Antenna design and RF layout guidelines

cypress com Document No 001 91445 Rev H 1 AN91445 Antenna Design and RF Layout Guidelines Authors Tapan Pattnayak, Guhapriyan Thanikachalam Associated Part Family CY8C4XXX BL, CYBL1XXXX, CY8C6XXX. cypress com Document No 001 91445 Rev H 1 AN91445 Antenna Design and RF Layout Guidelines Authors Tapan Pattnayak, Guhapriyan Thanikachalam Associated Part Family CY8C4XXX BL, CYBL1XXXX, CY8C6XXX.

Meandered Inverted-F Antenna (MIFA)

The MIFA antenna is a compact and efficient solution commonly used in human interface devices (HIDs), designed by Cypress for optimal performance in a small form factor of 7.2 mm × 11.1 mm This makes it ideal for applications such as wireless mice, keyboards, and presenters The recommended MIFA layout, illustrated in Figure 10, includes detailed top and bottom layer designs on a two-layer PCB, with a consistent antenna trace-width of 20 mils The RF trace width, denoted as "W," may vary based on the PCB stack spacing, impacting overall performance.

Bottom Layer (RF Ground Layer)

Orange: Top Layer Light Blue: Bottom Layer All dimensions are in mils

Transmission line 50 ohm to matching network

Light Blue: Bottom Layer All dimension are in mils

Antenna Design and RF Layout Guidelines

Antenna Feed Consideration

Table 2 presents the "W" value for various PCB thicknesses between the top and bottom layers of a two-layer FR4 substrate, which has a relative dielectric constant of 4.3, specifically for a coplanar waveguide model In this setup, the top layer features the antenna trace, while the bottom layer includes a solid RF ground plane Additionally, the remaining area of the bottom layer can serve as a signal ground plane for PRoC/PSoC and other circuitry Figure 11 illustrates the relationship between PCB thickness and "W" for a standard two-layer PCB.

Table 2 Value of “W” for FR4 PCB: Thickness Between Antenna Layer and Adjacent RF Ground Layer

Figure 11 Clarification of PCB Thickness

When designing a PCB trace for an antenna, the width requirement can be less stringent due to the trace's short length It is important to maintain the same width for both the antenna trace and its feed connection As illustrated in Figure 12, there are instances where the trace width feeding the antenna does not meet the wider specifications outlined in Table 2.

Figure 12 Antenna Feed Width for Short Trace

For long transmission lines, approximately 1 cm from the matching network to the antenna or back to the ANT pin of the PRoC/PSoC BLE device, Cypress advises using a transmission line (TLine) layout This layout should feature a specific width "W" over a bottom ground plane to ensure optimal feeding.

Note: See the coplanar wave guide calculator in Appendix B for the calculation of width for Coplanar transmission line

Only 3 mm trace to Antenna Transmission line width is not critical

Figure 13 plots S11 of the MIFA The MIFA has a bandwidth (S11 ≤ –10 dB) of 230 MHz around 2.44 GHz

Figure 13 S11 of the MIFA (Return Loss = –S11)

The complete 3D radiation-gain pattern of the MIFA at 2.44 GHz, illustrated in Figure 14, is essential for optimizing antenna placement in custom applications to enhance radiation in specific directions In this diagram, the antenna is positioned in the XY plane, with the Z-axis representing the vertical orientation.

Figure 14 3D Radiation-Gain Pattern for MIFA

The radiation pattern of a MIFA antenna mounted on a Pioneer Board is tested with a 30-degree angular resolution, utilizing metal connecting headers It is important to note that the radiation pattern may vary on a bare board compared to the illustrated example, which serves only as a guide for antenna positioning on a PCB To optimize antenna performance, it is recommended to measure a similar radiation pattern in your final product assembly to identify the best placement for the antenna.

Antenna Length Considerations

To optimize the antenna radiation impedance and frequency selectivity of the MIFA antenna, it is essential to adjust its length according to the thickness of the PCB Cypress provides recommended antenna lengths for different board thicknesses, as detailed in Table 3.

Table 3 Leg and Tip length

Figure 15 shows two MIFA antennas for two different board thicknesses Antenna designers should refer to Table 3 for adjusting the length of the MIFA antennas for a specific board thickness

When adjusting the original antenna, it’s important to start with its full length and then modify it based on the board's thickness, as shortening the antenna is easier than lengthening it Use Table 3 as a guideline for determining the final antenna length according to the specific board thickness rather than seeking an exact measurement For a quick tuning method, cutting the antenna length can be effective However, if the customer has sufficient space for a matching network component and the expertise for antenna tuning, Cypress recommends utilizing a matching network instead of simply adjusting the length.

Inverted-F Antenna (IFA)

The Inverted F Antenna (IFA) outperforms the Monopole Inverted F Antenna (MIFA) in terms of radiation efficiency, making it a superior choice when space permits While the IFA requires a larger area than the MIFA, its enhanced efficiency justifies the additional space needed for optimal performance.

The Inverted-F Antenna (IFA) is ideal for applications with constrained antenna dimensions, such as heart rate monitors As illustrated in Figure 16, the recommended IFA design features both top and bottom layer layouts on a two-layer PCB, with a trace width of 24 mils This IFA measures 4 mm × 20.5 mm (157.5 mils × 807 mils) and is optimized for an FR4 PCB with a thickness of 1.6 mm Notably, the IFA exhibits a larger aspect ratio (width to height) compared to the Monopole Inverted-F Antenna (MIFA).

Bottom Layer (RF Ground Layer)

The Gerber and brd files for a 1.6-mm thick FR4 PCB can be downloaded from the AN91445.zip file available at www.cypress.com/go/AN91445.

