Analysis of the damage of the protective device of the communication port in the surge test

1 Introduction

    Learning alone without friends is lonely and ignorant. This article takes the damage of the communication port protection device in the SURGE test as an example, analyzes the damage mechanism, analyzes the impact of different test methods, selects the test method for different application scenarios, and analyzes the points of attention in the SURGE protection design.

    2. Background introduction

    The communication port of a locomotive controller was subjected to a SURGE test of ±1KV, and the TVS tube of the communication port protection device was damaged.

    3. System networking

    Networking: A locomotive controller (EUT), connected to the AE equipment through the CAN communication line, the length of the interconnection line is 10m; Test port: CAN port, ±1KV, impedance 42Ω; Port protection description: EUT side CAN port TVS tube protection

    ①; AE side CAN port TVS tube protection

    ②; 15V power supply on EUT side connects capacitor C and TVS tube to ground

    ③; AE side has no jumper protection to the ground

    ④; Port injection method: Disconnect CAN-H and CAN-L connected to AE, but A-GND remains connected to AE equipment. Grounding instructions: CAN reference ground is A-GND, AE equipment communication reference and power reference ground is A-GND, AE is powered by DC15V of EUT, and DC15V reference ground is P-GND. The test network is shown in Figure 1.

1.png

    Figure 1 Network diagram Experimental phenomenon: It was found in multiple SURGE tests that the TVS tube of the EUT port was broken, but not every test can be reproduced. The more tests, the easier it is to reproduce.

    4. Analysis of the mechanism of TVS tube damage

    4.1 Path analysis of SURGE interference

    The reference ground of the CAN circuit is A-GND, the power reference ground of the AE device is A-GND, and the 15V power ground of the EUT is P-GND. Because the DC15V of the EUT supplies power to the AE, the A-GND and P-GND are connected Interconnect, and P-GND and PE have a jumper TVS tube and capacitor, so that SURGE noise flows to the ground along a low-impedance path. SURGE noise interference path: SURGE generator→Zin of CDN→communication TVS1 tube→communication ground wire Zline→power ground wire Zline→to PE jumper capacitor C1 and TVS3 tube→PE, as shown in Figure 2 below.

    Figure 2 SURGE noise loop schematic diagram 4.2 SURGE noise spectrum analysis 4.2.1 SURGE waveform According to the IEC 61000-4-5 2019 standard, SURGE's typical open-circuit voltage waveform parameter is 1.2/50us.

    See Figure 3:

    Figure 3 SURGE generator open circuit voltage waveform

    4.2.2 Analysis of SURGE's spectrum characteristics

    The BW bandwidth of SURGE described in IEC 61000-4-5 refers to the frequency bandwidth when the falling edge slope of the frequency domain waveform starts to reach -60dB/decade. The SURGE voltage waveform has a short front time and contains a wide frequency band. The BW of SURGE reaches 2MHz, but the main energy is concentrated in the lower frequency band. Amplitude spectrum analysis shows that SURGE presents low frequency characteristics (50KHz→-40dB), see figure 4 shown.

    Figure 4 Voltage surge (1,2/50 ?s): spectral response with ?f = 3,333 kHz 4.3 TVS tube damage analysis

    4.3.1 TVS tube damage mode

    There are two TVS tube damage modes:

    (1) The nominal power of a TVS whose single-shot energy exceeds the rated power is a single pulse energy applied to the TVS in a very short time. When the applied noise waveform energy is greater than the rated power, it will cause the TVS tube to burn out due to overcurrent and present a short-circuit failure mode. .

    (2) The accumulated energy exceeds the upper limit. The noise waveforms applied in the actual test usually appear repeatedly, so that the accumulated energy in a short time exceeds the upper limit, and the TVS tube will be damaged, showing a short-circuit failure mode.

    4.3.2 SURGE current analysis in TVS tube

    (1) TVS tube parameter comparison TVS1 and TVS2 tube models are PESD1CAN, their Ipp is 3A, 20uf pulse power is 200W, 50us pulse width power is about 130W; TVS3 tube model is SD05C, its Ipp is 24A, 20uf pulse power It is 350W, and the pulse width power of 50us is about 210W. The power and flow capacity of TVS3 is better than TVS1. Refer to Table 1 for the TVS parameter table.

    Table 1 TVS specification parameter table

    (2) Noise loop parameters According to the path analysis in Figure 3 and the parameter table in Table 1, the loop impedance parameters can be calculated as shown in Table 2.

    Table 2 Loop impedance parameters

    Symbol parameter line impedance (calculated under TVS1 tube 3A) meaning

    Zin42Ω2Ω differential mode +40Ω common mode signal line injection impedance

    Zline3.1Ω Zline=10m*1uH/m*2π*50KHz=3.1Ω line impedance 1uH/m

    Zc132ΩZc1=1/(2π*50KHz*0.1uf=31.8Ω to ground jumper capacitor C1, noise frequency 50KHz

    ZTVS123.3Ω ZTVS1=70V/3A=23.3Ω, model PESD1CAN (NXP) CAN circuit protection TVS tube

    ZTVS32.7Ω ZTVS3=8V/3A=2.7Ω, model SD05C connects TVS to ground

    (3) Circuit SURGE current calculation: Calculate and analyze under the condition of TVS1 tube 3A impedance: Isurge=Vsurge/(Zin +ZTVS1// ZTVS1+2*Zline+Zc1//ZTVS3)=16A Calculate under TVS1 tube short-circuit failure condition: Isurge=Vsurge/(Zin +2*Zline+Zc1//ZTVS3)=19.7A The current range of SURGE noise in the loop is between 16A~19.7A, and the TVS tube current flowing through CAN-H and CAN-L is each Around 8A-9.85A, it exceeds the Ipp (3A) of TVS1, but is less than the Ipp (24A) of TVS3. From the calculation and analysis, it can be known that TVS1 bears excessive SURGE current, and there is a greater risk of damage.

