Ferrite core transformer design step by step

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I am looking for practical design of transformers ( ferriet core). its considerations, selection calculations step by step.

for example

a transformer to work on PWM. where duty can vary from 30% to 99%.
frequency of operation is 60khz.
primary voltage is : 21.6V secondary peak voltage is 400V. the ouptut current should be 5Amp.

so if I take above as a input for ferrite core transfrmer design, then what will be the practical steps which need to be followed.

a) like based on Fq, selection of material.
b) size of core.
c) winding turn calculations with wire thickness.
d) something inductor value of the transformer coil( what is its significance)

I have tried to read other threads too they give some information but not enough to built your own. I have also tried to read online information.
but feel lost and at the end feels like back to the square one.
 
What are you building?
Sounds like a big transformer. Can the 21.6v source supply the 2000watt load your adding to?

Take the little bits of information you can find and put it all together. Like a mosaic.

a) I don't know
b) I don't know
c) The turns is a ratio, there are lot's of threads here for calculating the turns ratio of a transformer. Wire thickness will depend on the VA needed and I don't know. Bigger wire for more current.
d) I don't know what your saying. Do you want to know the inductance of this imaginary transformer?

The threads you have read must have some information, find specifications of already made transformers similar to yours, this will also lend some clues.

There is no "step by step" that I have come across, I have not looked hard. However I have seen the odd chart pop up now and again with lot's of useful information.
 
Oh, ohh... Toni's first post translates to me like:
...optimized transformer design, especially for higher power, cannot be taught by three sentences in a few postings.

I would support this view.
Despite their simple appearance magnetics are a complex design topic.
In addition to the fundamental theory + HF effects a transformer always is key component of a certain circuit. Depending on the circuitry you will need remarkably different designs. Last but not least safety isolation can be a serious concern.

But even if someone of the experienced members would agree on giving you many hours of support for your specific application, means doing the design for you and explain the reasoning behind... - without detailed description of the intended circuitry and mode of operation nobody can help you.

If you are looking more for a general learning about magnetics, then there is no way around books and further online search and practical verification of multiple designs.
 
@ dtproff

Thanks for the pointers. I managed to get the book you suggested. but still I am toddling to understand it.

friends,more helpful words on the topic would be welcome.

Tell you what I will do... If you will work with me over the next few days/to a week I will walk you through the design process step by step.

I need a couple things first...

Peak Power Rating
Input Voltage range
Output Voltage

I will then guide you through the various topology decisions, design choices and a detailed walk-through of the math and what it means.

How about it?

Tony
 
ChocoHolic is right on the complexity of the power supply.

There was a time several years ago that a few people spent some time keeping me on track so that I didn't go off in the weeds. I am especially thankful to a friend of mine who was one of the best magnetics engineers I have ever had the privelage of working with.

So maybe this is a chance to give back.

Tony
 
1. Check acceptable specific power losses per volume vs. freq and dB, probably you end up around 200mT@50kHz, 150mT@100kHz
2. Guestimate the core
3. dB= dV*dT/ 2*N1*Ae => calculate N1 and from desired turns ratio N2, Ae depends on a shape of a core and can be found in a datasheet
4. Check the window size and how thick the windings can be. Calculate copper losses for max. current P=(ro*lenght/S)*I^2
5. Now you may see if you've chosen too big or too small, if necessary start over again

hint: when core losses (specific losses*volume) and copper losses are comparable you've probably done a good job
 
If you follow the MathCAD file I sent, you will be able to run your converter at 370mT and there won't be any guesswork involved.

The power loss curves in the data sheets are for forwards and not Flybacks. And finally, none of the above make gap loss estimations (again see my MathCAD file) or estimations on the deltaT rise of the core given specific loss combinations.

I am not saying these calculations are perfect because I am absolutely certain that there are probably a few boneheaded mistakes in there somewhere. That being said, I developed these over the past 10 years or so and on average, the designs come within about 10% of actual full load results.

Rule one of magnetics is that you can run ANY power level with any core if you can get enough turns on the core with the correct wire gauge. Again, if you follow the MathCAD file you will find hints on the winding Window, skin effect, wire current density, Reverse recovery on the output rectifiers, Mosfet losses, ect...

Again... there isn't any guesswork in this. If you follow the design methodology all the way to the end, when you get done you have a professionally done power supply that is pretty much bullet proof because it was engineered that way from start to finish.