Orange: Top Layer Light Blue: Bottom Layer All dimension are in mils

The feed trace width “W” of the MIFA antenna is influenced by the PCB stack of the product Table 4 outlines the “W” values corresponding to various PCB thicknesses, measuring the distance between the top layer, which houses the antenna, and the bottom layer.

RF ground layer) for an FR4 substrate (relative dielectric constant = 4.3) for coplanar waveguide model

Table 4 Value of “W” for FR4 PCB: Thickness between Antenna Layer and Adjacent RF Ground Layer for 50-ohm

For antenna feeds with short traces measuring less than 3 mm, the trace width can be adjusted to match that of the antenna trace itself, as illustrated in Figure 12 For precise calculations of the width for coplanar transmission lines, please consult the coplanar waveguide calculator in Appendix B.

The bandwidth (S11≤ –10 dB) of the IFA is 220 MHz around 2.44 GHz, as shown in Figure 17

Figure 17 S11 of the IFA (Return Loss = -S11)

Figure 18 illustrates the qualitative radiation pattern of an Inverted F Antenna (IFA) in the XY plane, providing essential insights for optimal placement in custom applications to enhance radiation in specific directions While this overview presents only a qualitative direction, comprehensive radiation patterns across the XY, YZ, and ZX planes are available by contacting Cypress Technical Support.

Figure 18 Qualitative 2D Radiation Gain Pattern for IFA

For compact PCB designs, chip antennas are an effective solution, offering minimal size and reasonable performance These off-the-shelf antennas occupy little PCB space but can increase the bill of materials (BOM) and assembly costs, as they are external components that require separate purchasing and assembly The cost of chip antennas typically ranges from 10 to 50 cents, influenced by their size and performance characteristics.

When utilizing chip antennas, it's crucial to pay attention to the size of the RF ground, as it significantly impacts performance Adhering to the manufacturer's guidelines regarding ground size is essential Unlike PCB antennas, chip antennas cannot be adjusted by altering their length; instead, they necessitate an additional matching network for proper tuning, which further elevates the bill of materials (BOM) costs.

Cypress suggests chip antennas only for specialized applications that demand an extremely small PCB area.For such applications, Cypress recommends the Johansson Technology antennas mentioned below

The 2450AT18B100E has dimensions of 63 mils × 126 mils; the 2450AT42B100E has bigger dimensions of 118 mils × 196 mils but provides a better RF performance

The Cypress BLE module CYBLE-022001-00, equipped with the 2450AT18B100E antenna, has undergone thorough characterization for RF performance and pre-compliance testing To achieve optimal RF performance, specific layout guidelines must be followed for the chip antenna Key considerations for the placement and layout of the chip antenna are essential for ensuring effective operation.

1 Ground clearance around the antenna

2 Antenna placement for optimal radiation

4 Antenna matching network for bandwidth extension

Figure 20 and Figure 21 show the layout guidelines for the chip antenna from Johanson Technology 2450AT42B100E See their website for detailed guidelines for these antennas

Figure 20 Layout Guideline for Johanson 2450AT42B100E Chip Antenna

The layout features a 50-Ω transmission-line feed along with matching components, where the width of the transmission line is contingent on the board thickness, with the precise width specified in Table 4.

Figure 21 Johanson Antenna Layout Guideline for 24AT42B100E

The performance of chip antennas is significantly influenced by the size of the ground plane, necessitating a larger ground plane and increased spacing for optimal functionality For the 2450AT42B100E model, a minimum ground clearance of 0.8 mm from the antenna edge to the ground edge is recommended, although a clearance of 2-3 mm yields improved return loss.

The Johanson chip antenna (2450AT42B100E) exhibits directional radiation rather than isotropic behavior, with the direction of maximum radiation influenced by factors such as ground clearance and plastic assembly.

Figure 22 Radiation Pattern from Chip Antenna

Cypress recommends chip antennas for specialized applications requiring minimal PCB space, like nano Bluetooth dongles or ultra-compact modules The Johansson antenna is noted for its RF performance and pre-compliance with the Cypress EZ-BLE module Additionally, other chip antennas from reputable vendors such as Murata, Vishay, Pulse, and Taoglas are also available for use.

Wire antennas are traditional antennas characterized by their quarter-wave length conductors These antennas are mounted on printed circuit boards (PCBs) and extend above the PCB surface, reaching into free space while being supported by a ground plane.

They have excellent RF performance as they are exposed to space as a 3D antenna They have the best range and have the most isotropic radiation pattern

For BLE applications where space is limited, traditional antennas may not be ideal due to their size and height However, in scenarios where space is not an issue, these antennas excel in RF range, directivity, and radiation patterns, making them suitable for devices like smart home controllers that plug into walls It is essential to optimize the wire shape and size to fit specific industrial designs, allowing for bending to match the enclosure Careful manufacturing is crucial, as wire antennas can take on various shapes depending on the design requirements.

8 Only Johansson antenna is characterized; others are not.

Effect of Ground Plane

As explained before, a monopole PCB antenna requires a ground plane for proper operation

Figure 25 shows an example where a MIFA is placed on a PCB with a different ground plane size The PCB size varies from 20 mm × 20 mm to 50 mm × 50 mm

Larger RF ground planes effectively lower the resonant frequency, while improved grounding enhances return loss, which is essential for optimal PCB layout A well-designed ground for quarter-wave antennas ensures a closer alignment with theoretical performance, making this principle vital in antenna design for compact modules where space for ground clearance is limited.

Figure 25 Effect of PCB Ground Plane Size

Effect of Enclosure

To assess the impact of a product's plastic casing on antenna sensitivity, experiments were conducted using a wireless mouse The Cypress MIFA antenna was positioned within the mouse's plastic casing, and subsequent measurements were taken to evaluate the radiation pattern and return loss.

