Diode Reverse Recovery

Hello, I'm Sakov. This presentation is titled Diode Reverse Recovery. Let me start again with a description of a PN

diode

, say a silicon PN

diode

. We have two sections here, the N-type part and the p-type part, and in the middle we have the junction when there is a current in the forward direction. This is the direct address. There will be a charge stored in this junction area. Well, as long as there is current, there is a charge here that helps transfer charge carriers from one side to the other. If I change the direction of the current abruptly I will force the current in the

reverse

direction, okay as long as there is a charge at the junction there will still be current flowing i.e. you can have a

reverse

current in a diode as long as it has charge carriers. charge, however once the current is reversed, these charge carriers are actually swept out of the junction, you will get a layer of deposition and then the current will stop this process of reverse current and then

recovery

. how this current actually goes back to zero, it is indeed a reverse

recovery

phenomenon of a diode which we are going to talk about now, where do we find this situation?

Well, let's take for example a boost pwm converter, we have an inductor, a diode, a transistor. and an output section and in the off state of the transistor i.e. when the transistor is non-conductive, here there will be a current flowing through the inductor and the diode, this is a direct current and of course it will go to the section exit. now once the transistor is in the on state, there is a short circuit here. Here the inductor will be shown, but at the same time, as you will see here, a reverse voltage is imposed on this circuit and this loop now has a negative voltage is imposed on the DI so we can see here more clearly, the transistor is shorted , we have a positive voltage here and now there is a current that will go in this direction first, of course the forward current will die out, it will decrease and then you will have a short period of time where you will have a reverse current.

This is the reverse current that you will find in this boost converter as soon as the transistor turns on. When this is short here you will have this reverse current now this phenomenon is not limited to the Boost converter in fact any pwm converter will have this problem because any pwm converter is based on this generic cell which has this inductor diode and a switch in the case of a boost this is connected to ground here and this is the input and this is the output in the case of a rear converter this is the input this is the output and this is connected to ground so any pwm topology will have this situation of reverse current, in fact you will also have in an isolated topology like a full bridge, let's say a pwm converter and I am showing here a half bridge rectifier in this case, if let's say the top diode is conducting, this is in the forward direction, but then as this voltage here is reversed, we will have a reverse voltage in this loop here and consequently the forward current will decrease and eventually you will have a reverse current, so this phenomenon is very widespread and you will find it in many situations now in the characterization of the The reverse recovery phenomenon is usually given in this graph, okay, we are talking about the situation where you have a reverse voltage on this slope which is shown here more clearly and what we see in this graph which you have here This axis is the current and this is the moment we see here. the forward current this is the current that was flowing through the diode just before the short circuit occurred here is the reverse voltage imposed and now the current is starting to decrease with a certain ID now it is slope here the ID has nothing to do with the diode is a function of the voltage and inductances in this loop.

Now, whatever you do, there will always be some permanent inductance, so there will always be a slope here where this current decreases, the current decreases, approaches zero and then actually goes in the reverse direction, this is a current negative will occasionally reach a maximum value. The maximum value of reverse current and will begin to disappear. This is reverse recovery. Okay, associated with this, we have a certain amount of charge, of course, and as this current is dying at 10%, uh up to this point, this is the definition just a formal definition of the reverse recovery time, so the Reverse recovery time is defined from the time the current reverses to 10% of the maximum reverse value. is the inverse time we have q which is another parameter so here are the parameters we have the inverse recovery time we have the load and the IDT is very important if for example the inductances will be smaller this slope will be more pronounced and, in fact, the current will reach a much higher value.

This is the peak current will be much higher if the IDT is faster, so this graph depends on the operating point and the condition of the circuit that was used to measure it, so a different direct current will start. with you will have a different value here here or here and for different inductances you will have a different IDT and the suppliers must provide and they usually report what happens with this time and the selection value when the IDT is actually changing now another sensitivity which is very important realize is to the temperature as the temperature increases the reverse recovery time increases and the maximum value also increases so it is very important to realize this because the diode will heat up and the temperature will increase so you will have have to worry about what happens at high temperatures.

Now another point to note is that associated with this phenomenon, of course, there is a loss of energy because there are charges that are shed here in this circuit and of course there is an average current. The average current is the frequency of the charging time and multiplies the voltage that you have, so this is the amount of energy that is actually associated with this phenomenon and, for example, if you have 400 volts, 25 n two points of Q, which is pretty good for a 10 amp diode. and let's say 100 khz you have 10 watts so it's not zero it's a certain value that you have to keep in mind now where does this energy go well?

