Tester 'll Never " Satisfied "

Bluetooth® Manufacturing Test

1. Introduction Bluetooth technology is an exciting platform that offers wireless connectivity to an
expanding array of electronic devices – computers, laptops, personal digital assistants (PDAs), digital cameras, cell phones, and wireless headsets. Personal area networks can now be created on an ad hoc or semi-permanent basis with no cables or connectors and only minimal network administration efforts.
Bluetooth frequencies occupy the 2.4 GHz ISM (Industrial, Scientific, and Medical) band. This unlicensed portion of the radio spectrum is increasingly being filled by microwave ovens and other RF technologies, the most well-known being IEEE 802.11b. As the ISM band becomes more widely used, radio interference will no doubt increase. To counter this, Bluetooth technology uses several innovative techniques to provide
stable linkages, among them cyclical redundancy encoding, re-transmission of data packets, and frequency hopping at up to 1600 times per second. Bluetooth devices can achieve data rates of up to 1 Mb/s with the Bluetooth 1.2 standard and the Bluetooth 2.0 standard can achieve data rates of up to 3 Mb/s. Originally created as a simple cable replacement, Bluetooth connectivity has become much more – users are increasingly seeking wireless capability in their workplaces and homes. Building a test system for Bluetooth-enabled products rolling down a manufacturing line would be easy if there were no design or test restrictions on the engineer. In the real world, however, we must deal with issues such as product size, throughput (which affects time-to market), and cost. Addressing these issues and knowing the specific business strategies your company adopts will help you create a test system and plan that is efficient and cost-effective for your application. This application note introduces you to the Bluetooth manufacturing environment, cites the many good reasons to test, discusses high-level business considerations, and shows
you step-by-step how a Bluetooth test plan is created. The Appendices give you detailed information on test methods and conditions, implications of Bluetooth radio design, important manufacturing issues, and descriptions of Agilent products for Bluetooth testing. By learning the principles and procedures herein and customizing them to your use, you will be able to create a test plan that is best suited to your company, product, and target market. To help you in product development and testing, the Bluetooth Special Interest Group (SIG) maintains an official Bluetooth Web site, www.bluetooth.com, which is regularly updated with new and helpful information. Agilent’s own Bluetooth Web site can be found at www.agilent.com/find/bluetooth, which offers both technical information and product descriptions. A wide range of articles, press releases, application notes, and other information is also available on the Web site.

Bluetooth technology presents a major challenge to manufacturers. As a wireless
medium, it contains RF complexities and problems that wired systems do not. It comes
in many implementations, from wireless peripherals to local area networks, to cellular
phones. Moreover, market realities dictate that most Bluetooth devices must be
manufactured at both high volume and low cost. All of these factors will clearly
influence the approach engineers take not only to manufacturing, but also to test.
Several methods for adding Bluetooth technology to a product are currently available:
1. Creating the entire design, including Bluetooth capability, from the ground up
2. Buying Bluetooth integrated circuits or chips and designing them into the product
3. Installing pre-manufactured, pre-tested Bluetooth modules from other vendors
4. Buying hardware that contains the Bluetooth capability and complementing your
own design
In terms of testing, much overlap occurs in these approaches. Whether you are
manufacturing Bluetooth modules, larger sub-assemblies with Bluetooth capability,
or fully-functional products for the end user, many of the same test issues will apply,
including device verification and design/performance expectations. It is impossible to
state simply which alternative you should choose. It will depend on many factors, such
as economies of scale, availability of outside products, probability and type of process
errors, and the level of RF expertise in your company.
For high-volume products, you should strongly consider automating your Bluetooth
device testing. This usually involves developing custom fixtures and a sophisticated
materials-handling system, so that test connections can be made automatically.
The primary enabler of high volume is short test times (i.e. high throughput), with
a reasonable goal being less than 10 seconds per manufactured unit. For low-volume
products, you may be better served by choosing a manual connection (with or without
automated testing), where results come in a minute or two rather than a few seconds.
Another decision in high-volume operation is whether to test products in a single-site
(one at a time) or multi-site (several at a time) fashion. If multi-site, the issue then arises
of whether to test in sequence or in parallel. High volume products are almost always
tested in a multi-site fashion, with several devices under test (DUTs) being presented
to the test station simultaneously. These could be tested in parallel, which calls for more
complex fixtures, or sequentially, which reduces the test resources required but also
means greater calibration and switching complexity. Yet another consideration is the
type and diversity of products and fixtures the test station will encounter: will different
types and complexities of products be coming out in the future, putting more stress on
the test environment, or will products remain largely the same?
In short, if the test system and test suite meet your time requirements and support your
quality and profit goals, they can be considered successful.
In order to make a product, you need both materials and processes to put the product
together. Bluetooth products also incorporate a non-material content (bits) that governs
the product’s behavior.
The objective of making a product must be coordinated with that of making a profit. To
do this, you should strive for an optimum efficiency at which per-product manufacturing
cost falls well below the price of the product. Profit must take into account the costs
of materials, manufacturing processes, tests, and customer support (returns, repairs,
warranties, etc.)
A design or test engineer can do little about most of these, but minimizing manufacturing
cost by minimizing test cost is a constant and primary goal – balanced, of course, by
confidence in the product’s quality.

2. Bluetooth

Manufacturing Overview

Bluetooth technology presents a major challenge to manufacturers. As a wireless
medium, it contains RF complexities and problems that wired systems do not. It comes
in many implementations, from wireless peripherals to local area networks, to cellular
phones. Moreover, market realities dictate that most Bluetooth devices must be
manufactured at both high volume and low cost. All of these factors will clearly
influence the approach engineers take not only to manufacturing, but also to test.

2.1 The manufacturing environment

Several methods for adding Bluetooth technology to a product are currently available:
1. Creating the entire design, including Bluetooth capability, from the ground up
2. Buying Bluetooth integrated circuits or chips and designing them into the product
3. Installing pre-manufactured, pre-tested Bluetooth modules from other vendors
4. Buying hardware that contains the Bluetooth capability and complementing your
own design
In terms of testing, much overlap occurs in these approaches. Whether you are
manufacturing Bluetooth modules, larger sub-assemblies with Bluetooth capability,
or fully-functional products for the end user, many of the same test issues will apply,
including device verification and design/performance expectations. It is impossible to
state simply which alternative you should choose. It will depend on many factors, such
as economies of scale, availability of outside products, probability and type of process
errors, and the level of RF expertise in your company.

2.2 High-volume testing

For high-volume products, you should strongly consider automating your Bluetooth
device testing. This usually involves developing custom fixtures and a sophisticated
materials-handling system, so that test connections can be made automatically.
The primary enabler of high volume is short test times (i.e. high throughput), with
a reasonable goal being less than 10 seconds per manufactured unit. For low-volume
products, you may be better served by choosing a manual connection (with or without
automated testing), where results come in a minute or two rather than a few seconds.
Another decision in high-volume operation is whether to test products in a single-site
(one at a time) or multi-site (several at a time) fashion. If multi-site, the issue then arises
of whether to test in sequence or in parallel. High volume products are almost always
tested in a multi-site fashion, with several devices under test (DUTs) being presented
to the test station simultaneously. These could be tested in parallel, which calls for more
complex fixtures, or sequentially, which reduces the test resources required but also
means greater calibration and switching complexity. Yet another consideration is the
type and diversity of products and fixtures the test station will encounter: will different
types and complexities of products be coming out in the future, putting more stress on
the test environment, or will products remain largely the same?
In short, if the test system and test suite meet your time requirements and support your
quality and profit goals, they can be considered successful.

