Fullscreen HTML5Med Res (??MB)HTML4

Teach-In 2026 by Mike Tooley World of Wireless – An Introduction to Radio and Wireless Technology Part 12.7: Spectrum and vector network analysis I n the last part, we explained the principles and techniques that underpin a range of useful RF measurements, including voltage, power and frequency. Our HandsOn project featured an RF probe that can be used in conjunction with a conventional multimeter to make voltage measurements over the range of 100kHz to 100MHz. This month, we focus on techniques for analysing RF networks and devices, including S-­parameters, vector network analysis and spectrum analysis. S-parameters provide a common framework for RF circuit design, allowing engineers to measure and optimise RF devices and systems. They streamline measurement processes, enhance modelling accuracy and help engineers develop designs that are stable and reliable. Whether it’s adjusting antennas, creating filters or testing amplifiers, S-parameters are a fundamental RF engineering tool. Vector network analysis (VNA) is an indispensable tool in RF engineering, enabling accurate measurement of network and device characteristics over a wide frequency range. It ensures optimal performance in the design and manufacture of antennas, filters, amplifiers, couplers and isolators. Plane sheet Planeofsheet glass of glass Spectrum analysis (SA) is important when it’s necessary to investigate the components present in an RF signal, enabling an assessment of spectral purity and spurious content. Measuring RF networks & devices Using standard testing tools, measuring the effect of an RF signal on signals passing through a device can be quite challenging, often involving multiple checks of voltage and phase angle across different frequencies. The device you’re examining might be as simple as a resistive attenuator or as complex as a broadband amplifier. However, by considering the device as a ‘black box’—focusing solely on input and output rather than its internal workings—you can simplify the task. This approach makes RF circuit design easier for engineers. A simple network, such as an antenna or resistive load, may have only one port. With an antenna, RF power is received as an input and released as radiated energy. A load also receives RF power but transforms it into heat. Amplifiers, attenuators and filters are all examples of devices with two ports. In such cases, one port (usually Port 1) is designated the input and connected to a signal source. The output port (usually Port 2) is designated the output and connected to a load (or subsequent device). Since what happens between Port 1 and Port 2 can be very important, we need a way to describe and categorise it. There are several ways of doing this, but the S-parameters that we shall meet a little later provide us with arguably the most powerful and flexible method. Some basic two-port networks based on passive devices are symmetrical, which means that either port can be used as the input or output without changing how the network functions and what it does. Common examples of this are matched resistive attenuators and passive filters. In contrast, other two-port networks are asymmetrical and require specific connections for their input and output; swapping these ports will impact their operation. Amplifiers and multipliers are examples of asymmetric two-port networks. Devices like combiners, circulators and certain filters need a third port. As the number of ports in a network increases, more parameters are necessary to fully describe its characteristics. S-parameters S-parameters provide an efficient means of characterising the behaviour of both single- and multi-port Inciden Inciden t t Transmitted Transmitted networks. The required number of IncidentIncident Transm Transm itted itted Network or Network or S-parameters is determined by the device device d cted e total number of ports present: a t c Refle Refle Reflected Reflected single-port network can be described using just two parameters, a two-port network requires four and a three(a) Light (a)analogy Light analogy (b) RF (b) energy RF energy port network is represented by eight parameters and so on. Fig.7.1: like light, RF energy can be reflected. 14 Practical Electronics | May | 2026 S11 S11 S21 S21 DUT DUT DUT DUT PortPort 2 2 Reflected Reflected a1 a11 b b1 PortPort 1 1 S11 S11 Port Port 1 1 Incident Incident Port Port 2 2 S-parameters are complex vector Think of it like a boat crossing quantities that have both real and a lake while a steady crosswind is imaginary components. Since com- blowing. The boat’s motion relative plex numbers are similar to vectors, to the water and the wind’s effect they can instead be considered as a are two vectors acting at an angle to magnitude and an angle or phase. each other. The actual path of the S-parameters are named because boat is determined by the vector sum of the ‘scattering’ effect of energy of these effects. entering and leaving a network or Even though you may only see device under test (DUT). Fig.7.1 il- the boat moving across the water, lustrates this effect and shows how its final trajectory depends on both it’s analogous to light (which is also components. You cannot predict a form of electromagnetic energy) where it will end up by considering passing through a planar sheet of only one. glass. In electronics, AC currents and Reflection and scattering of in- voltages behave the same way. Their cident light occurs on one