Help understanding of SMPS transformer characteristics

[sahko123]

mmmmmmmm......


Depending on how wide the Max and Min loads can be apart from each other it might make designing/using an smps be a waste of time since the variation would probably be anywhere between 5W to 120W peak power usage. I think this basically forces me into using a mains transformer and a secondary side power regulation......

Would that wide a power range be too dangerous/impractical to operate the smps at?

[bdunham7]

Would that wide a power range be too dangerous/impractical to operate the smps at?

5 to 120W shouldn't be a problem for any garden-variety SMPS, 5W minimum load should be plenty for any general-purpose design.  It's done all the time and one of the features of the controller you've picked is low no-load power consumption.  Not using opto-isolator sensing might be the one thing that would make regulation with wide power ranges more difficult--why do you want to eliminate the opto?  Even cheap-ish, crap-ish commodity SMPS designs use them. 

Reading this has given me a bit more confidence. I think the combo I have going should work™.

I'd have even more confidence if you'd picked a controller from the list in the transformer datasheet!    But if you just use one of the schematics in the datasheet, mutatis mutandis, perhaps it will fire up.  If you do, I'd be very curious to see what frequencies it ends up operating at under various load conditions--I'm not sure that transformer will work well at 500kHz just based on its size.

[sahko123]

The driver the transformer recommends for 140W operation is basically not available anywhere but ebay and comes in a massive pentawatt package. Seems like this is more effort than its actually worth. I dont really want to buy an off the shelf smps cause then i'll have to pay for that myself otherwise work will pay for components.

[TimNJ]

Are you admitting to embezzling work funds on a public forum?  Anyway, I think there is a lot of useful advise already given in this thread - suggest you really read through it again. 140W power supply can be as little as $50USD for a decent one.

https://www.digikey.com/en/products/detail/mean-well-usa-inc/RPS-200-24/7706036

I'm not sure what size controller/MOSFET you are expecting. 140W flyback is going to require a heatsinked power MOSFET so a "Pentawatt" (TO-220) sounds about right.

[bdunham7]

The driver the transformer recommends for 140W operation is basically not available anywhere but ebay and comes in a massive pentawatt package. Seems like this is more effort than its actually worth.

https://www.newark.com/power-integrations/top247yn/ac-dc-conv-125w-40-to-150deg-c/dp/93K0787?gclid=Cj0KCQjwr-SSBhC9ARIsANhzu17ks2rNJSTx89mvWmTcd-Ki7aE53cgXuFd2HUTJT0OaKNEvAjZbUfUaAqtUEALw_wcB&mckv=s_dc|pcrid|486536532942|plid||kword|top247yn|match|p|slid||product||pgrid|113542825414|ptaid|kwd-10769551587|&CMP=KNC-GUSA-GEN-SKU-Semiconductors-Semiconductors-ICs&cmp=KNC-GUSA-GEN-SKU-Semiconductors-Semiconductors-ICs

https://www.farnell.com/datasheets/1498742.pdf

I don't think you're going to find a solution that is easier or cheaper to implement than that for a MyFirstSMPS^TM.  Just throw it together and it might work.  And a 140W flyback is going to need to dissipate 40W+ power unless you are very successful in your first try and somehow achieve some remarkable efficiency.  I'm sure there's a large heatsink and likely a fan involved no matter what you do. 

[sahko123]

Yeah going through the datasheet this TOP2XX series of switchers has got basically an entire design procedure and more than just the to-220 (penta-watt) package. Its literally the perfect controller/driver cause you can get it for power ratings from like 10W to 250W with active cooling don't know why I struggled to find a decent datasheet.

[sahko123]

Also what limits the power rating of a transformer? Is it a matter of how its driven or is there a characteristic inherent to the transformer that decides it's theoretical maximum limit for power or does it involve both?

I understand there is a saturation current but that wouldn't mean that if, for example a 4A saturation current with a primary side DCV of 340Vdc wouldn't give you a potential of 1360W wouldn't it?

So essentially to what extent is the topology a limiting factor and to what extent is the transformer the limiting factor?

[T3sl4co1l]

As mentioned, maximum power depends on loss.  You can crank more voltage by raising frequency (up to the peak power point*), but this raises core loss (significantly).  You can crank more current but the wire gets hotter, at least for transformers.  Inductors, you can't get much beyond saturation, as the nonlinear inductance is messier to work with and likely the core loss limit isn't much beyond.

*Depending on transformer design, like, winding strays.  At a high enough frequency, capacitance or transmission line effects dominate and gain and power handling will drop.  Also current flows will be very nonuniform over the windings, making heating even worse.

