Here is a boost converter (DCM) with waveform for three cases.
1. Exact ZCS (Zero Current Switching)
2. Transistor Mp turning OFF late
3. Transistor Mp turning OFF early

I have some questions hope anybody could explain.
When Mp turns OFF late, inductor current goes negative. So in this case, is Mn also OFF and the current will flow from output to input?
When Mp turns OFF early, as you can see in the picture below for the OFF period IL has two sections with different slopes. I am wondering why the second section is a straight line so VD is a constant pulse with magnitude VTHP as below? In this circumstance, inductor still has energy and Mp is turned OFF. So why the voltage VD is a constant (flat) as in the image?

 

Because Mp switches on when Vd rises to Vout + VTHP.

 

Because Mp switches on when Vd rises to Vout + VTHP.

So there is no need to consider the gate voltage Vp here?
Normally I see that to consider switch ON or OFF we need to look at VGS (NMOS) or VSG (PMOS).

 

So there is no need to consider the gate voltage Vp here?
Normally I see that to consider switch ON or OFF we need to look at VGS (NMOS) or VSG (PMOS).

That's correct. But I think in this case, it is assumed that vout >VG+VTH.

 

Could anybody explain why inductor current goes negative when Mp turns OFF late?

 

Two things to keep in mind. Vout > Vin and inductor current is proportional to inductor voltage.

 

Two things to keep in mind. Vout > Vin and inductor current is proportional to inductor voltage.

Thanks. I've got it now.

 

I am having problem hope somebody could explain it.
What are the reasons for PMOS Mp turning OFF early or late?
With fixed Vin and Vout, we can calculate the duration that inductor current falls from peak value to zero. So we can use this time to set PMOS ON.
Therefore, inductor current always reaching exactly zero. I don't understand why we need PWM here.
An other question is that why the PMOS is used instead of a diode?
I read that it is used because diode has high voltage drop.
This mean that Vthp is lower than diode drop?

 

In DCM mode the T_ON and T_OFF will need to change, if you what to keep the Vout unchanged. Because change in Iload will change the Vout, so to bring back the old Vout value the duty cycle must change.
As for the PMOS, When the PMOS is ON (in linear/triode region) his Rds(on) is very small. So the voltage drop across Rds(on) is much smaller than the diode forward voltage.

 

In DCM mode the T_ON and T_OFF will need to change, if you what to keep the Vout unchanged. Because change in Iload will change the Vout, so to bring back the old Vout value the duty cycle must change.

So this also used for Vin changes?
As Vin changes, the inductor discharging time also changes.
Is there a way to incorporate both controlling late and early turn off of PMOS and PWM?