Radiation detector, help needed!

(I’m about 2 days behind you in development of my HV board

I wrote an arduino program that lets me type in a pulse width and repetition rate, then generates pulses on one of the outputs.

This goes into a power MOSFET looking into a flash transformer. No (other) transistors, no oscillators, nothing but the micro and the MOSFET.

This has allowed me to experiment with pulse widths and repeat rates. Using this method I’ve been able to find the “sweet spot” for my transformer: if the pulse is any narrower it generates less output voltage, and if it is any wider it doesn’t generate more voltage.

Very little current is therefore wasted, so the system will have the a long battery time.

Eventually I plan to use the micro to adjust the output voltage by comparing against a voltage reference.

Contact me on my website if you’re interested in the pulse generator source code.

Diodes take time to reverse polarity. During this time they act like resistors and eat power.

Normal diodes take a relatively long time to reverse, and the excess power they dissipate can be enough to burn out the diode over time.

You can get fast recovery diodes with a much smaller reversal time. This results in less dissipated energy which is easier on the diodes, and less energy is wasted by the circuit.

In my system, replacing the 25KV diode with an ultrafast increases the output by 20 volts.

Ultrafast recovery diodes are only a little more expensive than regular diodes:

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=320579572234&ssPageName=STRK:MEWNX:IT

A friend pointed me to this:

http://tech.groups.yahoo.com/group/HomeRadLab/message/168

Before producing counters, you might want to measure the plateau of one of your tubes to be sure the HV is correct for your type of tube.

Just a note, I’m using an arduino to generate very short pulses into a 120:1 ratio transformer.

Using fast recovery diodes in a full wave rectifier generates several thousand volts with almost no wasted power (on a 5-volt supply).

So much voltage that I only have to pulse the thing 25 times/sec to generate 900 volts for my system.

You may be interested in this:

http://okianwarrior.com/Quenching/Quenching1.png

Add 1 to the last digit in the URL to get the other 3 pages or the article.

Looking over your schematic, I suspect that there is a problem.

I’d like to start a dialog with the designer. Please contact me from my website, or send a note to the registration E-mail for this blog.

I sent an E-mail to the address listed above and have received no response.

Your turn. Try contacting me here:

Barrabas (at) Barrabas (dot) MV (dot) com

I guess we haven’t been able to link up via E-mail.

I think there may be a problem in your schematic. A CCFL inverter usually puts out a lot of current – 10ma if memory serves. There is no current limiting resistor between the rectifier and the zener diodes. The diodes will
be bearing the entire current for all voltage over the zener voltage.

Your zeners add to 450 volts. Suppose the CCFL puts out 500 volts at 10ma. That means that the zeners have to shunt 50 volts at 100ma, which is a whopping 0.5 watts. The zeners are probably rated much lower than that – they will heat up and burn out very quickly.

You should limit the current going into the zeners. A 1 meg resistor between the zeners and the rectifier would limit the current to 0.45ma, and 50 volts at 0.45ma is .022 watts, which is much more reasonable for zeners to handle.

The GM tube needs very little current – 0.45ma is plenty.

For testing purposes, you can get cheap radioactive ore samples on eBay. I found this surprising, but it’s true.

Search for any of the common radioactive ores such as “gummite”, “uraninite”, “torbernite”, and so on. For example:

http://shop.ebay.com/?_from=R40&_trksid=m570&_nkw=uraninite

The 2nd one down shows 16,000 CPM and would cost $20. With a little searching, you can get reasonable samples for $5.

The CPMs listed can be from one of two devices: A pancake GM tube (Ludlum) will read 10x the value of the more common stick tube (CDV-700), so the sellers usually list the value taken by the more sensitive unit.

It’s getting pretty hard to find lantern mantles which contain Thorium. Due to public aversion to anything radioactive, manufacturers now use a different formula.

Both lantern mantles and smoke detectors emit alpha radiation (and a little gamma, but mostly alpha).

Some GM tubes will not register alpha, simply because the window is too thick. Alpha is blocked by just about anything, including a couple of inches of air.

This means that to measure alpha you need a sample that’s right up against the GM tube window, otherwise it’s blocked by the air.

Because of this, measuring alpha isn’t very useful for the Japanese folks – the radiation of concern over there is beta and gamma.

It’s a good idea though. I recently found this link which shows how to measure alpha using an uncapped power transistor:

http://www.youtube.com/watch?v=fKNNXvzo8UU

This will measure alpha. Add an old smoke detector and you have an interesting experiment – if you can eliminate the electrical noise.