Figure 26 Effect of Plastic Casing

Both Figure 25 and Figure 26 reveal some important observations:

 The resonant frequency shifts to a lower frequency when the antenna is placed near the plastic casing

The resonant frequency shift ranges from 100 MHz to 200 MHz, necessitating a retuning of the antenna to align it with the desired frequency band For effective antenna tuning, refer to the Guidelines for Antenna Placement, Enclosure, and Ground Plane.

In conclusion, increasing the ground plane size and plastic casing tends to decrease the resonant frequency of the antenna by approximately 100 MHz to 200 MHz

12 Guidelines for Antenna Placement, Enclosure, and Ground Plane

 Always place the antenna in a corner of the PCB with sufficient clearance from the rest of the circuit

 Always follow the antenna designer’s/manufacturer’s recommended ground pattern for the antenna Commonly used PCB antennas are variants of a monopole antenna Monopole antennas need solid ground for proper operation

 Never place any component, planes, mounting screws, or traces in the antenna keep-out area across all layers The actual keep-out area depends on the antenna used

When designing industrial products, it's crucial to avoid placing the antenna near plastic materials, as plastic has a higher dielectric constant than air This proximity causes the antenna to experience an increased effective dielectric constant, which in turn lengthens the electrical path of the antenna trace and lowers its resonant frequency.

 The battery cable or mic cable must not cross the antenna trace

To ensure optimal performance, the antenna should never be entirely encased in metal If the product features a metallic casing or shield, it is crucial that this enclosure does not obstruct the antenna Additionally, the antenna's near-field area must remain free of any metal to avoid interference.

 The orientation of the antenna should be in line with the final product orientation so that the radiation is maximized in the desired direction

 There must not be any ground directly below the antenna See Figure 14

 There must be enough ground at a distance (ground clearance) from the antenna and this ground plane must have a minimum width See Figure 10, Figure 15, and Figure 20

When designing an antenna, it is essential to include a provision for an antenna matching network, as various factors such as nearby materials, ground variations, and substrate differences can affect its impedance To ensure optimal performance, the antenna may require retuning if its impedance is not known.

The PI or T network consists of three components, featuring 0 ohms in series components and no load for shunt components This configuration enables the creation of any desired topology for an effective matching network in future applications.

When utilizing the matching network values from the antenna manufacturer, it is crucial to adhere to the specified trace length between the antenna and the matching network as outlined in the manufacturer's datasheet or reference design.

Ensure that the antenna matching network is tested with the final plastic enclosure installed and the product positioned in typical usage scenarios For instance, evaluate a mouse while it is held in hand and used on various surfaces such as a mouse pad, plastic, wood, metal, or the floor.

RF layout and antenna tuning necessitate a thorough comprehension of RF-specific principles, requiring greater focus compared to standard circuit layouts This section covers fundamental aspects of RF design, including transmission lines and their characteristic impedance.

The following concepts and terminologies need to be understood to design an effective RF layout

The impedance of RF circuits significantly impacts RF design compared to analog design, particularly at high frequencies Unlike low frequencies, where load impedance remains consistent regardless of measurement distance and trace width, high-frequency RF circuits exhibit varying impedance based on distance from the load, substrate material, and trace dimensions Consequently, RF traces must be carefully considered as critical design elements in RF schematics.

Transmission lines are essential components that transport electromagnetic energy along a specific route Common examples include coaxial cables, waveguides, and RF traces connecting the RF pin to the antenna Predominantly, RF traces are categorized as transmission lines, with microstrip lines and coplanar waveguides being the most common types.

The characteristic impedance (Z0) is a crucial property of a transmission line, defined as the ratio of voltage to current amplitudes of a wave traveling through a lossless medium In applications like Bluetooth Low Energy (BLE) operating at 2.45 GHz, a 50-ohm characteristic impedance is commonly employed for RF traces.

Figure 27 Equivalent Model of a Transmission Line

Z0, although a real number, does not represent the resistance of the RF trace In an ideal transmission line, energy dissipation and loss are absent, attributed to its characteristic impedance The equivalent model of a transmission line illustrates the relationship between distributed series inductance and distributed shunt capacitance.

Where L and C are distributed inductance and distributed capacitance respectively along an arbitrary length of the transmission line

The characteristic impedance (Z0) is influenced by several key factors, including the PCB material, substrate thickness, trace width, trace thickness, and the clearance between the RF trace and the ground fill While these parameters are frequently overlooked in traditional layout and design practices, they are crucial for effective RF design.

Figure 28 Representation of an Impedance Measurement Setup

Figure 28 illustrates a standard measurement setup for assessing the impedance of an RF circuit, where the impedance at a specific point on the RF trace is influenced by the trace's characteristic impedance, its distance from the load, and the load impedance, as outlined in the accompanying equation.

Where Z is the impedance measured at a distance l from the load, ZL is the impedance measured at the load (l = 0),

Z0, is the characteristic impedance of the transmission line, and β is the phase constant J is the reactive part of the impedance

Let’s check how the impedance changes in certain special scenarios

When measured at the load, l = 0, so Z becomes equal to ZL

Smith Chart

In RF design, the Smith chart is an essential graphical tool for plotting complex impedance and designing matching networks It facilitates the quick calculation of various parameters, including admittance, return loss, insertion loss, reflection coefficient, Voltage Standing Wave Ratio (VSWR), and transmission coefficient Additionally, the Smith chart allows for the calculation of impedance variations with distance from the load, making it invaluable for designing matching networks using RF stubs or passives.

