octoBox Throughput Test Application Note

Automated RF testbed measures MIMO-OTA throughput and Wi-Fi offloading

while controlling device orientation, path loss, multipath and interference
The octoBox® is a stand-alone completely isolated semi-anechoic wireless testbed supporting MIMO OTA
measurements of very high throughput links, including Wave 2 Wi-Fi using 160 MHz channels, 8x8 MIMO and
cellular/LTE technologies. Devices in the testbed are isolated from one another and from external interference. The
throughput software automates the test by programming RF attenuators, turntable, multipath, interference and test
traffic. The software generates graphical reports of throughput vs. range and vs. orientation.

OCTOBOX TESTBED HIGHLIGHTS

Stand-alone compact industry-standard benchmark

testbed used by SmallNetBuilder.com
Wi-Fi (802.11a/b/g/p/n/ac), cellular (GSM, UMTS,
LTE, LTE-Advanced), Bluetooth, ZigBee, etc.
Measures throughput and handoff/handover between
Wi-Fi and cellular/LTE networks
Excellent MIMO-OTA environment optimized for very
high throughput
Multipath emulation per IEEE 802.11
Supports up to 8x8 MIMO of any channel width in the
700 MHz to 6 GHz frequency range
Adds realistic multi-channel interference using the
iGen™ interference generator
Automatically controls RF attenuators, multipath
emulator, turntable and traffic generator, iterating
through measurements at each attenuation, rotation
and interference setting
Figure 1: octoBox STACK-38P-TT testbed

FEATURES & BENEFITS

Frequency range: 700 to 6000 MHz

Test automation software controls test conditions and provides graphical reporting
Formfactors include benchtop PAL and wheeled STACK configurations
Integrated
o Anechoic turntable rotates device during test for averaging of throughput vs. orientation
o MPE2 multipath emulator introduces realistic multipath per IEEE 802.11 channel models
o quadAtten RF attenuator emulates path loss due to distance or walls; attenuator dynamic range: 60
dB; step: 0.5 dB; 4 RF attenuators per module powered and individually controlled via single
Ethernet/PoE or USB; quadAtten is integrated into the MPE2 or provided standalone for PAL
benchtop testbeds
o iGen interference generator
o octoPal partner device for benchtop PAL testbeds

Throughput Application Note +1.978.222.3114 www.octoscope.com

What test results does the throughput automation software produce?

The throughput software controls programmable attenuators, turntable, interference generator and a traffic
generator. Test traffic can be generated using octoScope’s multiPerf, iperf or IxChariot to produce throughput vs.
path loss plots, such as the one shown in Figures 2 and 3.

Figure 2: Browser-based real-time plot of throughput vs. path loss. In this example two measurements are overlaid
on the same plot to show beamforming gains. Path loss incremented by 5 dB every 30 seconds.

The plot shown in Figure 2 is a real-time plot that updates every second in a browser as the test runs. The plot shown
in Figure 3 is an Excel plot based on saved data that is averaged over some specified time period for each point.

Figure 3: Saved Excel-based plot of throughput vs. path loss. If the turntable is used, every point on the plot can be
averaged vs DUT (device under test) rotation.

Throughput Application Note 2 www.octoscope.com

Another type of plot is the throughput vs. path loss vs. orientation plot, as shown in Figure 4. This polar plot can
pinpoint issues with the antenna field (Figure 4, right) by revealing low throughput at some orientations of the DUT
with respect to the test antennas.

Figure 4: Plot of throughput vs. path loss vs. orientation; nulls in the DUT antenna field can be seen at some
orientations of the DUT

How is the wireless test link formed?

To test a link of two or more real devices, you can use a STACK testbed (Figure 5). This testbed can demonstrate
how diverse chipsets or devices interact. We refer to one of the devices as the DUT (device under test) and the other
as its partner. The partner exchanges test traffic with the DUT. If the DUT is an AP, the partner is typically a client.

Figure 5: STACK-38P-TT throughput testbed block diagram (left); photos (right). The DUT is typically placed on a
turntable in the top box and the partner device resides in the bottom chamber coupled either via RF cabling or OTA.

Throughput Application Note 3 www.octoscope.com

To test a single DUT you can use a PAL benchtop testbed (Figure 6) with octoPal as the partner device. If the DUT is
an AP, octoPal can function as a client. If the DUT is a client (e.g. a phone, PC or tablet), octoPal can function as an
AP. Being completely isolated, octoPal does not need to be placed into an octoBox and can be attached on the side
of the octoBox, creating a compact yet fully functional benchtop testbed. iGen adds Wi-Fi or other common
interference for a comprehensive suite of tests in a realistic environment.

