Synchronous Buck Converter Vds and Ids Waveforms Analysis

  • Thread starter Ms FYP
  • Start date
  • Tags Tags
    Waveforms
14 replies · 6K views · last reply
Ms FYP
Messages
14
Reaction score
0
Hi,

I'm simulating the Synchronous Buck Converter as per attached, to obtain the Vds and Ids for the High Side switch and I'm using P-mosfet as the switch and Microsim PSPICE software to simulate the circuit. once it is done, i obtained the Vds and Ids waveforms as per attached.

my question is, why the Vds and Ids for the high side switch are not overlap to each other?
Is there something wrong with the circuit that i have designed?

really appreciate that if anyone could share some ideas regarding this problem..

many thanks..
 
Attachments
  • Schematic of a synchronous buck converter: 12V source Vi, PWM1 driving P-channel MOSFET M6 (IRFR9212) at the high side, PWM2 driving N-channel MOSFET M2 (IRFP250) at the low side, node vnode, 15u inductor L, 0.625u capacitor C, and 3.5 ohm load R.
    SBC.png
    16.7 KB · Views: 567
  • Simulator plot of two stacked waveforms over 233.2us to 235.4us: top panel shows V(M6:d,vnode) switching between about 0 V and 5 V; bottom panel shows Id(M6) in green and V(M6:d,vnode) in purple, with Id pulsing up to about 12 and V on a 0 to 4 scale.
    graf..vds n ids-1.jpg
    19.5 KB · Views: 568
Ms FYP said:
Hi,

I'm simulating the Synchronous Buck Converter as per attached, to obtain the Vds and Ids for the High Side switch and I'm using P-mosfet as the switch and Microsim PSPICE software to simulate the circuit. once it is done, i obtained the Vds and Ids waveforms as per attached.

my question is, why the Vds and Ids for the high side switch are not overlap to each other?
Is there something wrong with the circuit that i have designed?

really appreciate that if anyone could share some ideas regarding this problem..

many thanks..

Why is PWM1 returning to the inductor instead of ground?

The value of the inductor looks pretty low, but you appear to be running the frequency pretty high, so maybe it's okay.

The bottom transistor will see the flyback voltage go below ground -- are you driving its gate correctly to accommodate that?

Could you show a plot of PWM1, PWM2 and the input voltage to the inductor?
 
hi berkeman,

thanks for replying my post..

Oh sorry..i miss that one..it should be connected to ground..
Here is the PWM1, PWM2 and input voltage to the inductor.
and yes..i'm driving this circuit at high switching frequency, that's why the value of the inductor is small.

i have attached two plots for both of them.. one is when the PWM1 is returning to inductor instead of ground and another one is when it is connected to ground...it seem that they have kinda same waveforms...

thanks~
 
Attachments
  • Simulation plot with time in microseconds from 232 to 238.4 us: top trace V(vnode) pulses between 0 and about 8.5 V, bottom traces PWM2 pulsing between 0 and 5 V and PWM1 pulsing negative from 0 to -5 V; plot titled PWM connected to ground.
    PWM1_PMW2_vnode-input to inductor_when PWM1 is connected to ground.jpg
    22 KB · Views: 521
  • Simulation plot of two stacked waveforms from 225 to 231.5 us: top trace V(vnode) square-wave switching between about 0 and 90, bottom traces PWM2 at 0 to 5.3 V and PWM1 at 0 to -5.2 V, with the title PWM1 returning to inductor.
    PWM1_PMW2_vnode-input to inductor_when PWM1 is returning to inductor.jpg
    22 KB · Views: 533
Ms FYP said:
hi berkeman,

thanks for replying my post..

Oh sorry..i miss that one..it should be connected to ground..
Here is the PWM1, PWM2 and input voltage to the inductor.
and yes..i'm driving this circuit at high switching frequency, that's why the value of the inductor is small.

i have attached two plots for both of them.. one is when the PWM1 is returning to inductor instead of ground and another one is when it is connected to ground...it seem that they have kinda same waveforms...

thanks~

Weird. Could you please add the current through the inductor to the first plot (the one with the gate drive with respect to ground for both devices)?
 
The PMOS FET is shown upside down in that circuit.

Also, the L and C form a series tuned circuit at about 52 KHz without the 3.5 ohm resistor and, with the resistor, they produce a low pass filter that doesn't pass anything much above 80 KHz.You could consider exchanging the two FET positions, but it would be better to find a working circuit and just copy it.
 
Last edited:
vk6kro said:
The PMOS FET is shown upside down in that circuit and it needs the gate to go more positive than the source if it was connected that way.

Also, the L and C form a series tuned circuit at about 52 KHz without the 3.5 ohm resistor and, with the resistor, they produce a low pass filter that doesn't pass anything much above 80 KHz.
As your signal is at about 1 MHz, this will cause a problem.

You could consider exchanging the two FET positions, but it would be better to find a working circuit and just copy it.