2.3 Low cost orientation

In order to make a product, you need both materials and processes to put the product
together. Bluetooth products also incorporate a non-material content (bits) that governs
the product’s behavior.
The objective of making a product must be coordinated with that of making a profit. To
do this, you should strive for an optimum efficiency at which per-product manufacturing
cost falls well below the price of the product. Profit must take into account the costs
of materials, manufacturing processes, tests, and customer support (returns, repairs,
warranties, etc.)
A design or test engineer can do little about most of these, but minimizing manufacturing
cost by minimizing test cost is a constant and primary goal – balanced, of course, by
confidence in the product’s quality.

2.4.1 Completing product functionality

Bluetooth products are sophisticated RF communication devices with complex software
that drives their operation. One element of software is the ‘protocol stack,’ which is
usually downloaded during manufacturing. Other ‘soft’ components may also be required
for complete functionality – e.g., the unique Bluetooth identifier for each product. These
need to be installed during the manufacturing process at a time when they are most
supportive of the overall test approach.

2.4.2 Component or subsystem alignment

Bluetooth and other communication systems usually require some kind of component or
subsystem alignment before operating as a complete unit – e.g., crystal tuning for output
frequency accuracy, output power adjustment for battery longevity, power step accuracy
to meet specifications, and received signal strength calibration. All of these are part of
dynamically adjusting the total RF link characteristics, without which proper unit operation
in the field is impossible.

2.4.3 Performance verification

At some point, Bluetooth system designs must be measured for compliance to the
Bluetooth specification. Once the Bluetooth test specification has been met then it is
up to the manufacturer to decide what testing should be done on the device to ensure
that it is functioning correctly. This enables a large degree of flexibility when testing the
Bluetooth device. The Bluetooth device testing should create a high level of confidence
that the design will meet the Bluetooth standards. Among the parameters to measure
are modulation accuracy, sensitivity, power output, and various spectrum measurements.
Together with confidence in the overall quality (processes and materials), testing will
help guarantee that the product is reliable, has the proper operating range, and will
satisfy the customer.

2.4.4 Primary use case
Verifying that an electronic product operates at least approximately the way the
customer will use it – i.e., the ‘primary use case’ – may be critical to the ultimate
success of the product. In the case of a Bluetooth-enabled product, this means that it
must join a piconet (two to eight Bluetooth devices linked together) in its normal role
of master unit, slave unit, or both. It also means that it will be fully functional within
that context, performing as expected. Only full functionality tests (many of which are
non-Bluetooth) will be able to verify total performance.

2.4.5 Material defect and process error screening

Material defect and process error screening involves the identification of device
ailures due to aberrant performance because of shifts, long-term drift in components,
or process tolerance failures. While originating with the supplier of the components
(and its particular process deviations), such defects are still the responsibility of the
manufacturing test engineer – they must be addressed and eliminated.

2.4 Reasons to test

2.4.1 Completing product functionality
Bluetooth products are sophisticated RF communication devices with complex software
that drives their operation. One element of software is the ‘protocol stack,’ which is
usually downloaded during manufacturing. Other ‘soft’ components may also be required
for complete functionality – e.g., the unique Bluetooth identifier for each product. These
need to be installed during the manufacturing process at a time when they are most
supportive of the overall test approach.

2.4.2 Component or subsystem alignment

Bluetooth and other communication systems usually require some kind of component or
subsystem alignment before operating as a complete unit – e.g., crystal tuning for output
frequency accuracy, output power adjustment for battery longevity, power step accuracy
to meet specifications, and received signal strength calibration. All of these are part of
dynamically adjusting the total RF link characteristics, without which proper unit operation
in the field is impossible.

2.4.3 Performance verification

At some point, Bluetooth system designs must be measured for compliance to the
Bluetooth specification. Once the Bluetooth test specification has been met then it is
up to the manufacturer to decide what testing should be done on the device to ensure
that it is functioning correctly. This enables a large degree of flexibility when testing the
Bluetooth device. The Bluetooth device testing should create a high level of confidence
that the design will meet the Bluetooth standards. Among the parameters to measure
are modulation accuracy, sensitivity, power output, and various spectrum measurements.
Together with confidence in the overall quality (processes and materials), testing will
help guarantee that the product is reliable, has the proper operating range, and will
satisfy the customer.

2.4.4 Primary use case

Verifying that an electronic product operates at least approximately the way the
customer will use it – i.e., the ‘primary use case’ – may be critical to the ultimate
success of the product. In the case of a Bluetooth-enabled product, this means that it
must join a piconet (two to eight Bluetooth devices linked together) in its normal role
of master unit, slave unit, or both. It also means that it will be fully functional within
that context, performing as expected. Only full functionality tests (many of which are
non-Bluetooth) will be able to verify total performance.

2.4.5 Material defect and process error screening

Material defect and process error screening involves the identification of device
ailures due to aberrant performance because of shifts, long-term drift in components,
or process tolerance failures. While originating with the supplier of the components
(and its particular process deviations), such defects are still the responsibility of the
manufacturing test engineer – they must be addressed and eliminated.

2.4.6 Quality assurance

Quality assurance testing is directed toward:
• Integrity and improvement of the manufacturing process, including data gathering
and statistical process control
• Continued verification of correlated performances to confirm test plan assumptions
or derive new correlations
• Ongoing verification of product conformance to the Bluetooth standard or FCC/ETSI
regulations
The final four reasons for testing – performance verification, product functionality,
material and process defects, and quality assurance – are drivers of the cost
minimization process discussed above. The first two, completing product functionality
and product alignment, are considered “test” processes in that they add value similar
to that of material assembly processes. For example, defect screens should always be
regarded as temporary because they represent material or processes which can be
moved to a higher quality level. Quality assurance may take several forms, one being
full sample testing, either on or off the production line, to ensure that all processes,
materials, and designs are under control.

3.1 Business strategy and approach considerations

Strategic and business goals will affect everything done in a manufacturing process.
So a thorough understanding of one’s business environment and goals is prerequisite
to creating a test system and processes. Each business environment is unique, so no
attempt will be made to cover them in detail.
However, most design and test engineers will consider several common factors: product
type, target market, material automation, test time, test budget, product volume, floor
space, data handling for decision support, data handling for customer service, operator
requirements (user interface, ergonomics, safety, etc.), company infrastructure, and
business goals and objectives.
Each of these will place constraints on the test environment and processes – not to
mention the test engineer! For example, if your company markets a popular consumer
device for which high volume is necessary, test speed will be paramount. If equipment
cost is critical, functionality versus cost-of-test will be key. If credibility or reliable
performance under stress is critical, demonstrable performance under test will be the
goal. The most common and important of these objectives will be speed and cost.