Figure 6: PAL38-TT testbed. octoPal serves as the partner device eliminating the need for a second chamber.

octoPal is controllable via a web-browser or API, as described below. It is an embedded Linux device that
automatically runs a traffic endpoint, supporting several mainstream traffic generators, including octoScope’s
multiPerf, iperf, AT4 and IxChariot.

octoPal couples to the DUT over the air. Its RF ports connect via a quadAtten (Figure 6) or via an MPE2 (Figure 7) to
the antennas inside the DUT chamber.

Figure 7:
PAL testbed
configured
with MPE2 to
provide both
path loss and
multipath
between
octoPal and
the DUT.

Throughput Application Note 4 www.octoscope.com

How does the throughput measurement work?

The DUT and the partner device are connected either through a quadAtten or through an MPE2 multipath emulator.
Traffic generator software, such as octoScope’s multiPerf, iperf or IxChariot, is used to send traffic between the DUT
and the partner device via TCP/UDP/IP. If the DUT is a client (e.g. phone, tablet or PC), the partner is typically an
access point (AP) or a base station. If the DUT is an AP or a base station, the partner is typically a client. The traffic
flows between the partner and the DUT via an RF link formed by the octoBox testbed.

Figure 8:
Traffic is
controlled
by the
octoScope
automation
software,
which
supports
common
traffic
generators:
multiPerf,
iperf, and
IxChariot

Can I test cellular and LTE handover in the octoBox testbed?

You can run TCP/UDP/IP traffic between multiple pairs of devices. For example, when testing a phone, the traffic
endpoint can reside in the phone, as shown in Figure 9. In this example testbed, the phone can communicate with
any of the 3 eNBs or with the partner AP. The handoff/handover can be controlled by varying the path loss between
the phone and the 4 connected networks via programmable attenuators.

Figure 9: Example of a
handoff/handover test
configuration involving Wi-
Fi and cellular/LTE
networks.

Traffic is flowing on the RF

network that the phone
connects to (Wi-Fi in this
example).

Traffic can be iperf,

IxChariot or octoScope’s
multiPerf.

Throughput Application Note 5 www.octoscope.com

What is the path loss between the partner device and the DUT with RF attenuators set to 0 dB?

The path loss depends on:

1. Whether the devices are coupled conductively
or over the air (OTA)
2. Whether LOS (line of sight) or NLOS (non-line
of sight) conditions are configured
3. Operating frequency and channel width
4. TX power
5. RX gain
6. Test and DUT antenna alignment and antenna
patterns

Typically, the RF signals to the DUT are coupled OTA,

as shown in Figure 10. The number of antenna
elements can vary based on the MIMO configuration.
Multiband cellular/LTE antennas are also available for
Wi-Fi offloading test scenarios.
Figure 10: Example of octoBox over the air coupling;
DUT is on the anechoic turntable and surrounded by test
antennas for MIMO OTA signal coupling.

To set up LOS conditions, we use either a stand-alone quadAtten or the attenuators built into the MPE2. For NLOS
conditions, the multipath subsystem of the MPE2 is switched in-line with the programmable attenuators, as shown in
Figure 11.

Figure 11: MPE2 with RF switches set to switch

the multipath subsystem (4 MIMO paths) into the
RF path in-line with the 4 internal programmable
attenuators, forming an NLOS link. For LOS
conditions the switches are ‘flipped’ to the 4 bypass
lines, leaving only the attenuators in the RF path.

In a Wi-Fi test configuration where the partner device is coupled conductively and the DUT is coupled OTA, with the
two devices connected via LOS path with attenuators set to zero dB, the DUT’s RSSI is typically -20 to -30 dBm. This
is the top of the dynamic range of most Wi-Fi receivers.

When the MPE2 multipath subsystem is switched in (Figure 11), the wireless link matches an IEEE model of a typical
house with a broadband frequency response characterized by multipath nulls, as shown in Figure 12.

The nulls in the multipath response introduce an extra 25 to 30 dB of loss as measured by the receiver RSSI. This is
a typical loss across a wall in a house. In real life conditions, when devices are in the same room, LOS signal
dominates and when they are separated by one or more walls, NLOS signal dominates. The NLOS channel modeled
by the MPE2 multipath subsystem is shown in Figure 12.

How does the software transition from LOS and NLOS conditions while measuring throughput?

When making a throughput vs. range measurement, the octoScope software can increase LOS attenuation step by
step until it reaches the path loss of the multipath subsystem, for example from 0 dB to 30 dB. At this point, the
software switches the attenuators back to 0 dB and switches the multipath subsystem into the signal path so that the
RSSI at the DUT makes a smooth transition between LOS and NLOS conditions, as though we have been ‘moving’
the devices in the test link gradually away from one another and then across a wall from one another.

octoBox is the only wireless testbed on the market capable of this powerful measurement across the entire dynamic
range of the DUT and over both LOS and NLOS conditions.

