Switch Mode Power Supplies: SPICE Simulations and Practical Designs

Switch Mode Power Supplies: SPICE Simulations and Practical Designs
Christophe Basso February 2008
Corrections of typos, mistakes and errors found by readers (or by the author himself!)
Page xvi: Sorry Monsieur Balocco, your name was misspelled with two l. One is enough:
Dr. Didier Balocco from Saft Power Systems.
Page 4: equation (1-1) result is incorrect, it should be 140 (Fig. 1-1 is ok though)
Page 22: below equation (1-32): for a buck, V1 = Vin Vout and not Vout Vin.
Contributed by C. Denton, March 2008.
Page 57: to gain a comprehensive understanding of the buck-boost operation The word
buck is missing.
Contributed by Kulsangcharoen Ponggorn, June 2008.
Page 69, figure 1-50: I dont know what happened to the captions on the graph, but those on
the picture are certainly not at the right place. Here is the corrected graph:
2.00
= 0.1, Q = 5
= 0.5, Q = 1
1.50
1.00
3
4
5
7
9
10
6
8
1
2
500m
= 1, Q = 0.5
0
2.50u
Figure 1-50
7.50u
ti
12.5u
i
17.5u
22.5u
Page 75, figure 1-56: the picture is correct, however, the zeta captions are not: in both
cases, zeta is positive and greater than 0.
Page 92, figure 1B-15: the upper curve caption should be output step, 400 mA/s and not
stemp, 400 mA/ms!
Plot1
iout in amperes
Page 92, figure 1B-15: the amplitude caption in the figure is wrong, the 2 mV does not make
sense. The real capacitor contribution alone is 70 mV as corrected in the below figure.
500m
Output step, 400 mA / s
400m
300m
Iout(t)
200m
100m
LESL contribution
Plot2
vlesl in volts
40.0m
20.0m
0
40 mV
-20.0m
-40.0m
Plot3
vresr in volts
40.0m
20.0m
0
-20.0m
RESR contribution
40 mV
-40.0m
Plot4
vc in volts
5.08
C contribution
5.04
5.00
4.96
70 mV
4.92
Plot5
vout2 in volts
5.08
spike
5.04
Vout(t)
5.00
4.96
82 mV
undershoot
4.92
4.97m
5.05m
5.13m
time in seconds
5.21m
5.28m
Figure 1-B5: step, S1 = 400 mA/s (not 450 mA)
Page 167, CCM PWM switch CM listing

A capacitor has been purposely added between terminals C and P to create the sub
harmonic effects at half of the switching frequency. To obtain exactly the same curve as with
Ridley models, the capacitor must be connected after the voltage source VM, between nodes
cx and p accounting for the presence of this capacitor in the measured current. It was
previously connected between c and p, leading to a slight high-frequency deviation. The
correction has been included before print, the book listing is thus good.
Contributed by Ekrem Cengelci on May 2007.
Page 236: Eq. (2B-6) should be: T ( t ) = 1 2e t + e 2t
Contributed by Jason Neudorf, May 2008.
Page 268: Fig. 3-22, page 268: in the picture, Rload should be 2.5 and not 2.5 W.
Page 269: Fig. 3-23a, page 268: in the picture, fc should be 5 kHz and not 5 Hz.
Contributed by C. Denton, March 2008
Page 276: Fig. 3-28, page 276: in the schematic parameters list, fp2 should be 50k and not 50.

Page 288: The table describing the TL431 and TLV431 features an obvious typo. The
breakdown voltage of the TLV431 is not 1 V but 18 V!
Reference
TL431I
TL431A
TL431B
TLV431A
TLV431B
Vref
2.495 V
2.495 V
2.495 V
1.24 V
1.24 V
Ibias,min
1 mA
1 mA
1 mA
100 A
100 A
Precision
2% @ 25 C
1% @ 25 C
0.4% @ 25 C
1% @ 25 C
0.5% @ 25 C
Max voltage
36 V
36 V
36 V
18 V
18 V
Max current
100 mA
100 mA
100 mA
20 mA
20 mA
Contributed by Chalermkiat, June 2008.

