An Ionisation Chamber Shield for OpenRelief

Back in May I wrote about a low cost radiation detector I was prototyping for the OpenRelief project. Based on a simple ionisation chamber, initial testing used a bank of batteries for the bias power supply and this post details the construction of an inverter to replace these.

For details of the detector itself see the original post as none of the fundamental details have changed, and for more on OpenRelief see the interview with project leader, Shane Coughlan.

Just to recap, the bias voltage is applied across the ionisation chamber electrodes, between which tiny currents flow when ionising radiation enters the chamber. Using four PP3 batteries in series provided a bias of 36v, which is possibly suboptimal in addition to not being terribly convenient.

Once again I must give due credit to Charles Wenzel, who designed the circuit which is described here and that which the detector is based upon. The power supply schematic is below and for further details and alternative designs see Charles' excellent website.

Assembly

Note that a small change I made to the design was to substitute the MPSA18 transistor for a BC547, since the former proved difficult to source.

Since the detector is based around an Arduino compatible Nanode, the power supply was constructed on a prototyping shield. The top side of the shield can be seen in the picture at the top of this post, and the underside below.

This clearly isn't the neatest or most robust method of construction, but it's perfectly adequate for the purposes of a prototype.

Testing

With the shield completed I wanted to check the current consumption, which as can be seen below is around 360uA at 5v input, giving an output of around 90v. Since the current flow across the chamber electrodes will always be incredibly small — in the order of nano/picoamps — and the circuit is efficient, the power drawn by the shield should be barely noticeable.

Having confirmed that there were no shorts, the power supply operated as expected and did not draw too much current, it was time to assemble the detector ready for testing.

The Nanode sketch logs data to both the serial port and the Cosm Internet of Things platform, and so a USB adapter was plugged in and it was connected to the Internet .

On power up the background reading fluctuated around 36, rising to around 150 when a very weak alpha particle source was brought near to the mouth of the chamber. It should be noted that these values are entirely arbitrary and nothing more than a indication of the amount of current flowing across the electrodes — with no calibration of any sort.

Next steps

There still remains a reasonable amount of work to be done on the design of the detector itself. With the priority being to make it more stable in use as a simple alarm which can indicate when there is possible cause for concern. Beyond which it would be great to evolve it to the point where it can be calibrated to provide meaningful readings on par with a commercial design. This will require work on the chamber mechanical design, Nanode firmware, and perhaps processes for calibration.

In terms of the power supply shield:

• Explore potential for further optimisation, e.g. stability, simplicity / bill-of-materials

• Consider adding the ability to enable/disable the power supply from firmware (useful for calibration routines if not power saving?)

• Add a power indicator LED (would draw far more power than the whole circuit!)

• Add other things that are useful to have on a radiation detector, e.g. piezo sounder

• Turn into a PCB!

— Andrew Back