Article source: Notes from an Electronics Engineer
1. Summary
Magnetic beads are mainly composed of ferrite and coils, and the main principle of magnetic beads to suppress interference is to use high frequency to generate heat through resistance to consume interference. If the interference is strong for a long time, the beads may overheat and burn.
2. Problem description
Our products are used in industrial sites, the products are shipped about 10,000 units, after two months of operation, about 10 damaged equipment are returned from the customer, after R & D analysis, these damaged products are damaged in the same place, as shown in Figure 1 of the magnetic beads L1, L3, the appearance of the magnetic beads has obvious burn marks, but the pre-stage insurance, TV, and the power chip of the rear stage are not damaged.
Note: The power consumption of the product is about 24V, 0.3A, and the magnetic bead is 1200Ω/1A/L1206. There is no problem in derating the specifications of the selection.
Figure 1: A device with a damaged device
3. Cause analysis
Since the insurance, TVS, and power chip are not damaged, the situation of overcurrent can be basically ruled out, combined with the customer's on-site application, the customer uses DC24V power supply, and more than 50 AC contactors are mounted on DC24V at the same time for process control, and the action frequency of the contactor is about 1 time/s. After oscilloscope testing by field engineers, very high surge interference was captured in the field. AT THE POWER PORT, IT CAN TEST THE PULSE INTERFERENCE OF UP TO DC 57V AND ABOUT 60MHZ. When the contactor is disconnected, the interference disappears, indicating that there is interference with the contactor.
Figure 2: Waveform captured using the oscilloscope afterglow function
Let's first understand the internal structure of magnetic beads, which are composed of coils, ferrite cores, and external plating and packaging, as shown in Figure 3 below, magnetic beads are mainly composed of coil-wrapped multi-layer ferrite.
Figure 3: Magnetic bead diagram (left), physical diagram (middle), equivalent circuit diagram (right)
According to Figure 3, we can deduce the impedance calculation formula of the beads:
Expanding the formula yields:
Expanding the formula yields:
Take Murata's magnetic beads:MPZ1608B471ATA00 as an example, the parameters are known from the pdf document, R1 = 470 Ω, L1 = 8.6 uH, C1 = 0.2583 pF, R2 = 0.110 Ω. Comparing the two parameters, the original diagram of the specification is the same as the general trend drawn by Matlab, and the resonant frequency is the same, but the overall shape is still quite different. So why is this happening? The model of this bead is called the simple model, since it is simple, there are more complex, complex I didn't find a specific circuit model, but TDK gave a SPICE NETLIST, we can see some differences. I downloaded the SPICE NETLIST of the simple model and the complex model separately, and opened the files separately with txt. As you can see, there are only C1, L1, R1, and R2 in the simple model. Complex models are much more complex, with two for C, seven for L, and nine for R.
As you can imagine, complex models have more parasitic parameters and are more realistic devices. The curves in the datasheet should be derived from complex models.
Therefore, in the actual use process, we can directly use the frequency impedance curve provided by the manufacturer.
Figure 4: Comparison of specifications and MATLAB calculations
Figure 5: Simple and complex models of magnetic beads
As you can imagine, complex models have more parasitic parameters and are more realistic devices. The curves in the datasheet should be derived from complex models. Therefore, in the actual use process, we can directly use the frequency impedance curve provided by the manufacturer.
The reason analysis is so far, the reader may already know the answer, that is, in the case of long-term strong interference, the magnetic beads will always be in a state of energy consumption, and the high-frequency interference will be converted into heat energy consumption consistently, if the product is in a high temperature scene, the temperature will be superimposed, when the long-term heat is about the heat dissipation, the magnetic beads will continue to rise, and the final consequence is that the magnetic beads in Figure 1 are burned.