Page 293: Fig. 3-39, page 293: in the schematic parameters list, it should be fc = 1k and not
just 1 and below should be pm = 100 and not kpm = 100.
Page 322: Reference 3. Note from the author: its too old, cant find place. Obviously,
this sentence was left over during the editing process!
Page 411: Numerical application gives a cutoff frequency of 54 kHz and not 45 kHz.
Contributed by Kulsangcharoen Ponggorn, July 2008.
Page 418, equation 5-17:
This equation defines the diode conduction losses by:
Pcond ,diode = V f I d ,avg + Rd I d ,rms 2
(5-17)
In the above expression, Vf should be replaced by VT0 where VT0 represents the threshold
voltage of the junction. If you look at figure 1 (see the appendix at the end), the diode model
appears on the right side. It combines a constant voltage source, VT0, in series with a dynamic
resistor Rd. The total voltage drop Vf measured at a given current Id0, combines VT0 and the
dynamic resistor drop. If you use Vf as in Eq. (5-17), the term Rd is counted twice. Therefore,
the total average power dissipated by the diode is correctly expressed by:
Pcond = VT 0 I d ,avg + Rd I d ,rms 2 V f I d ,avg
VT0 depends on the diode technology. It is measured when the diode starts conducting (a few
hundred of A). The drop is roughly 0.4 V for Schottky and 0.6 V for a silicon diode.
The correction is as follows:
The conduction losses for a diode are given by
Pcond = VT 0 I d ,avg + Rd I d ,rms 2 V f I d ,avg
where Id,avg = average current in the diode

Vf = forward voltage at the considered diode current
(5-17)
VT0 = forward voltage at which the diode starts to conduct (0.4 V for a Schottky, 0.6
V for a silicon diode)
Rd = ..
Page 469: figure 5-14a:
The caption should say The CCM buck-boost converter
Page 562: equation 6-166
It should be R1 not R3:
R1 =
275 2 3
= 1.6 M
250u
Page 633: voltage-mode control table
The dc gain, Vout/Verror, should be:
NVin
(1 D )
V peak
The N is missing
Page 652: below 15 V, then there is no arm to let the driver there is no harm!
Contributed by C. Denton, April 2008.
Pd = VT 0 I d ,avg + Rd I d , rms 2 + DI R PIV V f I d,avg
(7-166)

Pd = VT 0 I d ,avg + Rd I d , rms 2 + DI R PIV V f I d,avg
(7-204)
Page 825, table 8-2:
In the where section, the definition for Sn has to be changed:

NVin Vout
Sn =
NRsense this is the on-time secondary slope reflected to the primary.
L
V
The equation given in the page, S n = in Rsense , corresponds to the magnetizing current slope.
Lp
It can, by the way, be considered as a free external ramp.
Page 834: small typo on the current Ivalley (or Ipeak) should be Ip,valley or Ip,peak :
PSW ,on =
I p ,valleyVbulk t
Fsw
eq. 8-125a
PSW ,on =
I p ,valleyVbulk,max t
Fsw =
PSW ,off =
PSW ,off =
I p , peakVbulk ,max t
1.56 400 45n

100k 470 mW
6
I p , peakVbulk t
Fsw =
Fsw
eq. 8-125c
eq. 8-125d
1.84 400 45n

100k = 1.65 W
2
eq. 8-125e
Contributed by T. Sutto, May 2008.

Page 856: equation 8B-12
2
N
The equation should be: Ceq = CL 2 3 || CL1 , the CL2 sub term is missing.
N2
For equations 8B-13c and 8B-13d, Fsw should be replaced by f, where f is a sinusoidal signal
of course.
Page 860: equation 8C-2
In the numerical application, the 104 is gone, but the result is correct:
Wa Ac =
Pout
250 10000
104 =
= 2.24 cm 4 eq. 8C-2
507 0.0005 0.11100000 400
K c K t Bmax Fsw J
Contributed by T. Sutto, May 2008.
Appendix 5X conduction diode loss calculations
Id
Id
Id
Rd
Id0
Rd =
dV f
Vf
dI d
VT0
Vf
VT0 Vf0
Figure 1: the equivalent diode representation.
In most of the examples, the conduction losses are simply calculated by Pcond = V f I d ,avg which
is ok since the total drop given by the data-sheets at a certain average current combines the
contribution of the dc source VT0 and the ohmic drop. However, what is the error if we do not
account for the ac ripple? Suppose we have a diode crossed by a full-wave signal like below:
23 A
Irms = 16.3 A
Iavg = 14.6 A
10 0
30 0
50 0
70 0
90 0
Figure 2: an example of a rectified current flowing in a diode.
The rms current is expressed by: I d ,rms =

