*=============Duty_Cycle_Plots=======================


One way to view the relationship between an output voltage and

the duty cycle that created it, is to do a duty cycle sweep during

a transient analysis. It is not hard to sweep from 0% to 100% over

a given length of time. A piece wise linear waveform which

ramps from zero to unity can be used to pulse width modulate a

square wave.


=======================================================================



dutyCycle.jpg


=======================================================================


VP        VP     0      DC     3.141592653589793

VT        VT     0      PWL    ( 0 0 1m .1m )

B_TRI     TRI    0      V =    acos( cos(v(VP)*v(VT)*2*1e6) )/v(VP) 


VD        VD     0      PWL    ( 0 0 1m  1 ) 

B_SW      PWM    0      V =    2*u( -v(TRI) +v(VD)) -1

B_ON      VON    0      V =    5*u( v(PWM) ) 



=======================================================================

A voltage ramp which represents time "VT" can be constructed. 

The time scale can also be scaled. It might be handy to create a

DC voltage "VP" equal to PI. An arcosine of a cosine can

create a very clean "TRI" which is a triangle wave.


The duty cycle voltage "VD" is set to go from zero to one.

This voltage then offsets the triangle wave and is limited

to produce the "PWM" pulse width modulated signal.


=======================================================================




Duty2.jpg



=======================================================================


The three plots above show how the duty cycle voltage "VD" 

offsets the triangle wave "TRI" at various times to 

to produce the "VON" pulse width modulated signal.


=======================================================================





Buck_Converter

*                 _  _  _

*  ___   /  V1   / \/ \/ \   ___     R1

* |VIN|_/  ______| () () |__|OUT|__/\  /\  /\__

* |___| S1   _|_    L1      |___|    \/  \/   _|_

*             ^     100u      |    C1         ///

* 1MHz    D1 /_\              |__||__ 3u

*             |   Freq = 9K      || _|_ 

*            _|_  Z=5.8Ohms         ///

*            ///


=======================================================================


A Buck_Converter almost looks like it is applying a

pulse width modulated signal to a simple LRC lowpass

filter. It that was the case, one would expect the 

output to track closely the duty cycle. 


But the switch and diode are making things a little nonlinear.


=======================================================================

Buck_Converter

*                 _  _  _

*  ___   /  V1   / \/ \/ \   ___     R1 = 50K

* |VIN|_/  ______| () () |__|OUT|__/\  /\  /\__

* |___| S1   _|_    L1      |___|    \/  \/   _|_

*             ^     100u      |    C1         ///

* 1MHz    D1 /_\              |__||__ 3u

*             |   Freq = 9K      || _|_ 

*            _|_  Z=5.8Ohms         ///

*            ///



  buck5k.jpg


                  plot      out vs vd




=======================================================================


One can call out a Y vs X plot of the output voltage "OUT"

versus the duty cycle "VD". This output voltage to

duty cycle relationship is a function of the size of

the output load resistor R1. In this case, the load resistor

is 50KOhms.



=======================================================================

Buck_Converter

*                         VL  _  _  _

*  ___   /  V1    RL 1m      / \/ \/ \   ___     R1 = 50K

* |VIN|_/  _______/\  /\  /\_| () () |__|OUT|__/\  /\  /\__

* |___| S1   _|_    \/  \/      L1      |___|    \/  \/   _|_

*             ^                 100u      |    C1         ///

* 1MHz    D1 /_\   1m  RD                 |__||__ 3u

*             |_/\  /\  /\_   Freq = 9K      || _|_

*          VD     \/  \/  _|_ Z=5.8Ohms         ///

*                         ///



When one sticks in some 1mOhm resistors in series with the diode

and the inductor, the nonlinear output voltage begins to make sense. 

Monitoring "V1-VL" shows the inductor current and "VD" is the diode

current. 


=======================================================================

buck60u.jpg


=======================================================================


At a time of 60us, the duty cycle is at 6%. The output "OUT" is very close

to zero. Inductor L1 is seeing 5volts across it for 6% of 10usec.

During that 600nsec, one would expect a 100uH inductor to ramp up to a

current of 30mA. When switch S1 is open, inductor L1 will be getting

its current from diode D1 which is 600mV. So the current will ramp down

about 8 times slower. And then no current will flow thru the inductor.