side of magnitudes alone are not sufficient; the glass, while a small proportion phase angle matters. The total curof the transmitted light is reflected rent at a node is found by vector light from objects on the other side addition, not simple arithmetic adof the glass, passing a second time dition, because each current has both through it and emerging back to- magnitude and direction (phase). wards the observer, as depicted in Fig.7.1(a). Incident and reflected waves Fig.7.1(b) shows a two-port RF S-parameters are best understood network or DUT represented as a when conceptualised as dynamic black box. Some energy is transmit- wave quantities rather than static ted and appears at the output while, values. Waves enter and exit a netdue to an imperfect match at the work through its ports, necessitating input, a proportion of the energy is a method to distinguish between reflected towards the source. The them. effect of reflected RF power was A vector network analyser (VNA) discussed earlier, in Part 5, which fulfils this role by separating incidealt with antennas, matching and dent waves from reflected waves, voltage standing wave ratio (VSWR). thereby enabling efficient and acS-parameters are ratios and there- curate characterisation of a network fore have no units. It’s important or device. to remember that they representS In Fig.7.2, waves entering and exit21 Transmitted complex quantities, meaning theyS ing the two-port network are labelled Incident Transmitted haveIncident both magnitude and phase. If 21 “a” and “b”, with subscripts 1 and b2 the input (Port 1) and you’re unfamiliar with athis, it can 2 indicating 1 seem confusing at first. a1 output (Port b2 2), respectively. At Port Transmitted Transmitted b2 b a22 a2 Port Port 1 1 Port Port 2 2 PortPort 1 1 Transmitted Transmitted a1 a11 b b1 DUT S12 DUT PortPort 2 2 (a) Signal applied to Port 1 with matched load at Port 2 Reflected (a) Signal applied 2 a1 to Port 1 with matched b2 load at Port Reflected a1 b2 DUT DUT b2 b a22 a2 S22 S22 S22 S22 Reflected Reflected Incident Incident S12 b1 a2 b1 to Port 2 with matched a2 load at Port (b) Signal applied 1 Incident (b) Signal applied to Port 2 with matched load at Port 1 waves. S 12incident, transmitted and reflected Incident Fig.7.2: a generic two-port network showing S12 Practical Electronics | May | 2026 Input Reflection Coefficient, S11, where (with a2 = 0): S 11= (b) Signal applied to Port 2 with matched load at Port 1 (b) Signal applied to Port 2 with matched load at Port 1 Reflected at Port 1 b1 = Forward at Port 1 a1 S11 tells us how much of the Transmittedatfrom 1 toThis Port 2 b2 signalS 21is=reflected the Port input. = Incident at Port 1 parameter is useful when evaluat- a1 ing impedance matching. With an Reflected Port 2 b1b2is zero, idealSmatch at theatinput = 22 = Incident at Port 2 a2 therefore S11 is also zero. Reflected at Port 1 b1 S 11= = Port(gain Forward Transmission or1 b1 Transmitted 2 to Port Forward at Port 1 from a1 S = = 12 loss), S21, where (withata2Port = 0): a2 Incident 2 S 21= Transmitted from Port 1 to Port 2 b2 = Incident at Port 1 a1 S21 tells us how much gain or loss Reflected at Port 2 b2 is This parameter is useful S 22present. = = Incident at Port 2 a2 when determining voltage gain (in the case of an amplifier) Reflected at Port 1 orb1loss (in S 11= = Transmitted from 1 b1 the an attenuator at Port Port 21 toorPort a1passive S 12 =case ofForward = a2 Incident at Port 2 filter). Transmitted from Port 1 to Port 2 b2 = S 21= Output Reflection Coefficient, Incident at Port 1 S22, a1 where (with a1 = 0): Reflected at Port 2 b2 = Incident at Port 2 a2 b of the sigReflectedus at how Port 1much S 11S22 = tells = 1 Port Transmitted from 2 toThis Port 1 b1 at Port 1 a1output. nal isForward reflected at the S 12 = = Incident at Port 2 parameter is useful when evaluat- a2 b ing impedance matching. With Transmitted from Port 1 to Port 2 an = 2 S 21= ideal matchIncident at the at output Port 1 b2 is zero, a1 S 22= therefore S22 is also zero. Reflected at Port 2 b2 S 22= = Reverse Transmission, Incident at Port 2 a2 S12, where (with a1 = 0): S 12 = b1 to Port 1 with matched a2 load at Port 2 (a) Signal applied Reflected (a) Signal applied b1 to Port 1 with matched a2 load at Port 2 Reflected Transmitted Transmitted 1, the incoming wave is labelled a1 and the reflected wave is labelled b1. At Port 2, the transmitted (outgoing) wave is a2 and the reflected wave from the load is b2. The ratios of these waves provide us with four S-parameters. Transmitted from Port 2 to Port 1 b1 = Incident at Port 2 a2 S12 tells us how much signal from the output appears back at the input. This parameter is useful when analysing isolators, circulators and directional couplers. Vector network analysers The measurement of S-parameters relies on matched impedances being present at all ports. Therefore, it is essential that both signal sources and loads are matched to a uniform impedance, typically 50Ω. It is also necessary to ensure that the VNA is calibrated before use. This generally involves the use of known short-circuit and opencircuit loads. Low-cost VNAs (some available for less than £100), like the one 15 Fig.7.5: fixed external attenuators are often necessary to avoid overdriving a device and to minimise the risk of overloading the VNA’s input. Fig.7.3: this inexpensive VNA can measure S11 and S21 from 50kHz to 1GHz. Fig.7.4: a NanoVNA plot of gain (S21) for a low-noise VHF preamplifier. shown