Hello, and welcome to the RF Basics and Getting Started presentation. This presentation serves as an overview of the basics of RF system design and crucial parameters to consider, digital communication theory, antenna matching, PCB layout best practices. Low power wireless devices are becoming common in various areas. To name a few, these include consumer networking, industrial remote monitoring, shipment and asset tracking, and many more. The following are some of the applications in the mentioned areas.

TI's portfolio is one of the industry's broadest. We have devices to support applications from very low frequency, such as RFIDs, to applications going up to microwave range, for example Wi-Fi and GPS. Some of the example applications are depicted here with corresponding devices from TI's portfolio.

RF basics-- transmitters and receivers. A typical RF communication system would contain an RF-IC, a crystal, balun, matching filter, an antenna, and a PCB on which all of these components are assembled. The RF-IC can be a transmitter, a receiver, a transceiver, or a system-on-a-chip, which essentially is a transceiver with an integrated microcontroller unit.

A typical wireless communication system consists of a transmitter and a receiver. The block diagram above of a transmitter converts the baseband signal to a digital bit stream, compresses the bitstream such that redundancy is minimized in the data into leaves and codes the data to facilitate robust signal recovery and error correction at the receiver. It then up-converts the baseband signal to passband with a modulator and transmits it over the air with a power amplifier and an antenna.

At the receiver, an inverse process is implemented to reliably extract the received information. The equalizer block shown removes the effect of the transmission medium on the received signal.

Below are the key equations governing an RF system design. Path loss equations relates received and transmitted power of signals to the frequency and distance between the transmitter and the receiver. GR and GT are the gains of the transmitter and the receiver.

Eb over No is the normalized version of signal-to-noise ratio in digital communication system. It relates energy in the bit of the signal to the bandwidth B and data rate R of the communication system. Shannon's channel capacity relation tells us how much information we can pack in a given bandwidth for reliable communication in a noisy environment.

Receiver architectures-- there are three basic receiver architectures-- the dual IF, the direct conversion receiver, and the low IF. In the dual IF receiver, the image frequency is easier to filter off, but the selection of intermediate frequency requires careful planning. It also requires high-Q, off-chip image reject filters.

The block diagram shows various stages the received signal has to travel before arriving at the A-to-D converter for baseband processing. The fundamental purpose of these stages is to increase the signal-to-noise ratio of the received signal, so that the back-end demodulator stages can reliably recover the digital data with acceptable bit at a rate. The spectrum of the signal shown below as it passes through the stages depicts the process mentioned.

The zero-IF architecture-- in this architecture, the received signal is the image itself. Some of the problems associated with direct conversion are the difficulty of DC offset removal, RF-LO coupling and self mixing, and power amplifier pulling.

The low-IF architecture has the advantage of not having the DC offset problem. This architecture is system-dependent, since a proper choice of IF is required. A typical RF transmitter is shown in this block diagram. The baseband processor prepares the bandlimited data and splits it into in-phase and quadrature channels.

After going through first up-conversion by frequency omega 1, the signal are added, filtered for harmonics, and prepared for second up-conversion by frequency omega 2. The last bandpass filter selects the required sideband and transmits it through the power amplifier. For maximum power transfer to the antenna, a matching circuit is also required. It is usually designed to mitigate harmonics as well.

Digital modulation and detection-- bits-to-waveform processing. Analog information is converted to binary stream by A-to-D converter. The binary sequences are gathered into groups-- for example, 000001 and so on. These sequences are equally likely to occur. The above pairs may be encoded as 11 to plus 3, 10 to plus 1, and so on. These analog levels are called alphabets.

The corresponding waveform can be converted to minus 3, minus 3, minus 1, plus 1, and so on. And these alphabets can be mapped to, for example, voltage levels. From transmission over an RF medium, these must be turned to analog waveforms. And to use bandwidth efficiently, these alphabets must be used to scale Nyquist pulses.

The four-level PAM sequence above, the analog waveform can be formed by choosing a pulse shape Pt, which is a Nyquist pulse, and can be used to scale the corresponding alphabets. The pulse train which is produced can then be sent over a passband medium using either amplitude shift keying, frequency shift keying, or phase shift key, or combination of either.