So the inductor, the switch, and the diode are in effect creating an

additional duty cycle as to how long is the inductor drawing current.


=======================================================================

Buck_Converter

*                         VL  _  _  _

*  ___   /  V1    RL 1m      / \/ \/ \   ___     R1 = 50K

* |VIN|_/  _______/\  /\  /\_| () () |__|OUT|__/\  /\  /\__

* |___| S1   _|_    \/  \/      L1      |___|    \/  \/   _|_

*             ^                 100u      |    C1         ///

* 1MHz    D1 /_\   1m  RD                 |__||__ 3u

*             |_/\  /\  /\_   Freq = 9K      || _|_

*          VD     \/  \/  _|_ Z=5.8Ohms         ///

*                         ///



buck550u.jpg



=======================================================================

At a time around 560us, the applied duty cycle is around 5us of 10us.

At a 50% duty cycle, the output "OUT" is around 4.13 volts. 

Now inductor L1 is seeing about 900mV when the switch is on and

will ramp up to around 55mA. When the switch is open, the diode

D1 will produce about 5V around L1, and the current will ramp 

down about 5 times faster. Again there is an additional duty

cycle in terms of how long is the inductor drawing current.




=======================================================================

Buck_Converter

*                         VL  _  _  _

*  ___   /  V1    RL 1m      / \/ \/ \   ___     R1 = 50K

* |VIN|_/  _______/\  /\  /\_| () () |__|OUT|__/\  /\  /\__

* |___| S1   _|_    \/  \/      L1      |___|    \/  \/   _|_

*             ^                 100u      |    C1         ///

* 1MHz    D1 /_\   1m  RD                 |__||__ 3u

*             |_/\  /\  /\_   Freq = 9K      || _|_

*          VD     \/  \/  _|_ Z=5.8Ohms         ///

*                         ///




buck750.jpg




=======================================================================

At a 75% duty cycle, the output "OUT" is close to 5 volts. 

Now inductor L1 is seeing a small voltage when the switch is on 

for 7.5us, and will be ramping up to around 17mA. When the

switch is open, inductor L1 see close to 6volt, and ramps

down much faster. 



=======================================================================


Buck_Converter

*                 _  _  _

*  ___   /  V1   / \/ \/ \   ___     R1 = 25

* |VIN|_/  ______| () () |__|OUT|__/\  /\  /\__

* |___| S1   _|_    L1      |___|    \/  \/   _|_

*             ^     100u      |    C1         ///

* 1MHz    D1 /_\              |__||__ 3u

*             |   Freq = 9K      || _|_ 

*            _|_  Z=5.8Ohms         ///

*            ///


But everything changes when the output resistor R1 is reduced

to 25 Ohms. Plotting the output voltage versus the applied

duty cycle shows a more linear relationship above a 50% duty

cycle. It turns out this happens when the current in inductor 

L1 never completely turns off.  



buck25.jpg


              plot      out vs vd




=======================================================================


Buck_Converter

*                         VL  _  _  _

*  ___   /  V1    RL 1m      / \/ \/ \   ___     R1 = 25

* |VIN|_/  _______/\  /\  /\_| () () |__|OUT|__/\  /\  /\__

* |___| S1   _|_    \/  \/      L1      |___|    \/  \/   _|_

*             ^                 100u      |    C1         ///

* 1MHz    D1 /_\   1m  RD                 |__||__ 3u

*             |_/\  /\  /\_   Freq = 9K      || _|_

*          VD     \/  \/  _|_ Z=5.8Ohms         ///

*                         ///


buck60u25.jpg



At a 6% duty cycle, everything is pretty much the same. The inductor

ramps up to about 30mA, and then back down again. 



