What magnetic design methodologies do you use?

Thread Starter

HRN88

Joined Jul 21, 2020
24
Hi guys, this post is related to share tips for magnetic design for SMPS like inductors and transformers for high frequency.
I'm a novice, I understand the magnetic theory and I've designed some inductors for buck and boost converter, nevertheless, I rarely achieve to match the number of turns with the calculated inductance and the gap.

Finally all my process is a mix of calculation try and error. Usually I use the method area product of a selected ferrites, that method is showed in the book "Transformer and inductor design, Colonel Wm. T.McLyman". Recently I´ve used an easy relation of Al (inductance factor) some manufacturers provide this value.

The approach is very different, and I don´t know which one is better, or if there are another methodologies.

So, what do you think about?
Which methodology do you use, can you share it?
Can you share some tips about your process design?

Cheers!
 

LowQCab

Joined Nov 6, 2012
2,342
Just a few days ago I was on a website claiming to be on "the cutting-edge" of Magnetics,
and they were claiming, ( in there own promotional material !!!!! ),
that there's no truly standardized measurements, and design procedures, that every one can agree on.

I'm just a life-long hobbyist, not a professional, but I've held this same view-point for
as long as I've had an interest in Magnetics, ( off and on for about ~10-years ).

It seems that the design procedure amounts to making the best estimation that You can,
then Cut-&-Try, Cut-&-Try, until You get the results You are looking for.

The biggest problem that I see, is that,
everybody wants their project to be microscopic, and operate at Mhz Frequencies, for higher efficiency.
OK, fine, I get it.
But the easy way is, ( very generally speaking ),
build it twice as big and heavy as You think it should be.

There is no substitute for "Core-Volume"............
Every manufacturer has their own prioritized "secret-sauce" Core-Materials and
their own special numbering and lettering system for those Core-Materials.
Some manufacturers do try to "standardize" things, but You still have to
just go by what they "claim" a particular Material is "best suited for".
None of the numbers or performance claims are "guaranteed".

Everything is a compromise.

Most, if not all, of the manufacturers do try to come up with superior performance products,
but as to whether or not their particular product lives up to their claims, is always open for discussion.
This is because there are so many variables to take into consideration.

It's virtually always better to go with a complete and tested design by a reputable manufacturer
than to attempt to cobble-up your own design.
But in certain circumstances, You may need a "one-off" design.
I'm still working out the details on a 12-Volt DC-DC 60-Volt, 8-Amp Output Supply,
because I can't find one commercially available,
it's not as easy as it looks,
the Transformer is 3-times the size I'd "like" it to be,
but that makes it cheaper, easier to build, less critical, and more stable overall.
It's capable of twice the power that I expect, that way nothing runs under heavy stress or heat.
I'm planing on epoxying together 2 commercially available Toroidal-Cores to
achieve "more than adequate" Inductance.
The simple calculations so far have come out to an astounding ~1.4-Volts-per-Turn,
and less than ~40ma Idle-Current, ( PWM-Push-Pull-Transformer ).
This has been achieved without "Doctorate-Level" Math calculations,
but I'll probably find 2 or 3 "unpredictable" problems before I'm satisfied with the performance.

I look at Magnetics as being a sort of "high-speed-mechanical-device" interfacing with Electronics,
kinda like a Motor,
only there's no moving parts that You can actually see working.
.
.
.
 

DickCappels

Joined Aug 21, 2008
9,092
I know one fellow who uses a resonant circuit that oprates at high current. He frequently complains that the resonant frequency of his circuit is significantly higher than suggested by the measured values of his L and C. The problem is that he measures the inductance of his L using only a few milliamps at a standard frequency but his circuit circulates amps at a much higher frequency. Inductors behave differently at different current levels and at different frequencies -harmonic content counts also. This leaves the designer making a best estimate similar to what LowQCab mentioned in his post #2.

There are many things that can make inductors perform differently from what you would expect using the datasheet.
 

