Design of SMPS with flyback,single ended, forward topology.

[Richard]

Hi. I'm not an electronic circuit designer, but may have to engage in the
art just a bit.

I'm just wanting some reassurance as to what my task will be.

I'm not sure what what wattage of power supply I'm needing, but I guess
somewhere between 50w and 70w.

The task before me is to make a SMPS with multiple outputs at 250v (150mA),
230v (100mA), 200v (100mA) 150v (50 mA) (that's 68 watts). Needs to be
somewhat regulated (well in the linear supply the 150v is a (tube) regulated
supply).

Now I want any designing to be simple, and from what understand it can be
simple, at least (I'm hoping) simple enough for someone like me who is not
an electronic circuit designer, but has at least might have half a brain to
grasp what might be involved.

Now according to this:

http://www.smps.us/Topologyselector.pdf

I'm seeking a SMPS with flyback, single ended, forward topology.

Okay, I think I'm looking to mess with a circuit similar to this:

https://www.onsemi.com/pub/Collateral/AND8076-D.PDF

Okay, now presuming I'm okay so far, what task have I got?

I think that my only task is in relation to the circuit contained at the
above link is to obtain a transformer, wind the appropriate number of
secondaries, rectify each, add the DC voltages in series and sort out the
smoothing. I beleive the regulation circuitry might already be designed,
although I don't know how good it will be in contrast to the regulation
provided by the old 150v gas-filled regulator. Hopefully no worse.

I'm not sure exactly how to do this, but at this stage I beleive that's my
overall task. IOW, I don't need to alter anything to the left of the
transformer on fig.13, all the changes are to the transformer secondary
-basically the output circuitry.

True? TIA.

 

[John Popelish]

Hi. I'm not an electronic circuit designer, but may have to engage in the
art just a bit.

I'm just wanting some reassurance as to what my task will be.

I'm not sure what what wattage of power supply I'm needing, but I guess
somewhere between 50w and 70w.

The task before me is to make a SMPS with multiple outputs at 250v (150mA),
230v (100mA), 200v (100mA) 150v (50 mA) (that's 68 watts). Needs to be
somewhat regulated (well in the linear supply the 150v is a (tube) regulated
supply).

Now I want any designing to be simple, and from what understand it can be
simple, at least (I'm hoping) simple enough for someone like me who is not
an electronic circuit designer, but has at least might have half a brain to
grasp what might be involved.

Now according to this:

http://www.smps.us/Topologyselector.pdf

I'm seeking a SMPS with flyback, single ended, forward topology.

You are on thin ice, already. Assumption 2 on the right side of this
guide specifies that it applies to output voltages less than 48 volts.

I would think that the most important selection criterion in your case
(that being the supply for a receiver if I remember, correctly) is low
RF generation and ease of filtering to control what ever noise is
created by the supply. The flyback topology is one of the noisiest.

Okay, I think I'm looking to mess with a circuit similar to this:

https://www.onsemi.com/pub/Collateral/AND8076-D.PDF

Okay, now presuming I'm okay so far, what task have I got?

I think that my only task is in relation to the circuit contained at the
above link is to obtain a transformer, wind the appropriate number of
secondaries, rectify each, add the DC voltages in series and sort out the
smoothing. I beleive the regulation circuitry might already be designed,
although I don't know how good it will be in contrast to the regulation
provided by the old 150v gas-filled regulator. Hopefully no worse.

Almost any active regulation scheme that controls the 150 volt output,
directly will almost certainly exceed the line and load regulation of
the gas regulator tube. Noise is another matter alltogether.

I'm not sure exactly how to do this, but at this stage I beleive that's my
overall task. IOW, I don't need to alter anything to the left of the
transformer on fig.13, all the changes are to the transformer secondary
-basically the output circuitry.

True? TIA.

Yes.

The flyback topology links the regulation of all outputs to the extent
that the windings are coupled (by similar volts per turn). Leakage
inductance, winding resistance and rectifier and filter impedance
degrade this common regulation

 

[Richard]

You are on thin ice, already. Assumption 2 on the right side of this
guide specifies that it applies to output voltages less than 48 volts.