Note the following in the diagram:

1 The left corner of the Smith chart indicates zero ohms and the right corner indicates open circuit

2 The circles touching the right corner are constant-resistance circles

3 The real part of the impedance is constant across all points in a constant-resistance circle

4 The curves between the right corners and the periphery of the Smith chart are constant-reactance circles

5 The imaginary part of the impedance is constant at all points along a constant-reactance curve

6 The circles in Smith chart that touch the left corner are constant-conductance circles

7 The real part of the admittance is constant along a constant-conductance circle

8 The curves between the left corner of the Smith chart and periphery of the Smith chart are constant-susceptance curves

9 The imaginary part of the admittance is constant along a constant-susceptance curve

10 The center of the circle is the Z0 point In this case, Z0 = 50 ohms.This is also the 20-millisiemens (mS) point

11 Two special circles are the 50-ohm circle and 20-mS circle

Impedance matching begins by transforming the impedance to align with either the 50-ohm circle or the 20-ms circle The next step involves adjusting the impedance from one of these circles to the 50-ohm point Additionally, the topology of the matching network is influenced by whether the impedance is located inside or outside these circles.

This application note provides guidance on designing a matching network utilizing Smith charts For additional details on Smith charts, please consult the user guides and tutorials available online, with links provided below.

 http://www.microwaves101.com/encyclopedias/smith-chart-basics

 https://www.youtube.com/watch?v=vDU5XnvZXwc

Figure 31 Smith Chart with Impedance and Admittance Circles

Impedance matching is essential for optimizing power delivery from an RF source to a load, such as in the case of PRoC BLE/PSoC BLE systems During transmission, the PSoC BLE acts as the source while the antenna serves as the load; conversely, during reception, the antenna becomes the source and the PSoC BLE is the load To ensure efficient performance, both the PSoC BLE and the antenna must be matched to a standard impedance of 50 ohms It is important to note that at RF frequencies, impedance varies with distance from the load or source, which can be visualized on a Smith chart where impedance rotates clockwise as one moves away from the load/source Figure 32 illustrates how impedance changes with trace length.

Figure 32 Smith Chart Depicting Impedance Change with Trace Length

The matching network must be adjusted based on the distance from the source or load When the measured impedance matches the characteristic impedance, it remains consistent regardless of the distance from the source or load.

To achieve optimal performance, it is essential to match the complex source impedance to the characteristic impedance with a matching network positioned near the source, while also aligning the load impedance to the characteristic impedance using a matching network located near the load This approach guarantees that the values of the matching network components remain consistent regardless of the trace length, provided that the source matching network is maintained close to the source and the load matching network is kept near the load.

For 2.4 GHz, most of the devices available are matched for 50-ohm impedance For this reason, Cypress uses and recommends a 50-ohm characteristic impedance for the RF trace

Impedance matching to 50 ohms can be achieved using two reactive passive components, such as inductors or capacitors, for any impedance except short and open circuits Although RF stubs can provide the necessary inductance and capacitance, they often require extra space on the PCB Due to size limitations, utilizing capacitors and inductors is the more efficient choice for impedance matching.

Incorporating a series inductor shifts the impedance clockwise along the constant resistance circle on the Smith chart The required inductance to adjust the reactance by a factor of XL can be determined using a specific equation.

Incorporating a series capacitor shifts the impedance on the Smith chart in a counterclockwise direction along the constant resistance circle To achieve this movement, the required capacitor value must adjust the reactance by a factor of XC.

Adding a shunt inductor moves the impedance along the constant conductance circle in an anticlockwise direction The inductor value needed to move the conductance by YL is

Adding a shunt capacitor moves the impedance along the constant conductance circle in a clockwise direction The capacitor value needed to move the conductance by YC is

To design an effective matching circuit using the Smith chart, the initial step involves adjusting the impedance to the 50-ohm or 20-mS circle Following this, the impedance should be further moved to the precise 50-ohm point With this foundational understanding, capacitors and inductors can be employed to achieve the desired impedance matching.

Figure 33 Smith Charts Depicting Impedance Changes with Addition of Reactance

Matching Network Topology

To effectively transform any impedance to 50 ohms, the topology of the necessary components is contingent upon the measured impedance This impedance measurement is best conducted using a vector network analyzer and should be taken at a location very close to the matching network for optimal accuracy.

When the measured impedance is located within the 50-ohm circle on the Smith chart, it is necessary to employ either a shunt inductor followed by a series capacitor or a shunt capacitor followed by a series inductor from the load The shunt component effectively adjusts the impedance to align with the 50-ohm circle, while the series component fine-tunes the impedance to reach the 50-ohm point.

Figure 34 Matching Network Topologies to Use When Impedance Is Within 50-Ohm Circle

When the measured impedance is located within the 20-mS circle on the Smith chart, it requires either a series capacitor followed by a shunt inductor or a series inductor followed by a shunt capacitor from the load By utilizing the series component, the impedance can be adjusted to align with the 20-mS circle, and subsequently, the shunt component can be employed to reach the 20-mS (50-ohm) point.

Figure 35 Matching Network Topologies to Use When Impedance Is Within 20-mS Circle

When the impedance falls outside the designated circles on the positive half of the Smith chart, it can be matched using either a series capacitor followed by a shunt inductor or capacitor, or by employing a shunt capacitor followed by a series inductor or capacitor from the load.

Figure 36 Matching Network Topologies to Use When Impedance Is Outside the Two Circles, on Positive Half of

When the impedance measured is located outside the two circles on the negative half of the Smith chart, it can be matched using a combination of inductors and capacitors This can be achieved by employing a series inductor followed by either a shunt inductor or shunt capacitor, or by utilizing a shunt inductor followed by a series inductor or capacitor from the load, as illustrated in Figure 37.