The average current is: I d ,avg =
I d , peak
2
2 I d , peak
In Eq. (5-17), we would assume the following losses:
Pd , avg = VT 0 I d ,avg + Rd I d , rms 2
(5-17)
Replacing by the rms and average definitions, we obtain:

Pd ,avg =
2VT 0 I d , peak
1
+ Rd I d , peak 2
2
(5-17b)
If we now neglect the rms current effect on the dynamic resistor, we have:
Pd , avg = V f I d ,avg = (VT 0 + Rd I d ,avg ) I d , avg
(5-17c)
Again, replacing by the definition, we have:

2 I d , peak 2 I d , peak 2VT 0 I d , peak 4
Pd ,avg = VT 0 + Rd
=
+ 2 Rd I d , peak 2
(5-17d)
If we compare equations 5-17b and 5-17d, the resistive factor is respectively wheigted by 0.5
4
and 2 , 0.405.
Suppose the peak current of figure 2 is 23 A, leading to:

I d ,rms =
I d ,avg =
I d , peak
2
2 I d , peak
23
= 16.3 A
1.414
(5-17e)
46
= 14.6 A
3.14
(5-17f)
We have selected a power diode whose characteristic appears on figure 3:
1.25
VT0 = 1.15 V
Rd =
dV f
dI d
250m
= 25 m
10
V f @15 A = 1.45 V
Figure 3: forward voltage versus direct current.
From the curve, we have extracted VT0 (1.15 V), the dynamic resistor (the slope of the blue
curve, 25 m) and the total forward drop at a 15 A current (1.45 V). Lets apply our previous
formulas with these numbers:
Pcond = VT 0 I d ,avg + Rd I d , rms 2 = 1.15 14.6 + 25m 16.32 = 23.4 W
Pcond = V f I d ,avg = 1.45 14.6 = 21.2 W
(5-17g)
(5-17h)
The second formula differs from the first by 9.4%.

Some manufacturers also give power curves in relationship to the peak to average current
coefficient Kf (figure 4). When this factor equals 1, this a dc current (no ac ripple). Otherwise,
with a signal as in figure 2, we would have:
Kf =
I d , peak
I d , avg
I d , peak
2 I d , peak
1.6
(5-17i)
For single wave rectification, this factor equals .

For a square wave signal with a duty cycle D, the peak to average current factor is:
I d , peak
I d , peak
1
(5-17j)
I d , avg
DI d , peak D
Figure 4 square-wave curve seems to imply a duty-cycle of 50% (Kf = 2).
Kf =
Figure 4: some manufacturers give power dissipation curves depending on the ac ripple
amplitude.

Switch Mode Power Supplies: SPICE Simulations and Practical Designs

Uploaded by

Copyright:

Available Formats

Switch Mode Power Supplies: SPICE Simulations and Practical Designs

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Switch Mode Power Supplies: SPICE Simulations and Practical Designs

Uploaded by

Copyright:

Available Formats

Switch Mode Power Supplies: SPICE Simulations and Practical Designs

Christophe Basso February 2008

Output step, 400 mA / s

Figure 1-B5: step, S1 = 400 mA/s (not 450 mA)

Page 167, CCM PWM switch CM listing

Contributed by C. Denton, March 2008.

Contributed by Chalermkiat, June 2008.

where Id,avg = average current in the diode

Page 633: voltage-mode control table

The dc gain, Vout/Verror, should be:

Page 693, equation 7-204:

Page 825, table 8-2:

In the where section, the definition for Sn has to be changed:

1.56 400 45n

1.84 400 45n

Contributed by T. Sutto, May 2008.

Contributed by T. Sutto, May 2008.

Appendix 5X conduction diode loss calculations

Figure 1: the equivalent diode representation.

Figure 2: an example of a rectified current flowing in a diode.

The rms current is expressed by: I d ,rms =

In Eq. (5-17), we would assume the following losses:

Pd , avg = VT 0 I d ,avg + Rd I d , rms 2

Replacing by the rms and average definitions, we obtain:

Again, replacing by the definition, we have:

Suppose the peak current of figure 2 is 23 A, leading to:

We have selected a power diode whose characteristic appears on figure 3:

Figure 3: forward voltage versus direct current.

Pcond = V f I d ,avg = 1.45 14.6 = 21.2 W

The second formula differs from the first by 9.4%.

For single wave rectification, this factor equals .

You might also like