How a Power Supply Works, Part 4

As I said, today we'll be looking at the power transformer, as well as the circuit known as the snubber.
Most people are probably familiar with transformers, but snubbers are mainly known to those who work with power supplies more closely.
The whole section is outlined in red in the photo, and I've circled the snubber in green.


You can also see it in the popular power supply design. In the photo, I've crossed out a diode that isn't part of the snubber.


And here's one in my homemade power supply. Its circuit is different, and I'll talk about that a little later.


I think many of you are familiar with the schematic of a typical flyback power supply; similar schematics often appear in my reviews.


Let's isolate the part I'll be discussing.
It consists of the snubber, transformer, input capacitor, and high-voltage transistor.


Let's remove the part that's unrelated to our topic. That leaves very few components, which I think will make the processes easier to understand.


So, what happens inside a switching power supply while it's operating?
First, the power switch turns on. Current flows through the path highlighted in red, and energy is stored in the transformer's core.


When the switch turns off, the polarity across the transformer windings reverses, and current starts flowing to the load.


But because the transformer and output circuits aren't ideal, a voltage spike appears on the primary winding and starts flowing through the snubber.
Look closely and you'll see that the winding starts, marked with dots, are oriented the same way relative to diodes D1 and D2. That's why these circuits don't conduct while the power switch is on.
The snubber's job is to absorb the parasitic spike that appears in the primary winding and thereby protect the high-voltage transistor. Some very cheap power supplies don't have a snubber at all, which is quite harmful because it reduces reliability.


In a typical power supply, this part of the circuit looks like this. Component values are chosen based on the transformer winding's inductance, the operating frequency, and the power supply's output power. I won't go into the calculation method—it's a fairly long topic—but I'll just say that the principle “the more, the better” doesn't apply here. The circuit needs to be optimized for the specific conditions.


Some of you probably noticed the diode in the snubber circuit and thought, “That looks familiar.”
And you'd be right. The closest equivalent is the protection circuit for a transistor that switches power to a relay coil. It performs a similar function here, preventing a voltage spike across the transistor when it switches off. By the way, replacing the diode in this circuit with a Zener diode should work better.


Since the diode-based version can't be used with a transformer, a resistor and capacitor are connected in series with it, or a TVS diode is used, as in this schematic.


Here's another new term: TVS diode. Don't be alarmed; a TVS diode is essentially just a Zener diode. But while a Zener diode is designed to provide a stable voltage, a TVS diode is designed to handle pulse current and dissipate power. Voltage stability isn't as important in this case.
It looks like an ordinary diode and can also be bidirectional, in which case the cathode isn't marked. The most common TVS diode series are P6KE and 1.5KE. The first has a peak pulse power of 600 W, the second 1500 W. More powerful ones are available, but we're not interested in those.


I'll rotate the schematic a little to make it easier to understand how it works. In circuits like this, TVS diodes rated for 200 V are most often used, for example, the P6KE200A.
This keeps the voltage across the transformer winding from exceeding 200 V. The voltage across the input capacitor is about 310 V.
That means the voltage across the transistor is about 510 V. In reality, it'll be a little higher because components aren't ideal, and the mains voltage can also be higher.


Datasheets for TOP-series ICs often showed exactly this kind of TVS diode connection.
This circuit has a sharper clamping characteristic: it doesn't clamp at all below 200 V, then tries to clip everything above 200 V. The capacitor circuit has a slightly different clamping characteristic, but in practice this isn't critical.


To reduce the power dissipated by the TVS diode, a capacitor can be connected in parallel with it.


Or you can combine the two circuits into a hybrid that includes all the components from both versions. This is often used in high-power flyback power supplies.


Sometimes an alternative transistor protection method is used: a TVS diode connected in parallel with the transistor. This approach is fairly uncommon and is more often found in power supplies with a low input voltage.


For example, this TVS diode configuration can be found in a PoE power supply, where the input voltage is not 310 V DC, but only up to 70 V.


Now let's move on to the transformer.
A transformer consists of a core and a bobbin. Sometimes the assembly also includes special clips that hold the core in place on the bobbin.


E-cores are most commonly used. If the power supply is a flyback design, as are the vast majority of inexpensive, low-power power supplies, there must be a gap between the two halves of the core. The gap can be made between the halves, or you can use a special core with a gap already formed in the center leg; in that case, the gap must be twice as wide.


Ferrite is usually used for the core. Branded cores may have a part number printed on them, so you can look up their specifications in the datasheet. Cheaper cores often have no markings.


The transformer windings are wound first, then the bobbin is fitted onto the core.


Winding small transformers is fairly simple.
First, we wind the primary.


Then the secondary, sometimes using two or more wires in parallel.


If there's a third winding, it's usually the PWM controller's supply winding, so we wind that too.


For safety, we insulate the entire assembly.


Next, we take the selected core. In this case, the center leg of one half has been shortened.


We put the whole assembly together. The core is usually glued, but I normally secure it with tape as well.


The result is a small, neat transformer. The one in the photo is rated for about 25-30 W.


This transformer is rated for up to 80-100 W. They're wound in a similar way, but with a few differences.


In transformers designed for low output voltage and high current, the output winding can be made either from litz wire or from copper strip.


The choice of primary winding turns depends directly on how well the transformer is wound. This isn't very critical for low-power transformers, but winding a high-power transformer incorrectly can have serious consequences.
The windings are usually arranged in three layers (when there are three windings): primary, secondary, and auxiliary.
But the coupling between the windings can be greatly improved by placing the secondary between the two halves of the primary.


It's also recommended to wind the wire neatly, turn by turn, rather than bunching it up, and to fill the entire bobbin evenly. Windings designed for high current are best made with several thin wires rather than one thick wire.

Problems that can occur in this section:
1. A short between turns if the high-voltage transistor fails.
2. The transformer overheats, its inductance then drops sharply, and the inverter transistor fails.
3. The snubber diode fails—very rarely.
4. The TVS diode partially fails. For example, a 200 V TVS diode starts clamping at 100 V. Nothing burns out, but the power supply stops working.