in Fig.7.3, measure just S11 and S21. Despite that obvious limitation, these two parameters can still be extremely useful, making such a VNA a useful and affordable tool for those on a limited budget. Laboratory-grade instruments can carry out measurements on multiple ports, but can be prohibitively expensive. For most amateur applications, a frequency range extending to 1GHz will be perfectly adequate, but it’s worth checking that an instrument can take many samples. A USB connection to a PC can be invaluable for analysing and displaying data using an external software package, such as the immensely popular NanoVNA Saver. The S-parameter data captured by a VNA can be processed and displayed in various ways, including a list of parameters showing real and imaginary quantities and charts with data plotted against frequency. Table 7.1 shows the first Table 7.1: partial S-parameter data for a dipole ant. few (of several hundred) # HZ S RI R 50 lines of data produced by a low-cost VNA when meas135000000 0.179377839 -0.305732042 uring the S-­ p arameters 135019822 0.176021755 -0.306564718 of the dipole antenna featured in Teach-In Part 135039644 0.172860488 -0.307987154 5. The data for this single 135059466 0.169660151 -0.309449017 port device is stored in an 16 s1p ‘snap’ or ‘Touchstone’ file. The data for two-port devices require correspondingly larger s2p files. The first line of data in Table 7.1 reveals the following: • The first column shows the test frequencies specified in Hertz (Hz) • The next two columns are real (R) and imaginary (I) quantities (the impedance and phase angle can easily be determined from these values) • The data has been normalised to an impedance of 50Ω (more of this a little later) Taking the first line of data in Table 7.1, we can determine the impedance of our dipole antenna at 135MHz by simply multiplying the real and imaginary quantities by the normalising impedance (50Ω). Thus, at 135MHz, the impedance of our antenna is approximately (8.95 – j15.3)Ω. The negative sign indicates that the antenna is capacitive at 135MHz, which is expected, as this frequency is below the antenna’s fundamental resonance. Fig.7.4 shows a typical VNA application in which the forward gain characteristic (S21) of a low-noise VHF preamplifier is plotted from S-parameter data obtained using an inexpensive NanoVNA. To prevent the signal from the NanoVNA overloading the preamplifier, a 20dB attenuator has been inserted in the preamplifier input. This means that 20dB needs to be added to the gain visible in Fig.7.4. The summary data for S21 thus shows that the preamplifier is delivering a gain of approximately 20dB – 2dB = 18dB. Fig.7.5 shows some SMA attenuators that provide an attenuation of 30dB at up to 25W (left) and 10dB at up to 2W (right). Although eminently suited for portable and handheld use, smaller Practical Electronics | May | 2026 VNAs often have displays of limited size. To overcome this problem and provide more advanced data processing and display, VNA data can usually be exported in s1p or s2p format to a computer using an SD memory card or USB link. Fig.7.6 shows RSplot being used to display s2p data obtained from a low-cost 6MHz filter. S11 (plotted in dB) is shown in red while S21 (also plotted in dB) is shown in blue. Markers at 6.4MHz have been superimposed on the characteristic. Notice how S11 is relatively flat, while S21 clearly shows how the attenuation increases significantly at around 6.45MHz. Using this data, it’s easily possible to determine filter bandwidth and shape factor. AppCAD is another excellent (and free) program for manipulating and displaying data from S-parameters. Fig.7.7 shows how this extensive package can be used to test and confirm the response of an HF linear amplifier. Notice how S21 has been selected in the “Parameters to Display” window and an offset of 30dB has been chosen in the “Plot Options”. Adjustments for the external attenuator (Fig.7.5) can then be made automatically. It’s evident from the frequency response that the quoted specifications have been exceeded, with the gain remaining reasonably constant over the HF range from 5MHz to 30MHz. Fig.7.6: using RSplot to show the frequency response of a 6MHz filter. The Smith chart Practical Electronics | May | 2026 Positive reactance +1.0 Fig.7.7: the powerful AppCAD package measuring the gain and frequency response of an HF linear amplifier. It has a quoted gain of 20dB from 5MHz to 30MHz. negative values of reactance (capacitive reactance) correspond to the lower semicircumference. Imaginary (reactive) components are preceded by the letter j (instead of the i used by mathematicians), the sign of which can be either positive or negative according to whether the reactance is inductive or capacitive. 