The power supply cannot be completely turned off. There is a leakage of 4~5MA. I hope an expert can give me some advice.

lyx_wq

The power supply cannot be completely turned off. There is a leakage of 4~5MA. I hope an expert can give me some advice. [Copy link]

 

I beg all masters and seniors to see where the circuit is configured incorrectly. The current problem is that there is a leakage of 4~5mA when the load (voice chip) is connected. Please give me some advice, thank you!

dcexpert

The diode between the DS of the MOS leaks over.

蓝雨夜

Pull up stronger?

bigbat

If RS is directly connected to "GND", the circuit should be able to be shut down, which means that there is a certain voltage on the "ground" of the RS control terminal. Try changing R37 to 10K

topwon

The connection method of PMOS in the figure (D to input, S to output) is used to prevent reverse connection circuit (at the same time, if the G pole is directly grounded or the input voltage is very high, a voltage divider resistor is connected between GS and R37 is removed). There is no control ability because the parasitic body diode between DS will be forward-conducted.

If you want to use PMOS as a power switch, you have to connect it in reverse, with S connected to the input and D connected to the output.

maychang

First, please clarify whether your load (voice chip) is connected to VCC or VCC3? In other words, which side is the power supply, VCC or VCC3?

lyx_wq

bigbat posted on 2019-12-27 10:33 If RS is directly connected to "GND", the circuit should be able to be turned off, which means that there is a certain voltage at the "ground" of the RS control end. ...

Hello, changing R37 to 10K does not work. Changing the IO port setting of the microcontroller from push-pull to open-drain does not work either.

lyx_wq

maychang posted on 2019-12-27 10:44 First, please clarify whether your load (voice chip) is connected to VCC or VCC3? In other words, which side is the power supply, VCC or VCC3?

Hello, thank you very much.

VCC is input, VCC3 is output

maychang

lyx_wq posted on 2019-12-27 11:54 Hello, thank you very much. VCC is input, VCC3 is output

Then what dcexpert said on the second floor, "It leaks through the diode between the DS of the MOS," is correct.

If you want to use a P-channel power MOS tube as a power switch, you must swap the drain and source of the tube in the first post.

maychang

lyx_wq posted on 2019-12-27 11:54 Hello, thank you very much. VCC is input, VCC3 is output

In fact, such a problem can be immediately seen as long as the parasitic diode inside the MOS tube is drawn when drawing the power MOS tube.

maychang

lyx_wq posted on 2019-12-27 11:54 Hello, thank you very much. VCC is input, VCC3 is output

For example, the drawing method in the datasheet of a typical P-channel power MOS tube will never be wrong, and the parasitic diode inside the body will not be ignored.

qwqwqw2088

The leakage current of the MOS tubes we usually use is very small, and some should be less than tens of uA.

Under ideal conditions, if no current flows when the MOS tube is turned off, even if the MOS tube has no channel, there is still a reverse saturation current between the drain and the source, which is the so-called leakage current.

qwqwqw2088

R37 is driven by G pole, 100K is relatively large, so the drive is a bit inadequate

chineseboyzxy

Comparing with the picture on the 11th floor, the poster just got the DS wrong. Our company has products that also use P tubes as switches, with a pull-up of 100K, and there is no problem at all.

maychang

lyx_wq posted on 2019-12-27 11:54 Hello, thank you very much. VCC is input, VCC3 is output

According to the explanation of topwon on the 5th floor, if the P-channel MOS tube in the circuit of the first post is connected in reverse, the poster's problem will be solved immediately.

lyx_wq

I followed the methods of the above users and tried them one by one. I reversed the polarity of DS of V3 and changed the resistor R37 from the original 100K to 10K, 1K, and 39K one by one, and grounded it to the gate rs_en. However, the problem was not solved. There was still a leakage current of about 5MA.

I would like to ask everyone to offer suggestions again. Thank you again!

topwon

lyx_wq posted on 2019-12-29 17:32 According to the methods of the above, I tried them one by one, reversed the polarity of V3's DS, and changed the resistor R37 from the original 100K to 10K, 1K, ...

You must first determine whether the MOS tube is turned off, that is, when the RS_EN pin is floating, whether the voltage of VCC3 is 0. Then analyze the cause of the leakage current, including the measurement method, the front and rear circuits, etc.

maychang

lyx_wq posted on 2019-12-29 17:32 According to the methods of the above, I tried them one by one, reversed the polarity of V3's DS, and changed the resistor R37 from the original 100K to 10K, 1K, ...

The easiest way to check your MOS tube is to remove the MOS tube and not connect it to any circuit on the circuit board. Short-circuit G and S together, and use the ohmmeter of the multimeter to measure the resistance between D and S (connect the red probe of the digital meter to D, and the black probe of the pointer meter to D). It should be infinite. Otherwise, your tube is damaged.

chineseboyzxy

lyx_wq posted on 2019-12-29 17:32 According to the methods of the above, I tried them one by one, reversed the polarity of V3's DS, and changed the resistor R37 from the original 100K to 10K, 1K, ...

The gate is grounded and it is conducting. . . .

lyx_wq

Thank you for your kind words and suggestions. I will test them one by one again and find the problem as soon as possible. Thank you again for taking the time to contribute good suggestions despite your busy schedule.