What is the reason for the flyback power supply duty cycle waveform

灞波儿奔

What is the reason for the flyback power supply duty cycle waveform [Copy link]

 

What is the reason for the uneven duty cycle waveform of the flyback power supply?

The flyback power supply circuit is made, with an input of 90V~265v and an output of 15V2A. The magnetic core EI30Ae=110mm, Lp=430uH, Np=37T, Ns=6T, the
operating frequency is about 60k, and it works in discontinuous mode.

The circuit diagram is as follows

The problem now is that the measured duty cycle square wave is not uniform. When the lowest input voltage is close to 90V, the waveform on the current sensing resistor Rs is very uneven.

Measure the GS waveform,

The unevenness is greater when the input voltage increases. What is the reason? Please help analyze and discuss it.

PowerAnts

Give me a reason to use C11. Are you full of yourself?

mcu5i51

Not only C11, but C12 is also abnormal, increasing the feedback time is too risky;

At 30W, the current sampling resistor feels too small. The limiting voltage of 3842 is 1V. This chip may be similar. The current limit of 4A (average 2A) is too large.

PowerAnts

mcu5i51 posted on 2022-3-14 15:05 Not only C11, but also C12 is not normal. Increasing the feedback time is too risky. At 30W, the current sampling resistor is too small. The limiting voltage of 3842 is 1V, ...

R14 is so small, the loop base is equipped with type I main pole compensation, and C12 is large, which will only affect the transient response and will not cause instability.

灞波儿奔

mcu5i51 posted on 2022-3-14 15:05 Not only C11, but also C12 is not normal. Increasing the feedback time is too risky. At 30W, the current sampling resistor is too small. The limiting voltage of 3842 is 1V, ...

Why does C11 affect the waveform change of the duty cycle?

What is the specific reason why C12 is also abnormal?

This capacitor is used in some circuits like this

灞波儿奔

PowerAnts posted on 2022-3-14 15:12 R14 is so small, the loop base is equipped with type I main pole compensation, and a large C12 will only affect the transient response and will not cause instability.

Remove C11, the duty cycle waveform does not change

PowerAnts

Baboerben published on 2022-3-14 15:34 Remove C11, the duty cycle waveform does not change

R10, C8, PWM frequency is about 350K, which is far from the 60K you said. Please mark all the parameters honestly first.

mcu5i51

PowerAnts posted on 2022-3-14 15:12 R14 is so small, the loop base is equipped with type I main pole compensation, and a large C12 will only affect the transient response and will not cause instability.

When disconnecting a large load, the reaction is slow and it will explode, and the current limit is too large

mcu5i51

灞波儿奔 posted on 2022-3-14 15:33 C11 can affect the waveform change of the duty cycle. What is the specific reason why C12 is also abnormal? This capacitor is used in some circuits...

It is OK to add capacitors to the 431 output, but you have to make sure the available capacity is small or large. It will oscillate within a certain range in the middle. Don't use it if you are not sure.

mcu5i51

Pay attention to D6 and D7, 4148, current 100ma, it is easy to have problems, at least add current limiting resistors.

maychang

The second picture is the voltage waveform across R11 (the current waveform in the MOS tube), right? The waveform on the gate of the MOS tube cannot be such a sawtooth wave.

mcu5i51

From the waveform, there is a frequency interference of about 20K, and it is suspected that there is a wiring problem

When debugging, first remove the transformer, and directly add about 17V input voltage from the IC output pin to the output rectifier to see the working status. If it is normal, there is a high probability of wiring problems, such as long feedback line, too long transformer output rectifier to filter line, voltage sampling point position, hot and cold ground

mcu5i51

Please also provide the waveform of R11 and the working load status. If the first one is, then it exceeds the protection current.

mcu5i51

If the second picture is the SG pin voltage diagram of the MOS tube, it is better to add a large heat sink.

PowerAnts

Following the 7th floor, R10=13K, C8=101, OSC=700K, PWM=350K! ! ! !

There is no need to read it until the host gives the correct schematic diagram with parameters that match the description.

se7ens

Isn't it normal for the drive waveform of a discontinuous mode power supply to look like this?

If you want to verify the drive waveform, you can increase the load and enter the continuous mode.

PowerAnts

se7ens posted on 2022-3-15 16:02 Isn't it normal for the drive waveform of the power supply working in discontinuous mode to look like this? If you want to verify the drive waveform, you can increase the load and enter the continuous mode to see...

The two waveforms posted by the OP are totally different. The first GS waveform has a maximum duty cycle of about 0.25, and the second drain current has a duty cycle of about 0.5, so I repeatedly asked him to give the correct parameters.

灞波儿奔

PowerAnts posted on 2022-3-15 16:10 The two waveforms posted by the OP are totally different. The first GS waveform has a maximum duty cycle of about 0.25, and the second drain current has a duty cycle of about 0.5, so I...

Thank you all teachers, I haven’t had time to read the posts these days.

This is because the duty cycle waveform intervals are uneven, which is exactly the problem that teachers need to help solve.

I remember that the waveform of the duty cycle should be the waveform on the current-sense resistor Rs when the input is 90V. The so-called duty cycle is different, and the waveforms taken are not necessarily simultaneous. The uneven spacing of the square waves needs to be solved. Do you have any suggestions?

灞波儿奔

PowerAnts posted on 2022-3-15 08:35 Following the 7th floor, R10=13K, C8=101, OSC=700K, PWM=350K! ! ! ! ! Before the original poster gives the correct schematic diagram with the parameters consistent with the description, ...

The parameters are R10 is 130K, C8 is 101

The duty cycle is 50%, and the theoretical frequency is 66khz, which is no problem.

The parameters are different. The actual measurement is like this waveform. This is related to the uneven size of the square wave distance. R10 and C8 may have been constantly debugged and changed when solving this problem.

灞波儿奔

PowerAnts posted on 2022-3-14 14:32 Give me a reason to use C11, are you full?

The reason why C11 is not used here, I hope the teacher can give me some advice patiently.