Amplitude shift keying-- in this modulation scheme, the amplitude of the carrier is switched between various levels depending on the number of waveform types. In case of two waveforms, where M is equal 2, the signal amplitude states would be either the magnitude of the signal or 0. The analytical expression of the ASK scheme is shown, where E is the signal energy and T is the signal period. The corresponding ASK spectrum and ASK modulated symbols are also shown.

For the frequency shift keying, the frequency of the carrier is switched between the number of levels of the transmitted waveform. For M is equal to 2, the modulated symbol waveform shows frequency changes at symbol transitions. Various forms of FSK are 2-FSK, 4-FSK, 2-GFSK, MSK, and so on.

Phase shift keying-- in this scheme, the phase of the carrier frequency is changed according to signal levels. For M is equal to 2, the phases are set to be 180 degrees apart.

Quadrature amplitude modulation is the combination of both ASK and PSK. It changes both amplitude and phase of the signal simultaneously. Various examples are shown in the diagram.

Demodulation and detection-- a digital receiver consists of a demodulation stage and a detection stage. Demodulation is the process of recovering the transmitted waveform, and detection identifies which waveform was transmitted. Frequency down-conversion portion can be any receiver architecture front-end discussed previously.

The front-end enables the signal to have enough signal-to-noise ratio to be detected reliably. The receiver filter performs away from recovery, and channel equalizer removes the effect of channel on the received signal, thereby reducing intersymbol interference. The baseband pulse zt is sampled at symbol period and presented to the decision-making block for accurate identification. Depending on the signal-to-noise ratio of the received waveform, the decision-making can be erroneous.

Bit error rate is a metric used to evaluate the performance of a digital receiver. The probability of error is determined directly from the normalized signal-to-noise ratio of the received signal. And analytical performance plots under various channel impairments are available for common modulation schemes.

There are two ways in which errors can occur in a binary decision-making. And error occurs if, for example, symbol one was transmitted, and due to channel noise and impairments, symbol 2 was identified, and vice versa. The analytical expressions of the probability distribution functions of such occurrences are shown.

Critical parameters-- transmit and receive. Transmit power is the RF power available at the antenna port of the transmitter. It is important in determining the communication range. A 6-dB increase in power doubles the range, assuming other factors are remain constant.

Transmit bandwidth is the range of frequencies in which a transmitter can radiate power. The bandwidth depends on modulation type and symbol rate. Occupied bandwidth is the bandwidth which confines 99% of the transmit power, although there are various other definitions.

Adjacent channel power is the power of RF signal which spills from the intended communication channel into an adjacent channel. This is measured in dBc relative to the intended channel power. This is a regulatory requirement based on local authorized regulations.

Harmonic suppression-- harmonic frequencies are generated unintentionally along with the desired transmit signal. The harmonics are 2, 3, 4, 5 times the fundamental. These are measured in either absolute value or relative value from the fundamental. Spurious emissions are also regulatory requirement based on local regulations.

Transmit frequency accuracy shows the accuracy of the transmitted frequency. It mainly depends on the crystal used in the circuit. This must be as per both the manufacturer and regulatory body specification.

Transmit spurious emissions are frequencies other than the fundamental and harmonics which are generated unintentionally in the transmitter. These signals are from the VCO circuits, crystals, high speed digital circuits, and other circuits in the transmit and the receive chains. These are classified as conductive emissions and radiative emissions. The measurements must be made in the RF shield room. This is also a regulatory requirement based on the local regulations.

Receiver, critical parameters-- sensitivity is the receiver's ability to detect the lowest RF signal power and produce an SNR at the detection stage for the desired bit error rate. This too influences the wireless communication range. A 6-dB improvement doubles the communication range while other factors remain constant.

The expression here shows the minimum power detected related to minimum signal-to-noise ratio. Noise figure of a system is defined as the change in the signal-to-noise ratio of a signal when it passes through it. Noise figure is a keyed receiver design parameter. And it is often used to design stages in front of a receiver.

Wireless systems can be modeled as linear only for small input signal levels. When inputs become large, nonlinear effects start to creep in and dominate system performance. To study the effect of large signal input to a system, the output can be modeled as an expansion of power series.

An analysis of the RF scenario with one interferer and a desired signal, after passing through the nonlinearity, it reveals the presence of a gain compression term and a desensitization term in the signal. A 1-dB gain drop at the output of a receiver when input is increased is called 1-dB decompression point.

