SMPS Full-Bridge Converter

[Hudson T. Clark]

I'm looking for a full-bridge converter project. I need a schematic
preferably to build from. A parts list would be nice to. Or maybe a
reference to a site with DIY SMPS. I'm looking to design a 1.3kW power
supply and I need to experiment first. Any articles, references, or books
with ready to build schematics would be wonderful! They don't need to be
1.3kW supplies but it would be nice to just get some experience building any
full-bridge converter.

 

[Winfield Hill]

Hudson T. Clark wrote...

I don't really have an application I just want to make a nice
power supply for my bench.

Your 1300 watt spec makes this a difficult project. At that power
level it should include a PFC input stage, for example. Why not
tone down a bit, say to the 50, 100 or even 200W level? You can
make an elegant full-bridge phase-shift 200W PWM supply. Multiple
output voltages, full metering, external control - these are more
useful features than raw power, unless you have an application that
needs raw power. In my lab, if the latter situation arises, I turn
to a Xantrex supply, usually obtained from eBay for about $250 up.
If you examine the guts of one of these supplies, you'll see how a
1000W+ supply should be made. Xantrex doesn't provide schematics,
but you can get repair manuals for similar Sorensen power supplies
to learn some of the details (but the older Sorensen models didn't
include PFC).


--
Thanks,
- Win

 

[Winfield Hill]

John Woodgate wrote...

Hudson T. Clark wrote,

Would you happen to know of a good book on power supply design?
I have taken courses and read books on AC and DC theory, digital,
analog, etc. I'm trying to learn about applications like power
supply. I'm trying to find a book that would help explain how
filtering is used in power supplies to turn rectified AC into DC.

Pages 45 and 46 of Art of Electronics even have the main equations.

That's in Chapter 1, which is just an introduction to electronics.
We have more extensive discussion in Chapter 6, power supplies.

Much of AoE is a bit TOO non-mathematical IMHO.

Agreed, that's being fixed somewhat in the next edition.

As for a good general-purpose in-depth book on power-supply design
for Hudson, there's Robert W Erickson and Dragan Maksimovic's
"Fundamentals of Power Electronics," 2nd ed.

And there's Abraham Pressman's tome, "Switching Power Supply Design,"
which is all about SMPS, which interests Hudson.


--
Thanks,
- Win

 

[Winfield Hill]

John Woodgate wrote...

Winfield Hill wrote...
John Woodgate wrote...
Hudson T. Clark wrote,

Would you happen to know of a good book on power supply design?
I have taken courses and read books on AC and DC theory, digital,
analog, etc. I'm trying to learn about applications like power
supply. I'm trying to find a book that would help explain how
filtering is used in power supplies to turn rectified AC into DC.

Pages 45 and 46 of Art of Electronics even have the main equations.

That's in Chapter 1, which is just an introduction to electronics.
We have more extensive discussion in Chapter 6, power supplies.

True, of course, but the question was specifically about filtering.

Chapter six, pages 325 to 331.


--
Thanks,
- Win

 

[Fred Bloggs]

Hudson T. Clark wrote:

I'm looking for a full-bridge converter project. I need a schematic
preferably to build from. A parts list would be nice to. Or maybe a
reference to a site with DIY SMPS. I'm looking to design a 1.3kW power
supply and I need to experiment first. Any articles, references, or books
with ready to build schematics would be wonderful! They don't need to be
1.3kW supplies but it would be nice to just get some experience building any
full-bridge converter.

Why limit yourself to a 1.3KW project? Why don't you go for 13KW?

 

[Winfield Hill]

Fred Bloggs wrote...

Hudson T. Clark wrote:
I'm looking to design a 1.3kW power supply ...

Why limit yourself to a 1.3KW project? Why don't you go for 13KW?

130kW would be more useful. Enough to power a sensible 175HP
motor in a hybrid vehicle.


--
Thanks,
- Win

 

[Winfield Hill]

John Woodgate wrote...

I don't wish to be tiresome or contentious, but the OP wrote:

I'm trying to find a book that would help explain how
filtering is used in power supplies to turn rectified AC into DC.

Page 325 is about filtering conducted emissions, a different subject.
Pages 329 and 330 include information on the filter capacitor (which
we used to call the reservoir capacitor, and that was less ambiguous),
but so are pages 45 and 46.