Not to the transistor because the transistor is shorted and usually the voltage across it is much less than full. voltage, so the amount of power dissipated here is small. The same goes for the DI, the diode has the same voltage drop when, in the forward current situation or in the reverse current situation, it is approximately the same voltage and is, for example, about a volt, so the associated power. here with this di is small, so where does this energy go well? As we'll see in a few minutes, there is a resonance circuit here and there are some losses associated with the parasitic resistance of this resonance circuit, so this energy is actually lost and of course, heating up and this is something you have to deal with. worry and also, as we will see, there is a problem of Emi, electromagnetic interference caused by this oscillatory type of phenomenon.

Another point to note is that a power Moser transistor does not have a diode in the body, it is built in, it is not something that is connected to it, it is just part of the construction and unfortunately this diode is a slow diode i.e. , it has a long reverse recovery time, so there will be a lot of reverse current, the reverse peak. the current will be high and the amount of load will be large, so this can cause some problems in various situations and I am bringing here just one example of a synchronous post converter.

Now, in a normal post converter, you have a transistor, you have a diode. when the transistor is off you have a current flowing through this diode now for low voltage applications, in case the output is low voltage say 5 volts the voltage drop across the diode will cause an appreciable amount of power that will be dissipated and improve efficiency. uh it will go down so one way to overcome this would be to put a transistor here so that q1 is on or Q2 is on and when Q2 is on current flows through the transistor and if the on RDS is small the voltage drop will be be small and therefore the energy dissipation will be small;

However, once Q2 is turned off, and of course q1 cannot be turned on at the same time, you have to wait a certain DeLay. Then the two will not conduct together, then you will have a short then Q2 turns off then the current will find its way through the diode and then after some delay when q1 turns on, here you have a reverse voltage imposed on the diet , so here you have the reverse recovery problem, so we find that in many applications, this body diode causes a lot of problems to remedy this, you can put a fast diode with a short recovery time in parallel to this body diode;

However, for this diode to really help, the voltage drop must be less than the voltage drop. Unfortunately, in the body diode, the voltage drop across the body diode is quite small, in some cases it is about 600 to 700 million volts, so if you just put in a fast silicon diode that normally has a drop voltage of, say, a volt or more, will not help you. can I put a shy diode a silicon shy diode uh because the voltage drop is very small on a shy diode we'll talk about it a little bit later and this will really help the situation because then the current will prefer to go through the shy diode instead of the body diode, even if it's shy Di uh, silicon shk di are available up to 100 volts, so a silicon shk di can't be used.

Well, we will talk about this a little later, now there is another issue that needs to be taken into account. This way recovery occurs, we can have a soft recovery and we can have an abrupt recovery or a fast recovery or a hard recovery and the difference is that when the ID here is very fast, a high voltage will be imposed on the leakage inductances. so you would rather have a soft recovery and here what happens due to the IDT of the recovery what we see here again is this reverse current which is now dying out, this is the voltage drop across the DI, this is when it is in the forward direction This is in the reverse direction and as I said the voltage drop is about the same here the current drops to zero the IDT is causing this high voltage to be generated and now if the diode is not conducting you actually have here a resonance. circuit, this is the capacitance of the diode when it is in the off state and this resonance circuit of course causes this oscillation, so you have two problems, one is the very high reverse voltage on the diode, which can actually cause a fault, so you need to protect the diode, we'll talk about that in a minute and then you have these illations that are causing electromagnetic interference Emi, so this is a real problem that one has to deal with now to stop these oscillations, what one can do and uh hold the reverse voltage high, you can use an RC snubber, this is a dissipative snubber and the idea here is that if you have a reverse current and then if the current drops to Z zero, you have sort of shunt so that the current will flow and so that there is no interruption of the current, it will continue to flow in this direction, okay, now there are some power values ​​uh.

Associated here, this is the power or the energy stored in the capacitor multiplied by the frequency is the power and this is the maximum power that you have. stored in the stray inductances, this is the maximum reverse current and this is the leakage, so this is the amount of energy associated with the leakage inductances um. Now, if you compare them, you can get an approximation of the capacitance you need to put in here. to absorb this energy, this is usually a first order approximation; in fact, you can also get a value for r. r has to be small enough so that when the DI is conducting, this circuit is discharged and ready for the next cycle now as I'.

