Questions about the actual overcurrent capability of NMOS and its impact current resistance after conduction

liyooduan

Questions about the actual overcurrent capability of NMOS and its impact current resistance after conduction [Copy link]

 

Dear Taoist friends: I have been working on digital circuits before, and now I have started to learn some analog knowledge. Recently, I have been troubled by the problem of MOS. Please give me some advice:

Take this NMOS from Vishay as an example:

(1) What is its actual current capacity? Many chips have large current ratings, for example, this chip ID (on) says 120A, and there is a provision in the notes that it is a pulse of <300us, so what is the actual current?

After actual experiments, we feel that this is related to the Soa curve. For example, if the chip is VGS>12V and the load current is 15A, it can work for a long time without any problems (the chip surface is also at room temperature). If the load current is slowly adjusted to 18A, the tin will melt instantly. This seems to be related to the intersection of the Limited by Rds(ON) curve in the SOA and the vertical axis. How to read this continuous current?

(2) Confusion about the Limited by Rds(ON) curve: I understand that the vds on the horizontal axis is the voltage after the MOS is turned on. Doesn't the larger the Vds, the greater the power consumed by Rds(on)? Why is the current greater? According to my understanding, the smaller the Vds, the greater the current should be. I am really confused here.

(3) Take this MOS as an example. We use it as a switch. When it is fully turned on, if there is a 30A surge current, the MOS will be damaged. But if the surge current is below 17A, there will be no problem. So are the 100us and 1ms surge pulse curves in the SOA only for the MOS that is being turned on, and not suitable for the MOS that is fully turned on? After it is fully turned on, the possible surge current pulse must be controlled below the VDS (voltage difference when fully turned on) corresponding to the Limited by Rds(ON) diagonal line? How do switching power supplies with so much current do this?

I think this may also be the confusion of many Taoist friends who have just designed this MOS. Please help me solve this problem. Thank you!

h472438470

Take this MOS as an example. We use it as a switch. When it is fully turned on, if there is a 30A surge current behind it, the MOS will be damaged. However, if this surge current

18870061684

Ha ha ha ha

18870061684

Take this MOS as an example. We use it as a switch. When it is fully turned on, if there is a 30A surge current behind it, the MOS will be damaged. However, if this surge current is

maychang

"For example, the ID (on) of this chip says 120A, and there is a provision in the notes that it is a pulse of <300us, so what is the actual current?"

The actual current is determined by external conditions such as your power supply voltage, load properties, etc. It may be 1A, it may be 2A... No one knows.

The current pulse 120A is marked as less than 300us. The 300us is the test condition, and 120A refers to the maximum current that the tube can withstand when the test condition is met.

After the current pulse passes, it takes a long time before the tube can withstand the next current shock. If the pulse current lasts longer than 300us, the tube's ability to withstand the current pulse will be further reduced.

maychang

"When VGS>12V, if the load current is 15A, there is no problem with long-term operation (the chip surface is also at room temperature). If the load current is slowly adjusted to 18A, the tin will melt instantly."

It is very unlikely that the solder will melt "instantly". I am afraid your measurements are wrong.

qwqwqw2088

The test conditions are marked on the manual.

In actual use, the junction temperature must be considered, so the maximum current value is equal to half of the maximum rated current of the MOS tube. When actually selected, it is one-third, and no more than half is safe. Sometimes there are also factors such as withstand voltage and junction capacitance.

maychang

"(3) Take this MOS as an example. We use it as a switch..."

Sorry, I don't understand this paragraph at all.

chunyang

The original poster's thinking angle is wrong, so the description seems confusing. In fact, the key to the problem is not difficult to understand, Ohm's law is enough. The key to the current carrying capacity of any MOS tube, including bipolar transistors, is temperature, because the tube still has resistance after conduction, and resistance will generate heat, and the heat resistance of the material is limited, and it will burn out if it exceeds the limit. It takes time for the temperature to rise. Under given environmental conditions, the greater the current, the shorter the time it takes for the temperature to rise to the limit tolerance temperature. In the example of the original poster's post, at 120A, the temperature rise reaches the limit after 300uS, which is the origin of this parameter. Then the definition of the maximum normal working current can be known, that is, at this current, the temperature rise of the tube will not exceed the limit tolerance temperature, and it can work for a long time.