Normally, a transformer is designed for nominal ratings, so, some tolerable heating or efficiency spec, at operating frequency, current, voltage and so on.  It can be used at other values, but gets proportionally worse in the process.  Flyback being the least critical, so the voltages there are kind of a suggestion, but you really want a proper part if you're using it more than say 2:1 off the suggested rating.  Or for forward, obviously if you need more voltage, you just can't get more than full duty cycle would give, but less is fine.

Tim

[sahko123]

So now I guess the last thing there to do is make the circuit with the transformer I found and the TOP271 controller and just see how much current I can pull out of it before it either pops or begins to drop voltage at an uncomfortable rate. 

[strawberry]

no voltage drop should happen (only due to wires) . overcurrent protection will shut off SMPS .
too big device TOP271 will have too high minimum current limit setting . core will go into saturation and MOSFET will POP instantly .
no chance pushing limits out of flyback without proper math .

[David Hess]

The saturation current is listed at 7.4A for I assume the primary, how would this relate to the maximum power I can get through the transformer?

The saturation current has nothing to do with power capability.

The maximum power depends on core losses which depend on frequency and peak flux.  There should be a series of curves from the manufacturer in the specifications.  Power is proportional to frequency, but because core losses are also proportional to frequency, peak flux must be lowered which lowers the power.

I get it would be easier and safer to buy an off the shelf power supply but the reason I'm trying to design one is to learn how to do so. If at some point I give up on designing one I will simply buy an off the shelf one (but end up learning very little to nothing). I have no issue with working at high voltages and have worked at higher voltages before.

The plan is to put together a design, prototype it, abuse and test and qualify it for the task, if it fails either fix whatever issue arose or if its inherently broken ditch it and buy an off the shelf one. At least this way I learn from my mistakes. The whole project is a learning experience after all.

Consider prototyping a design at a lower input voltage using a standard 60 Hz power transformer for isolation.  I also recommend wearing a face shield during testing if you use plastic power packages.

[sahko123]

The datasheets are very lackluster. I've found two different variations both of which a max peak current of 4A so I assume that the saturation current of the transformer. And going on the manufacturers website it states the transformer is 90W and not 120W/140W like other retailers say. So it probably isn't even worth laying money down for at this point

https://www.tme.eu/Document/078417cb0a3500f1b3b48fc4dd20480a/74060-Datasheet.pdf

https://www.farnell.com/datasheets/484382.pdf

I don't understand why this has to be so difficult.

[strawberry]

90W@66kHz , 140W@132kHz
inductance seems too high for universal line voltage at 132kHz

ETD39 DIY transformer would cost half that of MYRRA

[TimNJ]

Try simulating first. A lot of your questions and confusion comes from not really understanding what you are trying to do.

https://www.analog.com/en/products/lt8316.html#product-tools

Open the file:
LT8316 Demo Circuit - µPower No-Opto Isolated Flyback Converter (100-600V to 12V @ 1.5A)

You can visualize some of the standard flyback waveforms, adjust parameters and see what happens.

[jonpaul]

Rebonjour à tous
Just seeing this

Our experience since 1970s,  at high power SMPS transformers.....

Saturation is not approached, even at max primary volt seconds.

Power handling is set by worst case hotspot, and insulation class, (classe 105, 130, 150) and regulatory compliance

The old rule of thumb was design worst case load, etc, for equal copper loss and core loss. 
Nowadays that is not an optimal ratio,  but a good starting point.

Of course in a high frequency SMPS transformers or inductor  Pcu must include IsqR, skin and proximité effect, and consider the substantial temp coefficient of the copper.

Thus a transformer's design power rating is not quite as simple as mentioned.

Most ferrite core and materials specs include curves of,losses vs excitation and temperature, to,find,an optimal hotspot temperature as well  as flux densité in typical application.


Check spec of core matériel and core shape from manufacturer, like TDK, NiCera, Philips, Siemens

Old catalogues should have that information

Finally the only Simulator and CAD,we have used was,Intusoft, Magnetic Designer, available gratuit.

Bon courage


Jon

[sahko123]

Update:

I have fallen into the rabbit hole. I bought the Wurth Elektronik: Trilogy of Magnetics and oh boy.....

So for the initial project (A JBC T245 soldering iron base station) I ended up buying a power supply that is a good fit both physically and electrically. But I shall carry on falling into the rabbit hole by designing and winding a transformer based on my reading of the Trilogy of Magnetics (Amazing treasure trove on both power supply design and transformer winding). Probably at first I'll try get it to work at 48V input or something lower and work my way up designing for offline supplies and the likes.

I'm basically going to keep digging down until I reach the center of the earth.