5 +0 .2 0 Fig.7.8: the basic arrangement of a Smith chart. +2 .5 +0 0 Resistance 0.25 0.5 1.0 2 5 ∞ 25 -0. To get the best out of a VNA, it’s important to be familiar with a method of graphical representation known as a Smith chart. We will start by explaining how the Smith chart works before moving on to showing how a VNA can be used to perform some basic tasks. A typical application is in the design of antennas and the arrangements for feeding and matching them to a transmitter. Fig.7.8 shows the basic arrangement of a Smith chart. Although this may look somewhat daunting, the basic concept is simply that a complex impedance (an impedance comprising both real/resistive and imaginary/reactive components) can be mapped into a circular space in which the horizontal axis represents resistance (the real part) and the outer circumference represents reactance (the imaginary part). Positive values of reactance (inductive reactance) appear along the upper semi-circumference, while -0. -2 5 -1.0 Negative reactance 17 4 1.2 0.9 1.4 0.8 8 1.6 0.7 0.1 0.3 2 1.8 50 2.0 0.2 0 14 4 0 0.0 15 30 or 0.03 erat 1.0 Gen 0.47 0.2 0.48 50 20 10 5.0 4.0 3.0 2.0 1.8 1.6 1.4 1.0 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0.9 0.49 ±180 1.2 8.0 -170 0.02 7.0 1.0 0.03 0 -15 0.8 0.2 9 6 0.4 -30 40 -1 6 0 -1 3 0.6 0.3 3 6 0.1 4 0.3 1.2 -70 -80 0.15 0.35 1.0 7 7 0.0 0.14 0.36 0.37 -110 -100 -90 0.13 0.12 0.38 -1 0.8 1.4 -60 0.1 0.9 2 0.10 0.11 3 20 0.7 1.6 8 0.0 0.5 2.0 0.2 1.8 0 0.1 0.3 0.4 -5 1 4 3.0 0.4 0 9 0.1 0.4 0.0 8 0.4 2 0.0 9 1 0.4 0.40 0.39 any given value of characteristic impedance, the values plotted on a Smith chart must first be normalised by simply dividing the real/ resistive and imaginary/reactive parts by the value of characteristic impedance. So, for example, an impedance of 0.3 5 0.0 0.6 0 -j0.5 ati Fig.7.10: ve rea plotting cta normalised nce impedances on a (-j) Smith chart. 0.3 -4 5 0.4 0.2 1 4 0.0 1.0 4.0 0.47 0.28 0.22 5.0 -160 0.8 6.0 0.49 0.01 9.0 10 0.48 0.27 -20 0.2 0 0.4 g 0.23 Degre es 0.3 Ne 0.26 In 0.6 1 - j0.5 0.2 0.25 -10 0.4 0.1 ∞ 0.24 20 0.25 0.2 Coefficient 50 0.01 0.1 0.2 50 0 0.00 20 0.24 0.26 0.4 170 0.6 10 Reflection 0.00 0.27 Of 8.0 9.0 0.23 7.0 Angle 0.02 rd 0.8 160 6.0 10 1 0.22 0.28 5.0 1 + j 0.5 20 Towa 4.0 1.0 Resistance 0 9 0.2 0.4 6 1 0.2 0.6 0.8 Wavelengths 0 0.3 40 0.4 3.0 0.4 0.1 7 0.3 3 60 0 0.2 5 6 0.3 4 1 0.4 5 0.1 80 70 +j0.5 0 13 0.15 0.35 90 100 110 0.14 0.36 0.3 0.0 0.13 0.37 9 sit Po 0 12 0.12 0.38 0.1 6 3 0.4 0.5 0.0 0.0 eac 2 0.4 r ive 7 1 0.4 0.39 1.0 ce tan 0.40 0.6 8 0.0 Load 18 ) (+j 0.11 0.10 9 0.0 rd We briefly mentioned the need for normalising impedances in our VNA discussion. To cater for L C C L L C C L L C C L L C C L (e) Pure capacitance, C (f) Series L and C (d) Pure inductance, L (e) Pure capacitance, C (f) Series L and C (d) Pure inductance, L Fig.7.9: fundamental RF components plotted on Smith charts. (e) Pure capacitance, C (f) Series L and C (d) Pure inductance, L (e) Pure capacitance, C (f) Series L and C (d) Pure inductance, L Towa Normalisation f0 f0 f0 f0 Wavelengths If you’re unfamiliar with j-notation, just remember that j is used to indicate the reactive part of an impedance (j = √-1). If you’re not too happy with the notion of ‘imaginary’ numbers, just remember that when the sign of the j-term is positive, a component will be inductive and, conversely, when the sign of the jterm is negative, a component will be capacitive. All practical RF components and antennas exhibit complex impedances, with both resistance and reactance present, and thus can be represented using j-notation. Importantly, complex impedance will vary with frequency due to the changing values of reactance present. Consider the case of a component that can be represented by a resistance of 50Ω connected in series with an inductance of 100nH. At 100MHz, the reactance of the inductor will be approximately 63Ω, and the component will have an impedance that can be expressed as (50 + j63)Ω. At 140MHz, the component’s inductive reactance will increase to 88Ω, so the impedance will then be (50 + j88)Ω. Corresponding impedance values for a resistance of 50Ω connected in series with a capacitor of 20pF will be (50 – j80)Ω at 100MHz and (50 – j57)Ω at 140MHz. (c) Open circuit, R = ∞ (c) Open circuit, R = ∞ (c) Open circuit, R = ∞ (c) Open circuit, R = ∞ (b) Short circuit, R = 0 (b) Short circuit, R = 0 (b) Short circuit, R = 0 (b) Short circuit, R = 0 0.3 j-notation R R R R (a) Pure resistance, R = Z0 (a) Pure resistance, R = Z0 (a) Pure resistance, R = Z0 (a) Pure resistance, R = Z0 0.4 Some fundamental components are shown together with their Smith chart representation in Fig.7.9. A pure resistance identical to the system’s characteristic impedance is shown in Fig.7.9(a), while short and open circuit connections are respectively shown in Fig.7.9(b) and 7.9(c). Pure inductance and pure capacitance are represented by points on the circumference, as shown in Fig.7.9(d) and Fig.7.9(e). Because their reactance changes with frequency, they are shown as lines rather than dots, with the arrow on the line showing the direction of increasing frequency. Finally, Fig.7.9(f) shows a series combination of pure inductance and pure capacitance. Note that the impedance of this combination will become zero at the frequency of resonance, f0. This is the resonant frequency at which the inductive and capacitive reactance will become equal but of opposite direction and cancel each other out. (50 + j25)Ω will be plotted as (1 + j0.5)Ω. Conversely, an impedance of (50 – j25)Ω will be normalised to (1 – j0.5)Ω. These conjugate points are shown plotted in Fig.7.10. Standing wave ratio The standing wave ratio can be Practical Electronics | May | 2026 1.2 1.0 0.9 0.1 7 0.7 1.4 0.3 3 0.1 60 1.6 8 0.3 2 1.8 0.2 0 0.0 5 14 0.4 5 0.04 3.0 1.0 15 0 0.8 160 -170 50 20 10 5.0 4.0 3.0 2.0 1.8 1.6 1.4 1.2 1.0 0.9 0.8 0.7 0.5 0.4 0.3 0.2 0.1 0 0.8 1.0 -160 -150 0.04 3.0 0.46 40 -1 2.0 1.8 0.5 0.0 6 4 0.4 7 0.1 0.3 2 0.17 2 0.4 0.09 0.41 30 -1 0.2 8 3 0.4 0.0 8 2.0 0 1.8 1.6 1.4 5 0.0 5 0.4 0 0.4 6 15 0.46 rato r 0.03 Gene 0.02 Towa 0.01 170 ±180 50 20 10 5.0 4.0 3.0 2.0 Wavelengths 1.0 0.8 1.8 1.6 