Number 3 appears to be for a forward.. If that is the topology that is intended don't forget to calculate the Bac and Bdc of the design (assuming CCM operation) also remember that the leakage can also saturate the transformer core just the same way leakage in a common mode choke can be saturated. So you'll need to make sure you gap the core a bit. Gapping it a bit will also help compensate for any flux walking that occurs if the core didn't get fully rest on the last cycle. If its a forward, you'll need to make sure the output inductor is designed correctly. If he is doing audio then in all likelyhood he will be using a coupled inductor on the output to improve the cross regulation on the outputs. If someone is interested, you can look up some of my other posts and find the MathCAD for the Forward with the coupled inductor. Again... no guesswork involved.

Darkfenriz... no offense intended on my comments above. I just don't want anyone to think you just throw the stuff on the board and boom you are done.

Just throwing in my little bit of $0.02.

Tony
 
Thankyou all for the guidance.

well as Tony said first things first. we strat with more detail spect of transformer.
To start with we go with simple transformer.

1) step up transformer. for full bridge configuration.( not sure if you asked me for this but I guess no you didnt want this input at the moment.)
2) input voltage max DC = 12V
3) input voltage min DC = 10V
4) output voltage @ 10V DC in = 100V
5) Max Dutycycle is 90%.
6) VA = 100.
7) ie peak current = 1 amp @ output.( hope the peak current needed for calculation is for secondary).

Tony hope I have passed on necessary information.for my first lesson.:)

& I havent got the mathcad file. also to use the mathcac file do i need to download any specific Software.
 
Here's the first part...

A standard full bridge will not go beyond a 50% duty cycle. First we assign 45% for the duty cycle. Next we will select suitable MOSFET's and output rectifiers. There will also nee to be a current sense resitor.

This means that you will have 100W for the output, +20% for losses in the system and safety margin. so that 120W, this means 12A@10V.

Basic equation for the turns ratio of a full bridge n=(Vo+Vfd)/(0.9*(Vinmin-2*RdsonMOSFET*Imax(12A)-Vcurrentsenseresistor)) round this to the nearest turn.

For the MOSFETS you will have I2R losses that will be high so I would select a MOSEFET in the 20V range that has a <10mOhm Rdson so that you endup with about 1.2W on the FET. You will need 3-5W heat sinks for each. I imagine you will be in a TO-220 package.

The output rectifier will need to use Ultrafast parts. Assuming this is a center tapped output the rectifier will see 2X the max Vo. Assigning 25-50% voltage derating to the part means you need a 250-300V part. You have 1A/0.45 max current on each diode on the output but I would probably use 3A parts for thermal derating. You will be looking for something with a Trr in the <30ns range.

I'll let you do the parts selections and calculations above. Let me know what you choose and the the turns ratio. I'll take a look at your part selections and tell you what I think.

To anyone else reading the thread. Please review and comment anywhere you think we go wrong. I am curious as to what others have to say.

Tony
 
I am using microcontroller to generate PWM. and would be driving H-Bridge. the pwm would vary from 0% to 99%. but I just learnt that the max dutycycle allowed should be 90%. the 10 % should be for self discharge of the core.

using this variable PWM I would be generating Sinewave at the outputfilter of transformer.

I have been able to simulate the whole design using ideal transformer. but when it comes to real life I would need a real transformer.
 
Thanks Tony,

But lets assume I get a mosfet with is ideal on earth with Rdson of 0.00E. and so no power loss. and thus there is no losses in switching device. ( later I will adjust values of losses with the transformer.)

so we say for 100VA adding losses of 20% it comes to 120VA. and dutycycle of 45%. ie 12A @ 10V

next..... pls..

Will this mean 45% means Ton = 45% and Toff = 55% ?
 
Last edited:
Darkfenriz... no offense intended on my comments above. I just don't want anyone to think you just throw the stuff on the board and boom you are done.

Not taken ;)
You are talking about coupled inductor design (aka. flyback 'transformer', energy storage device).
I have given a formula for transformer as used in pushpull, halfbridge or full bridge topologies, which should be a good starting point for someone who's just asked how to design one on an internet forum.
 
Thanks Tony,

But lets assume I get a mosfet with is ideal on earth with Rdson of 0.00E. and so no power loss. and thus there is no losses in switching device. ( later I will adjust values of losses with the transformer.)

so we say for 100VA adding losses of 20% it comes to 120VA. and dutycycle of 45%. ie 12A @ 10V

next..... pls..

Will this mean 45% means Ton = 45% and Toff = 55% ?

You will use two synced PWM output 180 out of phase (hope i used the right digital terminology there). They must be locked together. Yes that is indeed 45% on and 55% off. You can do a phase shifted full bridge but that is a much more advanced control that I think you want to go for.

Whose simulation package are you using? PSIPCE (My favorite sim package is solder!:D)?

Tony
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.