Figure 37 Matching Network Topologies to Use When Impedance Is Outside the Two Circles, on Negative Half of

Tips for Matching Network

Use the following tips to minimize the gap between theory and practice in the matching network design:

 Measure the impedance at the same point where components have to be placed

 Calibrate the network analyzer setup with cables and connectors until the impedance measurement point

 Place the shunt components on the RF trace itself Do not use long traces to connect to the shunt components

 Choose capacitors that have a series resonant frequency at least twice the frequency of operation

 Choose inductors that have a self-resonant frequency at least twice the frequency of operation

 If the parasitic impedance data is available in the datasheet, use it to derive the actual reactance achievable with that component

 Use only high-Q components for both capacitors and inductors

Because the impedance is typically unknown during design time, a design with three components in a П or T fashion allows you to use all the possible topologies later

Antenna tuning is crucial for optimizing performance, as it ensures a return loss greater than 10 dB when viewed from the chip output towards the antenna within the desired frequency band This process is equally important when assessing the radio's impedance in receive mode, ensuring it remains at 50 Ω Achieving a return loss greater than 10 dB guarantees that 90% of the chip's power output is effectively transmitted to the antenna, while also ensuring that 90% of the received power is successfully transferred to the radio Both antenna tuning and radio tuning are collectively referred to as antenna tuning.

Maximizing power transfer involves matching the output impedance of the radio to the complex conjugate of the antenna impedance Typically, this is accomplished in antenna tuning by transforming the antenna impedance to 50 ohms and utilizing a balun.

In matching network design, passive components are utilized to achieve a 50-ohm impedance For a brief introduction to the principles of matching network design, please see Section 14 Additional resources and references on matching network design can be found in Appendix B.

Figure 38 Reference for Tuning and Matching Network

The 50-ohm reference point is linked to the network analyzer port, allowing for efficient antenna tuning During this process, the chip side is disconnected by removing the balun-matching components, and the antenna-matching components are also detached This 50-ohm standard is advantageous, as it aligns with the impedance of most conventional instruments, ensuring compatibility and ease of use.

In many applications utilizing Cypress MIFA, the antenna can be efficiently tuned to a 50-ohm impedance primarily through PCB length design, often requiring no additional components However, when using a non-50-ohm chip antenna, it may be necessary to incorporate two or more components to achieve the desired impedance, in accordance with the manufacturer's recommendations Overall, the radio side typically needs just two components to maintain a 50-ohm impedance in receive mode.

The following section describes the step-by-step procedure for antenna tuning by using a network analyzer For antenna tuning, you need to look towards the antenna.

Tuning Procedure

Antenna tuning is a crucial two-step process that ensures optimal performance Initially, the bare PCB is tuned to the desired frequency band Following the finalization of the industrial design, the tuning is reevaluated to account for the effects of the plastic enclosure and human body contact, which can detune the antenna and impact return loss.

To effectively tune antennas using a Network Analyzer, a basic understanding of the Smith Chart is essential This tool helps in analyzing key S parameters, including S11, which indicates return loss, and S21, which measures the forward transmission ratio For those seeking to deepen their knowledge, additional resources are available through the links provided below.

As the first step, the network analyzer is calibrated, and then the antenna is tuned by adjusting the matching network components and verifying the tuning in the Smith chart

The tuning procedure uses the following:

 Agilent 8714ES network analyzer (calibrated)

 Cypress CY5682 kit mouse as DUT

 A semi-rigid cable with 50-ohm characteristic impedance up to 5 GHz

 A high-Q RF component (this example uses Johanson kit P/N: L402DC)

The following major steps are required to tune the antenna:

2 Set up and Calibrate Network Analyzer

3 Tune the Bare PCB Antenna

4 Adjust Tuning with Plastic and Human Body Contact for Antenna

5 Tune the Radio Side by Putting the Chip in Receive Mode

Proper placement of the coaxial cable is crucial, as it can lead to variations in S11 of up to 3 dB To minimize these variations, the ground connection of the coaxial cable shield should be positioned as close as possible to the transmission line return path Below are the essential steps for preparing the ID.

1 Open the plastic casing and remove the batteries or power supplies

To ensure optimal performance, connect the coaxial cable directly to the RF out pin of the chip, and disconnect any existing connections from the chip Failing to do so may cause the balun to load the coaxial cable alongside the antenna, as illustrated in Figure 39.

3 Ensure that there is an exposed ground near to the tip of the coaxial cable Connect the sheath or the shield of the cable to ground

To calibrate the Agilent 8714ES network analyzer, connect a 3.5-mm calibration kit and press the 'cal' button after selecting the 3.5 mm option in the analyzer settings Alternatively, you can utilize other calibration kits, such as a type N calibration kit, for the calibration process.

2 Press the frequency button and set the start and stop frequency to 2 GHz and 3 GHz respectively, and then set the format to Smith chart

3 Press the marker button and set markers to 2.402 GHz, 2.44 GHz and 2.48 GHz

4 Press the cal button, select S11 on the network analyzer and then set it to ‘user 1 port calibration’

5 When prompted to connect the ‘open’ load, connect the “Open fixture” to the VNA and press ‘measure standard’

6 Connect the “Short Fixture” and press measure standard

To begin, connect the “Broadband load” fixture and select the ‘measure standard’ option The network analyzer will then compute the coefficient and represent the 50-Ω load as a point on the Smith chart, precisely labeled as “50, 0.”

8 Connect the tuning coaxial cable and set the electrical delay by pressing the ‘scale’ button and setting the electrical delay correctly

There are two methods to tune the antenna to bring it near 50 ohm

1 Length adjustment of the antenna if it is a PCB trace or a wire antenna, by cutting off the extra length

2 Use of a matching network (recommended practice)

To optimize PCB trace or wire antennas, it's recommended to initially extend the antenna length beyond the Cypress-recommended specifications This allows for easy adjustments by trimming the excess length to achieve resonance around 2.4 GHz This straightforward method does not necessitate any additional components, making it a practical solution for antenna tuning.

The matching network method is the preferred approach due to its flexibility for future EMI/EMC filtering and superior repeatability However, it does necessitate a certain level of expertise For assistance with tuning in high-volume manufacturing, reach out to Cypress Technical Support To tune the bare PCB using the matching network method, follow the outlined procedure.

