Optimal Passive Power Line Communication Filter for NB-PLC Applications

Abstract

Narrowband Power Line Communication (NB-PLC) involves transmitting data by overlaying a high frequency low amplitude signal (ranging from 9 kHz to 500 kHz) onto the low-frequency high amplitude signal (50 Hz to 60 Hz) of the power grid. While using the existing power grid for communication is convenient, it was not originally designed for this purpose, leading to challenges such as conducted emissions and infrastructure constraints. To overcome these technical obstacles, power line filters (PLFs) are a viable solution. The results of our research work, focusing on the optimization of PLFs for NB-PLC to ensure their design fits the needed use case while avoiding over-engineering, are presented in this article. Our study concentrates specifically on the filtering of PLC signal and conducted emissions up to 500 kHz. Building upon a PLC PLF extensively discussed in our previous work—which blocks the PLC signal in the CENELEC-A frequency band regardless of its placement within the electrical installation, sometimes leading to over-engineering—this research aims to adapt the filter order and components for a variety of real scenarios in CENELEC-A, FCC, and ARIB frequency bands. By characterizing different filters, our work provides tailored solutions for these scenarios and serves as a framework for future filter designs in PLC applications.

1. Introduction

At the end of 2022, more than 56% of electricity consumers in EU27+3 were equipped with smart meters (SM). This percentage is projected to surge to approximately 78% by 2028. As of 2022, over 165 million smart meters utilize Narrow Band Power Line Communication (NB-PLC) as their communication protocol in Europe, and this number is expected to increase significantly. NB-PLC will continue to be the preferred communication protocol between smart meters and Data Concentrators (DC) [1].

1.1. Spectrum Allocation and Introduction to NB-PLC

The spectrum allocation in Europe stipulates that SMs should communicate in the frequency band starting at 3 kHz and ending at 95 kHz, known as the Comité Européen de Normalisation Électrotechnique-A band (CENELEC-A) [2]. In other areas, PLC frequency bands have a broader spectrum allocation, starting in kHz range and going up to 500 kHz. Figure 1 presents how the PLC spectrum is allocated in regions that use PLC in SM, such as Association of Radio Industries and Businesses (ARIB) in Japan [3], Federal Communications Commission (FCC) in Unites States of America [4], and CENELEC [2]. Furthermore, Figure 1 categorizes the frequency spectrum into ultra narrow band (U-NB) [5,6,7], narrow band (NB), medium band (MB), and broadband (BB) segments [8,9,10,11], emphasizing how regulatory bodies distribute frequencies for PLC use.

PLC involves the usage of the existing power lines for data transmission [12]. This can be achieved by superimposing a relatively high frequency carrier signal onto the low frequency mains voltage. Figure 2 presents this fundamental concept on which PLC is based by using practical examples for a 40 kHz 1 Vpp PLC signal coupled onto a 50 Hz, 230 VRMS, mains voltage waveform.

To enhance comprehension of the PLC signal, Figure 3 offers a scaled-up graphical representation of the mains voltage waveform with the PLC signal superimposed. from this figure, we can see how the relatively high frequency low amplitude PLC signal is coupled on the high amplitude low frequency mains voltage. Upon examining Figure 3, one should observe how the relatively high-frequency PLC signal, with low amplitude, is coupled onto the high-amplitude, low-frequency mains voltage. Figure 4 is the frequency domain representation of the PLC signal coupled on the mains voltage whose time domain representation is depicted in Figure 3. Due to the significant difference in values, a dual-axis amplitude representation is used.

The electrical energy distribution grid was designed for energy distribution and not as transmission medium for communication. Though it has the benefit of relying on the transmission medium of an existing infrastructure, it is affected by the following types of factors: colored background noise, narrow-band noise, and periodic impulsive noise asynchronous to the mains frequency [13,14,15]. PLFs are a method for filtering out conducted noise from the NB-PLC frequency band, especially in the CENELEC A frequency band, which is prone to conducted noise because electromagnetic compatibility (EMC) standards specify relaxed limits below 150 kHz or in some cases no limits and most switched mode power supplies are switching in this frequency band [13,14,15,16,17,18].