3. The Test Process

depicts the flow of the manufacturing test process at a very high level. It provides
an excellent starting point to work through the complexities of the test process and
move into more detail. The illustration shows what happens at the factory level –
material flows in, is connected, tested, verified (compared to a benchmark), and then
shipped out after test results are stored to a database. When we approach the test
process this way, we have to consider not only the test plan and test station design –
the initial goal – but also product delivery to the station, the method of interconnection
to the station (not shown), specification comparison, data storage, and product exit.
In the diagram, the test plan is further expanded into product turn-on, initialize,
calibration, parametric, and functional. Product calibration and measurement issues
will be the chief drivers of overall test system design. Not shown but implied are test
environment issues, such as test station design, software drivers, and fixtures.

3.2.1 Product connection to test system (fixturing)

Since the test system contains the stimulus-and-response instrument(s) and support –
cables, interfaces, switches, etc. – a mechanism is required to connect the DUT to
the system. This is the test fixture. Depending on product complexity and functionality,
a fixture can be very simple or as sophisticated as the test system itself! Two factors
will dictate design of the fixture. The first is input/output (IO). IO connections are needed
for power, control, audio, and RF signals. The second factor is throughput, where the
decision to test in multi-up fashion or not is key. Issues affecting this decision include
whether switching is located in the test station or the fixture, speed of the connection,
automatic versus manual feed, isolation between IO lines, and fixture maintenance.
It is important for the manufacturer to know the loss factor from the test setup on
the manufacturing line. This can affect results if there is a high level of loss, so it is
important to compensate for any factors that may influence the test. An RF cable
connector can lose around 2 dB so this, with other loss factors, if not compensated for,
could produce a large fail yield of the product.
Another setup that is sometimes more practical when doing high volume manufacturing
is to use an RF link to test the device. This can cause problems because of interference
in the test environment. Consequently, when planning this type of setup, it is very
important to design a good shielded test cage where a good RF link can be established
between the tester and the test set. This setup will model a Rician channel so some
interference will still occur during testing. In a normal operating environment the Rician
channel is the type of interference that the device will normally operate.

3.2.3 Result comparison

A test verdict is given when the numerical result of each test is compared to a
pre-determined test line limit (TLL) – pass or fail. The TLL is considered part of the test
plan. It is established through an independent process, so it is discussed separately
(Appendix C). The process of setting a TLL takes into account production statistics,
desired yield, measurement uncertainties, and characteristics of design such as variation
over temperature or humidity.


3.2 The manufacturing test process
Product delivery Product connection Data storage Product exit Test plan Turn on Initialize Calibration Parametric Functional

Overview of possible stages in manufacturing environment.

Customers will judge the performance of Bluetooth products based on factors such as
range, transfer speed, and reliability of operation. Optimized testing in manufacturing
will ensure that their expectations are met. This section will identify a range of potential
tests which can be used during manufacturing. It discusses how to evaluate the
importance of potential tests and discusses the optimum conditions for using them.
Many of these tests can be implemented in test mode, which gives an engineer or
operator the ability to initialize and control a Bluetooth device over the RF or host
controller interface (HCI).
Until now, we have had primarily a factory perspective in describing the test process,
but testing is only the culmination of a long development process, which includes all
aspects of the value chain. Figure 2 depicts the factors that influence the creation and
implementation of an effective test plan.
Figure 2 shows that the test plan is based on broad manufacturing philosophies and
strategies, design topology, capabilities, limits, and idiosyncrasies of components, and
correlations between parameters to be tested, as well as the test coverage further up
the supply chain. When examined in detail, any product and test environment will reveal
potential trade-offs in yield, measurement uncertainty, and throughput which supports
an overall desired result. Certain factors in the test environment – such as infrastructure,
approach to fixturing, whether or not multiple DUT testing is used, and sampling rigor –
will also affect the test plan. There may have to be some iterative experimentation to
gain enough information to make decisions in these areas.

4. Factors of the Bluetooth

Test Plan

4.1 Factors in creating a test plan
Test point in supply chain Test plan Manufacturing philosophy Design opology Design conference Test environment
• Test design
• Calibration
• Equipment
• Maintenance
Factors in creating and implementing a Bluetooth test plan.
There are five main stages that can form a comprehensive Bluetooth manufacturing test
process. The starting point of testing will vary depending on the product and the stage
of testing.
1. Turn on
This is the first main stage of the test and it supplies the power to check for current
drain, integrity of output lines for connectivity, and any major circuitry faults.
2. Initialize
This stage involves the loading of the firmware for the Bluetooth chip control and setting
the Bluetooth address and DUT features. This can be done through control software
from the chipset vendor that can be automated into the manufacturing line. Verifying the
protocol stack is not a manufacturing test objective but it must be loaded correctly for
the device to operate correctly. Often in a high-volume manufacturing environment the
Bluetooth address is rewritten at the end of the manufacturing tests so that it is unique
to the device being tested.
3. Calibration
This is where crystal tuning will occur for the chip. The power calibration is also
performed at this stage. These calibrations can be performed with an Agilent N4010A
wireless connectivity test set combined with the chip control software. Some devices
are required to be reset after calibration so that the settings are permanently stored
in the devices memory. Other tests that are less commonly used at this stage are RSSI
calibration and IQ modulator calibration.
4. Parametric
This part of manufacturing will be the main focus of any Bluetooth product testing.
Verification of the device being tested is performed through any parametric measurement
that is critical to establishing and maintaining a link, so that data can be transferred
between devices. There are various different tests that can be chosen depending on
what area of the devices operation needs to be verified.
5. Functional
Functionality testing can be performed to simulate and verify the correct operation of
a Bluetooth device. Various functions can be tested such as audio and programmable
IO ports.

4.2 Bluetooth manufacturing stages

The general aim of a test plan in a high-volume manufacturing environment is to have
a test that can test the weak points of the Bluetooth device in the fastest time possible.
With the introduction of enhanced data rate (EDR), more complexity has been added
to the Bluetooth chip. This added complexity reinforces the need for fast and efficient
testing. The ideal test plan will be implemented in the shortest possible time and
cover all aspects of the Bluetooth chip design so that the manufacturer will have full
confidence that the device will function correctly.
A Bluetooth device can be tested at various stages in production, from chip foundry to
fully integrated device. If the manufacturer knows that a test will be conducted further
up the supply chain, it will be able to avoid doing any complex testing on this area of
the design. For example, at the basic IC level the chip maker will not be able to test the
whole functionality of the device, so instead the test coverage should be concentrated
on any flaws that would be shown at that level. As tests can be performed at different
levels in the production of the device it would be advised to test the Bluetooth device
when a new factor is introduced to the system such as crystal oscillators, capacitors,
or any protocol software. All calibration tests must be performed regardless of what
tests have been performed in previous stages as these are vital to the correct operation
of the device. As testing can occur at different stages of production, some tests are made redundant
because of the stage of production or because of testing further up the production
chain. Tests can also be made redundant if part of the chipset design guarantees its
performance. It is in the manufacturers’ interest to eliminate any redundant testing as
it will reduce the manufacturing test time.
When using a chip from one of the main silicon vendors it is useful to look at the data
sheet before constructing the test plan. The data sheet for the chipset can provide
valuable information when it comes to testing the limits of the device in a manufacturing
environment, as it can highlight the areas of weakness and strengths in a design.
Before assembling the test plan, a comprehensive test should be conducted on a
number of the Bluetooth devices to gain knowledge of the general trends that will
affect the device during testing. A good place to start when conducting comprehensive
testing is to use the test specification tests defined by the Bluetooth

4.3. Test plan factors
The general aim of a test plan in a high-volume manufacturing environment is to have
a test that can test the weak points of the Bluetooth device in the fastest time possible.
With the introduction of enhanced data rate (EDR), more complexity has been added
to the Bluetooth chip. This added complexity reinforces the need for fast and efficient
testing. The ideal test plan will be implemented in the shortest possible time and
cover all aspects of the Bluetooth chip design so that the manufacturer will have full
confidence that the device will function correctly.

