{"id":2134,"date":"2026-08-16T15:09:15","date_gmt":"2026-08-16T07:09:15","guid":{"rendered":"https:\/\/www.trafopsu.com\/?p=2134"},"modified":"2026-08-16T15:09:15","modified_gmt":"2026-08-16T07:09:15","slug":"why-flyback-transformer-air-gap","status":"publish","type":"post","link":"https:\/\/www.trafopsu.com\/pt\/why-flyback-transformer-air-gap\/","title":{"rendered":"Why Must a Flyback Transformer Have an Air Gap?"},"content":{"rendered":"<div class=\"wrap\">\n<header class=\"hero\"><\/header>\n<p>Many engineers new to switched-mode power supply (SMPS) design are confused when they first meet the flyback transformer in a textbook.<br \/>\nThe book says: <em>&#8220;A flyback transformer must store energy, so it needs an air gap.&#8221;<\/em> But an air gap is just a slice of air \u2014 its permeability is<br \/>\nfar lower than ferrite, so it conducts magnetic flux <em>worse<\/em>. Why would you deliberately carve a gap of air into the core? And if the gap lowers<br \/>\nthe inductance, doesn&#8217;t lower inductance mean <em>less<\/em> stored energy?<\/p>\n<p>Those two questions point at the same core misunderstanding: <strong>a flyback transformer is not a transformer \u2014 it is a coupled inductor.<\/strong><\/p>\n<nav class=\"toc\" aria-label=\"Table of contents\">\n<h2>Table of Contents<\/h2>\n<ol>\n<li><a href=\"#conclusion\">The Core Conclusion<\/a><\/li>\n<li><a href=\"#forward-vs-flyback\">Forward vs. Flyback: The Essential Difference<\/a><\/li>\n<li><a href=\"#what-gap-does\">What the Air Gap Really Does<\/a><\/li>\n<li><a href=\"#energy-in-gap\">Where the Energy Is Actually Stored<\/a><\/li>\n<li><a href=\"#example\">A Worked Example (EE16 Core)<\/a><\/li>\n<li><a href=\"#analogy\">The Reservoir Analogy<\/a><\/li>\n<li><a href=\"#cost\">The Cost of the Air Gap<\/a><\/li>\n<li><a href=\"#reference\">Air-Gap Requirements Across Components<\/a><\/li>\n<li><a href=\"#faq\">FAQ<\/a><\/li>\n<\/ol>\n<\/nav>\n<h2 id=\"conclusion\">1. The Core Conclusion<\/h2>\n<ul>\n<li>A flyback transformer is essentially a <strong>coupled inductor<\/strong> (store first, release later), not a &#8220;transformer&#8221; in the conventional sense.<\/li>\n<li>Energy-storage formula: <span class=\"formula\">E = \u00bd\u00b7L\u00b7I\u00b2<\/span>. The gap <strong>lowers the inductance L<\/strong>, but <strong>raises the saturation current I<sub>sat<\/sub><\/strong> dramatically; as long as I grows faster than L shrinks, the total <code>L\u00b7I\u00b2<\/code> stored energy actually <em>increases<\/em>.<\/li>\n<li>The gap is a trade-off: <strong>sacrifice permeability (lower inductance) to gain a much larger saturation current and a higher energy-storage ceiling.<\/strong><\/li>\n<li>The fundamental difference between flyback and forward: one is a <strong>warehouse of energy<\/strong> (flyback); the other is a <strong>conveyor of energy<\/strong> (forward).<\/li>\n<\/ul>\n<h2 id=\"forward-vs-flyback\">2. Forward vs. Flyback: The Essential Topological Difference<\/h2>\n<h3>Forward Converter<\/h3>\n<ul>\n<li>Primary and secondary conduct <strong>simultaneously<\/strong>; power transfers directly from primary to secondary.<\/li>\n<li>The core only <strong>transfers<\/strong> energy, it does not store it. The peak flux density B<sub>max<\/sub> is set by the <strong>volt-second product<\/strong>, independent of load current.<\/li>\n<li>A typical forward transformer <strong>normally needs no air gap<\/strong> \u2014 the gap only lowers permeability and brings no benefit. A small\/process gap appears only under DC bias, incomplete reset, or special design needs (not standard practice).<\/li>\n<\/ul>\n<h3>Flyback Converter<\/h3>\n<ul>\n<li>When the switch is ON: the primary winding stores energy like an inductor; the secondary rectifier is OFF.