1.2. Contributions Beyond the State of the Art

This article presents the continuation of the authors’ previous work related to passive PLF design, with results previously published in [16]. The results presented in this article expand the frequency band of the filters while adapting the filter order based on the NB-PLC frequency band used and existing attenuation in the network, thereby avoiding over-engineering.

The primary goal of the research presented in this article is to adapt the filter order with respect to the NB-PLC frequency used and the installation point. This involves the design, simulation, and measurement of filters of different orders and correlating these with the installation cases.

In previous work, the filter completely attenuated the PLC signal regardless of existing attenuation, resulting in an over-engineered, universal solution. This approach did not account for scenarios where the power grid’s inherent attenuation will contribute to the overall attenuation.

Compared to our previous work, we have expanded our testing to include actual NB-PLC modems for attenuation testing (Section 3.2) and introduced a new use case for filters in medium voltage (MV) connections (Section 2.1).

Using the appropriate filter order in relation to the installation point and frequency band yields the following benefits:

• Lower cost and size of the filter.
• Reduce the quiescent current it would draw.
• Reduce power losses caused by voltage drop over the series connected inductors.

Compared to research presented in other relevant articles:

• The research presented in [19] focuses on designing a filter for the CENELEC-C frequency band (120–150 kHz) with a relatively low current, making it suitable for mounting on individual devices. The design method employs equations for inductors and capacitors, and the filter is intended for PLC home automation and indoor use.
• The research presented in [20] focuses on a notch filter that is part of the coupling circuit connecting to the PLC transceiver. It utilizes active components and is designed specifically for the low-current signal path of PLC communication. The frequency band covered extends up to 9 MHz and the design method uses classical electronic circuit equations.
• The research outlined in [21] delves into the theoretical design of line traps for PLC. Compared to our article it lacks scalability concerning the installation location and practical measurements to validate the design. The article prioritizes the frequency range of the designed circuits over creating a solution capable of filtering the entire frequency bandwidth.

Our research covers the frequency band from 10 kHz to 500 kHz. The filter’s current rating makes it suitable for connecting households, with potential for higher current ratings. By using S-parameters in the filter design, we model the contribution of each component to the filter’s overall performance. Furthermore, we analyzed the filter’s integration into the network, aiming to correctly size the filter based on the connection case. The simulated and measured insertion loss was confirmed with practical measurements using NB-PLC modem using PLC-G3.

1.3. Article Structure

The article presenting the research activities is structured in the following manner:

Section 1 introduces PLC and highlights its relevance for SM communication. It lays the foundation for subsequent sections while also outlining the purpose and significance of the research.

Section 2 introduces the theoretical aspects, installation procedures, and measurement methods of PLFs in NB-PLC systems, taking into account the grid’s topology.

Section 3 presents the simulation and practical measurements of NB-PLC PLFs, simulates various power line attenuations, and discusses the use cases for each filter order.

Section 4 presents the concluding part of the research activity.

2. Optimal Passive PLF for NB-PLC Applications Theoretical Considerations

This section delves into the theoretical aspects of PLFs NB-PLC applications. It covers the installation of PLFs within the grid, considering the topology of the power grid. The subsections further discuss the methods used for PLF installation, design, and measurement within this context.

2.1. PLF Installation in the Power Grid

Three scenarios exist for integrating PLFs into the power grid. Figure 5, depicting a power grid featuring one transformer from high voltage (HV) to MV, along with two low-voltage branches, each equipped with two DCs to enhance data communication rates, was created to provide an overview of these three use cases:

• In filter case 1 between SM and LV power line to filter the conducted noise generated by the devices in the house, a filter is installed between the SM and the fuse box. The appropriate filter order and attenuation in this case can be determined by measuring the attenuation of segment denoted with “a” [22].
• In filter Case 2, a filter is positioned between two segments of the low-voltage power line. This setup SMs into two PLC areas while remaining connected to the same electrical grid branch. The division into two sections ensures that each SM utilizing NB-PLC connects to the corresponding PLC-DC. This type of PLF installation becomes essential when numerous SMs are attached to a branch and an extra DC is introduced to augment throughput [23]. The appropriate filter order and attenuation can be determined in this case by measuring the attenuation introduced by segments denoted “b” and “c”.
• Filter case 3 between two segments of MV power line so that the SMs will not communicate over the transformer. The appropriate filter order and attenuation can be determined in this case by measuring the attenuation introduced by segments denoted “d”, which also takes into consideration the attenuation introduced by the MV-LV transformers (Tr) [24,25].

