings
©2012 ACEEE Summer Study on Energy Efficiency in Buildings
1-62
Porter, Suzanne F.; Moorefield, Laura; May-Ostendorp, Peter. 2006. Final Field Research
Report. California Energy Commission
Roth, Kurt; Mckenney, Kurtis; Paetsch, Chris; Ponoum, Ratcharit. 2007. U.S. Residential
Miscellaneous Electric Loads Electricity Consumption. Washington, D.C.: ACEEE
Summer Study on Energy Efficiency in Buildings
©2012 ACEEE Summer Study on Energy Efficiency in Buildings
1-63