4.4. Testing the product in a production chain
A Bluetooth device can be tested at various stages in production, from chip foundry to
fully integrated device. If the manufacturer knows that a test will be conducted further
up the supply chain, it will be able to avoid doing any complex testing on this area of
the design. For example, at the basic IC level the chip maker will not be able to test the
whole functionality of the device, so instead the test coverage should be concentrated
on any flaws that would be shown at that level. As tests can be performed at different
levels in the production of the device it would be advised to test the Bluetooth device
when a new factor is introduced to the system such as crystal oscillators, capacitors,
or any protocol software. All calibration tests must be performed regardless of what
tests have been performed in previous stages as these are vital to the correct operation
of the device.
As testing can occur at different stages of production, some tests are made redundant
because of the stage of production or because of testing further up the production
chain. Tests can also be made redundant if part of the chipset design guarantees its
performance. It is in the manufacturers’ interest to eliminate any redundant testing as
it will reduce the manufacturing test time.

4.5. Researching the test plan
When using a chip from one of the main silicon vendors it is useful to look at the data
sheet before constructing the test plan. The data sheet for the chipset can provide
valuable information when it comes to testing the limits of the device in a manufacturing
environment, as it can highlight the areas of weakness and strengths in a design.
Before assembling the test plan, a comprehensive test should be conducted on a
number of the Bluetooth devices to gain knowledge of the general trends that will
affect the device during testing. A good place to start when conducting comprehensive
testing is to use the test specification tests defined by the Bluetooth SIG

This table shows the setup for all tests that can be performed on the Agilent N4010A
wireless connectivity test set in accordance with the Bluetooth test specification.
Test Packets/bits Channels/hopping Power level Packet type
EDR relative transmit power 1 L, M, H 2-DH1
EDR differential phase encoding 100 L, M, H 2-DH1
EDR CFSMA 4 L, M, H 2-DH5
EDR BER floor performance 8,000,000 L, M, H –60 2-DH5
EDR max input level 1,600,000 L, M, H –20 2-DH5
EDR sensitivity 1,600,000 L, M, H –70 2-DH5
Output power 1 Hopping –40 DH5
Power control 1 L, M, H –40 DH1
Modulation characteristics 10 L, M, H –40 DH5
ICFT 10 Hopping –40 DH1
Carrier drift 10 Hopping –40 DH135
Single slot sensitivity 1,600,000 L, M, H –70 DH1
Multi slot sensitivity 1,600,000 L, M, H –70 DH5
Max input power level 1,600,000 L, M, H –20 DH1

Another way to test the device before creating a test plan is to use a test program such
as the Agilent N4017A Graphical Measurement Application, Figure 3. Other tools can
also be used to measure the characteristics of the Bluetooth device. These include the
Agilent ESA Spectrum Analyzer and software tools such as the Agilent 89601A Vector
Signal Analyzer.
Once the comprehensive testing has been completed the results can be analyzed to
show where the main discrepancies between the devices occur. If all the devices tested
show similar results, then the engineer will be able test the chip design’s common
strengths and weaknesses, and create a effective overall test plan for the design.
The individual tests in the test plan can be chosen from the manufacturing test list in
Section 4. This will help match the known weaknesses of the design with the appropriate
tests for these areas.
As a Bluetooth device is made up of various parts, a sensible test strategy would be to
target specific areas within the chip by choosing suitable tests. Detailed in this section
is a list of all the test cases defined by the Bluetooth SIG in version 2.0+ EDR of the test
specification. Each test is given a brief description of its function and the corresponding
area of the Bluetooth device which is tested. This section should help you decide on
the tests to use in the test plan by explaining in greater detail what each test achieves.

TRM/CA/01/C – Output power
Tests a Bluetooth packet to find the highest and average power levels of the burst. This
test is useful as it is very fast to execute and it provides important information about the
transmitter power. It checks the operation of the power amplifier, which is an important
part of the circuit that can affect link budget, battery life, and incidence of failure. It is
tested with a hopping signal, so it also checks the hopping circuitry. The output power
test can be replaced if power alignment is conducted before the main test sequence.

5. Bluetooth Tests

5.1. Transmitter tests

The Agilent N4017A GMA showing measurement of a Bluetooth EDR waveform.