=======================================================================


Buck_Converter

*                         VL  _  _  _

*  ___   /  V1    RL 1m      / \/ \/ \   ___     R1 = 25

* |VIN|_/  _______/\  /\  /\_| () () |__|OUT|__/\  /\  /\__

* |___| S1   _|_    \/  \/      L1      |___|    \/  \/   _|_

*             ^                 100u      |    C1         ///

* 1MHz    D1 /_\   1m  RD                 |__||__ 3u

*             |_/\  /\  /\_   Freq = 9K      || _|_

*          VD     \/  \/  _|_ Z=5.8Ohms         ///

*                         ///


buck550u25.jpg


But at 56% duty cycle, the output "OUT" is around 2.6volts. Now the 

inductor L1 ramps between 37mA and 188mA. So the current in the inductor

never turns off, and the whole Buck converter is acting more like a simple

lowpass filtered 56% duty cycled 5V square wave.



=======================================================================


Buck_Converter

*                         VL  _  _  _

*  ___   /  V1    RL 1m      / \/ \/ \   ___     R1 = 25

* |VIN|_/  _______/\  /\  /\_| () () |__|OUT|__/\  /\  /\__

* |___| S1   _|_    \/  \/      L1      |___|    \/  \/   _|_

*             ^                 100u      |    C1         ///

* 1MHz    D1 /_\   1m  RD                 |__||__ 3u

*             |_/\  /\  /\_   Freq = 9K      || _|_

*          VD     \/  \/  _|_ Z=5.8Ohms         ///

*                         ///

buck570u25.jpg


At 76% duty cycle, the output "OUT" is around 3.6volts. Now the 

inductor L1 ramps between 108mA and 218mA. The whole Buck is 

looking more like a Pulse width modulated square wave that is 

LRC lowpass filtered. 


=======================================================================


buck25.jpg

              plot      out vs vd



Since the start of the linear region seems to suggest where

the inductor is drawing continuous current. Plotting output

versus duty cycle for three different output loads may 

be useful. 






===========Full_Netlist_For_Copy_Paste======================= 

Buck_Converter

*                 _  _  _

*  ___   /  V1   / \/ \/ \   ___     R1

* |VIN|_/  ______| () () |__|OUT|__/\  /\  /\__

* |___| S1   _|_    L1      |___|    \/  \/   _|_

*             ^     100u      |    C1         ///

* 1MHz       /_\              |__||__ 3u

*             |   Freq = 9K      || _|_ 

*            _|_  Z=5.8Ohms         ///

*            ///

VP        VP     0      DC     3.141592653589793

VT        VT     0      PWL    ( 0 0 1m .1m )

VD        VD     0      PWL    ( 0 0 1m  1 ) 

B_TRI     TRI    0      V =    acos( cos(v(VP)*v(VT)*2*1e6) )/v(VP) 

B_SW      PWM    0      V =    2*u( -v(TRI) +v(VD)) -1

B_ON      VON    0      V =    5*u( v(PWM) ) 


VIN       VIN    0      DC     5

*SXXXXXX  N+     N-     NC+    NC-    MODEL        <ON><OFF>

S1        VIN    V1     PWM    0      SWP

D1        ID     V1     DD

L1        V1L    OUT    100u 

C1        OUT    0      3u

R1        OUT    0      50K

RD        ID     0      1m

RL        V1     V1L    1m


.MODEL    SWP    SW(    VT=2.6m VH=.2  RON=10u  ROFF=100MEG)

.model    DD     D(     IS=3.15e-18 ) 


*TRAN     TSTEP  TSTOP  TSTART TMAX        ?UIC?

.tran    .01u    1m    0      .01us   

*=========Run_Sim============================================ 

.control 

run 

set       pensize = 2 

plot      v(vd)     v(von)/6  ylimit -.5 1.5

set       pensize = 3

plot      v(tri)   v(vd)  v(von)/5   xlimit .05m .08m  ylimit -.5 1.5

plot      v(tri)   v(vd)  v(von)/5   xlimit .49m .52m  ylimit -.5 1.5

plot      v(tri)   v(vd)  v(von)/5   xlimit .89m .92m  ylimit -.5 1.5


plot      out vs vd


plot      (v1-v1l)*10000 id*10000 von/50-250m out/10 v1/10   xlimit  .05m .07m  ylimit -.5 1

plot      (v1-v1l)*10000 id*10000 von/50-250m out/10 v1/10   xlimit .55m .57m  ylimit -1 1 

plot      (v1-v1l)*10000 id*10000 von/50-250m out/10 v1/10   xlimit .75m .77m  ylimit -1 1 

.endc 

.end