I've said this is just a first step because you don't really know exactly what the leakage is, you don't know exactly what the reverse PE current is, so you start with this and then unfortunately you have to go through some testing and error process to optimize the circuit and don't forget that you have to keep in mind that at high temperature the situation is worse, you have higher maximum current and therefore you have to prepare for this situation. Another problem is the maximum current. flowing through the DI which could damage the transistor if it is too high, so to buffer or lower it you can put an inductor that will speed up or slow down the IDT and therefore the maximum reverse current will increase. be less, however, there is now energy stored in this inductor, so it is actually a type of flyback arrangement where this energy is then pumped back to the output or the input, depending on the topology, unfortunately, There are some leakage problems between the primary. and secondary leakage inductances, which on their own are causing some problems, this solution is bulky and really notsuitable for high frequency maybe up to 100 KZ or maybe 50 Kilz above that you lose more than you actually gain now the best way to combat the problem of a reverse recovery is actually to use soft switching and I am showing here a resonant converter, this is lnc.

It is a resonant part of a network, it actually represents the load. There may be a transformer, a rectifier and this is reflecting. everything to the primary and representing the load by an equivalent R AC, as it is called, only one resistor to represent the load. Well, we have a complete bridge here and I'm showing here the case where q1 and Q4 conduct each other. Well, we have power. flowing in this direction and now we are walking at a frequency that is higher than the resonant frequency, therefore the current is lagging. I can see it here because since Q4 says Q4 is off, there's still power, there's still power here, okay?

I am turning off Q4 and q1 once I do that the currents of course will continue because of this resonance circuit so we will have a situation now where there will be a current through this diode with reverse current and this transistor Q4 turns off. this voltage here acts and it will actually increase until this diode conducts and here I'm showing this for a brief period where it will have a reverse current and then the current will find its way, uh, from here back and through this diode, so that we have a reverse current. recovery, but it's not really causing any problems because it's attracted to this resonance circuit and after a while it will die out and that's it and the voltage will drop due to the fact that there is still current, uh, charging or discharging this parasite or aggregate .

The capacitor and the diode will start to conduct, so this diode will conduct and this diode will conduct. Now we can convert Q3 and Q2 under zero voltage conditions, which of course is very good, so this soft switching scheme will solve the problem of reverse recovery and will also provide a situation where the transistor turns on with a zero voltage switching, which of course is very good. Another approach to combat the reverse recovery problem would be to use an inductor wound on an amorphous or saturable reactor or core. Now this amorphous magnetic material will have a hysteresis of this nature.

We have here a stip uh permeability, a high permeability section, so the inductance here is high and then we have a saturation, so the inductance is very low, so if we start with this point where. saturation occurs with low inductance, we go to lower current which is lower magnetic field and eventually we will again reach High inductance and this is the phenomenon that can be used to reduce the maximum reverse current and this is how we can do it We put this reactor in series with the diode, so that when this diode is conducting, we have a saturation of this element and there is no effect on the circuit.

Now when the voltage is reversed the current will start to die and then we will reach this high part of the inductance so this will reduce the maximum reverse current by reducing the rate at which the current changes and thus will alleviate the problem, so here is a simulation of the situation that we see, the direct current that we see. how the current reverse current is slowly building up and then it will recover and the diode voltage is like buffered and we don't have these high frequency oscillations, finally let me talk about the shy diode, now a short diode is a DI between a metal and a semiconductor, a timid di, has no reverse recovery because there is no charge stored in a junction, however it has junction capacitance and therefore there will be some reverse current, but this is not a reverse recovery phenomenon. it's just a current because of this capacitance, now the Silicon Shad can be manufactured to say 100 volts a little more, but a new generation of devices is the silicon carbide short key, this should be a short key.

I'm sorry, and they can be. made at high voltage of 000 volts and even higher so here is the comparison uh presented by St of different diodes this is the IDT of the circuit this is of course current and this is the moment now the blue is a simple di a basic di sees a long reverse recovery time. It has a high maximum reverse current. The red one is an ultrafast diode made by St and many other suppliers that has a much shorter recovery time and much lower maximum reverse current. The black is the Silicon Carbon Diode and as you can see there is some movement here but this is not a reverse recovery this is just due to the current through this capacitance and of course we see almost no current reverse, in this case.

You can use this car silicon diode to solve the body diode problem I mentioned above. Here is the circuit you can use. You need a forward diode here which is the same direction that the transistor is in contact and then I need this silicon carbide diode in parallel now the idea here is that when this transistor is conducting it has a current going in this direction now when there is a reverse current it doesn't go through the diode because the body diode because there is a diode Here in the reverse direction it will go through this silicon carbide.

Now obviously there is some power dissipation, additional power dissipation associated here, but that's not so bad if the power level is high, it's a high voltage and the drop across this diode maybe not. so high compared to the voltages in the circuit. This brings me to the end of this presentation. Thank you very much for your attention. I hope you find it interesting and that it will be useful to you in the future. Thank you.