Share a problem of MOS tubes being frequently damaged.

sfcsdc

Share a problem of MOS tubes being frequently damaged. [Copy link]

 

In a 24V system, I use PMOS as a switch, and occasionally the PMOS will be damaged.

I originally thought it was an accident, but later found out it was a design problem.

This is the circuit diagram of the problem: Q1 sometimes fails

This is the parameter of the MOS tube used:

When the PMOS is turned on , the voltage between GS is greater than the maximum value of GS , which makes the MOS tube easy to be damaged.

Then change the circuit to this

In this way, when the MOS tube is turned on and off, the voltage between VGS will not exceed the rated maximum value.

Since MOS tubes are voltage-controlled devices, when in use, in order to build up the voltage as quickly as possible, the gate drive resistance is often reduced or directly removed.

However, it is also necessary to pay attention to the circuit environment in which it is used. Each indicator cannot exceed the rated maximum value during the opening and closing process.

qwqwqw2088

Thanks for sharing your experience

I want to ask, did you add a 20K G-pole drive resistor? I feel this resistor is too large.

Also, I want to ask, before adding the resistor, when "Q1 sometimes fails", is 358 normal?

maychang

The power supply voltage is not specified in the text or in the figure. Without knowing the power supply voltage, it is hard to say whether the newly added resistor should be 20 kilo-ohms. It is possible that the newly added resistor is too large, which will prevent the P-channel MOS tube from entering the variable resistance area.

topwon

maychang posted on 2020-7-29 10:39 The power supply voltage is not specified in the text and the figure. Without knowing the power supply voltage, it is hard to say whether the newly added resistor should be 20 kilo-ohms. It is possible that the newly added resistor...

In a 24V system, I use PMOS as a switch, and occasionally the PMOS will be damaged.

-----It should be the case of 24V power supply. He improved it by using 10K and 20K voltage divider to make Vgs(TH)<Vgs=8V<12V.

maychang

topwon posted on 2020-7-29 13:33 In a 24V system, I use PMOS as a switch, and occasionally the PMOS will be damaged. -----It should be the case of 24V power supply. He improved it and used 10K and 20K...

According to the picture posted by the OP, the maximum allowable value of V GSS of the MOS tube used by the OP is 12V.

Such parameters are indeed rare, and the maximum allowable value of V GSS for most power MOS tubes is 20V.

If the power supply voltage used by the OP is 15V, there may be a problem that the MOS tube cannot be fully turned on.

led2015

The same thing happened in the power supply I designed before. I was most afraid that the chip would explode first, and the MOS would explode second.

topwon

maychang posted on 2020-7-29 14:01 The MOS tube used by the OP, according to the picture posted by the OP himself, the maximum allowable value of VGSS is 12V. Such parameters are indeed rare, and most power MOS ...

If Vgs=-5V, it should be enough to make 3401 conduct.

sfcsdc

qwqwqw2088 posted on 2020-7-29 08:55 Thank you for sharing your experience. I would like to ask, did you add a 20K G-pole drive resistor? I feel that this resistor is too large. I would like to ask, before adding the resistor, &l ...

Because this MOS does not switch frequently and is only used for logic control, it does not require a very high speed.

I used a larger resistor.

358 has never broken down.

sfcsdc

maychang posted on 2020-7-29 10:39 The power supply voltage is not specified in the text and the figure. Without knowing the power supply voltage, it is hard to say whether the newly added resistor should be 20 kilo-ohms. It is possible that the newly added resistor...

The voltage is 24V.

The Vth of this PMOS is only 1V

maychang

topwon posted on 2020-7-30 09:08 If Vgs=-5V, it should be enough to make 3401 conduct.

I didn't check the UT3401 manual.

Judging from the curve you posted, the tube voltage drop at -4.5V is at least one-third larger than that at -10V (55 milliohms vs. 42 milliohms), which will obviously increase MOS losses.

topwon

maychang posted on 2020-7-30 09:21 I didn't check the UT3401 manual. From the curve you posted, the tube voltage drop at -4.5V is at least one-third larger than that at -10V (55 milliohms vs. 42 milliohms...

Judging from the 1N4007 connected in series in the previous stage, the overcurrent capacity of this circuit should not exceed 1A, right? The difference in the on-resistance of 10 milliohms should be very small. . . . . .

maychang

sfcsdc posted on 2020-7-30 09:21 The voltage is 24V. The Vth of this PMOS is only 1V

The V GS(th) test condition is a drain current of 250 microamperes. The current is only 250 microamperes, which shows that the resistance of the MOS tube is quite large. With such a large resistance, it is difficult to say that it plays the role of "opening" the switch.

maychang

sfcsdc posted on 2020-7-30 09:21 The voltage is 24V. The Vth of this PMOS is only 1V

The test condition for R DS(on) is usually V GS of 10V. Of course, I didn't check the manual for this tube, so it may be slightly lower than 10V, such as 8V. You should have the manual for this tube in your hand to check whether it is true.

maychang

topwon posted on 2020-7-30 09:36 Based on the previous stage connected in series with 1N4007, the overcurrent capacity of this circuit should not exceed 1A, right? The difference in the on-resistance of this 10 milliohm should be very small. . . .

Yes, the 1N4007 is connected in series, which means the current will not exceed 1A. Even if the tube is not fully turned on, the power consumption will not be too large, so this difference can be ignored.

qwqwqw2088

sfcsdc posted on 2020-7-30 09:19 Because this MOS will not switch frequently and is only used for logic control, it does not require a high speed. I used a relatively large resistor. ...

If it is PWM drive and there is a frequency requirement, a large resistance may not work. The selection of the drive resistance requires actual debugging. If it is too small, the switching speed will be fast, affecting EMI. If it is too large, the efficiency will be affected. Different tubes should be selected according to the manual parameters.

maychang

topwon posted on 2020-7-30 09:36 Based on the previous stage connected in series with 1N4007, the overcurrent capacity of this circuit should not exceed 1A, right? The difference in the on-resistance of this 10 milliohm should be very small. . . .

Judging from the series connection of 1N4007, this circuit does not have high requirements on the tube voltage drop, otherwise a Schottky diode with a smaller tube voltage drop should be used.