Not wishing to be tiresome or contentious either, :&gt but "turning
rectified AC into DC" also involves the various additional items you
need to deal with the ac lines: switches, fuses, common-mode filters
for SMPS systems, such as the OP's, etc. What I do find missing from
our several-page discussion is the whole issue of transformer losses,
copper resistance and leakage-inductance, and its interaction with the
rectifier / filter cap, etc. Soar, droop and ripple vs load current,
etc., as well as the RFI produced from sharp snap-off voltage spikes
each cycle, as the rectifier diode finishes its reverse-recovery time.
We've devised an innocuous way to jam this stuff into the 3rd edition.


--
Thanks,
- Win

 

[Russ]

Winfield Hill wrote:

Fred Bloggs wrote...

Hudson T. Clark wrote:
I'm looking to design a 1.3kW power supply ...

Why limit yourself to a 1.3KW project? Why don't you go for 13KW?

130kW would be more useful. Enough to power a sensible 175HP
motor in a hybrid vehicle.

There'd be too much loss in the extention cord.

 

[Ross Herbert]

On Wed, 5 Jan 2005 22:30:45 +0000, John Woodgate
<jmw@jmwa.demon.contraspam.yuk> wrote:

SMPS are DIFFICULT! Except for the gurus who have designed hundreds.
It's much better to start with a transformer/rectifier/regulator
'linear' supply, but then I always say that.

I couldn't agree more...

From what I can glean from his posts the OP is a newbie to electronics
and doesn't have a good understanding of electronics components let
alone how to go about designing a high power SMPS. These things take
time to learn.

I would suggest that he first start with a relatively simple low power
linear design and perhaps in 3 or 4 years he may be ablew to tackle a
SMPS.

 

[John Popelish]

"Hudson T. Clark" wrote:

I'm looking for a full-bridge converter project. I need a schematic
preferably to build from. A parts list would be nice to. Or maybe a
reference to a site with DIY SMPS. I'm looking to design a 1.3kW power
supply and I need to experiment first. Any articles, references, or books
with ready to build schematics would be wonderful! They don't need to be
1.3kW supplies but it would be nice to just get some experience building any
full-bridge converter.

Have you looked at the application notes for some of the full bridge
converter control ICs?

http://focus.ti.com/lit/ds/symlink/ucc3895.pdf
http://focus.ti.com/lit/ds/symlink/uc1875.pdf
http://www.linear.com.cn/pdf/372212f.pdf
http://www.intersil.com/data/an/AN9506.pdf
http://www.powerdesigners.com/InfoWeb/design_center/Appnotes_Archive/u154.pdf

http://apec.engr.ucf.edu/publications/conferences/pesc/conf_28_5.pdf

--
John Popelish

 

[Hudson T. Clark]

Would you happen to know of a good book on power supply design? I have taken
courses and read books on AC and DC theory, digital, analog, etc. I'm trying
to learn about applications like power supply. I'm trying to find a book
that would help explain how filtering is used in power supplies to turn
rectified AC into DC. I understand how the rectifier works but I'm trying to
learn and understand how to design filters, for example to provide DC
voltage for one of these SMPS... Do you happen to know of any useful books
that clearly explains this? I understand its not a simple thing because most
authors tend to not really go into it in really heavy detail. Well thats
from experience. I have been trying to hunt down a good book but I'm having
problems. The authors tend to just put a capactor and choke after a
rectifier and assume I will understand the math behind it. Can someone
recommend a good book?

"Winfield Hill" <hill_a@t_rowland-dotties-harvard-dot.s-edu> wrote in
message news:crdr3t01sis@drn.newsguy.com...

Hudson T. Clark wrote...

I don't really have an application I just want to make a nice
power supply for my bench.

Your 1300 watt spec makes this a difficult project. At that power
level it should include a PFC input stage, for example. Why not
tone down a bit, say to the 50, 100 or even 200W level? You can
make an elegant full-bridge phase-shift 200W PWM supply. Multiple
output voltages, full metering, external control - these are more
useful features than raw power, unless you have an application that
needs raw power. In my lab, if the latter situation arises, I turn
to a Xantrex supply, usually obtained from eBay for about $250 up.
If you examine the guts of one of these supplies, you'll see how a
1000W+ supply should be made. Xantrex doesn't provide schematics,
but you can get repair manuals for similar Sorensen power supplies
to learn some of the details (but the older Sorensen models didn't
include PFC).


