Simple Digital Geiger Counter

Not requiring the use of any LCD but useable as a

standalone device.
Also allows accurate digital data acquisition

The simple galvanometer display provides simple and stable readings.

RETURN TO HOMEPAGE

This Geiger counter design presented here is compatible with many types of GM tubes having a
recommended operating voltage ranging from 300V to 600V. It requires no other exotic component
than the GM tube. It can operate with a 9V battery pack to be used as a handlheld device.

How does a Geiger counter works ?

Many literature can be found about the internet about the basic principle of radioactivity
detection by the Geiger-Mueller (GM) tube. If you are not familiar with the GM tube theory
you can refer to these interesting websites:

Geiger Counter Design (Website from the "Gamma Scout" Company)

Video from www.chipdip.ru which explains well the principle of GM tube
Exercises with a Geiger Counter from the Northern Virginia Community College

And finally, for people who cannot find GM tubes, this very interesting article from
Yasuyuki Onodera about a full home made Geiger counter, including a home made GM
tube.

As you can see in the above listed documents, depending on the GM tube that you use it
will be possible to detect only gamma particles, or gamma and beta particles, and more
rarely alpha, beta and gamma particles. Even if the GM tube is sensitive to several types of
particles, it is not able to discriminate these different types. For the most part of the "usual"
cases you won't have to place specific shields in front of the tube because detecting
simultaneously gamma and beta particles is still efficient to detect any eventual abnormal
radioactivity level. However, for analyzing suspicious samples (e.g. food) appropriate
conditions are required to reduce the ambient background radiation level. Refer to this
synthesis that I've prepared for this purpose.

DISCLAIMER

This project involves high volatges up to 600V. Therefore it is expected that you are experienced
enough to apply the relevant care to protect yourself against electric shocks. You engage your
own responsability by taking risks during the test and use of the hardware presented on this
pages.

SCHEMATIC
 Overview

As many other designs, the high voltage (HV) power supply used to set the GM tube in the middle
of the plateau region is based on a step-up transformer followed by a one-stage "diode pump" for
voltage doubling (also known as a one-stage "Cockcroft Walton circuit"). This provides the
advantage of being able to use any general purpose transfomer initially designed for transforming
220VAC into 6-9VAC. We just have to use it reversed. The poor transformer efficiency when used
in this way is not a problem at all because the needed current is very low. The more particular
feature of the HV generator is voltage regulation. For keeping the GM tube bias voltage as
constant as possible when batteries are ranging from 9.5V to 7.5V while avoiding the use of high
voltage zener diodes(1), the step-up converter features a regulation loop which automatically
adapts the duty cycle of the transformer driver to compensate power supply voltage variations.
Potentiometer "HV Adj" allows adjusting the regulated high voltage depending on the GM tube
specifications and on the transformer characteristics. It is better to tune the HV generator before
connecting the GM tube anode to avoid permanent triggering due to overvoltage. (This could
reduce the GM tube lifetime). The ideal way of tuning is to use a variable regulated power supply
and check that the GM tube voltage stays near the middle of the plateau region when the power
supply ranges from 7.5V to 9.5V. Some variations might be noticed, but they must be compliant
with the GM tube requirement sheet. If the compensation is not strong enough, the 120K resistor
biasing the feedback point "TP" with Vpp can be changed for a lower value(2). If the compensation
is too strong (inverse effect) the same resistor has to be increased(2). More generally, before
tuning the HV generator do not forget to mount "Rd" with the appropriate value(3).

The pulse counting system is based on a switching transistor (2N2222 with its filtering capacitors)
and on the PIC16F84 which not only counts pulses but also drives the buzzer (one pulse on RB0
induces one toggle on RB7). This allows us not to use an additional buffer (logic gate or
comparator) for this task. If no microcontroller is wanted, the analog part of the design can remain
used but a buzzer driver must be added, as well as a signal shape sharpener for any counting by
another external device.