Oh drat, I didn't see that assumption.

Looks then that the topology that I was thinking of adopting is no good for
my requirements.

Thanks for pointing this out.

 

[Richard]

You are on thin ice, already. Assumption 2 on the right side of this
guide specifies that it applies to output voltages less than 48 volts.

I suppose that's really significant.

If so, what topology have I recourse to in that chart? Given I'm seeking
250v out. Anybody know?

 

[The Al Bundy]

I suppose that's really significant.

If so, what topology have I recourse to in that chart? Given I'm seeking
250v out. Anybody know?

Maybe this site helps you a bit further:
http://henry.fbe.fh-darmstadt.de/smps_e/smps_e.asp

A flyback can also be used to get higher output voltages, but be aware of
the primary winding voltage (with oscillations!) when the output diode(s)
conduct. This voltage (~n*Vout) + the supply voltage will stand across the
switch. The switch must be able to handle this high voltage.

For example, in a TV a Flyback is used to generate the high voltage,
25..30Kv. So 250v should not be a problem at all if you choose the right
components.



Al

 

[John Popelish]

Oh drat, I didn't see that assumption.

Looks then that the topology that I was thinking of adopting is no good for
my requirements.

Let's say, probably sub optimal.

 

[John Popelish]

I suppose that's really significant.

If so, what topology have I recourse to in that chart? Given I'm seeking
250v out. Anybody know?

Here is a basic switching regulator tutorial that covers the simple
topologies.
http://www.national.com/appinfo/power/files/f5.pdf
Page 11 shows the push pull topology that is quite a bit quieter than
the flyback, and also has pretty good tracking between parallel
secondaries.

The half bridge version on page 13 is better suited for direct line
voltage input, because it produces only half the primary voltage
swing.

The full bridge version shown on page 15 is usually reserved for
supplies will above your 70 watt size.

I think I would go with the half bridge design, because of its well
utilized transformer, which has tight coupling between windings
compared to the gapped core transformer needed for the flyback
version. It is also potentially a lot quieter and more efficient.

 

[Rich Grise]

You are on thin ice, already. Assumption 2 on the right side of this
guide specifies that it applies to output voltages less than 48 volts.

The item I triggered on is that the box at that point says you're seeking
either flyback or single ended forward topology; you've conglomerated
them both into one. I'm no expert, but I'm pretty sure that they're
different.

I read about something called a "forback" in some NASA pub some years
ago, that supposedly combined all of the good features of both forward
and flyback topologies. I don't remember what it said about the less-
good features.

Good Luck!
Rich

 

[Richard]

Here is a basic switching regulator tutorial that covers the simple
topologies.
http://www.national.com/appinfo/power/files/f5.pdf
Page 11 shows the push pull topology that is quite a bit quieter than
the flyback, and also has pretty good tracking between parallel
secondaries.

The half bridge version on page 13 is better suited for direct line
voltage input, because it produces only half the primary voltage
swing.

The full bridge version shown on page 15 is usually reserved for
supplies will above your 70 watt size.

I think I would go with the half bridge design, because of its well
utilized transformer, which has tight coupling between windings
compared to the gapped core transformer needed for the flyback
version. It is also potentially a lot quieter and more efficient.

John

Nice tutorial link. I gather that in principle you could go for 250v with
flyback topology but folks would not recommend it. I'm not technically
aware of why this is so, why a voltage limit but accept the advice. Maybe
I'll come across the reason in some some tutorial.

Like someone pointed out flyback and single ended forward are I beleive two
seperate topologies.

Thanks for the knowleadgeable advice. I appreciate this.

 

[John Popelish]

Like someone pointed out flyback and single ended forward are I beleive two
seperate topologies.

Flyback stores energy in a magnetic field of the air gap in the core
and when the switch turns off, discharges this stored energy through
the secondaries. Flyback topology is used because it is simple. It
is limited by its inefficiency, and the large stresses it places on
switch, rectifiers and capacitors. They also have large and fast
voltage swings that couple high frequency currents from primary to
secondaries (or to shield) and those currents have to be carefully
dealt with to prevent them from becoming RF interference.

