Transformers and the ZVS driver


I've been looking for cheap, efficient ways to generate high voltage ever since I got interested in electronics. Ever since then, I've been looking for an easy, reproducible way to obtain various voltages and power levels for a number of different high voltage projects. I've found a number of methods, but never one that I could easily duplicate or apply to other situations to my satisfaction. Recently, I found out about the Mazili ZVS driver. It allows you to switch a center-tapped transformer at its resonation frequency, and can achieve impressive output powers and efficiencies. They are also available cheaply on ebay. I decided to pull the trigger and see what the fuss was about. After winding a couple of turns on a TV flyback and connecting everything up, I was greeted with a terrifying red-orange arc on the output wire similar to the device here


For my first experiment, I obtained some ferrite rods from ebay and 3D printed some spools to fit around it. 
Two prints later, I wrapped about 100 turns around one with magnet wire, 5+5 turns around another, and slipped them onto the rod:
5+5 turns (black), 100 turns (white)
Applying 9V to the driver resulted in the following output:

Since the ZVS driver only uses one coil at a time (5 turns) and the secondary is 100, that gives a ratio of 1:20, which means an input of 9V should result in an output of 180V. Since the mosfets have a voltage drop and there are coupling losses, I figured the 140V output was reasonably consistent. After trying it a second, smaller transformer with a ratio of 1:2, I decided it was time to start scaling. 


In summary, the objective of this project is to design a set of modular components that can be mixed, modified, and plugged together such that a wide range of output voltages can be produced, is reasonably small, can supply a decent amount of power, and does so at high efficiency. I decided my objectives were twofold:
  • Scale the driver circuit down (it was about 3 inches to a side). I wanted to aim for around 30 mm wide maximum, and half that in height. Length was flexible, but I'd want to minimize it where possible.
  • Scale the transformer up. I wanted to aim for around 2 kV output minimum, which meant I needed a minimum of 1.25 thousand turns on the secondary. With the same losses as my test, I'd need even more--about 1.6 thousand. 

Scaling Up

Since the objective is to minimize the amount of effort it takes to build a new HV system, I decided to invest in a cheap coil winding machine from ebay. I printed a larger bobbin on my 3D printer, and five-hundred turns later, I ended up with this:
It took about an hour
Every 150 turns or so, I stopped winding to apply a coat of electrical tape and nail varnish to reduce the chance of arc-over. After applying a final coating, I slipped the new coil around the core, which easily accommodated the new secondary.

I then applied power:
At these turn ratios, it should be producing around 1 kV, which looked consistent with the arc length. I didn't measure it because I didn't want to risk my multimeters, but I started looking into HV probes. Now that I was reasonably certain I could produce the transformers I needed, it was time to begin work on building out the electronics. 

Updates will be posted as they happen


Popular posts from this blog

UMPC Adventures: the Mix Yoga

The Wearable Computer--A Synopsis