    4.4 Experimental verification

    4.4.1 TVS tube impedance measurement

    Use a multimeter to measure the impedance of the TVS1 and TVS 3 tubes before and after the experiment, see Table 3. After many experiments, the impedance of the TVS1 tube became smaller and smaller, and eventually failed and short-circuited.

    Table 3 TVS tube impedance measurement

    TVS1TVS3

    Impedance value of test times Impedance value of test times

    Open circuit before experiment Open circuit before experiment

      SURGE experiment 65Ω SURGE experiment open circuit

    Secondary SURGE experiment 16Ω secondary SURGE experiment open circuit

    Three SURGE experiments 0.1Ω three SURGE experiments open circuit

    4.4.2 SURGE experimental waveform capture

    SURGE measured noise loop current peak value 19.2A, pulse width 55.3us, CAN-H and CAN-L TVS1 tube shunt 9.6A, which matches the theoretical calculation to verify the accuracy of the analysis. See Figure 5 for the measured waveform.

    Figure 5 Loop noise current

    4.5 Summary of the mechanism analysis of TVS tube damage

    (1) 15V reference ground P-GND and CAN reference ground A-GND are connected together, so that SURGE noise can flow to the ground through the P-GND to ground jumper;

    (2) The noise bandwidth of SURGE can reach 2MHz, but the general energy is concentrated in the low frequency band around 50KHz;

    (3) There are two types of TVS tube damage damage modes, short-circuit damage caused by over-rated or multiple energy superposition. This article is a typical RVS tube damaged by over-rated.

    (4) The theoretical calculation of the SURGE current of the noise loop corresponds to the actual measurement. Combined with the theoretical calculation, it can help the product design and selection of the protective device in the early theoretical model stage.

    5. From the product side solution analysis

    Based on the above analysis, to solve the TVS tube damage problem of the CAN communication port from the product side, it is to change the impedance distribution of the noise loop. There are three methods, see Table 4.

    Table 4 Analysis table of rectification design method

    Remarks on rectification design methods that can be implemented

    ◆Improve the flow resistance of TVS tube. Replace TVS1 tube with TVS3. The junction capacitance will increase by 17pf→350pf, which will affect the quality of communication signal.

    ◆Improve the impedance of P-GND and PE. ◆Remove the TVS tube to the ground. ◆Change the TVS tube into a 1MΩ resistance.

    ◆Line decoupling of noise loops, line grounding decoupling changes the impedance of communication, and there is a risk of common ground impedance

    6. Analyze the solution from the test end The different experimental methods of the SURGE test are shown in Table 5 below. Table 5 Different test methods

    The industry has different opinions on whether communication circuits such as CAN, 485, and 232 are balanced pairs, which leads to the use of different CDNs and different test results.

    (1) The definition of symmetry line: The balance line with the conversion loss of differential mode to common mode greater than 20DB is generally determined by the chip manufacturer.

    (2) CND asymmetric injection: A-GND is decoupled, the boss standard is 20mH, 50KHz impedance is 6.28KΩ, the loop SURGE current is about 0.3A, and the TVS tube works normally.

    (3) CND symmetry injection The standard does not specify whether A-GND is connected or not. However, in general, the A-GND experiment is connected by default, and only the symmetry line is injected. ◆A-GND not connected: The noise current of the SURGE loop is about 0.15A. ◆A-GND connection: the same as the initial test result without improvement.

    (4) Direct injection from the CAN port of the network. The AE device shunts the SURGE noise current through the network connection, but the current of each TVS tube is about 4.8A, which exceeds Ipp, and there is a risk of damage. To sum up the description of the test method, there are two solutions: ◆Do not connect A-GND, and carry out CDN injection; ◆Inject according to asymmetrical;

    7. Application analysis of different test methods

    In SURGE's non-shielded communication port test, companies have improved their own adaptability test methods based on IEC61000-4-5. There are three main ways, as shown in Table 6.

    Table 6 Test method of SURGE unshielded communication port

    EMC testing must be combined with the actual application scenarios of the product to customize suitable test methods to truly avoid product use risks on the design side. The application scenarios of the three test methods are as follows:

    ① Use the common mode impedance of the CDN to perform injection test of non-networked communication ports. It is mainly used in low-demand occasions, as long as the protective device is not damaged, such as secondary water supply.

    ②Using the common mode impedance of the CDN to carry out the communication port injection test of the network. It is mainly used for high requirements, considering the system's immunity to SURGE, rather than the single product itself, and requires that the communication cannot be wrong, such as production lines.

    ③ According to the standard recommendation, connect the CDN to the communication line for testing. Isolating EUT from AE is mainly for testing applications.

    8. Thinking and Enlightenment

    (1) The cause of TVS tube damage is that the impedance of the noise loop is too low, which makes the SURGE current too large, and the TVS tube exceeds the limit and is damaged. You can choose a TVS tube with high power and relatively small junction capacitance;

    (2) To increase the impedance to PE, you can remove the jumper TVS tube, reduce the jumper capacitance, or connect the jumper resistance in series to increase the impedance value, so that the SURGE voltage is mostly added to the jumper impedance;

    (3) Product design should conduct noise path analysis and SURGE current estimation to guide impedance distribution and device selection. Need to combine the actual application scenarios of the product to choose the test method. Don't copy the standard requirements completely, and lack systematic analysis.