Forward converters use the transformer more like an ordinary
transformer (not as a storage mechanism, but as a coupling mechanism
that produces an output voltage in proportion to the applied input
voltage) but drive it only one way. They need a mechanism (usually an
extra winding or an extra switch) to reset the core magnetization back
to near zero flux after each power pulse. But the core has no gap, so
the windings are easier to well couple.

Here is a basic tutorial on the single switch forward converter
topology.
http://www.onsemi.com/pub/Collateral/AND8039-D.PDF

The half bridge converter I mentioned earlier has the extra complexity
of a high side switch but gets almost twice the power through the same
sized transformer.

If I remember correctly, you are looking for an isolated supply that
will eliminate the big transformer in the present supply, so you will
need some topology that includes an isolation transformer.

 

[Richard]

Here is a basic tutorial on the single switch forward converter
topology.
http://www.onsemi.com/pub/Collateral/AND8039-D.PDF

The half bridge converter I mentioned earlier has the extra complexity
of a high side switch but gets almost twice the power through the same
sized transformer.

If I remember correctly, you are looking for an isolated supply that
will eliminate the big transformer in the present supply, so you will
need some topology that includes an isolation transformer.

Yep, ideally it looks like maybe I'm looking for circuit of a
single-switch-forward-converter that includes input rectification and
filtering.

The application note mentions the need for dialectric isolation from input
to output. I don't understand that. Is R10 and R11 negating the isolation
provided by T1 at figure 5, is that why there is no dialectric isolation?

 

[Richard]

Here is a basic tutorial on the single switch forward converter
topology.
http://www.onsemi.com/pub/Collateral/AND8039-D.PDF

Okay. That's good to understand the principles of converting a DC supply
into a DC output. Of course, this is at the heart of all SMPS. But off-line
converter or SMPS circuits will have include input rectification and
filtering and provide isolation.

So, if I'm rying to get pretty much the whole circuitry for a (single ended)
forward convertor, that leaves me trying to get circuits such as:

http://www.fairchildsemi.com/an/AN/AN-4134.pdf
http://www.onsemi.com/pub/Collateral/AND8076-D.PDF
http://www.fairchildsemi.com/an/AN/AN-9015.pdf

No doubt there are others. At least I'm now more aware of what to look for.

 

[John Popelish]

Yep, ideally it looks like maybe I'm looking for circuit of a
single-switch-forward-converter that includes input rectification and
filtering.

The application note mentions the need for dialectric isolation from input
to output. I don't understand that. Is R10 and R11 negating the isolation
provided by T1 at figure 5, is that why there is no dialectric isolation?

Exactly so. Leaving out the isolation mechanism makes it easier to
understand how the regulation loop operates, but requires eliminating
the line isolation.

Here is a paragraph from page 3:
"A cautious note must be now conveyed, this design
example is a non–isolated, high–voltage input power supply.
It is for example only and cannot be built for sale because it
does not meet the IEC (UL CSA or other) specifications for
dielectric isolation and for creepage (the distance along a
surface). To make this an off–line one transistor forward
converter, the input rectifier bridge, EMI filter, an
opto–isolated feedback circuit, and the transformer
would have to be built to IEC specifications."

 

[John Popelish]

Okay. That's good to understand the principles of converting a DC supply
into a DC output. Of course, this is at the heart of all SMPS. But off-line
converter or SMPS circuits will have include input rectification and
filtering and provide isolation.

So, if I'm rying to get pretty much the whole circuitry for a (single ended)
forward convertor, that leaves me trying to get circuits such as:

http://www.fairchildsemi.com/an/AN/AN-4134.pdf

This one does a pretty good pass through the magnetic and feedback
design processes. But it illustrates a design made with a highly
integrated controller that may not be easy to expand upon.

http://www.onsemi.com/pub/Collateral/AND8076-D.PDF

This is a flyback design.

http://www.fairchildsemi.com/an/AN/AN-9015.pdf

This is a pretty complete design that would be easy to scale down a
bit, and to change the output voltages. But the first one shows how
you link the output filter inductors to improve the cross regulation.