TRM/CA/02/C – Power density
Tests the power density in the frequency domain by taking 100 kHz bandwidth
measurements. It tests the power amplifier as well as the modulator and the hopping
circuitry. The power density test takes a long time to execute so it is not advised to include
this in a test plan.
TRM/CA/03/C – Power control
This tests the calibration of the power control to ensure that the power levels and
power control step sizes are within a specified range. If the device does not support power
control, this test does not have to be implemented. This test can be useful in a device
where battery life is maximized through dynamic level control as it tests the power
amplifier control.
TRM/CA/04/C – Frequency range
The power density is tested at the highest and lowest bands of the Bluetooth spectrum
to test the frequency range. The areas covered by the test are the hopping synthesizer
and the crystal tuning of the device. This test can be substituted for the –20 dB
bandwidth or the modulation index test as they all test for the same range of faults.
TRM/CA/05/C – –20 dB bandwidth
This tests the bandwidth of the device. It measures the –20 dB points on the output
spectrum. For the waveform to pass, the –20 dB points must be less than 1 MHz apart
for standard Bluetooth, or less than 1.5 MHz apart for EDR. This test is specified only
for standard operation in the Bluetooth RF test specification, but it is also useful for
testing an EDR waveform. This test is also useful for quality control of the modulator,
although this may be covered adequately in chip testing.
TRM/CA/06/C – Adjacent power bandwidth
This test measures the power in the adjacent channels from the channel where the
test transmission is taking place. The test checks the modulator for any out of band
transmissions. It is useful for checking the design but the test takes a long time to
execute. In a manufacturing environment, samples of the spectrum would be taken
instead, to minimize the test time.
TRM/CA/07/C – Modulation characteristics
This test measures the frequency deviation of the signal. It is one of the most important
tests to have in a manufacturing test plan as it tests for the waveform modulation
quality. A low modulation quality may result in poor sensitivity. A high modulation
quality will indicate a low spectral spread. The circuit areas this test addresses are
the pre-modulation filter and the modulator performance. This test also checks the
stability of the local oscillator circuitry.
TRM/CA/08/C – Initial carrier frequency tolerance
This verifies the accuracy of the transmitter carrier frequency. It also tests the transmit
burst function and the synthesizer settling time. This test can be used to check the
accuracy of the crystal oscillator in the circuit.
TRM/CA/09/C – Carrier frequency drift
This test checks and verifies the center frequency drift of a packet transmission.
For analogue designs, this test is important since there can be a large variation in
frequency drift. For digital systems using an IQ modulator, the design means the chip
is not susceptible to carrier frequency drift. However, if there is voltage pull from external
circuitry the carrier frequency could be affected.
TRM/CA/10/C – EDR relative transmit power
This test checks the difference in the power distribution between the two types of
modulation schemes. The power control and amplifier circuitry is tested, as well as the
switching between the different modulation types.
TRM/CA/11/C – EDR carrier frequency stability and modulation accuracy
This test checks that the modulation accuracy and frequency stability are within the
required limits set by the test specification. This test is useful to test the overall
modulation quality of the device. It is good for assessing the device as it neutralizes
the effects from inter-symbol interference. This test can also be used to test the crystal
oscillator accuracy.
TRM/CA/12/C – EDR differential phase encoding
This test assesses the modulator in the Bluetooth chip to ensure that it is encoding the
data correctly. This test is useful as it checks a specific area of the chip design, however
it is probably used more in design verification than in a manufacturing situation.
TRM/CA/13/C – EDR in-band spurious emission
Tests if the level of unwanted signal produced from the Bluetooth chip is below a limit
set by the modulation scheme used. This tests the modulator and filters in the chip.
RCV/CA/01/C – Single slot sensitivity
This test shows the minimum signal level needed to produce a maximum allowed
Bit Error Rate (BER) level when sending a one-slot packet. This test can be limited by
noise in the environment so test conditions must be taken into consideration when
constructing a manufacturing test. This tests the receiver demodulator performance.
Although the test specification defaults to “hopping off” for the sensitivity test, enabling
hopping is often worthwhile in order to test the hopping circuitry for receiving a signal.
RCV/CA/02/C – Multi slot sensitivity
RCV/CA/02/C tests the minimum signal level needed to produce a maximum allowed
BER level when sending a multiple-slot packet. The single and multi-slot sensitivity are
very similar tests and only differ in the number of packets that are being sent. If there is
not a significant difference in the results from the single and multi-packet tests, then only
one of these tests should be chosen for a test plan. Multi-slot tests normally produce
larger numbers of fail results so this should be used to assure better quality control.
RCV/CA/03/C – Carrier/Interference (C/I) performance
C/I performance measures the receivers BER after sending Bluetooth signals in parallel
on the co-channel or the adjacent channel to the received signal. The areas tested are
the filter and demodulator circuits. This is one of the most complicated tests to perform
as it requires a number of different pieces of equipment. Due to the complexity of the test
configuration this test is rarely used as a manufacturing test. The measurement is similar
to the sensitivity measurements, which can be used instead of the C/I performance test.
RCV/CA/04/C – Blocking performance
This test measures the blocking performance of the receiver by sending a continuous
interference wave. This measurement is again quite complex as it uses a similar setup
to the C/I performance test. The receiver’s blocking performance is guaranteed by the
design of the radio, so this test is not recommended for a manufacturing environment.
RCV/CA/05/C – Intermodulation performance
This test measures unwanted frequency components resulting from interaction of two
or more signals passing through a non-linear device. This characteristic is guaranteed
by the design of the device so it is not recommended for a manufacturing test.
RCV/CA/06/C – Maximum input level
This test measures the receivers BER performance when the input signal is at a
maximum power level. This test should not be included in the test plan if the device’s
basic function will never approach an overload problem during operation. This test is fast
so if there are any maximum power issues this test should be included in the test plan.

5.2. EDR transmitter tests
TRM/CA/10/C – EDR relative transmit power
This test checks the difference in the power distribution between the two types of
modulation schemes. The power control and amplifier circuitry is tested, as well as the
switching between the different modulation types.
TRM/CA/11/C – EDR carrier frequency stability and modulation accuracy
This test checks that the modulation accuracy and frequency stability are within the
required limits set by the test specification. This test is useful to test the overall
modulation quality of the device. It is good for assessing the device as it neutralizes
the effects from inter-symbol interference. This test can also be used to test the crystal
oscillator accuracy.
TRM/CA/12/C – EDR differential phase encoding
This test assesses the modulator in the Bluetooth chip to ensure that it is encoding the
data correctly. This test is useful as it checks a specific area of the chip design, however
it is probably used more in design verification than in a manufacturing situation.
TRM/CA/13/C – EDR in-band spurious emission
Tests if the level of unwanted signal produced from the Bluetooth chip is below a limit
set by the modulation scheme used. This tests the modulator and filters in the chip.

5.3. Receiver tests
RCV/CA/07/C – EDR sensitivity
This measurement tests the minimum signal level required to produce a maximum value
of BER. This can be limited by noise in the test environment so this must be taken into
consideration when constructing a manufacturing test. It tests the performance of the
filter and demodulator for an EDR signal.
RCV/CA/08/C – EDR BER floor performance
This test verifies that the receiver is below the BER maximum limit for normal test
conditions. It is a good general test for the standard operation of an EDR device as
it tests the receiver circuitry for general error conditions. This test could be removed
if functionality testing is used as part of the manufacturing test process.
RCV/CA/09/C – EDR C/I performance
This test is similar to the standard C/I performance test. It tests how co- and adjacent
channel interference affects the signal. This measurement is complicated as it requires
multiple pieces of equipment for conducting the test. It also tests the same areas of the
chip as the sensitivity tests. Due to these factors this test would not be recommended
as part of a manufacturing test plan.
RCV/CA/10/C – EDR maximum input level
This test is similar to the standard version of this test. It measures the receivers BER
performance when the input signal is at a maximum power level.
Another test that is not specified by the Bluetooth SIG test specifications is the EDR
guardband measurement. This is an important measurement as it checks that the
guardband time is within the 4.75 to 5.25 μs time period. The guardband allows time
for the Bluetooth device to switch between the standard GFSK modulation scheme,
to one of the EDR modulation schemes. It can be measured simultaneously with some
of the EDR tests, which means that it should not take any additional time to measure.
Table 2 shows a list of all 23 Bluetooth SIG qualification tests with a priority of high,
medium, and low. The priority is a measure of the general need and usefulness of a test
in a manufacturing environment. This list should be used as a guide in developing a test
plan, while giving consideration to the larger explanation of each test, which is provided
in Section 5. Appendix C contains further information on test equipment needed to
perform each test.

5.4. EDR receiver tests
RCV/CA/07/C – EDR sensitivity
This measurement tests the minimum signal level required to produce a maximum value
of BER. This can be limited by noise in the test environment so this must be taken into
consideration when constructing a manufacturing test. It tests the performance of the
filter and demodulator for an EDR signal.
RCV/CA/08/C – EDR BER floor performance
This test verifies that the receiver is below the BER maximum limit for normal test
conditions. It is a good general test for the standard operation of an EDR device as
it tests the receiver circuitry for general error conditions. This test could be removed
if functionality testing is used as part of the manufacturing test process.
RCV/CA/09/C – EDR C/I performance
This test is similar to the standard C/I performance test. It tests how co- and adjacent
channel interference affects the signal. This measurement is complicated as it requires
multiple pieces of equipment for conducting the test. It also tests the same areas of the
chip as the sensitivity tests. Due to these factors this test would not be recommended
as part of a manufacturing test plan.
RCV/CA/10/C – EDR maximum input level
This test is similar to the standard version of this test. It measures the receivers BER
performance when the input signal is at a maximum power level.