Appendix B outlines a systematic approach for designing a matching network after measuring impedance This section presents an example of tuning an antenna or radio using matching network components, assuming the reader has a basic understanding of the Smith Chart.

To optimize antenna performance, connect an 8.2-pF or 10-pF capacitor in series with the antenna, which effectively presents a 0 Ω impedance in the desired frequency band This configuration results in an antenna impedance of (100.36 –j34.82), represented as a dot on the Smith chart.

Figure 40 Smith Chart of Antenna Only

2 After determining the antenna impedance, use L-C components to bring it to 50-Ω impedance by performing an impedance transformation

Impedance transformation networks are designed to convert one impedance value to another without power loss, utilizing the impedance transformation characteristics of L and C resonating networks Most matching networks for Cypress MIFA or IFA can typically be achieved using just two components.

The matching network components can be effectively simulated using the open-source tool Smith V3.10 from the Bern Institute By incorporating a shunt capacitor of 0.45 pF in conjunction with a series inductor of 3.6 nH, the impedance is transformed to 50 Ω, significantly canceling out the imaginary component within the desired frequency band In the absence of precise component values, a 0.5-pF shunt capacitor and a 3.6-nH series inductor are selected for this application.

Figure 42 Moving to 50 Ω in Smith chart

The final schematic of the matching network illustrates the relationship between the antenna's impedance, denoted as ZL, when viewed through a 0-ohm resistor Additionally, Zin represents the impedance observed by a network analyzer that operates with a 50-ohm output impedance.

Simulation software provides initial estimates for component values, but actual values can vary significantly due to factors such as lead inductance, parasitic loading of pads, and ground return paths at 2.4 GHz These elements introduce additional parasitics that can alter the Smith chart substantially For optimal resonance in this application, it is recommended to use a 0.7-pF capacitor in conjunction with a series 1.2-pF capacitor, a common practice in 2.4-GHz RF tuning with standard components.

An explanation of this behavior follows:

The antenna impedance, initially observed through an assumed 0-ohm 8.2-pF capacitor, is affected by the parasitic lead inductance at 2.4 GHz Following the antenna, the ground return path introduces additional parasitics, contributing to the overall inductance seen by the antenna To achieve proper tuning, capacitance is added; however, this creates a discrepancy between theory and practice, as the Smith chart indicates a movement akin to adding inductance instead Figure 45 illustrates the final Smith chart with the actual components used.

Figure 45 Smith Chart with Real Components

From Figure 45, it is clear that all the marker point 1, 2, 3 representing 2402 MHz, 2440 MHz, and 2480 MHz are close to the (50,0) point on the Smith chart This shows a good match

The return loss is plotted for the following component values A return loss greater than 15 dB is good enough for this application

Figure 46 Return Loss with Real Components

As seen Figure 46, the return loss is greater than 15 dB for the marker 1, 2 and 3

Microstrip Line

A microstrip line consists of a signal trace positioned on top of a substrate, with a ground plane located underneath the substrate The characteristic impedance of a microstrip line is influenced by several key factors, as illustrated in the cross-section shown in Figure 48.

 Dielectric constant of the substrate (εr)

 Thickness of the RF trace (T)

Figure 48 Cross-Sectional View of Microstrip Line

Microstrips are easy to build, simulate, and manufacture, making them a popular choice in circuit design They exhibit a higher effective dielectric constant than coplanar waveguides for the same substrate, resulting in a more compact layout This efficiency in design contributes to their widespread use in various applications.

CPWG (with Bottom Ground)

A CPWG is similar to a microstrip, but it has copper filling on either side of the RF trace with a gap between them, as shown in Figure 49

Figure 49 Cross-Sectional View of a CPWG with Bottom Ground Plane

The characteristic impedance of a CPWG depends on the following factors:

 Dielectric constant of the substrate (ε r )

 The gap between the trace and the adjacent ground fill (G)

 Thickness of the RF trace (T)

CPWG may be preferred over microstrip for the following reasons:

 It provides a better isolation for RF traces and a better EMI performance

 It makes it easier to support the grounding of shunt elements on an RF trace

 It reduces cross-talk with other traces

 It has a low loss at very high frequencies compared to a microstrip line.

RF Trace Layout Considerations

The following are the guidelines for RF trace design:

 Choose the right kind of transmission line (microstrip or CPWG) when calculating the trace width needed for a 50-ohm characteristic impedance

 Ensure that the RF trace has a 50-ohm characteristic impedance Use impedance calculators to calculate the trace width and gap needed for a given stackup

 The characteristic impedance must be constant throughout the trace Therefore, maintain a constant width for the

RF trace For the CPWG, maintain a constant gap between the RF trace and the adjacent ground

 For the CPWG, ensure that the gap between the grounds in the top layer is less than the height of the substrate; otherwise, the trace will be predominantly microstrip

 For the CPWG, ensure that the ground pour area on either side of the trace is wider than the gap between the grounds

Some of the commonly made mistakes in the design of RF trace and the correct way to do them are represented in Figure 50 and Figure 52

Figure 50 Common Mistakes in RF Trace

 Ensure a clean, uninterrupted ground beneath the RF trace without any other traces crossing the RF trace to allow a proper return path for the RF currents

 Maintain the shortest possible length for the RF trace because the traces and the substrate below attenuate the

RF signal proportionate to the length

To ensure optimal performance in RF trace design, it is crucial to avoid sharp bends; instead, opt for smooth, curved bends to maintain a consistent width If right-angled turns are necessary, consider mitering them to enhance signal integrity, as illustrated in Figure 51.

Figure 51 Mitering of a Right-Angled Turn This ensures that the impedance is continuous across the bends

M is the width of mitering

W is the width of the RF trace

H is the height of the substrate

To ensure optimal impedance matching in RF designs, it is crucial to avoid stubs or branching in the RF trace, as these can introduce reactive impedance When following reference designs, components must be placed precisely as indicated, since any deviation, such as branching off the RF trace to add a shunt component, can disrupt impedance matching Consequently, using the same component values as in the reference design may not yield the desired results in a modified design.

