Introduction

Though somewhat outside of the scope of the Fusor project we have decided to create a webpage dedicated to gamma spectrometry as a sub-page of the Monitors page.

With some restrictions, the technique of gamma spectrometry can become of interest when we use e.g. metal foil activation by neutrons for measuring neutron output of the Fusor.
The details of how to perform this measurement can be found in reference 1.

The basic setup for gamma spectrometry consists of a scintillator-photomultiplier unit (the detector), a high voltage power supply, a pre-amplifier (preamp), an analog-to-digital (ADC) converter and a computer for analysis by means of a suitable program.

The parameters of importance for measuring radioactivity are amount, time and energy. Of these parameters measuring energy is the most tricky part.

For professional purposes the setup of gamma spectrometry equipment was usually done by connecting a number of NIM-modules (NIM stands for Nuclear Instrumentation Module), which were powered in the so-called NIM-bin, a storage crate for NIM modules containing a multiple voltage symmetric DC power supply. The NIM modules used for gamma spectrometry are a pre-amplifier placed very near to the detector, respectively followed by and connected in the NIM-bin: a high voltage (HV) power supply, a (pulse shaping/spectroscopy) amplifier a pulse height analyzer (PHSA) an analog to digital converter (ADC), a computer system and a program. Everything after the amplifier is the multichannel analyzer.

The pre-amplifier converts the primary charge pulses into voltage pulses, where the incoming charge is proportional to the output voltage. The amplifier amplifies the pulses up to a range from 1 to 10V. The mutichannel analyzer evaluates the incoming pulses with respect to the height (voltage), which is relative to the energy and sorts them in consecutive pulse-height channels.

For the amateur with an interest in gamma spectrometry the setup of necessary equipment is nowadays a lot simpler thanks to enthousiasts who took the time and effort to write software, which made it possible to use the sound card of a personal computer for sampling the pulses from the scintillation detector and digitise them at 48 kHz or sometimes 96 kHz sampling. The software can process the signal by calculating the RMS value, filtering out and skipping bad pulses and display them on screen as e.g. a pulse height histogram. Examples of this (free) software can be found for Marek Dolleiser's program PRA (Pulse Recorder and Analyser) in reference 2, the excellent from Italy originating open source program Theremino in reference 3 and the less known from Japan originating BecquMoni program in reference 4. Another excellent program, though not free, is FitzPeaks, which can be found in reference 5, but it must be mentioned here that for our purposes the NaI program of Fitzpeaks is the most appropriate and it can be found on their webpage in the right hand bottom corner.

Of all these programs we shall take Theremino as our reference program because it has excellent characteristics and it is by far a favourite amongst gamma spectroscopists, though the website is a horror to find your way on and moreover to understand the writing because apparently a Google translator has been used to translate from the original Italian language into English. Another disappointing issue with Theremino is the fact that they make clever designs (which are said to require special components) but for some of them (in particular the PMT-adapter) it is impossible to get hold of the circuit boards and components. There is no webshop connected to the Theremino website and the (official) Chinese and European outlets (warehouses) do not sell the PMT-adapters or printed circuit boards. Instead you must send an e-mail to a guy in Italy who might or might not sell you a kit or an assembled set (after sending my request I never got an answer from this guy). It must be admitted that the Theremino website does give access to the (Eagle) files which will make it possible to order this pcb from a pcb manufacturing service, but than you must order a minimum of ten boards and who needs ten boards for a project?
On occasions it will be possible to source this PMT-adapter from a commercial seller in the USA but the price is (in our opinion) quite excessive, USD 225, and for shipping the small package from USA to Europe a ridiculous USD 62.50 will be asked. We are aware of the prices of all the used components inside the unit and we know that the time of building the unit is a couple of hours and therefore we do not consider the selling price as fully jusitified.
In our opinion the intention of the open source project Theremino should be to encourage enthousiasts to enjoy their hobby with a minimum of costs. That means that the creative minds behind the website should ensure that proposed pcb's will be easily available at a decent price to enable also enthousiasts with a modest budget to exercise their hobby. But that is just my opinion.

We shall proceed to discuss the required parts for our gamma spectrometer in the following paragraphs.

Equipment

Scintillation Crystal

The detector consist of two parts, connected together: a scintillation crystal and a photomultiplier of which the scintillation crystal has the function to convert gamma radiation into light and of which the photomultiplier has the function to convert light into electrons and to amplify these significantly into a measurable electric (charge) pulse.