If the interferer was modulated by an AM signal, the expression here shows how the AM off the interferer is transferred onto the desired signal. This is called the cross modulation.

If two interferers are passed through a nonlinear system, the output consists of an intermodulation product resulting from mixing of two components. The resulting IM products are shown in the expression. A problem arises if the desired signal is at the same frequency as that of the third-order IM product.

Receiver-- critical parameters continued. Gain compression and saturation, it is the measure of receiver's ability to detect maximum input signal level to produce the required signal-to-noise ratio. It is also measured in dBm. This level depends on the 1-dB compression point of the front-end of the receiver.

The dynamic range is a measure of receiver's ability to detect a range of input signals from the minimum power level to the maximum power level. The adjacent channel rejection, also called the selectivity, is a measure of receiver's ability to reject an unwanted adjacent channel. It is measured in dBc relative to the desired channel. This depends on the specification of the IF channel filter.

LO leakage and receive spurious emissions are frequencies which are leaked unintentionally from an antenna port of the receiver. These signals are from the local oscillator's crystals high speed digital circuits and other circuits of the receiver. These are classified as conductive emissions and radiative emissions. The measurements are done in an RF shield room for accurate results. This is a regulatory requirement based on the local regulations.

Crystals provide reference frequency for local oscillator. The important characteristics of the filter are price versus performance trade-off, tolerance both in the initial spread and aging over temperature. Crystals provide reference frequency for local oscillator and the carrier frequency.

Using inaccurate crystals will require larger receive bandwidth to compensate for frequency drift. Large receive bandwidths result in more noise in the receiver chain and thus reduce sensitivity. Accurate crystals are expensive. Thus choosing a crystal is a trade-off between cost and performance. The accuracy of a device operating at 868 megahertz with a plus/minus 10 ppm crystal is around plus/minus 8.68 kilohertz.

Wider receive filter bandwidths mean reduced sensitivity and close-in selectivity. Wider receive filter bandwidths allows cheaper, that is less accurate, crystals to be used. For true narrowband systems, [INAUDIBLE] and [INAUDIBLE], the crystal accuracy is set by regulations. The crystal drift can cause the received signal to fall outside the receiver bandwidth and cause packet error rates.

The receiver will have to increase the receive bandwidth for the signal to be received. And this causes increased in noise figure. There are ways employed by some of the TI's devices that can be used to compensate for this drift without causing degradation in sensitivity.

Feedback to PLL-- writing 030x to the register shown enables feedback to the PLL. With this feature, the programmed receive filter bandwidth is increased by bandwidth by 4. However, the noise bandwidth is not increased. And therefore, the sensitivity remains unchanged. This feature, however, allows the use of less tolerant crystals. The test results of the experiment are shown here.

antennas-- antenna a crucial component in the RF signal chain. And its correct design and matching can relax RF system requirements by improving overall performance. The quality of match is specified in terms of the voltage standing wave ratio of the interface.

Resistive and dielectric losses reduce the efficiency of the antenna. Whereas, standing waves on the RF line reduce power coupling to the antenna. And in case of receiver, it increases its noise figure. Radiation is produced by acceleration or deceleration of charges. If the charge is accelerated due to an electromotive force or due to discontinuities, such as termination, bend, curvature, electromagnetic radiations occur.

Some of the critical parameters of antenna are the radiation pattern, which is the way in which the RF energy is distributed or directed into free space by the antenna. Isotropic antenna intermediates electromagnetic energy equally well in all directions. The isotropic antenna is a hypothetical antenna, but the model serves as a conceptual standard against practical antenna.

Antenna gain is referred to the antenna's effective radiated power to the effective radiated power of some reference antenna. If it is referenced to an isotropic model, the gain in specified dBi. If the gain has been compared to a standard dipole, it is referred to as dBd. The gain of a dipole is around 2.2 dB higher than the isotropic antenna. A 6-dB dropp in antenna gain results in half the communication range.

Antenna polarization refers to the directional behavior of the electric vector of the electromagnetic field emanating from the antenna. When an antenna is oriented horizontally with respect to ground, it is said to be horizontally polarized. When the antenna is oriented vertically with respect to ground, it is said to be vertically polarized.