5.5. Other Bluetooth tests
Another test that is not specified by the Bluetooth SIG test specifications is the EDR
guardband measurement. This is an important measurement as it checks that the
guardband time is within the 4.75 to 5.25 μs time period. The guardband allows time
for the Bluetooth device to switch between the standard GFSK modulation scheme,
to one of the EDR modulation schemes. It can be measured simultaneously with some
of the EDR tests, which means that it should not take any additional time to measure.

5.6. Priority of Bluetooth tests

Test priority for manufacturing.
Bluetoothtest Priority
TRM/CA/01/C – Output power High
TRM/CA/02/C – Power density Low
TRM/CA/03/C – Power control Medium/Low
TRM/CA/04/C – Frequency range Low
TRM/CA/05/C – –20 dB bandwidth Medium
TRM/CA/06/C – Adjacent power bandwidth Medium/Low
TRM/CA/07/C – Modulation characteristics High
TRM/CA/08/C – Initial carrier frequency tolerance High
TRM/CA/09/C – Carrier frequency drift High
TRM/CA/10/C – EDR relative transmit power High
TRM/CA/11/C – EDR carrier frequency stability and modulation accuracy High
TRM/CA/12/C – EDR differential phase encoding Low
TRM/CA/13/C – EDR in-band spurious emission Low
RCV/CA/01/C – Single slot sensitivity High
RCV/CA/02/C – Multi slot sensitivity High
RCV/CA/03/C – Carrier/Interference performance Low
RCV/CA/04/C – Blocking performance Low
RCV/CA/05/C – Intermodulation performance Low
RCV/CA/06/C – Maximum input level Medium
RCV/CA/07/C – EDR sensitivity High
RCV/CA/08/C – EDR BER floor performance plan High
RCV/CA/09/C – EDR C/I performance Low
RCV/CA/10/C – EDR maximum input level Medium

When testing a Bluetooth device it is wise to cover both the EDR and the standard
Bluetooth 1.2 tests, as they are vital for testing different areas of the chip. None of the
eight Bluetooth EDR tests test for hopping, so a standard Bluetooth test is needed to test
this aspect of the Bluetooth device. When testing an EDR device, some of the tests in EDR
and standard Bluetooth modes can overlap, which means that one test can be used to
check all three of the modulation schemes specified in the Bluetooth 2.0 standard. In
Section 5 there is an explanation of each of the Bluetooth SIG qualified tests. This can give
a good indication as to which test to include in a manufacturing test plan, enabling you to
make an educated choice on the tests you want to include in a manufacturing test plan.
Test plans will vary depending on the design of the Bluetooth device. an example manufacturing test plan for a Bluetooth 2.0 device. The aim of
this test plan is to test all the main functional blocks of the Bluetooth chip, in the fastest
time possible.
When constructing the test plan, it is a good idea to start with the EDR tests. EDR is a more
recent addition to the Bluetooth standard and it is likely that there will be less characterization
data available for the EDR portions of the design. Selecting the EDR tests first will enable a
broad range of testing to be accomplished, and any gaps in test coverage (for example, in testing
the hopping circuitry) can be addressed by the inclusion of some Bluetooth 1.2 tests.
EDR relative transmit power tests the power amplifier and the effectiveness of switching
between modulation schemes. This test is a key measurement for EDR, so should be
included in most test plans.
EDR carrier frequency stability and modulation accuracy tests the modulator in addition
to the crystal oscillator.
The hopping circuitry should also be tested, however, none of the EDR tests incorporate
hopping. One of the four standard Bluetooth 1.2 transmitter tests should therefore be
selected to test the hopping circuitry. These tests are output power, power density, ICFT,
and carrier frequency drift. When deciding on which of these to use, it is important to
understand what parts of the device are going to be the most susceptible to failure,
refer to Section 5 for more information. Performing each of these tests could be useful
although there is some overlap with the EDR tests so it is important to carefully select
which tests to perform.
The receiver circuitry can be tested using either the EDR BER floor performance test or EDR
sensitivity test. Although the RF test specification does not mandate that hopping is used
for Bluetooth 1.2 or EDR receiver sensitivity tests, it is recommended that hopping is used in
order to test the hopping circuitry in the receiver. In the example test plan, the Bluetooth 1.2
multi-slot sensitivity test was included, with hopping enabled, as the multi-slot test is more
useful for quality control purposes, as described in Section 5.3. The EDR BER floor performance
test was also included in the test plan as a general test of EDR receiver operation.

6. Creating the Test Plan

6.1. Combining Bluetooth standard and EDR tests
When testing a Bluetooth device it is wise to cover both the EDR and the standard
Bluetooth 1.2 tests, as they are vital for testing different areas of the chip. None of the
eight Bluetooth EDR tests test for hopping, so a standard Bluetooth test is needed to test
this aspect of the Bluetooth device. When testing an EDR device, some of the tests in EDR
and standard Bluetooth modes can overlap, which means that one test can be used to
check all three of the modulation schemes specified in the Bluetooth 2.0 standard. In
Section 5 there is an explanation of each of the Bluetooth SIG qualified tests. This can give
a good indication as to which test to include in a manufacturing test plan, enabling you to
make an educated choice on the tests you want to include in a manufacturing test plan.

6.2 Choosing test plans for Bluetooth 2.0 EDR devices

An example of a Bluetooth EDR test plan
Test Packets/bits Channels/hopping Power level Packet type
EDR relative transmit power 1 L, M, H 2-DH1
EDR differential phase encoding
EDR CFSMA 1 L 2-DH1
EDR BER floor performance 300,000 M –60 2-DH1
EDR max input level
EDR sensitivity
Output power
Power control
Modulation characteristics
ICFT 1 Hopping –40 DH1
Carrier drift
Single slot sensitivity
Multi slot sensitivity 100,000 Hopping –70 DH5
Max input power level

The same principles apply in the pre-testing and calibration of a Bluetooth 1.2 device to
those needed to test an EDR device, so the main change in testing will be in the test
plan. When testing a standard Bluetooth 1.2 device that doesn’t use the EDR capability
there are 15 test cases to choose from.
Of the 15 possible test cases, there are six test cases that are classed as high priority.
These six tests are shown below. All these tests can be performed by an Agilent N4010A
Wireless Connectivity Test Set. Refer to Table 2 for more details about the pricing of
the tests.
Table 4 is an example of a generic test plan for a standard Bluetooth device that does
not use EDR. The test is designed to cover all of the major sections of a Bluetooth device
in a very short time span, so it should only be taken as a starting point for a test plan.
The variables of each test are important since each one can be tailored to suit an individual
device plan. For further information about each test chosen please refer to Section 5.
All the modulation types used in Bluetooth 2.0 devices employ frequency hopping. The
effect of the hopping is that the spectral density of the modulation scheme is widened.
This has an obvious effect on some of the spectral measurement tests.
All EDR tests, and the majority of Bluetooth 1.2 tests, are conducted with frequency
hopping switched off. This is done to ensure that specific parts of the chip are working
correctly. It is therefore recommended that a test be conducted that has frequency
hopping involved so that the RF synthesiser can be tested.
The hopping in a Bluetooth system is defined as a slow hop system, because the hopping
rate is less than the symbol rate. This slow hopping enables the device to spread its
spectrum across an 80 MHz range. This means that it is much more difficult for a signal
to fail over the full range of the spectrum. When hopping is on, the power verses channel
measurement, such as the one shown in Figure 4, gives a good indication as to which
channel to test when constructing the test plan. The overall frequency response will
almost always have some variation because certain frequency characteristics in the
design will not be flat. For test purposes, the channels which have maximum and minimum
variations would be a good choice for test channels.