<\/li>\n<li>When the switch is OFF: the energy stored in the core is released to the load through the secondary winding.<\/li>\n<li>Primary and secondary <strong>do not conduct at the same time<\/strong> \u2014 this is the essential difference between flyback and forward.<\/li>\n<li>The work cycle is <strong>store \u2192 release<\/strong>, which is why it should be called a coupled inductor.<\/li>\n<\/ul>\n<h2 id=\"what-gap-does\">3. What the Air Gap Really Does: Let the Energy &#8220;Get In&#8221;<\/h2>\n<ul>\n<li>Without a gap, the B\u2013H curve is <strong>very steep<\/strong>: ferrite permeability \u03bc is typically 2000\u20133000, so a tiny current drives B straight to saturation. Once saturated, L collapses, current runs away, and energy simply cannot be held.<\/li>\n<li>With a gap, the effective magnetic path lengthens and the B\u2013H curve <strong>flattens<\/strong> (effective \u03bc drops to 1\/5 ~ 1\/10 of the original): at the same current, B is lower and farther from saturation, so the allowable maximum current rises sharply.<\/li>\n<\/ul>\n<table>\n<thead>\n<tr>\n<th>Par\u00e2metro<\/th>\n<th>No gap<\/th>\n<th>With gap (reasonable)<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Permeability \u03bc<\/td>\n<td>2000\u20133000<\/td>\n<td>200\u2013500<\/td>\n<\/tr>\n<tr>\n<td>Inductance L<\/td>\n<td>High<\/td>\n<td>Medium<\/td>\n<\/tr>\n<tr>\n<td>Saturation current I<sub>sat<\/sub><\/td>\n<td>Low (~1 A)<\/td>\n<td>High (~5 A)<\/td>\n<\/tr>\n<tr>\n<td><code>L\u00b7Isat\u00b2<\/code> stored energy<\/td>\n<td>Small<\/td>\n<td>2\u20135\u00d7 larger<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h2 id=\"energy-in-gap\">4. The Energy Is Actually Stored Mainly in the Gap&#8217;s Field<\/h2>\n<ul>\n<li>From a field perspective, the energy density is <span class=\"formula\">w = \u00bd\u00b7B\u00b7H<\/span>.<\/li>\n<li>The gap&#8217;s permeability \u03bc is far below ferrite, so at the <strong>same B<\/strong> the gap&#8217;s H is much larger \u2014 meaning its <strong>energy density is far higher than the core material&#8217;s<\/strong>.<\/li>\n<li>The gap volume is small, but per unit volume it stores far more energy than the core \u2014 that is exactly why opening a gap makes the <em>total<\/em> stored energy larger. (It is not &#8220;the air storing energy&#8221;; the gap simply lets the core tolerate a larger current and therefore store more.)<\/li>\n<\/ul>\n<h2 id=\"example\">5. A Worked Example (EE16 Core)<\/h2>\n<ul>\n<li>No gap: L = 1 mH, I<sub>sat<\/sub> \u2248 1 A \u2192 <span class=\"formula\">E<sub>max<\/sub> = \u00bd \u00d7 1 mH \u00d7 1\u00b2 = 0.5 mJ<\/span><\/li>\n<li>0.3 mm gap: L = 200 \u00b5H, I<sub>sat<\/sub> \u2248 5 A \u2192 <span class=\"formula\">E<sub>max<\/sub> = \u00bd \u00d7 200 \u00b5H \u00d7 5\u00b2 = 2.5 mJ<\/span><\/li>\n<li>Conclusion: <strong>stored energy rises ~5\u00d7<\/strong> \u2014 sacrifice inductance to gain a larger saturation current and a higher storage ceiling.<\/li>\n<\/ul>\n<h2 id=\"analogy\">6. The Reservoir Analogy (memory aid)<\/h2>\n<ul>\n<li><strong>No gap:<\/strong> the reservoir is shallow (high \u03bc); a little water and it overflows (saturates).<\/li>\n<li><strong>With gap:<\/strong> the reservoir is dug deeper (lower effective \u03bc); it holds far more water (larger current, more energy) before overflowing.<\/li>\n<li>The energy is of course held <em>in the reservoir<\/em>, but <strong>without that deeper basin it could never hold this much water.<\/strong><\/li>\n<\/ul>\n<h2 id=\"cost\">7. The Cost of the Air Gap<\/h2>\n<table>\n<thead>\n<tr>\n<th>Cost<\/th>\n<th>Effect<\/th>\n<th>Mitigation<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Increased leakage inductance<\/td>\n<td>Voltage spike on the switch; needs RCD snubber<\/td>\n<td>Optimize winding layout (sandwich primary around secondary)<\/td>\n<\/tr>\n<tr>\n<td>EMI<\/td>\n<td>Conducted\/radiated noise rises (fringing field couples to windings &amp; PCB)<\/td>\n<td>Shield winding, Y-capacitor<\/td>\n<\/tr>\n<tr>\n<td>Higher copper loss<\/td>\n<td>More turns needed for the same L<\/td>\n<td>Use higher-B<sub>sat<\/sub> core material; distributed gap (powder core) relieves fringing but costs more<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<div class=\"note\">Note: The <strong>fringing flux<\/strong> at the gap produces leakage inductance and EMI \u2014 a key reason flyback EMC is harder to tame than forward.<\/div>\n<h2 id=\"reference\">8. Air-Gap Requirements Across Magnetic Components<\/h2>\n<table>\n<thead>\n<tr>\n<th>Componente<\/th>\n<th>Gap needed?<\/th>\n<th>Reason<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Forward transformer<\/td>\n<td><span class=\"tag no\">\u274c Normally no<\/span><\/td>\n<td>Stores no energy ideally; B<sub>max<\/sub> set by volt-second; process gap only under bias\/reset issues<\/td>\n<\/tr>\n<tr>\n<td>Flyback transformer<\/td>\n<td><span class=\"tag yes\">\u2705 Must<\/span><\/td>\n<td>Stores energy; must avoid saturation<\/td>\n<\/tr>\n<tr>\n<td>PFC boost inductor<\/td>\n<td><span class=\"tag yes\">\u2705 Yes<\/span><\/td>\n<td>Large DC bias; needs saturation resistance<\/td>\n<\/tr>\n<tr>\n<td>Buck output inductor<\/td>\n<td><span class=\"tag maybe\">\u26a0\ufe0f Depends<\/span><\/td>\n<td>Small-power ferrite needs a gap; large-power powder core has built-in distributed gap<\/td>\n<\/tr>\n<tr>\n<td>Transformador LLC<\/td>\n<td><span class=\"tag no\">\u274c No<\/span><\/td>\n<td>Mostly transfers energy; leakage\/magnetizing inductance join the resonance<\/td>\n<\/tr>\n<tr>\n<td>LLC external resonant inductor<\/td>\n<td><span class=\"tag maybe\">\u26a0\ufe0f By design<\/span><\/td>\n<td>Large AC current; judge by core and current margin<\/td>\n<\/tr>\n<tr>\n<td>Common-mode choke<\/td>\n<td><span class=\"tag no\">\u274c Usually no<\/span><\/td>\n<td>Differential fluxes cancel; suppresses common-mode; real leakage exists, evaluate if diff-mode bias is large<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<div class=\"callout\"><strong>One-line summary:<\/strong> A flyback transformer <em>must<\/em> have an air gap \u2014 not because the gap &#8220;stores&#8221; energy, but because <strong>without the gap, the energy simply cannot get in<\/strong>: the core saturates at a very small current. The gap sacrifices permeability to win a larger saturation current and a higher storage ceiling.<\/div>\n<h2 id=\"faq\">9. FAQ<\/h2>\n<div class=\"faq\">\n<details open=\"\">\n<summary>Why does a flyback transformer need an air gap?<\/summary>\n<p>Because a flyback transformer is really a coupled inductor that must store energy in the core. Without a gap the ferrite B\u2013H curve is very steep, so a small current drives the flux to saturation; once saturated the inductance collapses and energy cannot be held. The gap lowers effective permeability, flattens the B\u2013H curve, and lets the winding carry a much larger peak current before saturating.<\/p>\n<\/details>\n<details>\n<summary>What is the difference between a flyback and a forward converter?<\/summary>\n<p>In a forward converter the primary and secondary conduct at the same time and power transfers directly; the core only transfers energy and normally needs no air gap. In a flyback converter the primary stores energy while the switch is on and the secondary releases it while the switch is off; primary and secondary never conduct together, so the transformer acts as a coupled inductor and must store energy, which is why it needs an air gap.<\/p>\n<\/details>\n<details>\n<summary>Where is the energy actually stored in a gapped flyback transformer?