Hello, Sotiris here! 👋

🚀 First off, I want to show you the error I made when designing this transformer using the standard equations. Then we’ll continue with the PSFB transformer design, using our online tool. We will discuss the winding specs in this newsletter.

How many of you know how to calculate the turns of a PSFB transformer? 🧐  

This is Faraday’s equation, applied to the transformer primary winding:

Now the first thing we need to choose as a starting point is the effective duty cycle.

👉   Don’t get confused 😵, the driving signals’ duty cycle of each MOSFET in a PSFB are always fixed at 50%, the only thing that changes is the phase between the switching waveforms.

👉   That way, the time the transformer swings positive/negative is controlled. So, the transformer input waveform duty cycle varies theoretically up to 100%. In practice dead times and propagation delays decrease the maximum duty cycle achievable.

Usually, application notes suggest operating at 60-70% effective duty cycle. Having selected an effective duty cycle, from the equation above, we can find the turns ratio (n) easily.

👉   Going back to Faraday’s equation, notice that A_e is known from the core spec sheet, so the last thing to choose is the peak swing flux density.

And that’s where the confusion started…

🔑   The way I think about Faraday’s equation is the following:

“How many turns a winding with ΔB_pk flux swing density_, with a core of A_e effective area, with that many Vs (volt-seconds) applied to, has?”

Looking at the primary voltage/current waveform:

Notice that in each half-period the polarity changes. So, to apply Faraday’s law we look at the first half-period (notice the T/2 in the equations).

Question: “What happens with ΔB_pk?”

Answer: “When the systems is in steady state the flux density B, swings from -B_pk to +B_pk”

BUT, I got confused because I imagined the system just being “turned ON” and then I looked at the first half period. I said that the flux density will swing from 0T to B_pk so ΔB_pk=B_pk, which is not wrong, but the effective duty cycle is not the same, as in steady state. We are talking about just the first period(s), where the system is in transient mode. That was the mistake.

Lesson for me: Always check the conditions in steady state!

Missing a division by 2 changed everything in the transformer design 😵… Thankfully the tool calculated the correct B_pk and made me think and go over the equations again!

💥 PSFB transformer design now💥

Design stage 4:

Looking at the selections I made back in 2019 I see some good and some bad choices now. A good choice is that I used the PSPS (prim.1/2 – sec.1 – prim2/2 – sec. 2) configuration for my transformer which minimizes severe proximity losses (which are still a lot..).

👉 I see low current density in the primary windings which is always good, although I could have pushed for more using a smaller wire.

👉I see that the tool warns me that this winding config may not fit the bobbin, which I did found out in real life 🤭

 Looking at the secondary winding:

👉 Tοo much current density there, should have gone for a bigger Litz wire, the problem was the bobbin height wouldn’t allow for this. The total losses are high, and a hotspot of 86°C don’t get me excited to be honest. We will see…

Now take a note at the leakage inductance which is also calculated. A calculation of this parameter is crucial for the PSFB topology to ensure ZVS conditions, but there is no way to know this with a couple of equations. You just have to build it and see…

✅ Using our AI tool will enable you to have a pretty accurate prediction on this parameter. Enough to make a decision on a design, using or not a “shim” inductor beforehand, to be precise.

✅ Please check out here to see how we use artificial intelligence and classical models to get accurate results!

NEXT ->ORDERING SAMPLES!

All we need to do having reached design stage 4 is to hit verification/sample order.

✅   Then our engineers examine your projects specs. We care about manufacturability and red flags in the parameters specced in the tool.

✅   An engineer assigned to your project will review the design and give the final ok to move forward to sample creation, or he will propose changes shared with you, which will can be adopted in the final design in agreement with you!

The nice thing is that you can save time and money avoiding iterations because of mistakes. Iterations take weeks when you have to make purchases for materials every time…

A magnetics expert engineer will easily spot the mistakes.

Why not have someone experienced at this take care of it, sending you the finished thing with a spec sheet at your facilities?

🚀 🚀   Frenetic labs have a very nice stock of every part needed to design an inductor/transformer with the tool. When you build a prototype, you care about moving into production as well! The team is here to ensure smooth transition from the samples stage to volume manufacturing! 🚀 🚀 

It’s like ordering pcbs from a fab house VS making them at home. There is just no comparison...❗

WHAT’S COMING NEXT WEEK?

👉 Next week we’ll check out the sample transformer we designed in these last newsletters!

👉As always, if you have a question or an idea, let me know at sotiris.zorbas@frenetic.ai

Previous Newsletters: here!

LAST BUT NOT LEAST...

In previous Newsletters I told you about the High Frequency Magnetic Training Program.

The first session (duration is one month) finished last week. Check out what one of the participants said... 😍 

The next session starts on June 1st., yes..very soooooon!

We have still 4 available seats, do you want to get the Frenetic SuperPowers?? All details here. Hurry up! 🚀

 👋  Until next week 👋 

Sotiris Zorbas, MSc 

Power Εlectronics Εngineer