1.4 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.00 0.4 5 14 0 0.0 5 0.3 0.47 0.40 0.1 0.6 0.11 -5 1 0.0 -120 0.8 0.9 0.12 0.38 Gen erat or 0.03 0 1.0 0.13 0.37 0.02 rd 1.2 0.36 0.10 0.2 0.49 0.47 0.03 0.7 0.35 -100 -90 0.4 0.3 7 0.0 1.6 30 -1 0.6 0.14 170 0.48 -80 0.15 ±180 0.00 0.01 0.34 -170 0.49 0.48 0.02 -110 160 0.6 0.33 -60 1.4 -70 0.16 Towa 3.0 0.33 -60 0.47 0.8 1.2 0.17 0 1.0 -70 1.0 2 rd 1.0 5.0 0.9 8 0.3 0.01 0.8 6.0 -110 0.16 0.34 -80 0.15 0.14 0.35 0.39 0.36 -100 -90 0.13 0.37 0.12 0.38 -120 0.8 0.2 0.48 7.0 0.7 0 0.1 1.2 8.0 0.6 -5 1 0.49 9.0 10 0.5 9 6 0.0 0.1 0.1 9 0.3 Wavelengths 50 3.0 0.4 0.00 0.2 4.0 0.6 0 110 MHz 0.23 – j0.65 0.3 0 -4 40 -1 0.8 0.1 0.4 -150 4.0 0.29 0.2 0 0.4 0 0.3 -4 0 0.4 4 0.4 5.0 0.2 -160 6.0 1.0 0.21 0.04 0.28 0.22 0.2 9 5 0.0 7.0 -20 0.3 -30 0.2 1 0.2 0 5 0.4 8.0 s Load 0.28 0.22 0.03 0.27 0.23 Degree rd -20 -30 0.46 50 In 9.0 10 0.6 0.00 0.1 0.3 0.01 0.2 0.04 0.4 0.49 0.26 0.24 -10 Towa 0.6 2 0.25 Coefficient 20 0.4 0.2 0.25 0 Wavelengths es 0.27 0.23 Degre 0.02 0.2 0.4 0.26 50 0.2 0.24 Reflection 20 In 0.48 0.3 0.6 10 0.26 -10 Load 0.47 0.4 10 0.1 0.23 8.0 9.0 Of 0.25 Coefficient rd 0.3 120 MHz 0.3 + j0 0.22 7.0 0.27 0.25 0.24 0 Towa 0.6 6.0 Angle Reflection Wavelengths 20 0.2 0.24 0.26 50 5.0 20 10 20 4.0 1.0 0.28 Of 10 0.29 0.23 0.27 Angle 8.0 9.0 0.8 0.21 20 7.0 30 0.28 0.22 6.0 0.4 0.10 0.11 0.39 0.40 0.09 0.41 Fig.7.11: circles of constant SWR on a Smith chart. 0 0.3 0.29 5.0 0.2 0.1 0.6 130 MHz 0.75 + j0.41 4.0 1.0 1.5 1 40 0.21 30 0.8 2 0.4 0 0.2 0 0.2 0.6 3 0 0.3 40 0.4 1 0.4 0.3 4 50 9 13 2.0 2.0 0.16 0.34 70 0.6 1.4 0.8 3 0.4 6 0.15 0.35 80 0.5 1.0 0.6 12 0.0 1 0.3 0.5 0 0.0 50 110 0.42 7 0.3 2 0.2 0.4 4 8 0.14 0.36 90 0.1 10 0 13 0.1 9 0.1 6 0.0 0.41 0.08 0.3 3 60 1.8 3 0.4 0.40 100 1.6 12 0.1 7 1.2 0.8 0 0.0 70 0.09 0.37 0.38 0.39 0.34 0.13 0.12 0.11 0.10 0.16 0.35 80 0.7 0.42 7 0.9 110 0.15 0.36 90 100 0.08 0.14 0.37 0.38 0.39 0.40 0.41 0.13 0.12 0.11 0.10 0.09 Fig.7.12: an impedancefrequency plot. 3 0.0 8 0.4 2 0.4 Fig.7.13: a typical VNA analysis of a VHF antenna. easily shown on a Smith chart by superimposing circles of constant SWR on the chart, as shown in Fig.7.11. Note that a unity value of SWR corresponds to a circle of zero radius (ie, a point at the exact centre of the Smith chart), while infinite SWR corresponds to the outer perimeter of the chart. When used as a tool to investigate matching, the RF designer will invariably strive to ensure that load impedances are as close to the centre of the Smith chart as possible. Being able to see this all going on in realtime means making adjustments is a real breeze! Impedance-frequency plots Smith charts are often used to Practical Electronics | May | 2026 visualise the behaviour of antennas and devices over a range of frequencies. Fig.7.12 shows an example of the analysis of a component in which the normalised impedance varies from (0.23 – j0.65)Ω to (0.75 + j0.41)Ω over a frequency range extending from 110MHz to 135MHz. Note how the antenna becomes purely resistive at 120MHz and how the SWR of the antenna has fallen below two above 130MHz. Antenna analysis A low-cost vector network analyser (as in Fig.7.3) can provide a quick check on the characteristics of an antenna and its feeder. Fig.7.13 shows some typical results of the analysis of a 130MHz quarter-wave antenna in a 50Ω system. The analyser’s frequency sweep has been set from 120MHz to 135MHz with markers displayed at 120MHz, 125MHz and 130MHz. From Fig.7.13, we can quickly determine that: • The antenna’s SWR is less than 2 over the entire scan range • The antenna is purely resistive at 123MHz and 133MHz • The lowest SWR (approximately 1.6) is achieved at 130.5MHz; the best match to a 50Ω system is obtained at this frequency. To help you get up to speed with a VNA, the low-cost RF Demo Kit 19 shown in Fig.7.14 provides you with a variety of different devices for experimentation. This kit is available from several online suppliers and includes a handy set of PCB-mounted devices and circuits suitable for attachment to your VNA. The Demo Kit includes short-­ circuit and open-circuit terminations as well as a low-power 50Ω load. Spectrum analysers Fig.7.14: this useful RF Demo Kit will help you get up to speed with a VNA. Fig.7.15: this inexpensive handheld spectrum analyser operates from 100kHz to 5.4GHz Fig.7.16: a typical screen grab from the spectrum analyser in Fig.7.15, showing the fundamental, 2nd and 3rd harmonic components in the output of a 7MHz transmitter. 20 Spectrum analysers are invaluable tools for examining signals, enabling users to observe and interpret the output spectrum of RF sources including oscillators, synthesisers and transmitters. By displaying the amplitude of signals across a wide range of frequencies, these instruments allow for the identification of fundamental signals, harmonic components, noise and any unwanted or spurious emissions. Whether assessing the performance of transmitters, diagnosing interference, or ensuring compliance with regulatory standards, a spectrum analyser can provide you with a clear and immediate insight into the behaviour of a wide variety of RF systems. They are available in both benchtop and handheld versions, making them ideal for use both in the laboratory and in the field. Laboratory-grade spectrum analysers can be prohibitively expensive, but in recent years, smaller portable instruments (as in Fig.7.15) have become increasingly affordable, yet still offer a reasonable