transformer design for smps

An isolated synchronous buck converter is a kind of forward converter, and that's what this is.

That auxiliary winding on the transformer supplies a DC voltage that's roughly a diode drop and some resistance loss below 1/4 of the input voltage. Every time the main transistor turns on, the dotted end of the auxiliary winding jumps up to (turns ratio) * (input voltage), more or less. The 200 ohm resistor insures no huge current spikes, the diode rectifies, and the result is an unregulated power supply to the chip.

--
Tim Wescott
Control system and signal processing consulting
www.wescottdesign.com

n V y V e

Now, a forward converter usually has some kind of reset winding right? But, would I be right to think that this circuit doesn't need the reset winding because there are bidirectional switches that allow for some current to keep flowing into the secondary, to keep that flux going and not violate anything? I think that's right... so I think I'll approach my transformer design then assuming I'm designing for a forward converter... Vo =3D nDVin

much thanks, appreciate the help

wait, I just looked at this circuit again, and i thought that in order to keep the flux flowing through the transformer core that current had to be able to flow into one of the dots at all times, but when i analyze this circuit for the switching transistor open, I believe this would force Si4800DY to turn on and Si4884DY to turn off, which would break the circuit for the bottom end of our second winding... so I don't see current able to flow into any dot of this transformer during the off cycle

I'm not sure what you're thinking. The flux in the core is equal to the sum of the turns * current for all the windings -- so as far as the core is concerned, the flux is the same whether there's 1A in a 1-turn primary, or 0.1A in a 10-turn secondary.

In a forward converter current comes out of the secondary (or secondaries) proportional to the current going into the primary, for the most part, except for magnetizing current.

--
Tim Wescott
Control system and signal processing consulting
www.wescottdesign.com

e e

I was just pointing out that there needs to be a path for current to flow into the transformer in order to support the flux in the core at all times. When current flows into the dot of a winding it's able to support the flux. This schematic has an error, during the off portion of the PWM there is no way for current to enter a dot, no way for current to support the flux.

Because it is a flyback. Don't model the 'transformer' like a transformer but like an inductor! The transformer has an air gap in which the energy builds up. When the primary FET opens, the inductor wants to keep the current flowing. The only way is through the secondary coil.

--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------

It's not a flyback- the relative 'polarity' of the windings is wrong for that.

It's a resonant-reset forward converter- the 'missing' current flows are through parasitic capacitances.

Best regards, Spehro Pefhany

--
"it's the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

I think you're right. The extra output inductor should have rung a bell :-) IMHO its not the easiest design to start with. The OP needs

5W output power. A flyback is the easiest topology to achieve his goal.

--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------

If he wants to give it a try, here's a relevant article:-

Best regards, Spehro Pefhany

--
"it's the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

"The Journey is the reward"

eff.com

thanks for the article, I believe it's a forward converter as well... I've let myself slip into a long back and forth in another forum with someone who keeps arguing it's a flyback. That person doesn't care about the inductor at the output

but i've analyzed the circuit, it's a buck derived topology Vo =3D DVin, it can't boost

have you guys designed a transformer before? there seems to be a lot to it

The dots on the windings tell you immediately by inspection that it's not a flyback.

Yes, a few (including flybacks), but it's been a while. You can pick up what you need in an intensive weekend once you have the info in hand, getting the cores and bobbins usually takes longer. I agree with Nick Coesel that a (lower frequency) flyback is going to be an easier way to go.

Best regards, Spehro Pefhany

--
"it's the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

IMHO you can think of a forward converter like a buck/boost converter with the transformer doing a current conversion between input and output. This helps to figure out the output inductor.

I did a couple of months ago. This application note has a very clear recipe on how to determine the core size and number of windings:

I ordered a whole bunch of core size and material types and wound several prototypes to get a feeling what would be the best solution.

One thing the remember: a big part in the transformer design procedure is driven by the number of windings a core can hold.

You'll also need a current probe. Otherwise you'll flying blind.

I made a couple of current probes myself using this transformer loaded with a 5 Ohm resistor which gives me 100mV/A.

You'll need to use stranded wire to keep the inductance low. Having multiple probes allows you to see various currents throughout the circuit.

Its fun if you have the time.

Anyway, I would use a flyback topology with an off-the shelf transformer for your project.

--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------

"The Journey is the reward"

eff.com

Thanks for the info. I'm not really looking for easy I'm looking for efficiency, do you think I would gain efficiency with a flyback topology? I've been reading that they are not as efficient as other dc/dc converters, but if they were more efficient than the one I'm pursuing I'd give them another look. If the efficient one happens to be easier that of course is nice, but it's not as important as efficiency.

Its easier because you can use available parts (very little math involved) and cheaper due to fewer parts (the transformer is both transformer and storage inductor). The efficiency doesn't matter much.

5% of 5W is 250mW difference. I'd look into the Topswitch devices I mentioned earlier. Many transformer/inductor manufacturers have off the shelf transformers for these chips.

--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------

AFAIK a resonant converter (zero voltage switching / zero current switching) is the most efficient converter of all. Over 95% efficiency should be possible. But I recall there are some issues with these converters regarding transient behaviour. Maybe someone came up with a better topology.

--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------

co

rth=3D

Thanks, I've looked through power integrations stuff, and had no luck finding parts that can handle my input voltage range 70V-105V

You need to dig a little deeper. For instance the minimum operating voltage for this device is 50V:

All their examples and components values are based on off-line converters but that doesn't mean they can't be used for a DC-DC converter. Perhaps you should try and contact their technical support.

--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------

s

you know i see a lot of stuff like this, and they have the 85Vac -

120Vac input range, which mutiply by sqr root of 2 gives me a Vmin input 120V, which is too high for my Vmin =3D 60V need.... but I don't know, maybe it would work? I'd like it to be within spec of the datasheet though, within normal operating range