As for the relationship between VDS and current, of course, the two are positively correlated. The on-resistance of the tube is determined, and according to Ohm's law, the greater the current flowing through, the greater the voltage difference between the two ends.

Regarding the last question in the original post, if you understand what I said before, you won't be confused. It must be all about temperature rise, so think in terms of Ohm's law.

呜呼哀哉

This safe operating area curve is indeed used to determine the current size under different working conditions. Regarding your second question, why does the current increase with the increase of Vds? In fact, this section is the DC working area, and the slash part is Vds/Ron.

liyooduan

maychang posted on 2021-9-12 18:38 "For example, the ID (on) of this film says 120A, and there is a provision in the notes that it is a pulse of <300us, so what is the actual current? 』 The actual current...

What I want to ask is: Under normal temperature conditions, within the maximum voltage range allowed by the voltage MOS, if there is no heat sink for the MOS, and the VGS exceeds the maximum value of VGS (th) in the chip manual, based on your own experience, what is the maximum current that such a chip can pass (I know there are many situations that determine this current, I just want to know the most common situation); rather than what you said, this depends on whether I use 1A or 2A. Because at the beginning of the design, it is generally necessary to preliminarily evaluate whether the MOS selected is suitable.

maychang

liyooduan posted on 2021-9-13 15:32 What I want to ask is: Under normal temperature conditions, within the maximum voltage range allowed by the voltage MOS, if no heat sink is added to the MOS, the VGS exceeds the VGS in the chip manual...

You should take a good look at chunyang’s reply on the 9th floor.

maychang

liyooduan posted on 2021-9-13 15:32 What I want to ask is: Under normal temperature conditions, within the maximum voltage range allowed by the voltage MOS, if no heat sink is added to the MOS, the VGS exceeds the VGS in the chip manual...

On the first floor, you asked, "So what is the actual current?" The actual current and the current allowed to pass are two different things.

On the 11th floor, what you are asking is "under normal temperature conditions, within the maximum voltage range allowed by the MOS, if no heat sink is added to the MOS, and the VGS exceeds the maximum value of VGS (th) in the chip manual... generally, what is the maximum current that such a chip can pass?"

You need to distinguish between single current pulses, multiple current pulses, and continuous current.

liyooduan

qwqwqw2088 posted on 2021-9-12 18:47 The test conditions are marked in the manual. Sometimes the junction temperature must be considered in actual use, so the maximum current value is equal to half of the nominal maximum of the MOS tube. The actual...

Thank you for your reply, you answered the first question I asked:

We also tested this 6-pin NMOS from ST; it was used in a slow-start circuit with TI's TPS2491, and the slow-start power-on was started normally when the electronic load was set to 20A.

As the test progresses, the setting current of the electronic load remains unchanged, while the current on the input line of the MOS measured by the current clamp will gradually increase. This should be the reason why the Rds(on) gradually increases after the MOS slowly works.

After testing for about 90 minutes, the current clamp showed that the current increased to 22A. In this case, the MOS instantly melted the tin on the pad and burned. Before this happened, the MOS was not too hot to touch.

This situation is stable, and we reproduced it three times using this method.

So, in this case, is 20A already the limiting current at room temperature without a heat sink?

I hope this can help clarify the doubts.

liyooduan

chunyang posted on 2021-9-12 20:00 The author's thinking angle is wrong, so the description seems confusing. In fact, the key to the problem is not difficult to understand, Ohm's law is enough. Any MOS tube also includes...