level of performance. For example, the tinySA analyser shown in Fig.7.15 provides a frequency coverage extending from 100kHz to 5.4GHz, making it suitable for examining a wide range of signals in the LF, MF, HF, VHF and UHF bands. Like the VNA that we met earlier, this handy device also supports a USB connection to a PC for charging and also for advanced data analysis and visualisation. Even ‘proper’ benchtop instruments from brands like Rigol can be purchased starting at a few hundred pounds, although the cost goes up quickly if you start looking at the more capable instruments (eg, with a wider bandwidth). Fig.7.16 shows a typical application for a low-cost spectrum analyser. This screengrab shows the output frequency spectrum of a home-built 7MHz transmitter. The frequency scan (horizontal axis) is set to cover the range from 1MHz to 50MHz with markers (dotted vertical lines) every 5MHz. Practical Electronics | May | 2026 The fundamental (marked “1”) is shown at 7.09MHz, at a measured level of -3.1dBm, corresponding to a power level of 0.49mW at the input of the spectrum analyser. To prevent overloading the spectrum analyser’s input and to provide a matched load for the transmitter, a matched 30dB attenuator was connected between the transmitter’s output and the spectrum analyser input. Consequently, the true output power delivered by the transmitter will be 30dB greater than the power received by the spectrum analyser, ie, (30dBm – 3.1dBm) = 26.9dBm. This is equivalent to 0.49W from the transmitter. As with a VNA, the ability to transfer data from the spectrum analyser to a PC opens further options for analysing and presenting data. Fig.7.17 shows an investigation of the output spectrum from a handheld digital VHF transceiver. Note how the signal (at 145.145MHz) appears modulated by an unwanted component at 5MHz, resulting in an output spectrum that extends at 5MHz intervals for at least 50MHz on either side of the carrier! The spectrum analyser’s scan range can be adjusted to permit further investigation close to the carrier frequency where spurious content is at a level that exceeds the specified limits for the equipment under test. For example, setting the scan range from 130MHz to 160MHz will permit further examination of all spurious signals having a level of less than -60dB relative to the carrier. Spectrum analysers can also provide a means of examining the RF spectrum over a very wide frequency range. This feature is invaluable when searching for sources of noise and interference that might cause difficulties with radio reception generally. Fig.7.18 shows the spectrum received by a tinySA over the range 100kHz to 1GHz, enabling simultaneous display of the entire LF, MF, HF and VHF spectrum in real time. In Fig.7.18, we’ve indicated the principal services present, as well as a strong locally generated carrier at 400MHz. Signal types can often be recognised by their shapes (due to the modulation method) viewed on the spectrum diagram. This can help you quickly distinguish AM, SSB, FM and digital signals. Hands-On: 5W load & power meter This month’s Hands-On project is a 5W load and power meter designed for measuring the output Practical Electronics | May | 2026 Fig.7.17: PC applications can further analyse spectrum data. Here, tinySA is showing unwanted (spurious) signals on the output of a handheld digital VHF transceiver. SW broadcast VHF FM broadcast DAB broadcasting Aircraft DTV broadcast Cellular telephone Isolated carrier Noise floor -96dB Fig.7.18: using a handheld spectrum analyser with a telescopic antenna. Thanks to its wide frequency coverage, we can view the LF, MF, HF and VHF spectra simultaneously. Fig.7.19: the circuit of our 5W load and power meter. SK1 RF input R1 100Ω C1 100pF R3 10kΩ R2 D1 100Ω BAT42 C2 10nF of low-power HF transmitters. The instrument is self-contained and its power indications are shown on an analog meter scale. This makes it ideal for adjusting RF circuits and systems where it’s easier to see the movement of a pointer over a calibrated scale than make sense of the fluctuating digits found on a conventional digital display. Although intended primarily for use with transmitters for the 80m (3.5MHz), 40m (7MHz) and 20m (14MHz) bands, the instrument P1-1 RV1 50kΩ P1-2 can be used at frequencies of up to 30MHz. The load can dissipate 6W (see later) and this makes it ideal for use with low-power ‘QRP’ transmitters, which can normally deliver up to 5W. The circuit of the load and power meter board is shown in Fig.7.19; the required 50Ω load resistance is obtained by connecting two 100Ω resistors (each rated at 3W) in parallel. The RF voltage developed across R1 and R2 is sensed and rectified by Schottky diode D1. 21 Resistor R3 and capacitor C2 form a low-pass filter, and the rectified DC output is developed across C2. Calibration adjustment is provided by RV1, and the output available at P1 is connected to a low-cost moving coil meter. The component layout for the 5W load and power meter (viewed from the top) is shown in Fig.7.20 while the corresponding track layout (viewed from below) is given in Fig.7.21. The required 18 track breaks can be made using a spot face cutter or small drill bit, and the six links on the upper side of the boards are made using short lengths of tinned copper wire. Note that R1 and R2 will run warm, so it’s important to ensure that you leave space around (and under) these resistors to permit unrestricted airflow. Figs.7.20 & 21: top and underside views of the 5W load and power meter board. Fig.7.22: the completed 5W load and power meter board. Fig.7.23: internal wiring of the 5W load and power meter. 