To assess the additional attenuation introduced by the power line, we utilized the propagation Equation (1), which characterizes the attenuation of transmission lines, specifically power lines in this context. Equations (2) and (3) define the terms used in Equation (1), providing the characteristic parameters for each type of cable analyzed [26]. In Section 3.3, we present the attenuation simulation results for CENELEC-A and FCC frequency bands, spanning from 10 kHz to 500 kHz for both copper and aluminum solid core cables of various cross-sections.

$${\mathrm{V}}_0\left(\mathrm{z}\right)=\left|{\mathrm{V}}_0\left(0\right)\right|{\mathrm{e}}^{\mathrm{ȷ}\mathsf{\phi }}\ast {\mathrm{e}}^{-\mathrm{ȷ}\mathsf{\beta }\mathrm{z}};\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{p}\mathrm{a}\mathrm{g}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{e}\mathrm{q}\mathrm{u}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}$$

(1)

$$\mathsf{\gamma }=\mathsf{\alpha }+\mathrm{j}\mathsf{\beta }=\sqrt{\left(\mathrm{R}+\mathrm{j}\mathsf{\omega }\mathrm{L}\right)\left(\mathrm{G}+\mathrm{j}\mathsf{\omega }\mathrm{C}\right)};\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{p}\mathrm{a}\mathrm{g}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{c}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{t}$$

(2)

$$\begin{array}{c}\mathsf{\alpha }=\mathrm{R}\mathrm{e}\left\{\mathsf{\gamma }\right\};\mathrm{a}\mathrm{t}\mathrm{t}\mathrm{e}\mathrm{n}\mathrm{u}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{c}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{t}\\ \mathsf{\beta }=\mathrm{I}\mathrm{m}\left\{\mathsf{\gamma }\right\};\mathrm{p}\mathrm{h}\mathrm{a}\mathrm{s}\mathrm{e} \mathrm{c}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{t}\\ \mathrm{L};\mathrm{i}\mathrm{n}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{e} \mathrm{p}\mathrm{e}\mathrm{r} \mathrm{u}\mathrm{n}\mathrm{i}\mathrm{t} \mathrm{l}\mathrm{e}\mathrm{n}\mathrm{g}\mathrm{t}\mathrm{h}\\ \mathrm{G};\mathrm{conductance}\mathrm{per}\mathrm{unit}\mathrm{length}\\ \mathrm{C};\mathrm{capacitance}\mathrm{per}\mathrm{unit}\mathrm{length}\end{array}$$

(3)

2.2. PLF Design Method

For designing the filter, we have employed the usage of scattering parameters (S-parameters). S-parameters for each component have been acquired using a vector network analyzer (VNA) and simulations were conducted using RF-SIM99 simulation program.

Touchstone files provide pairs of the real and imaginary parts at each measurement frequency point, thus characterizing the component measured. In this case S₁₁, single port measurements, have been used for component level characterization and for assessing the attenuation of the filter S₂₁ and S₁₂ [27]. Figure 6 graphically represents a two port generic device (which can be the filter) having the following S-parameters specific to a bidirectional network [28]:

• S₁₁—reflection coefficient at port 1
• S₁₂—transmission coefficient that defines reverse gain
• S₂₁—transmission coefficient that defines the forward gain
• S₂₂—reflection coefficient at port 2
• a₁, b₁ incident voltage wave at Port1 respectively Port2
• a₁, b₁ reflected voltage wave at Port1 respectively Port2