Thank you for your reply. After your explanation, I realized that I was thinking in the wrong direction. When MOS is fully turned on, RDS(On) is a constant value (of course it will increase with the temperature of the device itself, so let's do a qualitative analysis here). As the current increases, VDS=RDS(ON)*I will naturally increase. However, the larger the VDS, the greater the power consumption of MOS.

maychang

liyooduan posted on 2021-9-13 15:32 What I want to ask is: Under normal temperature conditions, within the maximum voltage range allowed by the voltage MOS, if no heat sink is added to the MOS, the VGS exceeds the VGS in the chip manual...

For a single current pulse, the heat dissipation conditions have been given on the 11th floor: ambient temperature, no heat sink, and the MOS tube is fully turned on (the manual says the gate voltage is 10V). If a large current passes through the MOS tube in a very short time (300us in the manual), the heat generated by the current cannot be dissipated into the air or the heat sink in such a short time, and even the surface of the tube shell cannot have a significant temperature rise. The heat generated by the current will only increase the temperature of the tube core (this is called a near-adiabatic process in physics). This is how the 120A/300us marked in the manual comes from.

maychang

liyooduan posted on 2021-9-13 15:32 What I want to ask is: Under normal temperature conditions, within the maximum voltage range allowed by the voltage MOS, if no heat sink is added to the MOS, the VGS exceeds the VGS in the chip manual...

The above is the case of a single current pulse. If it is a continuous current, the heat generated by the current must be balanced with the heat dissipated into the air, that is, the heat generated by the current over a period of time is equal to the heat dissipated into the environment. The heat flows from the die to the environment, relying on the thermal conductivity from the die to the heat sink and then to the environment. The stronger the thermal conductivity (for example, the larger the heat sink), the more heat can be dissipated for the same difference between the die temperature and the ambient temperature, and the greater the continuous current allowed to pass (the greater the power allowed to generate heat from the die).

maychang

liyooduan posted on 2021-9-13 15:32 What I want to ask is: Under normal temperature conditions, within the maximum voltage range allowed by the voltage MOS, if no heat sink is added to the MOS, the VGS exceeds the VGS in the chip manual...

The heat dissipation conditions you gave on the 11th floor are that there is no heat sink and the tube is fully turned on. I don't know what package the tube you use is. If it is a TO220 package, the general heat dissipation power can only be 1W. If it is a TO247 package (larger than the T220 package), it is estimated that it can be 2W (this is what you said on the 11th floor "rely on your own experience"). According to the heat dissipation power and the MOS tube on-resistance (the second row from the bottom to the top of the table on the 1st floor), the maximum continuous current can be calculated.

liyooduan

maychang posted on 2021-9-12 18:47 "(3) Take this MOS as an example, we use it as a switch..." Sorry, I don't understand this paragraph at all.

For the third question, as shown in the figure above, a battery is used to supply power to a high-power regulated current. When the battery is turned on, the battery charges the power input capacitor, and the surge current is too large, causing the battery to enter the overcurrent protection state.

In order to solve this problem, a slow start circuit is added in the middle. After the slow start, the current can be limited and the MOS is fully turned on normally. When the power switch is turned on, the power supply works. If the capacitive load behind is large

It will cause a surge current at the power input end, which will instantly damage the MOS. As shown in the SOA in the figure below, when we reduce the capacitive load at the back end, turn on the power switch, and control the surge current to about 20A, the MOS will not have any problems.

So the question is: when the MOS is fully turned on, there will be a surge current again. If the surge current is 80A for 1ms,

So at this time, should the VDS of the SOA horizontal axis be the value of 80A*RDS(on)? Or should it be the previous stable conduction current*RDS(on)? I personally think it is the former.

liyooduan

maychang posted on 2021-9-13 16:16 On the 11th floor, the heat dissipation conditions you gave are that there is no heat sink and the tube is fully conductive. I don’t know what package the tube you use is. If it is a TO220 package, one...

Thank you. This package power consumption is really useful. I just want to know this data. I have never worked with this MOS before. I really have a lot to learn from you.