22 Building the 5W load & power meter When the circuit board is complete (see Fig.7.22), it should be mounted in a suitable enclosure and interconnected as shown in Fig.7.23. When positioning the circuit board, adequate clearance must be left above it to accommodate the meter movement. To avoid the somewhat tedious task of manually fitting the SMA and BNC connectors, we recommend the use of ready-made coaxial cables to link the power meter board to the panel-mounted BNC input connector. These ready-made cable assemblies are available at a reasonable cost from online suppliers. To provide convenient power indications, it is necessary to replace the meter scale. A suitable replacement scale is shown in Fig.7.24. This can be copied and printed at a size that will allow it to be glued directly over the meter’s existing scale. Careful removal of the meter’s bezel involves gently prising the transparent cover upwards and away from the meter’s body. The completed 5W load and power meter is shown in Fig.7.25. If you prefer not to open and replace the meter’s scale, you can use the existing 15V DC scale in conjunction with the calibration chart shown in Fig.7.26. Practical Electronics | May | 2026 1 2 3 4 5 6 7 W Fig.7.24: the replacement meter scale. Calibration and testing Calibration of the power meter is straightforward and simply requires an RF source capable of delivering 5W at a frequency between 3MHz and 10MHz. If you have an RF power meter at hand, you can use it as a reference. Simply adjust RV1 for identical readings. Alternatively, the power meter can be calibrated using an RF voltmeter—including the RF probe featured in last month’s Hands-On project. Connect the voltmeter across R1 or R2 and adjust the RF source for an indicated voltage of 15.8V RMS with a power of 5W applied (this power will be shared equally between the two parallel-connected load resistors). Now carefully adjust RV1 so that the power meter scale reads exactly 5W if you’ve replaced the scale, or 10V if you’ve decided not to replace it. This completes the calibration process; the instrument is ready for use. The power meter exhibits a VSWR of 1.05 at 7MHz. This represents a reasonably good impedance to a 50Ω source. However, due to the stray reactance associated with the simple construction method used, the load will become increasingly 5W load calibration curve 6.0 Fig.7.25: the completed 5W load and power meter. imperfect as the frequency of the applied power increases. We recommend that the power meter should be limited to a maximum of 30MHz, at which point the VSWR will be in the region of 1.2. S-parameter data imported into App­CAD (see earlier) yields the S11 VSWR plot shown in Fig.7.27. This clearly shows the rapid rise in VSWR above 10MHz. Derating All practical loads require derating. Theoretically, the two 3W resistors (R1 and R2) in this project should be capable of dissipating 6W. Unfortunately, it’s necessary to progressively derate an RF load resistor because, in practical use, resistors are unable to dissipate their maximum rated power indefinitely without overheating (without forced airflow). When subjected to continuous high power, such as the 6W theoretical maximum for the resistors used in this project, the resistors will become excessively hot if operated at this level for extended periods. With all types of load, excessive heat can lead to premature failure, changes in resistance value, or even damage to surrounding components. Therefore, derating ensures reliable operation by reducing the allowable power dissipation as operating time or ambient temperature increases, preventing thermal stress and maintaining the working life of the components. 5.0 Power (W) 4.0 3.0 2.0 1.0 0.0 0 2 4 6 8 Voltage indicated (VDC) 10 12 Fig.7.26: the 5W load & power meter calibration chart. Fig.7.27: an AppCAD VSWR plot from S11 data for the 5W load & power meter. Practical Electronics | May | 2026 23 30 5.0 25 4.0 20 Load power (W) Load power (W) 6.0 3.0 Safe operating area 15 Safe operating area 2.0 10 1.0 5 0.0 0 2 4 6 8 Elapsed time (minutes) 10 Fig.7.28: the recommended derating curve for the 5W load. Our derating curve shown in Fig.7.28 indicates recommended safe operation limits: 6W (the maximum rated power) for up to one minute, 4W for four minutes and 3.5W for ten minutes. Just remember to keep within the ‘safe operating area’ marked on the chart! Hands-On: A 25W load If you require a load with a better rating (for example, to test VHF transceivers), we have a second Hands-On project to put this right. This improved load is conservatively rated at 25W and is suitable for operation up to 450MHz. The circuit of the 25W load is shown in Fig.7.31. The load uses a single 50Ω 100W RFP1109 ‘microwave resistor’ mounted in a subFig.7.30: the mounting details for the load resistor. 