The mathematical definition of the parameters is conducted using Equations (4)–(10).

$$S_{11}={\displaystyle \frac{b_1}{a_1}}$$

(4)

$$S_{12}={\displaystyle \frac{b_1}{a_2}}$$

(5)

$$S_{21}={\displaystyle \frac{b_2}{a_1}}$$

(6)

$$S_{22}={\displaystyle \frac{b_2}{a_2}}$$

(7)

$$\left(\begin{array}{c}{b}_1\\ b_2\end{array}\right)=\left(\begin{array}{cc}{S}_{11}& S_{12}\\ S_{21}& S_{22}\end{array}\right)\times \left(\begin{array}{c}{a}_1\\ a_2\end{array}\right)$$

(8)

$$b_1=S_{11}{a}_1+S_{12}{a}_2$$

(9)

$$b_2=S_{21}{a}_1{+S}_{22}{a}_2$$

(10)

2.3. PLF Measurement Methods

In order to assess the performance of the filter attenuation, two types of measurements were selected:

• VNA transmission measurements ($S_{21}$ and $S_{22}$) for NB-PLC frequency band up to 500 kHz. In this article, this evaluation method will be termed as “measurement setup 1” and will be elaborated upon in Section 2.3.1.
• PLC-G3 modem communication testing in both CENELEC-A and FCC frequency bands. In this article, this evaluation method will be referred to as “measurement setup 2” and will be elaborated upon in Section 2.3.2.

2.3.1. Measurement Setup 1

The measurement setup from Figure 7 offers a complete attenuation measurement, starting as low as 30 kHz and going up to 1 MHz, by injecting a variable sinusoidal signal at one of the ports and measuring the signal it receives at the other port, subsequently computing Equations (2) and (3). To exclude the influence of the couplers from the NB-PLC filter performance measurement, a pass-through calibration, bypassing the filter, was performed.

The following equipment was used for constructing measurement setup 1:

• NanoVNA-F V2, which is used in various research activities [29,30,31].
• PLC coupler LonWorks Model: 78200 [32].

2.3.2. Measurement Setup 2

This testing method in both CENELEC-A and FCC frequency bands is a communication method that mimics real-world field conditions more accurately compared to measurement setup 1. This is achieved by utilizing actual PLC modems, thereby accounting for signal modulation, amplitude, and modem receiver sensitivity. However, a drawback of this method is its inability to provide a precise attenuation value; instead, it yields a communication channel quality metric.

Measurements are conducted by configuring one PLC modem in transmit mode at one end of the filter and setting the other PLC modem on the opposite side in receive mode. Communication feasibility is then assessed, and the transmitted and received frequency spectra are measured using a spectrum analyzer.

The subsequent apparatus was utilized in the construction of measurement setup 2 presented in Figure 8:

• Spectrum analyzer Keysight™ N9320B [33].
• PLC Modem, MAX79356CAEVK1 [34].
• PLC coupler LonWorks™ Model: 78200 [32].

3. Optimal Passive PLF for NB-PLC Application Experimental Results

This section primarily focuses on the development and assessment of optimal passive PLFs for NB-PLC applications. It involves simulating the performance of proposed PLFs, conducting measurements to validate their effectiveness in a laboratory environment replicating real-world conditions, and exploring practical use cases while considering the appropriate filter order in relation to the installation point within the power grid.

Throughout this section, the complexity of the order filter designed in previous research published in [16] will be reduced, thus adapting it to the use case. The starting point is the 13th order filter from Figure 9, whose design is fully presented in [16].

To enhance clarity regarding the filters under analysis in this section,

Table 1. Filter order.

Filter	Components Removed
13th order	none
10th order	L7, L8, C5
7th order	L7, L8, C5, L5, L6, C4
4th order	L7, L8, C5, L5, L6, C4, L3, L2, C3

has been generated. The filter order will serve as a reference point, with components being removed relative to the filter depicted in Figure 9. The essential components needed for the filter to maintain attenuation and provide immunity against network transients are: thermally protected varistor (TMOV), R1, R2, R3, C1, L1, and L2 [35].

