Isolation Transformer Design Guide

An isolation transformer transfers energy through a magnetic field rather than through a conductor, so the two sides have no electrical connection. That single property is what makes it a safety component, a noise barrier and, in some designs, the whole reason the product can be certified.

What Isolation Provides

The primary and secondary windings are separated by insulation whose thickness and quality determine the voltage the barrier can withstand. The breakdown voltage of the insulation is the first number in the specification, and it is tested at production.

The transformer also blocks direct current and low frequency common mode voltage between the two sides. That is what allows a measurement on a high potential to be read safely, and what breaks a ground loop in an audio or instrumentation system.

Isolation is not perfect. Capacitance between the windings couples a small high frequency current across the barrier, and that coupling is the reason an isolation transformer alone does not remove all common mode noise.

insulation Classes and Construction

The insulation between the windings is what separates a functional transformer from a safety one. A reinforced barrier uses two separate layers of insulation, one of which is rated for the full voltage on its own.

Triple insulated wire allows the primary and secondary to share the same winding window without a physical barrier, because each conductor carries its own insulation. It is used in small transformers and in switch mode designs.

The margin tape at the ends of a winding serves the same purpose. The creepage distance between the windings along the surface of the former has to be maintained, and the tape is what keeps the wire away from the edge.

creepage and Clearance

creepage is the distance along the surface of the insulating material between two conductive parts, and clearance is the distance through the air. Both have minimum values set by the working voltage and by the pollution degree of the environment.

The values are specified in the safety standard rather than chosen by the designer, and they are the reason a transformer for a mains application is physically large. A design that reduces the creepage to save space will not pass the test.

Slots and ribs in the former increase the creepage without increasing the size. A slot interrupts the surface path, so the distance along it is longer than the straight line, and the technique is used widely in small transformers.

leakage current and Capacitance

leakage current flows through the capacitance between the windings and through the insulation resistance. At the mains frequency it is small, and at the switching frequency of a converter it can be large enough to be a problem.

An electrostatic shield between the windings, connected to a defined point, intercepts the capacitive coupling and diverts it. The shield is a foil or a single layer of copper that carries no current and takes the common mode current to the reference.

The shield reduces the common mode current but it does not eliminate it, because the shield itself has capacitance to the secondary. Where the residual matters, two shields with the second connected to the secondary reference are used.

common mode Noise and Its Path

In a switching supply the primary side switches at a high frequency and the secondary is connected to a load that may be earthed. The capacitance between the windings carries a current into the secondary circuit, and that current returns through the earth path.

common mode chokes on the input and output, the electrostatic shield, and a careful layout of the return paths all reduce the current. The objective is to give the common mode current a defined path that does not include the load or the measurement.

Measuring the common mode current is a matter of a current probe around the conductors that leave the product. A high reading indicates a path that should be redirected rather than a transformer that is faulty.

Core, Frequency and Losses

A transformer core is chosen for the frequency and for the flux density. At fifty hertz a laminated steel core is used, and at a hundred kilohertz a ferrite core, because the eddy current losses in steel would be prohibitive at that frequency.

The core loss rises with frequency and with flux density, while the copper loss rises with the square of the current. The design point balances the two, which is why a transformer that is efficient at full load may be hot at no load.

Saturation sets the maximum flux and therefore the maximum volt second product per cycle. A transformer that saturates in a switching converter produces a current spike that destroys the switching device, and the margin has to cover the worst case duty cycle.

Winding Arrangement and Coupling

A split primary with the secondary between the halves improves the coupling and reduces the leakage inductance. That matters for a switch mode transformer and not at all for a mains isolation transformer.

The leakage inductance appears in series with the winding and stores energy that has to be absorbed by a snubber when the switch turns off. Reducing it is a matter of interleaving the windings, which costs money and reduces the isolation distance.

The winding order also affects the capacitance between the primary and the secondary. Putting a shield between them and separating the windings physically reduces the capacitance at the cost of a larger leakage inductance.

Thermal and Mechanical Design

The transformer is often the largest heat source in a power supply, and its thermal path is through the core and the former to the board. A transformer specified by its temperature rise rather than by its rated power is easier to apply.

The mechanical fixing has to survive the weight of the component. A transformer that is held only by its pins will eventually crack a solder joint, and a bracket or a strap is standard practice above a certain size.

Vibration and mechanical noise are produced by magnetostriction in the core at the operating frequency. A transformer that buzzes audibly at fifty or a hundred hertz is normal to a degree, and a loose core makes it much worse.

Verification and Faults

Verify the isolation with a hipot tester at the specified voltage, applied between the primary and the secondary with the core and any shield connected as the standard requires. The test is destructive to a marginal unit, which is the point.

Measure the insulation resistance with a megohmmeter at the working voltage. A low reading indicates moisture or contamination in the insulation, and it is the failure that appears after a humid period rather than at the factory.

A transformer that fails the hipot test at production is usually a winding that has reached the edge of the former or a barrier that has been displaced. The release checks that keep such a component consistent are collected in our PCB design release checklist, the assembly points in judging PCB quality, and the layout measures that keep the two sides apart in our guide to mixed signal board design.

FAQ

Does an isolation transformer remove all noise? No. Capacitance between the windings couples high frequency common mode current across the barrier. An electrostatic shield reduces it.

What is the difference between creepage and clearance? Creepage is the path along the insulating surface and clearance is the path through air. Both have minimum values in the safety standard.

Why is my isolation transformer hot at no load? Core loss. It rises with frequency and flux density, so a transformer designed for full load efficiency can still run warm with nothing connected.