0 12 0 2 4 6 8 Elapsed time (minutes) 10 12 Fig.7.29: the recommended derating curve for the 25W load. stantial diecast enclosure. This not only provides effective screening for the load but also provides a very effective means of dissipating the generated heat. If necessary, to cater for higher pow­ ers (up to 50W), the enclosure can be augmented with one or more external heat dissipators. Suitably rated 50Ω 100W ‘microwave resistors’ are available from several online suppliers. Assembling the 25W load The load resistor should be mounted as close to the SO-239 connector (SK1) as possible, as shown in Fig.7.30. Heat dissipation will be further improved by the application of a small amount of thermal paste between the resistor’s metal mounting lugs and the inner surface of the enclosure. Solder the resistor directly to the centre pin of SK1, but avoid applying excessive force when bending the resistor’s connecting strip. Testing and derating If an inline VSWR/RF power meter is available, it should be used to check the 25W load using an arrangement like that shown in Fig.7.32. The VSWR should be less than 1.05 at all frequencies up to 450MHz, and the load should be capable of sustaining a power of 30W for short periods and 12.5W indefinitely. If required, a further heat dissipator can be bolted to the exterior of the enclosure. This should ideally be fitted near the mounting bolts fitted to R1 and as close to the load resistor as possible. Our recommended derating curve for the 25W load is shown in Fig.7.29. This chart assumes that no external heat dissipator has been fitted. We recommend a maximum of one minute when 30W is applied, 20W for five minutes and 15W for ten minutes. The prototype was placed on test with 15W applied for several hours at an ambient temperature of 20°C. At the end, the surface temperature SK1 RF input R1 50Ω Fig.7.31: the circuit of the 25W load. 24 Practical Electronics | May | 2026 Fig.7.32: testing the 25W load with the cover removed. At 145MHz, the indicated VSWR is 1.00. of the enclosure appeared warm, but not excessively hot. Parts List – 5W load and power meter Coming up! Part 8 will introduce frequency changing methods, exploring multipliers and mixers with practical examples for testing in the popular Tina simulation package. Our Hands-On constructional project will be a 50MHz digital frequency meter. See you then! PE 1 small ABS enclosure, 100 × 68 × 50mm or larger 1 25 × 64mm piece of stripboard (9 × 24 holes) 1 BAT42 Schottky diode (D1) 1 2-pin male 0.1in/2.54mm header (P1) 150kΩ miniature preset potentiometer/‘trimpot’ (RV1) 1 PCB-mounting female SMA connector (SK1) 1 15cm panel-mount BNC female to SMA male coaxial cable adaptor 1 15V DC analog voltmeter with zero adjuster [eg, Velleman 7015] 4 brass or nylon M3 × 10mm hex tapped spacers 4 M3 panhead machine screws 4 M3 countersunk head machine screws Parts List – 25W load 1 diecast enclosure, 110 × 60 × 30mm 1 SO-239 female coaxial connector (SK1) 1 RFP1109 (or similar) 50Ω 100W resistor 2 M3 countersunk head machine screws 2 M3 hex nuts and shakeproof washers 1 small tube or packet of thermal paste Capacitors 1 10nF (C2) 1 100pF (C1) Resistors (all ¼W axial, ±5% or better) 1 10kΩ ¼W axial, 5% or better (R3) 2 100Ω 3W axial resistors (R1, R2) The Fox Report, continued from page 13... ... radio web pages and the Sounds App) use AAC-LC coding at a range of qualities up to 320kbps. However, Freeview and Freesat max out at 192kbps, and DAB reduces rates to the best cases of 160kbps or 192kbps for Radio 3. So what data rates does Freely use for BBC Radio? Netgem replied to my question with the question: “Please can you clarify what you mean by data rates?” After clarifying, I heard nothing back on this from Netgem and Freely simply ignored my questions on data rates. Fortunately, Manhattan was on the ball and clarified, “All four national BBC radio stations use the same 320kbps high-quality stream on Freely....by default.” I also asked about the delay viewers and listeners will experience when switching between streamed channels, now and in the future when more and perhaps higher quality channels are added. Says Manhattan, “It does vary a Practical Electronics | May | 2026 little depending on circumstances, but we aim for around two seconds maximum.” Says Netgem, “The target delay between Freely TV channels is three seconds. For radio channels it’s five seconds: three seconds to load, and then two seconds to load the BBC Sounds App by pressing the Red Button.” Adding more TV and radio channels “will not impact the target delay times”, says Netgem. Freely acknowledges the need for “enhancements specifically aimed at boosting performance on lowerspecification hardware” but will give no indication of what these “enhancements” are or what the target delays are, now or in the future. Freely has pledged to “futureproof live TV for the streaming age”, so I asked if Freely can assure people who buy first-generation receivers that they will not suffer increased streaming channel switching delays as more channels are added: We continue to evolve the channel tune speed across all of our Freely device...As part of these efforts, we recently completed a series of enhancements specifically aimed at boosting performance on lower-specification hardware, bringing all Freely viewing experience in-line. We’re also finalising work to improve overall UI [user interface] responsiveness across Freely, with specifically identified opportunities to optimise channel tuning further. Freely is continually undergoing improvements and enhancements, with regular updates throughout the year. The number of channels on the Freely streaming platform is evolving and growing at pace. Freely viewers can also enjoy new and enhanced features added throughout the year. Over time, the platform is enhanced and the viewing experience will only be improved... The viewing experience will only be improved as more chanPE nels are added. 25