Papabravo

Joined Feb 24, 2006
18,982
I know one fellow who uses a resonant circuit that oprates at high current. He frequently complains that the resonant frequency of his circuit is significantly higher than suggested by the measured values of his L and C. The problem is that he measures the inductance of his L using only a few milliamps at a standard frequency but his circuit circulates amps at a much higher frequency. Inductors behave differently at different current levels and at different frequencies -harmonic content counts also. This leaves the designer making a best estimate similar to what LowQCab mentioned in his post #2.

There are many things that can make inductors perform differently from what you would expect using the datasheet.
If you have ever swept an inductor on a VNA and realized that it will turn into a capacitor you won't ever be able to look at them the same way again.
 

atferrari

Joined Jan 6, 2004
4,650
I know one fellow who uses a resonant circuit that oprates at high current. He frequently complains that the resonant frequency of his circuit is significantly higher than suggested by the measured values of his L and C. The problem is that he measures the inductance of his L using only a few milliamps at a standard frequency but his circuit circulates amps at a much higher frequency. Inductors behave differently at different current levels and at different frequencies -harmonic content counts also. This leaves the designer making a best estimate similar to what LowQCab mentioned in his post #2.

There are many things that can make inductors perform differently from what you would expect using the datasheet.
I recall reading here in AAC that certain manufacturer stated in the datasheet the actual current used to characterize their inductors.
 

Thread Starter

HRN88

Joined Jul 21, 2020
24
Just a few days ago I was on a website claiming to be on "the cutting-edge" of Magnetics,
and they were claiming, ( in there own promotional material !!!!! ),
that there's no truly standardized measurements, and design procedures, that every one can agree on.

I'm just a life-long hobbyist, not a professional, but I've held this same view-point for
as long as I've had an interest in Magnetics, ( off and on for about ~10-years ).

It seems that the design procedure amounts to making the best estimation that You can,
then Cut-&-Try, Cut-&-Try, until You get the results You are looking for.

The biggest problem that I see, is that,
everybody wants their project to be microscopic, and operate at Mhz Frequencies, for higher efficiency.
OK, fine, I get it.
But the easy way is, ( very generally speaking ),
build it twice as big and heavy as You think it should be.

There is no substitute for "Core-Volume"............
Every manufacturer has their own prioritized "secret-sauce" Core-Materials and
their own special numbering and lettering system for those Core-Materials.
Some manufacturers do try to "standardize" things, but You still have to
just go by what they "claim" a particular Material is "best suited for".
None of the numbers or performance claims are "guaranteed".

Everything is a compromise.

Most, if not all, of the manufacturers do try to come up with superior performance products,
but as to whether or not their particular product lives up to their claims, is always open for discussion.
This is because there are so many variables to take into consideration.

It's virtually always better to go with a complete and tested design by a reputable manufacturer
than to attempt to cobble-up your own design.
But in certain circumstances, You may need a "one-off" design.
I'm still working out the details on a 12-Volt DC-DC 60-Volt, 8-Amp Output Supply,
because I can't find one commercially available,
it's not as easy as it looks,
the Transformer is 3-times the size I'd "like" it to be,
but that makes it cheaper, easier to build, less critical, and more stable overall.
It's capable of twice the power that I expect, that way nothing runs under heavy stress or heat.
I'm planing on epoxying together 2 commercially available Toroidal-Cores to
achieve "more than adequate" Inductance.
The simple calculations so far have come out to an astounding ~1.4-Volts-per-Turn,
and less than ~40ma Idle-Current, ( PWM-Push-Pull-Transformer ).
This has been achieved without "Doctorate-Level" Math calculations,
but I'll probably find 2 or 3 "unpredictable" problems before I'm satisfied with the performance.

I look at Magnetics as being a sort of "high-speed-mechanical-device" interfacing with Electronics,
kinda like a Motor,
only there's no moving parts that You can actually see working.
.
.
.
I agree with you. At the beginning of my career I supposed that the magnetic design was a exact task, do the maths, select a ferrite core, select the wire size and start to do turns like a crazy jaja. For a long time I was confused why the final inductance doesn't match with the turns calculated. Now with more experience I understood that you say, Its a process try and error, cut the wire and re-cut again, play with the gap, etc.

Due to my area of work, I can't afford to select manufacturer inductors o transformers, everything needs to be designed here, for a specific applications.

Cheers
 
Top