In the variant of mobile device proposed here, the mean of display does not require to buy a
specific LCD which might be hard to find. A simple galvanometer is directly driven by the
microcontroller by the mean of a duty cycle modulated signal passing through a low pass filter.
Your galvanometer might require more or less current. This is why the low pass filter resistor
marked "4K7" might be decreased down to 1K5 (as shown in brackets) if you don't succeed in
getting the pointer at its maximum when adjusting the potentiometer "Scale Adj" (detailed further).
Depending on the selected scale (switch "SCALE" on RA2) the galvanometer can show a linear or
logarithmic response to the count rate (See the scale aspect of my galvanometer to adapt your
logarithmic graduations to the conversion law). The count rate is continuously measured over 10
second periods. Therefore the value displayed by the galvanometer is refreshed every 10
seconds, when the LED connected to RA4 flashes. This combines the simplicity of an analog
display with the stability and repeatability of a true digital counter. In linear mode, the maximum
galvanometer deviation (duty cycle at 100% and potentiometer "Scale Adj" properly adjusted)
corresponds to 25.5cps (1530cpm). In logarithmic mode pulse counts ranging from 0.1cps to
1500cps (over a 4,5 decade range) can be displayed. Depending on your GM tube characterized
sensitivity, you can also graduate the galvanometer in units of your interest (Becquerel, Curie,
Roentgen/h, Gray or Sv/h...)

For accurate measurements of low level radioactivity, pulse counts over a time interval of more
than 10 seconds are expected. For example the averaging of several 10s measurements or the
counting over a global 3 minute period can be necessary (refer to the prepared synthesis). This is
why a serial data output is available on RB6. For connection to a computer this 5V level signal has
to be translated to RS-232 voltages (+/-9V min) by using for example this interface(4) also usable
as a regulated power supply for the Geiger counter when it is not used as a mobile device (See an
example of construction). Note that the device sends actual serial data and not just countable
"dummy bytes". Therefore any simple computer program can easily handle the generated true
data streams at 1200 bauds, whatever are the operating system and the realtime constraints. To
obtain details about the format, as well as example programs (source codes and executable files
for DOS/Windows/Linux), you can download this archive.

Notes:

(1) High voltage zener diodes are not so common, and furthermore their power dissipation can
easily reach 1W when only biased at 2mA.
This is not only a dissipation problem but also a waste of energy from the batteries.
(2) After having eventually changed the resistor, potentiometer "HV Adj" must of course be
adjusted again.
(3) Measure the DC resistance Rdc of the transformer primary coil (former secondary coil in its
planned use) and apply Rd = 6.Rdc
(4) If using very old PC computers (e.g. 286, 386...) voltage levels required on the RS-232 line
might be more than +/-9V.
In this case you may have to use another kind of interface circuit than IC MAX232.

Note about the GM tube serial load resistor "Rp":

Respecting the GM tube manufacturer's recommendations is very important to ensure a correct

lifetime. If using standard 0.25W resistors for that purpose, split Rp into two equal resistors Rp/2 in
series. This is not a problem of power dissipation but a question of maximum allowed voltage per
resistor to guarantee a correct insulation.
Example for the russian GM tube SBM-19 that I use: Typical load resistor of 10 Megohms
specified => Minimum 2 x 4.7Meg in series.

Construction details

As it can be seen on the schematic, the high voltage components surrounded by a dashed line
delimiting the "HV part" area shoud be mounted preferably out of the main circuit board. This
guarantees a better insulation resulting in a safer use and in the absence of leakage currents
which could exist between adjacent circuit tracks and might cause functional problems, taking into
account this kind of voltages. Therefore the HV part is connected to the main board via 4 signals:
Ground, point "TP", transformer output, and GM tube cathode. Here is my example of
implementation:

The HV components can easily be seen on the left side of the box (as viewed on the picture)
under the transformer and are welded on an old-fashionned mounting barrette. This provides high
insulation. More than two high voltage capacitors can be seen because I did not have 100nF but
47nF capacitors that I had to use in parallel. The small PCB featuring an SMD component (behind
polystirene) is simply my available replacement of a DIP18 packaged PIC16F84. It is absolutely
not necessary to use this kind of package.
The piece of polystirene is only intended to keep wires away from the GM tube to avoid low
voltage circuit disturbance by the high voltage pulses. The wires travelling from the high voltage
circuit to the tube terminals are placed so that they form the smallest possible loop to avoid
excessive device sensitivity to radio interference (non-ionizing radiations).