 Do not place any other traces close to and parallel to the RF trace This causes mutual coupling of the signals between traces

 Do not place test points on the RF trace They act as stubs and affect impedance matching

Figure 52 Stubs, Test Points, and Parallel Traces

Four-Layer PCB

Cypress highly advises utilizing four-layer boards for all RF designs due to their advantages in providing a complete ground and power plane, as well as facilitating simpler signal routing For optimal performance, it is recommended to follow a specific stackup configuration for these four-layer PCBs.

Top layer RF IC and components, RF trace, antenna, decoupling capacitors, and other signals

Bottom layer Non-RF components and signals

A complete power plane layer provides low resistance and distributed decoupling capacitance alongside the ground plane The RF trace width required for a 50-ohm characteristic impedance is influenced by the substrate thickness between the RF trace and the underlying ground plane, which can vary among PCBs of the same board thickness due to differences in metal layer spacing from various manufacturers It is advisable to consult with the PCB vendor to obtain the stackup before proceeding with the design If switching to a new PCB vendor that offers a different stackup, it is essential to recalculate the RF trace width necessary for achieving a 50-ohm impedance and adjust the layout accordingly.

Two-Layer PCB

Two-layer boards are ideal for simpler, cost-sensitive applications, and should be kept as thin as possible The width of RF traces, which is essential for maintaining characteristic impedance, is directly related to the substrate height Consequently, thicker PCBs (over 0.8 mm) lead to wider RF traces, complicating signal routing.

RF traces also trigger spurious parasitic wave modes

For routing the power supply, use thick traces on the top layer only

Use the following plan for two-layer boards:

Top layer RF IC, all components, RF trace, antenna, decoupling capacitors, power, and other signals

Bottom layer Solid ground plane

If it is not possible to have a complete ground plane at the bottom, try to ensure a complete ground plane below the entire radio section

In RF PCB design, the ground layer plays a crucial role as it serves as the return path for RF signals beneath the RF trace To ensure optimal RF performance, this return path must remain uninterrupted and as wide as possible Interruptions in the ground plane can lead return currents to take alternative, smaller paths, creating current loops that introduce unwanted inductance This adversely affects the impedance matching between the radio and antenna, resulting in significant attenuation of the RF signal Additionally, a narrow ground plane beneath the RF trace may not function effectively as a microstrip, leading to increased signal leakage.

Ground Plane Considerations

 Do not have traces running across the RF trace in the ground plane It is better to keep a layer completely dedicated for ground, even for two-layer PCBs

To optimize the design, fill the unused regions in the top and bottom layers with ground material, ensuring a connection to the ground plane through multiple vias These vias should be spaced no more than one-twentieth of the wavelength corresponding to the operating frequency.

Using two-layer boards for CSP packages is not advisable due to the challenge of routing signals through the second layer, which complicates the design of a continuous ground plane necessary for RF signals.

 Do not have split grounds unless you can ensure that no current loops are formed in the ground for the current in the return path

 Allow a wide ground plane beneath the RF trace Narrow ground planes permit parasitic modes of transmission and increase leakage

The combination of the bottom and top ground planes, along with the vias connecting them, effectively shields all traces, leading to a notable enhancement in both EMI and EMC performance.

To mitigate unwanted electromagnetic interference (EMI) from the power plane, it is advisable to cover its corners with via holes that connect to ground planes on both sides This technique effectively reduces EMI emissions through the edges of the board.

Decoupling capacitors are essential for power supply systems as they filter out noise from integrated circuits (ICs), preventing interference with other devices In radio applications, power supply noise can elevate phase noise in frequency synthesizers, leading to degraded signal quality This noise can also induce instabilities in RF output, causing unwanted interference and spurious emissions that may surpass regulatory limits Additionally, in receivers, increased noise results in higher packet errors and diminished sensitivity.

To effectively filter out noise at various frequencies, it is often necessary to use multiple capacitors in parallel Capacitors exhibit the lowest impedance at their self-resonant frequency (SRF), making them most efficient around this frequency For optimal noise isolation, it is crucial to identify all noise-frequency components and consult the capacitor datasheet to select capacitors with an SRF that closely aligns with these frequencies.

To effectively support the sudden in-rush current demands of an integrated circuit (IC) during RF transmission or reception, it is advisable to use a large capacitor The required capacitance value is determined by the in-rush current (I), the allowable voltage drop (dV), and the duration (dt) of the current surge The capacitance (C) can be calculated using the appropriate formula, ensuring optimal performance and stability in the circuit.

For example, to support an in-rush current of 20 mA for 15 às for a maximum 300-mV drop from 3.3 V, the capacitance needed is 1 àF

For optimal performance of PSoC 4 BLE and PRoC BLE, it is essential to use a 0.1-µF capacitor on all power supply pins, along with a 1-µF bulk capacitor for each net, specifically one for VDDD, one for VDDA, and one for VDDR Additionally, a 10-pF decoupling capacitor should be placed on pin 15 for the QFN package and pin J6 for the CSP package to effectively filter PLL noise from the power supply.

For optimal performance of the PSoC 6 MCU with Bluetooth Low Energy (BLE) connectivity, it is advisable to utilize specific capacitor values: a 3.3 µF capacitor for VRF, a 2.2 µF for VDCDC, a 4.7 µF for VBUCK1, and a combination of 0.1 µF and 10 µF decoupling capacitors for VDD_NS Additionally, all other power supply pins should be equipped with 1 µF and 0.1 µF capacitors To ensure effective decoupling, low ESR capacitors are recommended.