Throughput Application Note 6 www.octoscope.com

Figure 12: octoScope’s patented multipath emulator subsystem is unique in that it is broadband (DC to 6 GHz) and
matches the power delay profile of the IEEE channel model B, the model of a typical house.

With the MPE2 in the testbed, the total dynamic range of the RF link is about 90 dB. With only the quadAtten in the
testbed, the dynamic range of the RF link is 60 dB.

Why should I add controlled interference into the throughput testbed?

In real-life installations, particularly in apartment or office buildings, there may be 30 or more Wi-Fi networks in-range
of one another. The spectral skirts of adjacent channels can impact throughput, particularly if the signal in the test
channel is attenuated with respect to the interfering ACI (adjacent channel interference). Adding ACI or CCI (co-
channel interference) to the throughput testbed lets you qualify the impact of interference on throughput. The iGen
can generate common sources of interference, including Bluetooth, baby monitors, radar and other sources.

Traffic Interference,
example of 2 Bluetooth frequency hopping
adjacent channel interference
interferers

Throughput Application Note 7 www.octoscope.com

 Interference can be injected into the DUT chamber using
the iGen via antennas. One or more iGen can be added
to the testbed to emulate interference from neighboring
Wi-Fi networks and from other unlicensed devices. For
example, with 3 iGens, you can inject ACI, CCI and non-
Wi-Fi interference, such as Bluetooth, baby monitor or
cordless phones.

The iGen can emulate radar interference for 802.11h

DFS (dynamic frequency selection) testing of the channel
switching process upon radar detection.

iGen is controlled and powered from a single

Ethernet/POE (power over Ethernet) cable, making it
easy to integrate into the octoBox wireless testbed Figure 13: iGen interference signal is typically
coupled into the DUT chamber via dipole antennas.

iGen has two subsystems: Wi-Fi and a frequency synthesizer, as shown in Figure 14 on the left. The Wi-Fi
subsystem is based on the most popular Qualcomm chipset that can be programmed to function as any legacy device
to replay captured traffic in any of the 802.11a/b/g/n/ac formats. The frequency synthesizer subsystem emulates
On/Off Keying (OOK) waveforms to model common interference, including pulsed waveforms, such as radar;
frequency hopping waveforms, such as Bluetooth and baby monitors; or swept waveforms such as microwave ovens.

Figure 14: iGen

block diagram and
photo

iGen can mount

iGen block diagram over the quadAtten
or another iGen
module

iGen has a built-in RF switch to switch between the traffic and waveform interference and a built-in RF attenuator to
control the power of the interference.

Do the octoPal, quadAtten and iGen modules couple uncontrolled surrounding interference into the testbed?

All octoBox building blocks, including octoPal,

quadAtten and iGen are completely isolated from
Figure 15: quadAtten
outside interference.
and iGen modules
feature filtered Ethernet
To maintain isolation, these modules have filtered and USB ports to
Ethernet and USB connections for power and control. maintain complete
isolation of the testbed.
Their enclosures are precision-machined and
extensively gasketed.

Throughput Application Note 8 www.octoscope.com

How does octoPal work as a partner device?

octoPal is based on a popular Qualcomm chipset and can act as a client, AP or a test instrument. It is controllable via
a browser-based user interface and an open API (applications programming interface). The API enables you to
automate and easily sequence through important performance tests in the ideal conditions and in the presence of
controllable impairments. The octoPal datasheet provides further details.

Figure 16: octoPal browser-based user interface showing the available settings including the 802.11 interface type,
channel, priority, number of MIMO streams and other useful settings for a variety of test scenarios.

Throughput Application Note 9 www.octoscope.com

Is the iGen also controllable via a bowser?

The iGen interference generator is also browser and API controllable, allowing you to create powerful automated test
scenarios and comprehensive test suites. See the iGen user interface screen shots below and refer to the iGen
datasheet for further details on the iGen interference gen

Figure 17: iGen browser-based user interface showing traffic setting menu on the left and waveform interference
settings on the right.

Throughput Application Note 10 www.octoscope.com

Summary of octoBox testbed benefits
octoBox testbed offers three important benefits and enables you to:

1. Reduce test time from weeks to Complete isolation and repeatable RF environment minimizes
hours time-consuming open-air testing. Test automation accelerates
data collection and improves test coverage and product quality.

2. Demonstrate highest achievable Ideal MIMO environment for highest possible throughput
performance Supports latest technologies, such as 160 MHz 802.11ac,
802.11ax, MU-MIMO, Beamforming, and beyond

3. Take meaningful measurements on Exercise programmable range of condition from best MIMO
real devices environment to challenging real-life impairments.

To learn more, view our instructional videos featuring the octoBox test configurations.

CONTACT
octoScope, Inc.
305 Foster Street
Littleton, MA 01460
Tel: +1.978.222.3114
sales@octoscope.com

Throughput Application Note 11 www.octoscope.com