I've chosen to put the battery packs inside the box and to mount plugs for an external power
supply on the side protecting parts. I've added the protecting parts not only for that purpose but
also for hiding the GM tube extremities which are secured by the box itself where the GM tube is
hold by a shock absorbing foam:

The GM tube terminals are connected to wires by the mean of old springs cut and formed to fit
with the situation. Load resistors are located as close as possible to the anode, so that the anode
circuit (at very high impedance) is not subjected to radio interference collected by its wire.
However it is not guaranted that bringing for example a cell phone close to the device will not
cause disturbance such as false counting. But this has nothing to do with radioactivity, although
strong non-ionizong radiation is also quite bad for your brain ;-).

Since my GM tube is sensitive to gamma and also hard beta radiations, I've made an aperture in
front of it for not stopping beta particles:
The tube is sensitive all along its body. It is not a window GM tube. The efficiency is maximum
when incidence of particles is perpendicular to the tube.

Setting up and checking

As previously specified, it is better to adjust the high voltage level before connecting the GM tube.
A common digital multimeter able to measure 1000VDC can be used for this measurement. The
HV generator withstands a permanent load of 4Meg without dropping down and it is expected that
the multimeter set at 1000VDC will show a greater impedance.

It is also better to operate first the device without the microcontroller inserted. (I suppose that
anyway you use a DIP18 socket for this component).

When the high voltage output is at its ideal value and correctly compensated, the GM tube can be
connected (while the power is switched off, of course!). After connection and power-up, a +5V DC
voltage subjected to short negative pulses (50-200us) must be visible on RB0. Pulse occurence
must approximately correspond to the specified GM tube background.

Turn "Scale Adj" toward ground and connect a jumper wire beetween pins Vdd and RB1 of the
microcontroller socket. Then turn "Scale Adj" smoothly back until the galvanometer pointer
reaches the maximum position. The galvanometer is then calibrated.

Switch the power down, insert the microcontroller and restart the device. A test beep must be
heard on start-up and the LED must emit two flashes. Then the counting immediately begins while
the buzzer produces audible clicks on pulse detections. After the 10 first seconds, the LED flashes
once again and the pointer shows the background radiactivity level (small move). Every 10
seconds the pointer position is updated according to the last counting result.

Specific and alternative components - Microcontroller programming

* GM tube: Depending on the country where you live, it might be difficult to buy GM tubes online. I
found a quite large choice of GM tubes at http://gstube.com. This russian supplier also ships to
Western Europe and US and I've been fully satisfied of the product good condition and delivery
time. Among the available products on his website I'd better recommend tubes having a length of
more than 15cm, sensitive at least to hard beta radiation, and for which characteristics are
available. They are more expensive than tiny and cheap tubes but much more sensitive. This is
logical.
These kinds of window-less tubes with thin metallic walls must not be subjected to external
pressure. Handle them with care to avoid crushing them !

* Others: If you don't succeed in finding the transistor or passives specified in the schematic,
you'd have to take care of the following parameters for alternate components:

- BC337: NPN, Icmax > 50mA, Hfe > 60

- 2N2222: NPN, gain-bandwidth above 100MHz, Hfe > 80, Icmax > 50mA, switching transistor
recommended
- Piezoelectric buzzer: Any kind of ceramic disk based piezo buzzer (without built-in oscillator). It is
just a question of noise.
- PIC16F84: PIC16C83, PIC16C84, PIC16F84A (other PIC microcontrollers if adapting the source
code).
- Crystal: The provided PIC source file is fitted for a 2.097MHz crystal.
You'll have to adapt timer ratios as well the serial transmission routine in case of use of another
frequency.
However, if you don't care about serial data, choosing a frequency close to 2MHz keeps the
device usable without code change:
=>New linear full scale value on galvanometer = (2.097MHz / Actual Frequency) x 25.5cps.
=>New logarithmic scale on galvanometer: Slightly offset, but this might not be noticeable.
=>New counting interval = (2.097MHz / Actual Frequency) x 10 seconds.

If you have an exact 2.00MHz crystal, I leave this replacement routine to help you in adapting the
serial transmission to this crystal frequency.

The ready-to-use HEX file is available, but adapted for 2.097MHz.