3.1. Proposed PLFs Simulation

Simulation was conducted using the RFSim99 program, with the S-parameters of each component measured using the VNA as input. The simulation frequency band is between 9 KHz and 1 MHz. Figure 10 is the 13th order passive filter simulation diagram from Figure 9 as it was simulated in RFSim99.

Figure 11 displays the simulated S₂₁ magnitude plot for the 13th-order filter and its de-rated versions. The −80 dB line represents the experimentally determined attenuation level at which communication ceases for PLC-G3 modems in both the CENELEC-A and FCC frequency bands. This level was determined by interposing a rotary attenuator between the two PLC couplers, rather than the filter, in measurement setup 2.

By analyzing Figure 11, the following was observed:

• 13th order filter crosses −80 dB at 25 kHz
• 10th order filter crosses −80 dB at 26.7 kHz
• 7th order filter crosses −80 dB at 54 kHz
• 4th order filter crosses −80 dB at 192.9 kHz

Figure 12 provides a detailed view of the S₂₁ simulation results of the 13th order filter and its de-rated versions with focus on CENELEC-A frequency band.

Given the attenuation levels at each frequency, the following conclusions can be drawn out of the simulation:

• Both the 13th and 10th order filters excel at filtering out signal and noise within the CENELEC-A band. However, the 13th order filter is considered over-engineered since the additional attenuation it provides is unnecessary for NB-PLC applications.
• The 7th order filter demonstrates attenuation effectiveness within the FCC band, but at least one CENELEC-A sub-band remains unfiltered.
• The 4th order filter proves relatively ineffective; communication via PLC-G3 remains operational in both frequency bands.

3.2. Proposed PLFs Measurements

The de-rated 13th order filter was tested following the procedures outlined in Section 2.3.1 and Section 2.3.2. Figure 13 illustrates the integration of both measurement setups 1 and 2 into a combined test setup as well as the 13th order passive filter. Measurements were conducted for each filter order within both the CENELEC-A and FCC frequency bands. Additionally, attenuation measurements (S₁₂ and S₂₁) were taken from 30 kHz up to 1 MHz. Subsequently, the filter order was decreased in accordance with Table 1, and further measurements were carried out.

To evaluate the PLC communication performance through the filters, the Zeno PLC kits from Maxim™ MAX79356CAEVK1 [34] were utilized, employing the configuration detailed in

Table 2. PLC-G3 Transmitter configuration used for measurements.

Parameter	Setting	Comments
Standard	CENELEC or FCC	Selectable depending on the frequency band
Tone Map	3F for CENELEC-A/FFFFFF for FCC	Selectable depending on the frequency band
Modulation	ROBO	4 times repeated binary or differential Phase Shift Keying
Type of modulation	Differential
Packet size	30	Payload size in bytes
LD Gain	6	Line Driver gain
IAFE Gain	3.5 dB	Integrated Analog Front End gain
Auto PDC	Disabled	Phase detection counter

for transmission.

3.2.1. Measurements Performed on the 13th Order Filter

The 13th order filter depicted in Figure 9 has been measured using the test setup from Figure 13, which can be configures as Setup 1 and Setup 2 from Figure 7 and Figure 8. Figure 14 displays the CENELEC-A frequency spectrum measurements, while Figure 15 illustrates the FCC frequency spectrum measurements for the 13th order filter in both figures. Trace V1 shows the PLC signal at the modem’s output in transmitter (TX) mode, measured at the filter’s input. Trace V2 displays the frequency spectrum at the modem in receiver (RX) mode, measured at the filter’s output. It is evident that no PLC signal is detected at the output of the filter. Both types of PLC-G3, CENELEC-A and FCC, are completely filtered out by the 13th order filter. Additionally, the attenuation S₂₁ and S₁₂, measured with the VNA, as shown in Figure 16, indicates an average measured attenuation of −82 dB. This measurement is constrained by the VNA’s measurement capabilities, with a noise floor of −82 dB. The −80 dB attenuation was recorded at 42 kHz. Communication between the PLC-G3 modems was not possible in both the CENELEC-A and FCC frequency bands. From a technical standpoint, the simulated S₂₁ and S₁₂ values depicted in Figure 11 accurately represent the actual attenuation. However, due to the inability to establish communication at attenuations below −80 dB, it was concluded that the VNA measurements are also relevant.

