Wednesday, July 6, 2011

Ozone Generator

I've found that the fat cables that came with the robot arm I got from the Makery are out-gassing a horrid stink of damp basement. Whelwick is starting to smell like an old barn. To address this I'm going to take a two phase approach. I'll need to address the source of the odor, the cables, and also, mostly for fun, I'll build an ozone generator to eradicate the airborn odor through oxidation of whatever fumes it's been putting in the air here.

To treat the cables I intend to soak them in bleach solution for a few days. Before I do this, I want to leave the ozone generator running for a while, to see how effective it is at neutralizing the odor causing particles.

The simplest way to generate ozone, at least for me, is to use the corona discharge method to convert atmospheric oxygen, O2 to ozone, O3. I'll be using a 12kV 30mA neon sign transformer. This operates at the line frequency of 60Hz, which isn't great for ozone production, but it does work. Another way to do this is to repurpose the flyback transformer from an old CRT (I have half a dozen if anyone needs such a device). The flyback can be driven from a simple 555 or microcontroller circuit to produce a high voltage, high frequency signal which is more effective than 60Hz for ozone production. It's also smaller.

To produce the corona discharge I'm going to try using conductive plates separated by an insulator. In this case, aluminum ductwork tape and glass. First I cut a bit of glass from a piece of scrap.



I then applied the tape to both sides and trimmed it clear of the edges to avoid arcing around the edges.



The corona discharge occurs at the edges of the plate, not where they overlap, so I want as much edge as I can get. Overlap provides capacitance, which for this application, I don't need or want, so I want a plate pattern with no overlap and plenty of edge. I went with an 'E' pattern. It's easy to cut.

I decided to use copper wire from a length of Romex 10/3 as connectors, so I could easily support the panel with the connectors. To maximize the contact area I made the wire long and gave it a right angle bend to prevent it twisting.



I secured the connectors with some Gorilla Tape (it was what was handy, and it's fun to use).



The corona panel is then connected to the transformer. For testing I've attached the neon sign transformer to a variable autotransformer so that I can start out at a low voltage and ramp up the output of the neon sign transformer. Too much voltage can puncture the glass dielectric, which will shatter the glass (another reason for the tape, to contain the glass). I don't think there is much danger of this, since I'm not reaching voltages where this should be a problem, but if there are flaws in the glass it is possible.

The wiring for the neon sign transformer also has a safety gap configured. The spacing of the gap is set so that any voltage much above the 12kV the transformer secondary is designed for will cause an arc across the safety gap to protect the transformer.



To verify that it is generating ozone, and just in case it bursts, I put a container over the corona panel until I ran it up to full voltage and let it sit for a while.



The corona is fairly faint unless your eyes are dark adapted. To get visualize it I've set the camera up for long exposure in the dark. My cameras are cheapie point-n-clicks, so you'll have to put up with the not-so-great quality. here it is from both sides.




Here's what happens if you up the voltage but forget to open up the safety gap.



From the corona images it is apparent that the sharp points of the 'E' pattern are producing a lot of corona (the shape of an electric field around sharp points encourage dielectric breakdown and corona discharge, this is why bulbous, polished surfaces are so popular in high voltage labs). If you look closely though, you can see that there is discharge along the edges as well, as was intended. I don't know which is better. Regardless, it generates enough ozone. The staticy odor of ozone is strong after running for a couple of seconds of running.

I guess that qualifies this as a success.

No comments: