Table of contents
The Kvarts DRSB-01 (Кварц ДРСБ-01) is a simple consumer Geiger counter. It does not feature a display of any kind like most modern Geiger counters do, but instead each particle detected by the tube make a very characteristic "click". It was manufactured in the early 1990's and is not made any more, but you can still find it commonly on places like eBay. I got mine for about 15€ a few years ago, but unfortunately prices have skyrocketed recently after the events in Fukushima brought back the reality that is radioactivity into everyone's minds.
A Geiger counter measures only ionizing radiation. Furthermore, as this device uses a metal shielded tube, only relatively energetic particles will be detected, that is mostly gamma rays or high energy beta rays. Is is not calibrated, and thus does not give a Sievert dose rating, but it can show relative differences of radioactivity between the background radiation and a higher activity by a correspondingly higher click frequency.
The device has two LED indicators, a green one marked ФОН (background) indicating normal background radiation, and a red one marked ВНИМАНИЕ (attention) indicating a level warranting attention.
Let's open it up.
The enclosure is hold by a single screw sealed in some wax. It looks like my unit has the serial number #003711 and was manufactured in week 6 of 1992 :
Inside the case is a single sided circuit board holding the high voltage power supply, a counter and a timebase. Standard Russian components are used. The К561ИЕ14 is the equivalent to the CD4029 (Presettable Binary/Decade Up/Down Counter), and the К561ЛЕ5 is equivalent to the CD4001 (Quad 2-Input NOR Gate)
On the top is the Geiger–Müller tube itself, an SBM-20U in this case. Some other versions of the DRSB-01 contain other types of tubes, probably depending on availability. In this case the high voltage power supply will be modified. That's probably why R2 is soldered on little posts.
To figure out how to modify the circuit a schematic is needed. I was able to find the schematic online, but unfortunately I could not find the source. Credit goes to the original authors :
The tube is charged with a potential of about 400V. When an ionizing particle strikes the tube, it briefly becomes conductive and a current passes through R4, creating a voltage across it. This voltage is amplified by Q2 and then is fed to the binary counter U2. The green LED is attached to output Q1, and the red LED is attached to output Q4. U1 is wired as an astable mutivibrator oscillating at about 1/2 Hz, and is connected to the preset enable input of U2. The preset is hard wired to all zeros, so the effect is that the U2 counter is reset twice a second. This means that to show a red "Attention" indicator, the counter must detect at least 8 particles per half second.
I want to make this device more useful by adding an USB port which will allow quantitative display of particle counts and long term logging of radioation levels. Due to space constraints inside the case I will use a mini USB port. I also need a microcontroller with USB support. The ATtiny44, which I already used in my USB NES Pad project, will do nicely. Type B mini-USB ports are not that easy to find, and I don't have one at hand, so an old multi memory card reader that barely ever worked will generously donate his own :
I also extract the AME8805 voltage regulator as well as the 12MHz quartz and a few filtering capacitors from the donor board.
The mod schematic. The COUNT signal must be connected to the point between U2 pin 15, the collector of Q2, and R6. To be able to power the device from the USB port only, I added a 1N4004 diode in series with the 3.6V supply. This diode has a drop of about 0.6V, just perfect to provide 3.0V equivalent to the batteries. The 3 filtering capacitors are low ESR ceramic type. They can also be replaced by a 10uF electrolytic + 20nF MKT in parallel.
The mod is mostly independent from the original circuit. It only takes its input from the clock signal on pin 15 of U2, and provides 3V to allow operation without batteries. Note that if you have batteries in your device when the USB port is plugged-in, the ON/OFF switch must remain in the OFF position. Otherwise current might flow into the batteries and they tend not to like it.
Now let's assemble the circuit in such a way as to make it fit in the little free space we have inside the enclosure :
The circuit in prototype phase and assembled dead bug style (kinda).
One of the worst pains of USB is the need for drivers (particularly under Windows). However there is one class that does not need a driver for any OS : the HID class. The USB designers seem to have had the simplicity of implementation and use in mind when writing this specification. May they live long and prosper. So I implemented the USB device as a standard HID device. Plug it in your computer, and tada! It is instantly recognized and useable. The devices identifies itself as an LED with a 16 bit counter attached. A bit strange, but it does not seem to pose any particular problem.
Unless you used the USBtinyISP way explained above, you'll need to burn specific fuses for the circuit to work : low 0xdf, high 0xdd, extended 0xff.
Here's what the device looks like after the mod. The exterior aspect of the device is not damaged. The switch on the side is to turn off the speaker, no need to hear constant clicking when using the detector for long-term monitoring.
To test the operation of the mod, I made a small Windows software in Java. Accessing the USB port in Java proved to be a little tricky due to the complete lack of support for USB in the language, but thanks to the HID class and this blog post it was relatively easy :
The dose values are the result of the multiplication of the detection count by a constant calibration factor that depends on the tube and the nature of the radiation. The SBM-20U tube datasheet gives calibration values for strontium 90 and cobalt 60. For this test I just used a random value.
I collected one hour worth of detection events, it might be interesting to perform some statistical analysis on this data. The potential as a random number generator is quite low however considering the low event rate.
Here is the source code for this software. It's a maven project, unzip and run 'mvn compile exec:java'.
One of the main reasons I built this mod was to monitor ambient radiation levels and be able to notice if a giant radioactive cloud flies over the city. So here it is. For simplicity I decided to write a Munin plugin for this, since I already had the infrastructure in place and it is a perfect tool to draw nice graphs. Here is the script :
#!/bin/sh if [ "$1" = "autoconf" ]; then echo yes exit 0 fi if [ "$1" = "config" ]; then echo 'graph_title Geiger counter' echo 'graph_args --base 1000 -l 0 ' echo 'graph_period minute' echo 'graph_vlabel Counts per minute' echo 'graph_category environment' echo 'graph_info Counts per minute' echo 'geiger.label counts' echo 'geiger.draw LINE2' echo 'geiger.type DERIVE' echo 'geiger.min 0' echo 'geiger.max 65535' exit 0 fi event=/dev/input/by-id/usb-Bidouille.org_Kvarts_DRSB-01_USB-event-if00 od $event -t u2 --width=24 -N 24 | head -1 | awk '{printf "geiger.value " $12 "\n"}'
The device id is hardcoded, it's probably a bad idea since it might change between kernel versions or for any reason. Ideally the device name should be grep'ed from /proc/bus/input/devices. However it does work :
Well the graphs do actually show some pretty big variations; this is unexpected as the background radiation should be fairly constant. Actually I may have damaged the tube during testing by inadvertently applying about 700V instead of the recommended 400V. However some parts of the graph look suspiciously like exponential decay, so there may be more to this. If you have any explanation, please comment below.
Sorry, won't do ;-)
That's all folks.