Fourth, the solution
The designer adds magnetic beads to the power supply port, the main purpose is to filter the high-frequency interference in the high-frequency tens of Mhz ~ hundreds of Mhz, and at the same time consider the effect of interference suppression, we can use the filter mode of varistor + common mode inductance + capacitance + TVS, as shown in Figure 6, the varistor is placed at the front end, mainly considering that the varistor has a large flow capacity, and it is easy to do hundreds of A, but the response time of the varistor can reach tens of nS, and the interference above the nS level is still powerless. The response time of the TVS can reach the nS level, but the flow capacity of the TVS is smaller than that of the MOV, so it is necessary to increase the common-mode inductance between the MOV and the TVS, and the common-mode inductance and the capacitors at the front and back end can form a decoupling circuit, which can flatten the higher spike pulses and reduce the pressure on the TVS. The response time of TVS is nS, which can theoretically cope with the interference frequency of 1Ghz, and the power chip that needs to be protected in the later stage also has a decoupling capacitor, which rarely occurs in high-frequency interference in actual conduction applications.
Figure 6: Interference can be greatly reduced with protection and decoupling.
Figure 7: Complete DC port input protection scheme
The first stage is the filtering scheme, which is composed of Figure B, in which C5 and C7 are composed of C1812/1nF/2kV ceramic capacitors, and the main reason for choosing such a large package is that when doing the surge test, because the capacitor is not an ideal capacitor, there is ESR inside, and the capacitor will heat up under high voltage, so it requires a larger package capacitor to improve the flow capacity, C6 is a differential mode filter, and the main flow path is varistor R1 and the electrolytic capacitor of the later stage, so the package C0805 can withstand voltage 100V.
The second level is the protection scheme, which is composed of Figure A, the flow capacity of the self-recovery fuse and the varistor should be comparable, otherwise, when the fuse is much larger than the pressure sensitive flow capacity, it is possible that the pressure sensitive has begun to heat up seriously, and there may be a short circuit and fire, but the fuse has not been disconnected, which will lead to a fire accident. When the pressure sensitive is much larger than the fuse flow capacity, the normal interference pulse may cause the fuse to be disconnected and the circuit cannot work. The main purpose of replacing TVS in this position is that the flow capacity of the pressure sensitive is many times larger than that of TVS for devices of the same package size, and the cost performance is very outstanding.
The third stage is the filtering scheme, as shown in Figure C, which is mainly composed of capacitance and common mode inductance, because the response time of the varistor is slow (relative to TVS), it is at the us level, while the response time of the TVS at the back stage is at the level of nS, and the main disadvantage of the front-end varistor is that the protection is not accurate, there is residual voltage, so the common mode inductor damping effect is required, the high-frequency spike pulse is flattened, and then discharged to the negative electrode through the bleeding channel of the electrolytic capacitor and the ceramic capacitor.
The fourth level is the protection scheme, as shown in Figure D, due to the slow response time of the front-end varistor and the residual voltage, it is necessary to increase the TVS for the final protection to prevent over-frequency pulses from entering the post-stage circuit, where the TVS can use 600W.
The fifth stage is the filtering scheme, as shown in Figure E, the filtering is placed in the TVS post-stage, which provides the final filtering for the power chip of the post-stage and has been used for energy storage.
V. Summary
Although many engineers are convinced that the supplier swears that the beads can pass through XXA, it is easy to ignore the structural defects of the beads and use a lump of ferrite to wrap a smaller coil, and it is an energy-consuming device, which can easily lead to heating and burning if the heat is much higher than the heat dissipation in a high-temperature environment. If the reader is still confident, it is recommended to unravel a magnetic bead to see if the size of the coil inside can cope with the current you need.
Hardware notebook, learn circuit design, PCB design, simulation, debugging and EMC knowledge together
For more dry articles, please click to follow:
If you understand the components, you will understand half of the circuit
Let's learn together:
Hardware notebook|add group
Disclaimer: This account remains neutral to the statements and opinions of all original and reprinted articles, and the push articles are only for readers to learn and communicate. The copyright of articles, pictures and other copyrights belong to the original author, if there is any infringement, contact to delete.