Panteltje's tritium decay experiment page

This project, the acquired data, and its asm code code is released under the GPL license

The purpose of this experiment is to verify the rumors that radioactive decay varies over time with possibly a yearly cyclus.
I decided to measure 'radioactive decay' in a slightly different way, to avoid environmental effects on photo multiplier setups.
Photo multipliers are extremely sensitive to external magnetic fields, and are also great cosmic ray counters.
One simple example of radioactive decay are the tritium lights, as for example used in fishing to attract fish.
These lights have a usable lifetime of about 12 years.
The idea is to measure the light output of such a light for one year, and see if apart from the normal decay curve there are any cyclic effects.
I ran this simple experiment for one year, and here you find the acquired data, pictures of the setup, circuit diagrams, and source code.
No analysis of the acquired data yet, working on it.
Here is the setup;

A picture says often more than a thousand words, so here you go

All starts with two photo detectors mounted against a tritium light tube:

The tritium light tube is glued to a metal heatsink that is kept at a steady temperature to within half a degree centigrade, just a little above the maximum expected environmental temperature (not counting freak global warming).
Luckily that did not happen..
Photo detectors and tritium tube side view:

The photo detectors together with the measurement electronics, voltage references, and temperature sensor are mounted within a small black light proof box:

This box is mounted on a 100x160 mm bigger circuit board, that sits in an metal case;

Thermal insulation is provided, to reduce power required for heating the measurement setup, the black box is offset from the main board too:

More thermal insulation, a beeper warns if the experiment needs attention (one year alarm, temperature alarm, power failure alarm, etc etc):

The other side of the main board with batteries for backup if a power failure happens.
There is also a RS232 serial connection for the PC, an alarm reset switch, 2 warning lights (one a blue flashing light):

The circuit diagrams, this is the black box electronics:
I make these sketches and then solder it together, do not try anything like that if you do not have not a lot of experience designing and building electronics, and writing code, as then it will not work:

The same goes for the main board electronics:

One year passed, got the data out, cleared the EEPROMs, and restarted the experiment for an other year long run, did not calibrate the clock....:

ASM source code, made available under the GPL, this was assembled with gpasm-0.13.7 beta running on Linux:
Click here to download tri.asm, the source code for the PIC 18F14K22 in the black box.
Click here to download mon.asm, the source code for the PIC 18F14K22 on the main board.

The user interface presents a menu, here the choices: menu

The one year acquired data from 2012 to May 14 2013 with this setup:
Click here to download one year of acquired data (2012 to May 14 2013).

The one year acquired data from May 14 2013 to May 14 2014 with this setup:
Click here to download one year of acquired data (May 14 2013 to May 16 2014).

The one year acquired data from May 16 2014 to May 16 2015 with this setup:
Click here to download one year of acquired data (May 16 2014 to May 16 2015).

I have had many reactions via email, thank you all for the effort.

Here are some things I receved:
Romain Clement de givry made an Exel table from the data:

Paul McGill did an analysis of the second year data:
Interesting in this is that there seems to be a similar seasonal variation as in the original experiment here:
Evidence for Correlations Between Nuclear Decay Rates and Earth-Sun Distance

Click here for some programs I make available under the GPL.

Click here to send mail