 

[Richard]

This one does a pretty good pass through the magnetic and feedback
design processes. But it illustrates a design made with a highly
integrated controller that may not be easy to expand upon.

Actually right now this seems to be my preference:

http://www.fairchildsemi.com/an/AN/AN-4134.pdf
Describes a stet-by-step design procedure for off-line forward convertors
using a "reference" circuit (Fig 1).

http://www.fairchildsemi.com/products/analog/fps_designer_software.zip
Which is made easier if you use this power switch design software. (Includes
AN4100 instruction booklet).

http://www.fairchildsemi.com/collateral/AN-4134.xls
Design aid file for off-line converters.Possibly could use this instead of
power switch design software above. Not examined this.

http://www.fairchildsemi.com/whats_new/fps.html
Fairchilds new product FPS.

http://www.tdk.co.jp/tefe02/e140.pdf
Ferrite for switching power supplies.

http://www.fairchildsemi.com/collateral/fps_psg.pdf
Power switch selection guide.

 

[Richard]

http://www.fairchildsemi.com/an/AN/AN-4134.pdf
Describes a stet-by-step design procedure for off-line forward convertors
using a "reference" circuit (Fig 1).

This is correct.

http://www.fairchildsemi.com/products/analog/fps_designer_software.zip
Which is made easier if you use this power switch design software.
(Includes AN4100 instruction booklet).

This is incorrect I beleive. This application software for SMPS design is
for use with multiple output SMPS *flyback* type. The program initializes
at start up with the sample design of a 19W set top flyback SMPS as shown in
AN4106 (Multiple Output Swtched Mode Power Supplies Using Fairchild Power
Switch). See the default start-up conditions in appendix A in AN4100 the
instruction book. (Not sure why the schematic in the application does not
show multiple outputs as per figure 11 in AN4106.)

http://www.fairchildsemi.com/collateral/AN-4134.xls
Design aid file for off-line converters.Possibly could use this instead of
power switch design software above. Not examined this.

This is the design aid for use with AN4134, involving design of off-line
*forward* convertors.

I don't think you can use that niffy designer software application to design
a forward converter using the reference circuit shown in AN4134. Unless
somone knows better.

 

[John Crighton]

Hi. I'm not an electronic circuit designer, but may have to engage in the
art just a bit.

I'm just wanting some reassurance as to what my task will be.

I'm not sure what what wattage of power supply I'm needing, but I guess
somewhere between 50w and 70w.

The task before me is to make a SMPS with multiple outputs at 250v (150mA),
230v (100mA), 200v (100mA) 150v (50 mA) (that's 68 watts). Needs to be
somewhat regulated (well in the linear supply the 150v is a (tube) regulated
supply).

Now I want any designing to be simple, and from what understand it can be
simple, at least (I'm hoping) simple enough for someone like me who is not
an electronic circuit designer, but has at least might have half a brain to
grasp what might be involved.

Now according to this:

http://www.smps.us/Topologyselector.pdf

I'm seeking a SMPS with flyback, single ended, forward topology.

Okay, I think I'm looking to mess with a circuit similar to this:

https://www.onsemi.com/pub/Collateral/AND8076-D.PDF

Okay, now presuming I'm okay so far, what task have I got?

I think that my only task is in relation to the circuit contained at the
above link is to obtain a transformer, wind the appropriate number of
secondaries, rectify each, add the DC voltages in series and sort out the
smoothing. I beleive the regulation circuitry might already be designed,
although I don't know how good it will be in contrast to the regulation
provided by the old 150v gas-filled regulator. Hopefully no worse.

I'm not sure exactly how to do this, but at this stage I beleive that's my
overall task. IOW, I don't need to alter anything to the left of the
transformer on fig.13, all the changes are to the transformer secondary
-basically the output circuitry.

True? TIA.

Hello Richard,
here is an article you might like to have a read through
in addition to all the other information you have gathered.
http://www.shlrc.mq.edu.au/~robinson/Information/Inverter_4W.html
It shows how to build a low power inverter for valve radios
using a salvaged controller IC TL494 from old computer
power supplies. You could play around and beef this one
up a bit to suit your needs.
Regards,
John Crighton
Sydney