3.2.2. Measurements Performed on the 10th Order Filter

The 13th order filter has been de-rated to a 10th order filter in accordance with Table 1 and it has been evaluated using the combined test setup from Figure 13. Figure 17 depicts the CENELEC-A frequency spectrum measurements, while Figure 18 illustrates the FCC frequency spectrum measurements for the 10th order filter. In both figures, Trace V1 shows the PLC signal at the modem’s output in TX mode, measured at the filter’s input. Trace V2 displays the frequency spectrum at the modem in RX mode, measured at the filter’s output. It is apparent that no PLC signal is detected at the output of the filter. Both types of PLC-G3 signals, CENELEC-A and FCC, are effectively filtered out by the 10th order filter. Additionally, the attenuation, measured with the VNA from Figure 19, indicates an average measured attenuation of −82 db. This measurement is limited by the VNA’s capabilities. The attenuation of −80 dB was recorded at 49 kHz. Communication between the PLC-G3 modems was not achieved in either CENELEC-A or FCC frequency bands.

As a result of the assessments, it can be concluded that under practical, close to field conditions, the 10th order filter is sufficient to filter out noise and PLC-signal in both CENLEC-A and FCC frequency bands.

3.2.3. Measurements Performed on the 7th Order Filter

The 13th order filter has been de-rated to a 7th order filter in accordance with Table 1 and it has been evaluated using the combined test setup from Figure 13.

Figure 20 depicts the CENELEC-A frequency spectrum measurements, while Figure 21 illustrates the FCC frequency spectrum measurements for the 7th order filter. In both figures, Trace V1 shows the PLC signal at the modem’s output in TX mode, measured at the filter’s input. Trace V2 displays the frequency spectrum at the modem in RX mode, measured at the filter’s output while Trace V3 is the noise floor. By comparing traces V1, V2, and V3, it is visible that a small portion of CENELEC-A PLC signal is detected at the output of the filter; however, the FCC PLC-G3 signal is filtered out by the 7th order filter. The attenuation, measured with the VNA from Figure 22, indicates an average measured attenuation of −80 dB, while the −80 dB attenuation was achieved at 60 kHz. Communication between the PLC-G3 modems is functional in CENELEC-A band

As a result of the assessments, it can be concluded that under practical, close to field conditions, the 7th order filter is sufficient to filter out noise and PLC-signal in FCC frequency band, and that with an additional ~5 dB attenuation, it will also filter the CENELEC-A band.

3.2.4. Measurements Performed on the 4th Order Filter

The 13th order filter has been de-rated to a 4th order filter in accordance with Table 1, and it has been evaluated using the combined test setup from Figure 13. Figure 23 depicts the CENELEC-A frequency spectrum measurements, while Figure 24 illustrates the FCC frequency spectrum measurements for the 4th order filter. In both figures, Trace V1 shows the PLC signal at the modem’s output in TX mode, measured at the filter’s input. Trace V2 displays the frequency spectrum at the modem in RX mode, measured at the filter’s output, while Trace V3 is the noise floor. By comparing traces V1, V2, and V3, it is visible that both the CENELEC-A and FCC PLC signals are detected at the output of the 4th order filter. The attenuation, measured with the VNA from Figure 25, indicates an average measured attenuation of −78 dB, while the −80 dB attenuation was achieved at 224 kHz. Communication between the PLC-G3 modems is functional in both CENELEC-A band and FCC frequency bands.

3.3. PLF Installation in the Power Grid Attenuation Simulation

This section presents the simulated attenuation levels that power lines can introduce in the three installation scenarios described in Section 2.1. The results obtained will be used to analyze the relationship between filter order and installation.