Power Supply Decoupling

Note the following best practices when laying out the power supply traces:

 Place the components as close to the supply pin as possible

 Place the smallest-value capacitor closest to the power supply pin

To optimize circuit performance, position the decoupling capacitor on the same layer as the integrated circuit (IC) If placing all capacitors on the same layer is not feasible, prioritize the placement of capacitors with smaller values.

 The power supply should flow through the decoupling capacitors to the power supply pin of the IC Avoid using supply vias between the component and the pin

 Use separate vias to ground for each decoupling capacitor Do not share vias

 For four-layer boards with a separate power plane, use separate vias for each power supply pin to the power plane

It is recommended not to share the vias

 Some of the commonly made layout issues related to power supply decoupling are shown in Figure 53

Figure 53 Power Supply Decoupling Mistakes

Vias play a crucial role in ensuring signal connectivity in multilayer boards, but if not utilized correctly, they can introduce significant parasitic effects, particularly at RF frequencies Sharing a via between different circuit sections can elevate common-mode noise, while a ground via positioned far from a shunt component alters the impedance perceived at the trace, leading to impedance mismatches Additionally, at high frequencies, the parasitic inductance associated with vias can lead to substantial impedance, affecting overall circuit performance.

The following guidelines help ensure a proper RF layout:

 Use plenty of vias spaced not more than one-twentieth of the wavelength of the RF signals between ground fillings at the top layer and inner ground layer

 Place ground vias immediately next to pins/pads in the top layer Place more vias whenever feasible More vias in parallel reduce the parasitic inductance

This section discusses the non-ideal behavior of capacitors and inductors at high frequencies, providing guidance for selecting the appropriate components for various applications, including matching networks, DC blocks, crystals, and power supply decoupling.

Capacitors

All capacitors contain parasitic resistance, parasitic capacitance, and parasitic inductance apart from the intended capacitance Figure 54 shows the theoretical model of a typical capacitor

C is the capacitance for which the capacitor is designed The reactance (Xc) because of the capacitance (C) and the reactance (XCp) because of the parasitic capacitance (Cp) are

The reactance of a capacitance diminishes as frequency increases, while the parasitic capacitance (Cp) remains low, resulting in high reactance at lower frequencies Consequently, since Cp is in parallel with the primary capacitance, it exerts minimal influence at low frequencies.

A change in current flowing through a capacitor leads to a variation in the surrounding magnetic field, which is partially induced by the conductors This process generates an electromotive force (EMF) that opposes the current change, resulting in parasitic inductance The reactance associated with this parasitic inductance plays a significant role in circuit behavior.

The reactance of the parasitic inductance increases with frequency Usually L is a very small value in capacitors; so, at low frequencies, XL is negligible

R is the effective series resistance of the capacitor It is usually a very low value

The effective impedance of the capacitor is

At low frequencies, XCp is very high and the effective impedance is

At low frequencies, the circuit is predominantly capacitive Xeff is almost same as XC But as the frequency increases,

As the frequency increases, the capacitive reactance (XC) continues to decrease while the inductive reactance (XL) increases At a certain frequency, XL equals XC, resulting in the capacitor's impedance matching the resistance (R) This specific frequency is known as the series resonant frequency (SRF) of the capacitor.

When selecting capacitors for impedance matching, it is crucial to ensure that the self-resonant frequency (SRF) is significantly higher than the operating frequency This approach guarantees that the capacitor's reactance is primarily determined by its specified capacitance value, minimizing the impact of parasitic inductance on effective reactance.

When selecting capacitors for decoupling, it is advisable to choose values with a self-resonant frequency (SRF) that closely matches the noise frequency you intend to decouple This approach guarantees that the noise encounters a low-impedance path to ground, enhancing the effectiveness of the decoupling process.

At a higher frequency, the reactance XCp becomes equal to the reactance of the other arm (which is mostly equal to

XL now) At this frequency, the capacitors behave like an open circuit This frequency is the parallel resonant frequency Avoid using capacitors at their parallel resonant frequency

The quality factor (Q) of a capacitor (C) is the ratio of the reactance of the capacitor to its resistance (R) at a given frequency (f)

High-Q capacitors exhibit minimal undesired resistance, making them essential for RF circuits operating at specific frequencies Utilizing capacitors with a high Q value is crucial, as it prevents the loss of RF energy, which can otherwise dissipate as heat due to the capacitor's resistance.

 Use only C0G/NP0 capacitors for components of the matching network This ensures that the matching network does not change across temperatures

When selecting capacitors for crystal loads, it is essential to use only C0G/NP0 capacitors to maintain consistent clock timing and RF frequency across varying temperatures For further information on crystal selection and tuning techniques, refer to AN95089 – PSoC 4/PRoC BLE Crystal Oscillator Selection and Tuning Techniques.

 For the matching network, choose capacitors that operate well below their SRF

 Use only high-Q capacitors for the RF circuit

When selecting decoupling capacitors, C0G capacitors may not be necessary due to their high accuracy Instead, X5R or X7R capacitors are commonly used, depending on the required temperature range It's essential to choose low ESR capacitors to ensure effective decoupling performance.

 For decoupling capacitors, choose the component values that have an SRF at the noise frequencies

 It is recommended to use smaller components (0402 or 0201), as they have less parasitic reactance

When incorporating a DC block into an already matched RF trace, it is advisable to select a capacitor with a self-resonant frequency (SRF) near the operational frequency and a low equivalent series resistance (ESR) This is because the capacitor's effective reactance approaches zero at its SRF, ensuring optimal performance in the circuit.

So it does not alter the impedance matching

Định dạng
Số trang	60
Dung lượng	2,74 MB

Tiêu đề	Antenna Design and RF Layout Guidelines
Tác giả	Tapan Pattnayak, Guhapriyan Thanikachalam
Trường học	Cypress
Chuyên ngành	Antenna Design and RF Layout
Thể loại	application note