The efficiency of the antenna is the ratio of power radiated to the total power applied to the antenna port. The antenna voltage standing wave ratio is the quality of the antenna match and is measured by the return loss between the antenna and the receiving or the transmitting circuit. Antenna is a function of the environment it operates in. And the matching condition can change accordingly.

The causes of VSWR or to vary sometimes by a large amount and, hence, can disrupt the noise figure at the front end of the receiving system. Commonly used antennas are the PCB antennas, whip antennas, chip antennas, wire and helical antennas.

Communication link example and typical RF environment-- in a case where there's a single transmitter and a single receiver, the range and the quality of the link depends only on the transmit power and the receiver sensitivity. When no other transmitters or interferers from other sources are present, maximum communication range is achievable with best noise figure and high gain of the receiver. In this scenario, only the noise figure and gain are important. And saturation and filtering may not be critical for this case.

In the case where there are other interferers in the vicinity of the receiver, the desired transmitter is located far away from the receiver. The adjacent channel interferer transmitters are located near to the receiver. The receiver should have high dynamic range to receive the weakest desired signal and high level interferer rejection. The receiver should also have high adjacent channel rejection, should have good phase performance at adjacent channels.

A typical RF environment which is close to reality looks something like this, where the communication link is to be established between a receiver and a distance transmitter and there are various other transmit sources in the vicinity of the receiver. In this case, the desired transmitter is located far away from the receiver. The receiver should have high dynamic range to receive the weakest desired signal in the presence of high level of interference.

The receiver should have high adjacent channel rejection. It should have a very good blocking performance. It should have very good phase noise performance and should have very good linearity performance, in which case IP2 and IP3 should be high. It should also have good power detection mechanism to control the automatic gain control loops and generate RSSI accurately.

RF hardware design considerations-- the design review process helps facilitate first pass successes of the design. It helps in achieving expected performance, quicker design to market time, and ensures good experience with TI products. A typical design flaw is shown in the flow chart.

A typical schematic when downloaded from TI website consists of an RF-IC with balun filters, antenna connectors, decoupling capacitors, and crystals. A typical PCB looks something like this. It has digital traces, transceiver plus microcontroller or could be an SoC, crystal oscillators, LNA and power amplifier, vias on the PCB, matching network, PCB antennas, and SMA connector for conducting measurements in the RF lab.

TI radios require matching circuits from output to the load. This could be a discrete balun with various portions of it shown in this figure. When laying out the RF portion of the circuit, it is best to follow the approach shown in Figure A. This would ensure best RF performance. Figure B will introduce impedance change. And figure C will cause imbalance that may effect RF performance.

Here, we see that the RF trace and the ground plane just underneath it. This illustrates that return current flow is directly below the signal trace. This creates the path of least impedance.

If a ground plane is broken under the signal trace, the return current path becomes longer. It adds Inductance, which produces undesired performance. Detour of current will cause magnetic flux linkages, which will manifest as extra inductance. Also, the electric and magnetic fields would extend further from the circuit, causing interference.

LPRF design usually come with two-layer or four-layer PCB boards. Two-layer PCB designs are cheaper and can achieve comparable performance to a four-layer board but requires careful routing and component placement. Four-layer designs provide dedicated RF ground plane and component decoupling of the power plane. In a four-layer board, it is recommended to have the first layer for component and signal, the second layer for ground only, the third layer for power, and the fourth for ground and signal routing.

A common mistake may occur when giving the final PCB design to the board manufacturer house by not defining the layer mapping properly. This can cause impedance changes. And the final performance of the design will suffer. The recommendation is to send a readme file along with the design collateral to the fabrication house.

Ground plane-- a continuous ground plane provides easy connection to the ground by allowing via drops. To reduce board radiations and coupling, fill the unused area on top plane and connect to the ground below with vias. Via fencing also reduces unwanted board edge radiations.

Power supply pins of the devices are bypassed with ground plane using a capacitor, which provides low impedance path for high frequency noise. Transient demand of current by active devices generates high frequency harmonics, undesired coupling between the power supply lines. And therefore, it reduces the performance.

Inductors on the PCB have associated with them a magnetic field that can couple with other components and can affect the performance in undesired manner. Do not place parallel inductors close to each other unless it is required to utilize the field coupling. If required, place inductors perpendicular to each other to minimize the coupling effects. TI reference designs are designed carefully considered such coupling effects. And thus, it is recommended to follow the component placements and orientation on TI reference designs.