For simulating the attenuation levels vs. distance and at different frequencies, we have implemented Equations (1) and (2) presented in Section 2.1 into a Matlab script. The power line cable characteristic were taken from cable datasheets or deduced using per unit length formulas of capacitance, inductance and resistance [36].

Simulation was performed using the following parameters and considerations:

• Power line cable core made out of solid copper or aluminum having cross sections of: 14 mm², 40 mm², 100 mm², 200 mm².
• Simulation stops when the PC signal level reaches 0 dBμV, although communication will stop at 40 dBμV as determined in Section 3.2.
• PLC signals with frequencies of 10 kHz, 100 kHz, 200 kHz, and 500 kHz are transmitted at a level of 120 dBμV at the beginning of the power line.
• There is no multipath propagation and the power line impedance is 2 Ω [37].

The filter use cases and attenuation based on each use-case as defined in Section 2.1 and graphically represented in Figure 5:

• Case 1 filters are connected through 14 mm² to 40 mm² cables with lengths of up to 100 m. The simulated power line impedance for Case 1 filters is provided in Figure 26.
• Case 2 and Case 3 filters are typically connected through wires having cross sections in the range of 100 mm² to 200 mm², while lengths are in excess of 1000 m. The simulated power line impedance for Case 2 and Case 3 filters is provided in Figure 27. Additional attenuation of up to −20 dB is introduced by MV-LV transformers for filters, which are installed as in Case 3 [24,25].

3.4. PLFs Use Case and Filter-Order Discussion

After analyzing the simulation results and practical measurements from Section 3.1, Section 3.2 and Section 3.3, we observed the following:

• A 13th-order filter can be used in all three filter use cases, but it is over-engineered.
• A 10th-order filter provides sufficient attenuation for all three filter use cases, regardless of the additional attenuation introduced by the power lines.
• A 7th-order filter is adequate for filtering the FCC frequency band, regardless of the additional attenuation introduced by the power line. However, it allows NB-PLC signals in the CENELEC-A band to pass at power line distances of less than 100 m. This filter provides enough attenuation for Case 2 and Case 3. It can potentially be used as a Case 1 filter in installations, which have other attenuation sources contributing to the overall attenuation.
• A 4th-order filter is suitable only for Case 3 installation, where the length of the line causes significant power line attenuation.

4. Conclusions

This article presents our research on design methodologies and practical measurements for the development and evaluation of optimal PLFs, with a particular emphasis on the CENELEC A, FCC, and ARIB NB-PLC communication bands. Our study focuses exclusively on the filtering of PLC signal and conducted emissions. By characterizing different filters, we aim to provide insights that can aid in the design of new filters for PLC applications. The theoretical and practical methodologies employed in our research, as outlined in this article, constitute a comprehensive framework for the design of optimal PLFs tailored to specific installation use cases.

The selection of the suitable filter order was determined through simulations and measurements of S-parameters, with further validation provided by PLC-G3 tests. Optimizing the filter order in accordance with the installation point and frequency band offers several notable advantages. First, it leads to a reduction in both cost and size of the filter, making it more efficient and economical to implement. Additionally, by minimizing the filter’s quiescent current draw, it enhances overall energy efficiency, contributing to sustainable operation. Moreover, the optimization helps mitigate power losses stemming from voltage drop across series-connected inductors, thereby improving the performance and reliability of the system. By prioritizing the appropriate filter order, these benefits collectively enhance the functionality and effectiveness of the filter, ensuring optimal performance in diverse applications.

We have demonstrated that the attenuation from power lines significantly contributes to the overall attenuation, allowing for a reduction in filter attenuation and filter order. This reduction can lower costs and decrease power consumption.

Future research directions include exploring alternative filter designs, such as resonant topologies, and new materials to enhance key characteristics like attenuation, power consumption, and impedance. Additionally, expanding the study to encompass real-world deployment scenarios and long-term testing will provide deeper insights into the practical applications and potential improvements of PLFs in various PLC communication bands. Investigating filtering for broadband PLC applications is also of significant interest. The framework used in our research will aid these future studies by providing a foundation for design, simulation, and measurement methodologies tailored to specific installation use cases.