Transformer Corner part 2

In Transformer Corner part 1, I introduced one of the key parts of the Oscilloclock – the HV transformer, and tried to illustrate some of the concepts and history behind it.

Next, let’s explore the Prototype’s power supply configuration. This will tell us a lot more about the transformer I had to wind!

Power supply design

My greedy little Oscilloclock wanted lots of different voltages…

  1. 6.3V AC for the CRT heater
  2. 60-80V DC for the blanking circuit, feeding the CRT grid
  3. 5V DC to drive the blanking optoelectronics
  4. 300V DC to drive the horizontal and vertical amplifier circuits
  5. 3kV DC to accelerate the electron stream to unfathomable speeds

A key constraint is that the first 3 outputs have to float at cathode potential, around -3kV with respect to the power supply’s input. Hence they need to be isolated from the input and from the other outputs.

The most common way to obtain lots of isolated voltages at different potentials, is to use a flyback topology circuit, and a transformer with multiple secondary windings – one for each output needed.

Typical multiwinding flyback circuit - as used in production model Oscilloclocks

Typical multiwinding flyback circuit – as used in production model Oscilloclocks

Well, winding such a complicated transformer by hand would have been way beyond me!

So for my Prototype, I decided to use DC-DC converter modules to provide all the lower-voltage isolated outputs. I could then focus all my efforts on the 300V and 3kV HV outputs. Since these two are both tied to a common ground, I could use a single-secondary flyback transformer producing 150V and use voltage doublers off that to get to 300V and 3kV.

My flyback circuit and transformer now looked something like this:

Flyback circuit used in the Prototype

Simpler topologies?

But I still balked at winding my own transformer. Wasn’t there an easy way?

Yes! There is another type of circuit, called buck-boost, which produces a single output, just as I needed. And this topology does not use a transformer!! It only uses a (single-winding) coil, and there are lots of off-the-shelf inductors that would do the job!

Simpler 'boost' topology circuit -takes an off-the-shelf inductor!

Simpler ‘boost’ topology circuit – takes an off-the-shelf inductor!

So, why didn’t I use this magic boost topology?

It turns out the LM2588 switcher that I planned to use has a maximum switch voltage of only 65V – and in a boost topology, this output drives the final output directly. Since I want 150V output, I had to use the flyback topology. (Or hunt for an alternative regulator that could handle 150V – if one exists.)

Always check the specs!

Always check the specs!

Post-script: Much later, I learned that you CAN use a boost topology to supply high voltages! The trick is to feed the output of the switcher into another transistor that can handle the higher voltages… A topic for another post, another day.

Now… figuring out the transformer’s specs

Remember my quirky analogy to the twirling skater? The faster the twirl, the smaller the size – but limited by the physical characteristics of the skater.

Well, in a transformer, one key characteristic that directly governs the twirl frequency is called the primary inductance. The lower this value, the faster it can twirl. I think of this as a kind of ‘transformer inertia’. (Engineers among you will cringe!)

If we can determine this inductance, it will tell us a lot about the materials we need and how to wind the trafo. But don’t worry! This page will have no mathematical formulae whatsoever. I’m going to cheat.

My switching regulator was an LM2588 from National Semiconductor. I figured the manufacturer must have a tool for computing component values. And they did! Although National Semiconductor had been taken over by Texas Instruments, the old design software called “Switchers Made Simple” (SMS) had been fully incorporated into TI’s WEBENCH® Design Center . A quick free registration and away I went…

Using WEBENCH was relatively simple: I punched in my desired voltage (9V input to 150V output) and current (25mA), then selected my topology and device. At some point I selected the operating frequency, then a click, and I had values for the various components, including the primary inductance and winding ratio! Voila.

Step 1 - specify required voltage and current

Step 1 – specify required voltage and current

Step 2 - Show alternate topologies

Step 2 – Show alternate topologies

Step 3 - Select the LM2588

Step 3 – Select the LM2588

Step 4 - Voila! Review the circuit

Step 4 – Voila! Review the circuit

Step 5 - Check the transformer specs

Step 5 – Check the transformer specs

Putting it all together!

Here is everything I knew about the transformer at that point:

  • operating frequency would be 151kHz
  • primary inductance needs to be around 80uH
  • primary DC resistance should be 0.08 ohms
  • winding ratio should be about 1:14
  • windings would have to be reasonably insulated from each other
  • the secondary would need to supply around 25mA of current

In the next installment, I explain how I used these specs to determine the materials and wind the transformer myself. Drum roll please….