It is recommended that the end user copies the reference design as close as possible, understand the PCB board properties, the PCB stack-up, the PCB dielectric properties. TI reference designs include a document that specifies the manufacturing and the fabrication process.

Some of the best practices for PCB layout are listed here. Verify that the stack matches the reference design. Changing the layer and spacing stack-up will affect the matching in the RF signal path. There are various application notes on TI website that can be referenced for proper PCB layout, especially in the RF signal chain.

0603 mills discrete parts are not recommended because of the size and the parasitic values. Ensure that the bypassing decoupling capacitors are as close as possible to the power supply pins and that they are meant to bypass those pins only. Ensure that decoupling is done pin capacitor and power via.

Verify that the RF signal path, as well as the component arrangement, matches the reference design as quickly as possible. The crystal oscillator should be as close as possible to the oscillator pins of the IC. And long traces should be avoided if possible.

Verify that the under-the-device power pad layout is correct. The solder pads in the mask should match. And the opening size should ensure correct amount of paste. Vias should be the correct number and mask tented to ensure that they don't suck up all the solder.

Important considerations for antennas-- if using an antenna from a TI reference design, be sure to copy the design exactly as drawn and check if the stack-up in the reference design matches your stack-up. Changes to feed line length of the antenna will change the input impedance match. PCB and chip antennas should not have any ground plane under them.

Any metal in close proximity, plastic enclosure, and human body will change the antenna's input impedance and resonant frequency, which must be considered in the design and matching of the antenna. For multiple antennas on the same board, use antenna polarization and directivity to isolate. For chip antennas, verify that the spacing from the orientation with respect to ground plane is correct as specified in their datasheet. If the design uses a battery, such as a coin cell, the battery will act as a ground plane and should be, therefore, not placed under the antenna.

Antenna matching procedure-- in antenna matching, only inductors or capacitors should be used. Matching circuits, distributed or lumped, have losses due to limited quality factor. If an antenna has a reasonable initial resonance, the improvement obtained from the matching circuit will compensate for the losses it introduces. The block diagram here shows a typical matching circuit prototype.

The analytical expressions of the antenna match are shown here for two scenarios-- one where RL is less than RS and the other where RL is greater than RS. Depending on the measurements in the lab, either one can be used.

Antenna matching can also be done using graphical methods, such as Smith chart. And sometimes, it is more convenient than using analytical expressions. The L-network prototype shown here depict that narrowband match using the L-type prototype can always be achieved.

To match a compact PCB helical antenna, we take an example at 868 megahertz. This antenna should be large in size, but it has been purposely designed to be as compact as possible to fit onto a small PCB. Since this antenna requires a load in order to be resonant at 68 megahertz, then the effects of correct antenna match will be more crucial. Step one is to properly calibrate the vector network analyzer and extend the ports to the measurement plane. This is necessary to make accurate impedance measurements of the antenna.

Over here, we have measured the S11 of the unmatched antenna to be 15 minus j70 at 868 megahertz. From the S11 plot, we see that it has a nice resonance at 1.5 gigahertz. But it is way off at the desired 868 megahertz.

Using one of the analytical equations from the previous slides, the theoretical values and the result of the L-Match are shown here. The values turn out to be 17.1 nanohenry and 5.45 picofarad. The simulation results are shown to be perfectly matched antenna.

However, these values of 17.1 nanohenry and 5.45 picofarad do not exist in reality. So realistic models are chosen and simulated for desired performance. And once these values are confirmed, they are put on the boards and tweaked again.

Most likely, the measurements on the PCB are still off and are tweaked on the board to give final acceptable antenna performance. In this case, the final values that were chosen were 18 nanohenry and 4.7 picofarad. Fred The comparison of the theoretical simulated results and the measured results are shown. The VSWR of the design is well within the acceptable limits of the antenna match at 1.35.

The antenna performance also depends on the final product in which it's going to be incorporated. If the product is enclosed in a metal casing or plastics and it's operating close to a human body, the resonances and the impedance of the antenna changes. And therefore, the matching circuit becomes invalid. So it is important that the final matching of the antenna is done in the environment it is going to operate in.