<\/summary>\n<p>Mostly in the magnetic field of the gap. Energy density is w = \u00bd B H. The gap has far lower permeability than ferrite, so at the same flux density B the field strength H in the gap is much larger, giving a much higher energy density than the core material. The gap does not &#8220;store energy by itself&#8221; \u2014 it lets the core tolerate a larger current, so the total stored energy rises.<\/p>\n<\/details>\n<details>\n<summary>Does the air gap reduce the inductance?<\/summary>\n<p>Yes. The gap lowers the inductance L. But it raises the saturation current Isat far more. Stored energy is E = \u00bd L I\u00b2, so as long as Isat grows faster than L shrinks, the total L\u00b7Isat\u00b2 stored energy increases. Example: an EE16 core with no gap gives about 0.5 mJ, while a 0.3 mm gap drops L to 200 \u00b5H but raises Isat above 5 A, lifting stored energy to about 2.5 mJ, roughly 5\u00d7 more.<\/p>\n<\/details>\n<details>\n<summary>Can you avoid the mechanical air gap?<\/summary>\n<p>Yes, by using a low-permeability toroid or a powder core (alloy powder core such as Sendust, High-Flux, MPP, or iron powder). These materials have a built-in distributed air gap and strong saturation resistance, so no separate machined gap is needed. A gapped ferrite is still the common choice for many flyback designs because of cost and loss trade-offs.<\/p>\n<\/details>\n<details>\n<summary>Is leakage flux the same as leakage inductance?<\/summary>\n<p>No. The escaping field at the gap is called leakage flux. Leakage inductance is produced by the spacing between the primary and secondary windings and exists even without grinding a core gap; it is measurable. They sound alike but are different concepts. The fringing flux at a gap does increase leakage and EMI, which is why flyback EMC is harder to tame than forward.<\/p>\n<\/details>\n<\/div>\n<h2>10. Explore Our Magnetic Components<\/h2>\n<p>TrafoPSU designs and manufactures custom magnetic components for SMPS, automotive, industrial, and medical applications.<\/p>\n<p class=\"note\">\n<\/div>","protected":false},"excerpt":{"rendered":"<p>Many engineers new to switched-mode power supply (SMPS) design are confused when they first meet the flyback transformer in a textbook. The book says: &#8220;A flyback transformer must store energy, so it needs an air gap.&#8221; But an air gap is just a slice of air \u2014 its permeability is far lower than ferrite, so [&hellip;]<\/p>","protected":false},"author":3,"featured_media":2140,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[96,97,95,99,98],"class_list":["post-2134","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-application","tag-air-gap","tag-coupled-inductor","tag-flyback-transformer","tag-magnetic-components","tag-smps"],"_links":{"self":[{"href":"https:\/\/www.trafopsu.com\/pt\/wp-json\/wp\/v2\/posts\/2134","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.trafopsu.com\/pt\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.trafopsu.com\/pt\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.trafopsu.com\/pt\/wp-json\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"https:\/\/www.trafopsu.com\/pt\/wp-json\/wp\/v2\/comments?post=2134"}],"version-history":[{"count":2,"href":"https:\/\/www.trafopsu.com\/pt\/wp-json\/wp\/v2\/posts\/2134\/revisions"}],"predecessor-version":[{"id":2141,"href":"https:\/\/www.trafopsu.com\/pt\/wp-json\/wp\/v2\/posts\/2134\/revisions\/2141"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.trafopsu.com\/pt\/wp-json\/wp\/v2\/media\/2140"}],"wp:attachment":[{"href":"https:\/\/www.trafopsu.com\/pt\/wp-json\/wp\/v2\/media?parent=2134"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.trafopsu.com\/pt\/wp-json\/wp\/v2\/categories?post=2134"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.trafopsu.com\/pt\/wp-json\/wp\/v2\/tags?post=2134"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}