--
Thanks,
- Win

 

[John Woodgate]

I read in sci.electronics.design that Winfield Hill <hill_a@t_rowland-
dotties-harvard-dot.s-edu> wrote (in <crgqsl01uu2@drn.newsguy.com&gt
about 'SMPS Full-Bridge Converter', on Wed, 5 Jan 2005:

John Woodgate wrote...

Winfield Hill wrote...
John Woodgate wrote...
Hudson T. Clark wrote,

Would you happen to know of a good book on power supply design?
I have taken courses and read books on AC and DC theory, digital,
analog, etc. I'm trying to learn about applications like power
supply. I'm trying to find a book that would help explain how
filtering is used in power supplies to turn rectified AC into DC.

Pages 45 and 46 of Art of Electronics even have the main equations.

That's in Chapter 1, which is just an introduction to electronics.
We have more extensive discussion in Chapter 6, power supplies.

True, of course, but the question was specifically about filtering.

Chapter six, pages 325 to 331.


I don't wish to be tiresome or contentious, but the OP wrote:

I'm trying to find a book that would help explain how
filtering is used in power supplies to turn rectified AC into DC.

Page 325 is about filtering conducted emissions, a different subject.
Pages 329 and 330 include information on the filter capacitor (which we
used to call the reservoir capacitor, and that was less ambiguous), but
so are pages 45 and 46.
--
Regards, John Woodgate, OOO - Own Opinions Only.
The good news is that nothing is compulsory.
The bad news is that everything is prohibited.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk

 

[Jeroen]

"Winfield Hill" <hill_a@t_rowland-dotties-harvard-dot.s-edu> schreef in
bericht news:crtuhm02ghm@drn.newsguy.com...

Hudson T. Clark wrote...

<snip>

Our advice is to start out at low power levels and learn as you go.

Talking about serious power, a company I worked for had among their products
a helicopter jump starter powered by 3 phase 380V. It could supply 1500A@24V

I never bothered to find out how this device exactly worked, I believe it
chopped up the high voltage with massive thyristors and rectifying it with
big green diode's that looked like stacked plates.

I used to be amazed by the heat the dummy load produced and how hot the
wrist thick cables got

Jeroen

 

[Hudson T. Clark]

That's interesting. I haven't read all of the books just specific areas to
get an idea if this is something I can even handle. I was also thinking
about noise? Wouldn't it be catastrophic if noise was present on the gates
while switching several amperes?

"Winfield Hill" <hill_a@t_rowland-dotties-harvard-dot.s-edu> wrote in
message news:crtuhm02ghm@drn.newsguy.com...

Hudson T. Clark wrote...

I have my A.S. degree in electronics. I have experience with
microcontrollers, and I don't need to build a linear regulator
because I have repaired them before. ... I have bought EDN
power supply design, and switching power supply design. They
gave me a good understanding of how they work, but I'm just
empty on projects. Yes I am unexperienced at designing SMPS.
But anyone with a fundamental AC and DC theory education can
understand this paticular application, and I do because I have
taken the time to start reading about SMPS.

Certainly help yourself to a first project. But our advice is to
stick to low power levels at first. This is simply to protect
you from burning out all kinds of stuff and becoming discouraged.

Consider a few points you should have learned in your reading.

If you make a 1.3kW supply, the currents involved will be 20 times
more than in a 65-watt supply, all other things equal. Resistive
power losses, which create heat, given by P = I^2R, will be 400
times higher, unless you can reduce R by say 50x, so the loss will
only be 8x higher. But reducing R by 50x means the gate switching
capacitance of the large power FETs will be 50x higher, leading to
all kinds of problems, compounded by a need to switch faster.

Recall that every time you switch a current you create an inductive
voltage spike, given by V = L dI/dt. When you make a 1.3kW supply
its dI term is 20 times more than a 65-watt supply, all other things
equal. Furthermore, with high capacitance FETs and large currents
being switched, you'll learn about switching losses and the need to
reduce switching times in large power supplies. This means dt must
be smaller, say 10x faster than you could safely get away with in a
65W supply. Therefore the V = L dI/dt formula tells us the induced
inductive spikes will be 200x larger in a hypothetical 1.3kW supply.

For example, a 12V battery powering a 1.3kW supply. In an H-bridge
we'll be switching more than 108 amps. We determine that to protect
our FETs they have to switch in 30ns. That's 3.56 x 10^9 A/s. Now,
if our FETs have 25nH of total source-lead inductance, which is very
little and requires effort to achieve, the source-voltage spike will
be 89 volts. Which far exceeds the 20V gate-voltage failure limit.