6.3. Choosing a test plan for Bluetooth 1.2 devices
The same principles apply in the pre-testing and calibration of a Bluetooth 1.2 device to
those needed to test an EDR device, so the main change in testing will be in the test
plan. When testing a standard Bluetooth 1.2 device that doesn’t use the EDR capability
there are 15 test cases to choose from.
Of the 15 possible test cases, there are six test cases that are classed as high priority.
These six tests are shown below. All these tests can be performed by an Agilent N4010A
Wireless Connectivity Test Set. Refer to Table 2 for more details about the pricing of
the tests.

6.4. Frequency hopping
All the modulation types used in Bluetooth 2.0 devices employ frequency hopping. The
effect of the hopping is that the spectral density of the modulation scheme is widened.
This has an obvious effect on some of the spectral measurement tests.
All EDR tests, and the majority of Bluetooth 1.2 tests, are conducted with frequency
hopping switched off. This is done to ensure that specific parts of the chip are working
correctly. It is therefore recommended that a test be conducted that has frequency
hopping involved so that the RF synthesiser can be tested.
The hopping in a Bluetooth system is defined as a slow hop system, because the hopping
rate is less than the symbol rate. This slow hopping enables the device to spread its
spectrum across an 80 MHz range. This means that it is much more difficult for a signal
to fail over the full range of the spectrum. When hopping is on, the power verses channel
measurement, such as the one shown in Figure 4, gives a good indication as to which
channel to test when constructing the test plan. The overall frequency response will
almost always have some variation because certain frequency characteristics in the
design will not be flat. For test purposes, the channels which have maximum and minimum
variations would be a good choice for test channels.

An example of a non-EDR test plan
Test Packets/bits Channels/hopping Power level Packet type
Output power 1 Hopping –40 DH1
Modulation characteristics 5 M –40 DH1
ICFT 1 Hopping –40 DH1
Carrier drift 5 M –40 DH1
Single slot sensitivity
Multi slot sensitivity 100,000 Hopping –70 DH5
Agilent N4017A GMA showing a power verses channel graph.

With the Bluetooth 2.0 EDR standard, two new types of modulation scheme were
introduced. These new modulation types, /4 DQPSK and 8DPSK, are in addition to the
standard GFSK modulation scheme. The addition of two new modulation schemes has
increased the complexity of the chip design. Unlike GFSK modulation, /4 DQPSK and
8DPSK do not have a constant transmission envelope, and therefore these schemes can
exhibit the effects of non-linearity in the transmissions. The Bluetooth Special Interest
Group has defined eight new tests in the 2.0 test specification to cover these new
modulation schemes.
The type of modulation scheme used for EDR tests depends on the device being tested.
The Bluetooth standard defines the 8DPSK modulation type as being an optional
enhancement to the Bluetooth standard, so it would only be used in the test plan if
the device being manufactured supported it. If both types of modulation are supported,
then it would be recommended to use the 8DPSK modulation scheme since it has the
highest data rate. Using 8DPSK you could conclude that if the test was successful, then
the other modulation types would also work.
Once the tests for the test plan have been selected, the order that the tests are performed
must be chosen. The ideal test plan would have the tests that are most likely to fail at
the start of the test so that a faulty Bluetooth device can be removed from the production
line, to avoid wasting time by conducting further tests.
Another consideration when defining the order of tests is that some tests use a common
signal and can be measured simultaneously. For example, output power and modulation
characteristics can be measured at the same time using an Agilent N4010A test set.
Tests should be performed simultaneously whenever possible, as this can significantly
reduce time in the manufacturing process.
Once the individual tests have been chosen, the test parameters must be selected.
Test parameters are an important factor as they can cause test times to vary by a large
degree. The main factor that affects test time is the number of bits or packets that are
processed for the test. The Bluetooth qualification tests can have up to eight million
bits processed for one test. Using this many bits is unrealistic for a manufacturing
environment since the test would take too long.
For manufacturing test, the key is to find a balance between time for the test and the
confidence that the test will produce valid results for that device. When planning the
parameters it is important to look back at the research conducted on the device to know
its weaknesses, so that if a section of the design is more likely to fail, more testing
should be conducted on this part of the plan. If the number of bits is reduced for a test
plan then the parameters for passing the test should also be changed so that the test
gives an accurate result as possible. (Refer to Appendix C for information about test line
limits (TLL).)
When deciding upon which packet type to use, each test in the plan has to be considered
individually. The majority of tests will use a one-slot packet to ensure that the fastest
possible transmission time is achieved.
When testing with hopping switched off the standard practice is to test the lowest
middle and highest frequencies in the Bluetooth band. These are 2402, 2441, and
2480 MHz respectively. When constructing a manufacturing test plan, just one of these
bands is normally chosen. The chosen band depends on which frequency is most likely
to produce the greatest number of errors. It is general practice to test at the weakest
point of a design, as this is most likely to show up any errors during the manufacturing
process.

6.5. Modulation types
With the Bluetooth 2.0 EDR standard, two new types of modulation scheme were
introduced. These new modulation types, /4 DQPSK and 8DPSK, are in addition to the
standard GFSK modulation scheme. The addition of two new modulation schemes has
increased the complexity of the chip design. Unlike GFSK modulation, /4 DQPSK and
8DPSK do not have a constant transmission envelope, and therefore these schemes can
exhibit the effects of non-linearity in the transmissions. The Bluetooth Special Interest
Group has defined eight new tests in the 2.0 test specification to cover these new
modulation schemes.
The type of modulation scheme used for EDR tests depends on the device being tested.
The Bluetooth standard defines the 8DPSK modulation type as being an optional
enhancement to the Bluetooth standard, so it would only be used in the test plan if
the device being manufactured supported it. If both types of modulation are supported,
then it would be recommended to use the 8DPSK modulation scheme since it has the
highest data rate. Using 8DPSK you could conclude that if the test was successful, then
the other modulation types would also work.

6.6. Test order
Once the tests for the test plan have been selected, the order that the tests are performed
must be chosen. The ideal test plan would have the tests that are most likely to fail at
the start of the test so that a faulty Bluetooth device can be removed from the production
line, to avoid wasting time by conducting further tests.
Another consideration when defining the order of tests is that some tests use a common
signal and can be measured simultaneously. For example, output power and modulation
characteristics can be measured at the same time using an Agilent N4010A test set.
Tests should be performed simultaneously whenever possible, as this can significantly
reduce time in the manufacturing process.