Hi vk6kro,

thanks for your replied..
i will try to exchange the position for both of the FETs and see the results..

many thanks~
 
Ms FYP said:
Hi vk6kro,

thanks for your replied..
i will try to exchange the position for both of the FETs and see the results..

many thanks~

No, your highside FET looks okay. It's your lowside FET that looks upside-down. Quiz Question -- Why?
 
A P-Channel MOSFET should have its drain more negative than the source.

In this case, it is more positive.

So, why do you feel it is OK?

[PLAIN]http://cnx.org/content/m1029/latest/4.48.png
 
Last edited by a moderator:
berkeman said:
Weird. Could you please add the current through the inductor to the first plot (the one with the gate drive with respect to ground for both devices)?

hi berkeman,

sorry..just came back from class..here is the current through the inductor for the inductor for the gate drive with respect to ground.
 
Attachments
  • Simulation plot of a buck converter with three stacked traces over time from 209.0us to 214.5us: IL triangular inductor current between about 830mA and 950mA, I(L) vs. V(vnode) pulses near 10V, and PWM1 and PWM2 gate drive waveforms between 0V and 5.3V with inverted timing.
    PWM1_PMW2_vnode-input to inductor_IL_when PWM1 is connected to ground.jpg
    23.5 KB · Views: 489
Most of that current is DC. The actual change in current is (950 mA - 825 mA) or 125 mA.

Did you measure the voltage across the 3.5 ohm resistor?
 
vk6kro said:
Most of that current is DC. The actual change in current is (950 mA - 825 mA) or 125 mA.

Did you measure the voltage across the 3.5 ohm resistor?

hi vk6kro,

yes, i have measured the voltage across 3.5 ohm resistor and attached is the waveforms.
hmm..may i know, is the duty ratio actually effected the switching of the PMOS FET? i mean.. can it become the reason why the Vds and Ids are resulted as shown in previous attachment? this is because, this circuit is driven with duty cycle of 0.25 and 1MHz switching frequency. so the expected output voltage should be 3v.

thanks..
 
Attachments
  • Simulation plot of a switching converter: top panel shows Vo ripple between about 3.097 V and 3.122 V over 196 to 202 microseconds; bottom panel shows PWM2 and PWM1 gate drive square waves swinging between 5.2 V and -5.3 V.
    PWM1_PMW2_vo_when PWM1 is connected to ground.jpg
    21.8 KB · Views: 485
If the PMOS is connected as shown, won't the device's body diode conduct when the switch is supposed to be off?

Re: the low-side switch quiz question, the source is grounded so as long as the driver can overcome the threshold voltage of the device, the current configuration should be fine, no? Also re the suggested swapping of the FETs, I can't see any good reason to use a PMOS as the synchronous rectifier?

Maybe the OP could post his updated schematic/waveforms with 1) the high-side PMOS connected with its source towards Vi and drain towards vnode, and 2) PWM1 properly grounded.
 
gnurf said:
Maybe the OP could post his updated schematic/waveforms with 1) the high-side PMOS connected with its source towards Vi and drain towards vnode, and 2) PWM1 properly grounded.

hi gnurf,

many thanks for your replied..
here i updated the schematic and waveforms for the circuit. and yet..i still obtained the same vds and ids results..
 
Attachments
  • Schematic of a synchronous buck converter: 12V input Vi, high-side MOSFET M6 (IRFR9212) and low-side MOSFET M2 (IRFP250) driven by PWM1 and PWM2, switch node feeding inductor L 15u, output capacitor C 0.625u and load resistor R 3.5.
    updated SBC.png
    18.2 KB · Views: 530
  • Simulation plot of a buck converter: top trace IL ramps between about 0.8A and 1.0A over 216.6 to 220.4us; middle trace I(L) pulses to about 90 during high-side on-time; bottom traces PWM2 and PWM1 switch between 5.0V and -5.0V with complementary timing.
    updated waveforms_IL_vnode_PWM1_PWM2.jpg
    21.1 KB · Views: 527
  • Simulator plot of two stacked waveforms over 310.9 to 314.6 us: top trace is a 0 to 60 V square wave with a flat-top high level near 50 V; bottom trace shows a 0 to 12 plot with a purple and a teal curve.
    Vds ids.jpg
    19.2 KB · Views: 502
gnurf said:
Maybe the OP could post his updated schematic/waveforms with 1) the high-side PMOS connected with its source towards Vi and drain towards vnode, and 2) PWM1 properly grounded.
Please try again.
 
By the way, it seems to me that you're pushing the cart before the horse a little bit here. Why not try the simplest solution first? That is, replace the synchronous rectifier (M2) with a free-wheeling diode so that you have a conventional buck converter.

When you're familiar with that and it's waveforms (e.g., why is the diode needed, why does the inductor current ramp up and down, how is the dutycycle related to the output voltage, etc), you can add another piece to the puzzle: the synchronous rectifier. Just leave the diode as it is (it'll provide a conduction path during the deadtime) and place M2 in parallel. Now you can compare the switch-node ("vnode") waveform with the one you got with only a diode and hopefully be understand better how the synchronous buck converter improves efficiency compared to its conventional counterpart.