The way you'll experience this is that the converter may work fine at
low load levels, but as you turn up the power, it'll suddenly fail.
No doubt with substantial explosive accompaniment. When you look to
find the part failure, you'll discover there are many failed parts,
because as one fails it takes out others, making it hard to determine
the exact failure sequence.

Our advice is to start out at low power levels and learn as you go.


--
Thanks,
- Win

 

[Winfield Hill]

Hudson T. Clark wrote...

... I was also thinking about noise? Wouldn't it be catastrophic
if noise was present on the gates while switching several amperes?

With so much fierce stuff going on, be thankful that noise, in the
sense we normally think of it, isn't a factor. Consider a MOSFET
gate, with a typical gate capacitance of 2nF. Think of a typical
50MHz equivalent switching bandwidth. The impedance of 2nF at 50MHz
is under two ohms. As you know, low impedances like that won't be
sensitive to "noise" pickup, especially considering the high voltage
levels required to change the on/off state of a power MOSFET.


--
Thanks,
- Win

 

[Winfield Hill]

Hudson T. Clark wrote...

I have my A.S. degree in electronics. I have experience with
microcontrollers, and I don't need to build a linear regulator
because I have repaired them before. ... I have bought EDN
power supply design, and switching power supply design. They
gave me a good understanding of how they work, but I'm just
empty on projects. Yes I am unexperienced at designing SMPS.
But anyone with a fundamental AC and DC theory education can
understand this paticular application, and I do because I have
taken the time to start reading about SMPS.

Certainly help yourself to a first project. But our advice is to
stick to low power levels at first. This is simply to protect
you from burning out all kinds of stuff and becoming discouraged.

Consider a few points you should have learned in your reading.

If you make a 1.3kW supply, the currents involved will be 20 times
more than in a 65-watt supply, all other things equal. Resistive
power losses, which create heat, given by P = I^2R, will be 400
times higher, unless you can reduce R by say 50x, so the loss will
only be 8x higher. But reducing R by 50x means the gate switching
capacitance of the large power FETs will be 50x higher, leading to
all kinds of problems, compounded by a need to switch faster.

Recall that every time you switch a current you create an inductive
voltage spike, given by V = L dI/dt. When you make a 1.3kW supply
its dI term is 20 times more than a 65-watt supply, all other things
equal. Furthermore, with high capacitance FETs and large currents
being switched, you'll learn about switching losses and the need to
reduce switching times in large power supplies. This means dt must
be smaller, say 10x faster than you could safely get away with in a
65W supply. Therefore the V = L dI/dt formula tells us the induced
inductive spikes will be 200x larger in a hypothetical 1.3kW supply.

For example, a 12V battery powering a 1.3kW supply. In an H-bridge
we'll be switching more than 108 amps. We determine that to protect
our FETs they have to switch in 30ns. That's 3.56 x 10^9 A/s. Now,
if our FETs have 25nH of total source-lead inductance, which is very
little and requires effort to achieve, the source-voltage spike will
be 89 volts. Which far exceeds the 20V gate-voltage failure limit.

The way you'll experience this is that the converter may work fine at
low load levels, but as you turn up the power, it'll suddenly fail.
No doubt with substantial explosive accompaniment. When you look to
find the part failure, you'll discover there are many failed parts,
because as one fails it takes out others, making it hard to determine
the exact failure sequence.

Our advice is to start out at low power levels and learn as you go.


--
Thanks,
- Win

 

[Hudson T. Clark]

I will have to take that into consideration. As you know I am new and I'm
not all that experienced. I never took into consideration the impedence of
switching mosfets... Thanks for the information.

"Winfield Hill" <hill_a@t_rowland-dotties-harvard-dot.s-edu> wrote in
message news:cs28ol01q9v@drn.newsguy.com...

Hudson T. Clark wrote...

... I was also thinking about noise? Wouldn't it be catastrophic
if noise was present on the gates while switching several amperes?

With so much fierce stuff going on, be thankful that noise, in the
sense we normally think of it, isn't a factor. Consider a MOSFET
gate, with a typical gate capacitance of 2nF. Think of a typical
50MHz equivalent switching bandwidth. The impedance of 2nF at 50MHz
is under two ohms. As you know, low impedances like that won't be
sensitive to "noise" pickup, especially considering the high voltage
levels required to change the on/off state of a power MOSFET.


--
Thanks,
- Win