6.7. Test parameters

When testing the receiver circuitry the main measure of quality is the BER. The BER
can be affected by many issues, from the design to the operating environment. When
choosing parameters for the test plan it is important that the type of data being sent in
the packet is known since the data can cause different levels of errors. Because of this
it is wise to use the same data load that the Bluetooth test specification uses for its
qualification tests.
BER is a statistical analysis of the transmission accuracy; the number of bits that
is sampled greatly influences the result. This becomes a problem when planning
manufacturing tests, making it difficult to find the balance between the time for test
and the accuracy of the test result. Even in a well designed Bluetooth chip there will
always be some random noise that causes a BER. As some of these errors will be
random, it makes it difficult to scale down the number of bits that are tested without
changing the BER pass level.
Some manufacturers may scale down the BER test parameters of the Bluetooth SIG
using ratios to reduce the time for the measurement. This technique is the most basic
of all BER reduction methods so there is an element of error associated with it. Other
more complex statistical BER test time reduction may be used for a higher degree of
accuracy. This technique works by using the formula shown below. This formula allows
the test to have threshold pass rates so that if there are very few errors then the test
will pass the device quickly or if the device is near the fail threshold it will be tested
with a larger number of bits to ensure an accurate result. These limits can be written
into a manufacturing program and can be measured using a test set such as the
Agilent N4010A test set.
The formula can be used to work out what BER level to set as a pass depending on how
many bits are sent correctly. The formula can be solved empirically using a computer.
E is the total number of error bits when working out error values. If there are no errors
in the data, the second term equals zero, which greatly simplifies the equation. N is the
number of bits that the system is transmitting. K is the range of error numbers which
determines how many times the second part of the formula is calculated. The CL is the
confidence level of the test given by a percentage as to how correct the test results are.
The formula to calculate the confidence level is:
This formula is the probability that the transmitted BERT will be less than the BER limit,
R. It is up to the manufacturer to choose the desired confidence level of the device
during testing.
Functional testing is used to test a Bluetooth chip for a specific use, such as audio
capabilities or data transfer. When deciding on functional testing, it is vital for the test
engineer know what the main function of the device will be so that an appropriate
test can be included in the test plan. Once the main function is known then a functional
test can be constructed to simulate the device in operation. The tests can range from
a simple data link connection to audio quality testing. Many simple functional tests are
covered by the Bluetooth SIG test cases, so they do not have to be covered by additional
functional testing. Setup information for data connection and audio functionality tests

6.8. BER measurements
When testing the receiver circuitry the main measure of quality is the BER. The BER
can be affected by many issues, from the design to the operating environment. When
choosing parameters for the test plan it is important that the type of data being sent in
the packet is known since the data can cause different levels of errors. Because of this
it is wise to use the same data load that the Bluetooth test specification uses for its
qualification tests.
BER is a statistical analysis of the transmission accuracy; the number of bits that
is sampled greatly influences the result. This becomes a problem when planning
manufacturing tests, making it difficult to find the balance between the time for test
and the accuracy of the test result. Even in a well designed Bluetooth chip there will
always be some random noise that causes a BER. As some of these errors will be
random, it makes it difficult to scale down the number of bits that are tested without
changing the BER pass level.
Some manufacturers may scale down the BER test parameters of the Bluetooth SIG
using ratios to reduce the time for the measurement. This technique is the most basic
of all BER reduction methods so there is an element of error associated with it. Other
more complex statistical BER test time reduction may be used for a higher degree of
accuracy. This technique works by using the formula shown below. This formula allows
the test to have threshold pass rates so that if there are very few errors then the test
will pass the device quickly or if the device is near the fail threshold it will be tested
with a larger number of bits to ensure an accurate result. These limits can be written
into a manufacturing program and can be measured using a test set such as the
Agilent N4010A test set.
The formula can be used to work out what BER level to set as a pass depending on how
many bits are sent correctly. The formula can be solved empirically using a computer.
E is the total number of error bits when working out error values. If there are no errors
in the data, the second term equals zero, which greatly simplifies the equation. N is the
number of bits that the system is transmitting. K is the range of error numbers which
determines how many times the second part of the formula is calculated. The CL is the
confidence level of the test given by a percentage as to how correct the test results are.
The formula to calculate the confidence level is:
This formula is the probability that the transmitted BERT will be less than the BER limit,
R. It is up to the manufacturer to choose the desired confidence level of the device
during testing.

6.9. Functional testing
1
E (N X BER)k
N = [– In (1 – CL) + In ()] BER k–0 k!
CL = PROB [BERT < R] given E and N
After the test plan been selected a manufacturing engineer can then decide on the test
equipment needed for the design of the whole test system. The majority of the Bluetooth
calibration and manufacturing test can be carried out using an Agilent N4010A test set.
Other tests may be chosen that require additional equipment such as spectrum analyzers
and power meters. It is the test engineer’s responsibility to design the system and
choose the equipment that best fits the test need. (Refer to Appendix C for a full list
of Agilent equipment for the Bluetooth qualification tests.)
If a device fails in production, a more comprehensive test plan performed outside of
the production line should be used. This secondary test should be written to thoroughly
exercise the device to find where the fault lies. Once the area of the fault has been
detected, the manufacturer may choose to repair the device and return it to the start
of the high volume test line.

7. Other Important

Manufacturing Issues
For detailes on the Bluetooth SIG tests covered in this application note, please refer to
the Application Note 1331-1 Bluetooth Measurement Fundamentals. This appendix will
describe measurement setup for common tests or calibrations that may be performed
during manufacturing but are not detailed by the Bluetooth SIG.
There are a number of different calibration tests, of which many are vendor-specific.
Detailed below is a method for calibrating two of the most generic parameters.
Crystal tuning
Crystal tuning is vital for the operation of the circuit since it sets the operating frequency
of the device so that it is transmitting on the correct frequency. This can be performed
using an N4010A test set. The N4010A is set to RF analyzer mode where it can measure
the frequency offset and average power.
Power calibration
Power calibration is set up in much the same way as the crystal tuning using an N4010A
test set in RF analyzer mode.
There are two main ways that data and audio connections can be made: through HCI
loopback and using audio profiles. Audio connections are established in Synchronous
Connection Orientated (SCO) state. Data and audio routing can be established using HCI
loopback using a N4010A test set. When this mode is enabled, the audio quality of the
Bluetooth device can be tested using an audio analyzer program. As illustrated in
Figure 5, the test setup is similar to the standard Bluetooth test setup for test mode.
There are two HCI loopback modes in which an N4010A Bluetooth test set can operate.
Local loopback sends back any data and command packets to the device without the
host controller altering any data. This type of test can check the audio interfaces of
the Bluetooth device. Remote loopback sends back all data that is transmitted and no
analogue signal is recovered. The test is executed this way so that the signal can only
be degraded by the DUT.
Another way to test Bluetooth audio is by using a profile that describes the procedure
a device must adhere to so that it can communicate with another device. A profile
testing setup is shown in Figure 6.