Different materials can be used for detecting gamma radiation, such as a transparant plastic, named Bicron BC-412, but also grown crystals of doped materials such as NaI(Tl) (image 1).
These are naturally grown sodium iode crystals which have been doped with traces of thallium, which has been added at a mole ratio of approximately 0.1%.

Image 1: Russian NaI(Tl) crystal with dimensions 30 mm diam. and 95 mm length (© supplier Ukrania)

Because the plastic scintillators have a low sensitivity for the lower gamma energies, we prefer to make use of a NaI(Tl) crystal. These can be obtained relatively cheap usually from Russian or Ukranian sellers on eBay but care should be exercised that the crystal is in a usable state. This means that the NaI(Tl) crystal should be milkwhite and not yellowish or even brownish. A yellow/brown color comes from free Iodine that has been released and migrated into the crystal latice. The NaI crystals are hygroscopic and therefore they are contained in a hermetically closed metal shielding. Also the NaI crystal should not have cracks or broken off chips, though a single mild flaw does no harm much.

The open part of the crystal (where the NaI surface can be seen) is called the optical window and is highly polished. Plastic scintillators usually have all sides polished and one of them is chosen as the window. The other ones are the open outer faces and these should be covered.
For optimising the detection of the tiny light flashes from entering gamma's, it is advised to cover these open outer faces of the plastic crystal with e.g. white teflon tape in order to enhance reflection.
For further optimising the transport of photons into the photomultiplier an optical coupling compound is applied between the crystal optical window and the photomultiplier optical window.

The characteristics of a NaI(Tl) crystal are:

Density (g/cm²)	3.67
Time constant (ns)	230
Luminescence wavelength (nm)	420
After glow	0.5-5%
Relative light intensity	100
Refractive index	1.85

Table 1: Characteristics NaI(Tl) Crystal

For gamma rays with an energy below 10 MeV, the main effects of interaction between the incoming gamma ray and the material of the NaI scintillator crystal are photoelectric absorption, compton scattering and pair production.
The process of photoelectric absorption, the dominant reaction below 400 keV, is the direct interaction between the incoming γ-ray and NaI, especially the Iodine atom. The photon is totally absorbed by the atom and the energy is transferred to an outgoing electron from a shell of the atom. The energy is the factor determining from which shell the electron will be excited. The orbital electron that will be removed is called the photoelectron. The atom may remain in a higher energy state until the vacant orbital place is occupied by a shifting orbital electron (de-excitation of the atom) followed by emission of low-energy photons in the blue end (near UV) of the visible spectrum. For the NaI crystal this is at 420 nm, a wavelength that can be seen (detected) by a suitable photomultiplier.
Basically, the function of the scintillation crystal is linear conversion of gamma's into visible light pulses.

The detection efficiency of the scintilation crystal is expressed by

where
N = count rate under the photo peak
T = time of measurement
A = activity of the radioactive source
I_γ = relative intensity of all energies of the radioactive source.

A further parameter of the scintillation crystal is the energy resolution expressed by

E_r (%) = (FWHM / Ch_no) x 100%

where
FWHM = full width at half maximum, the width of the photo peak in eV or MeV at half the peak height
Ch_no = energy of the photo peak position, usually expressed as the channel number.

An example of the latter equation is calculated by assuming a measured FWHM of 9.15 keV at a photo peak position in the 122 keV channel, which yields an E_r of 7.5%.
A NaI crystal with almost ideal dimensions of 75 mm diameter and 75 mm thickness can achieve an energy resolution of about 6% for the 662 keV γ-ray from Cs-137.

The light decay time constant in Nal is about 0.23 µs. Typical charge-sensitive preamplifiers translate this into an output pulse rise time of about 0.5 µs.

The relation between the dimensions of the NaI crystal and the efficiency comes from the stopping power of high Z of NaI(Tl) for γ-rays. Measuring higher gamma energies will require a thicker layer of the scintillation crystal, whereas a thin layer of scintillation crystal (e.g. 1 mm thick) will enable to detect low energy X-rays. Therefore, an optimum crystal thickness exists for a certain γ-ray energy.
Increasing the diameter of the scintillation crystal increases also the solid angle under which the detector sees the source. This will increase the detection efficiency. It should however be noted that the outer diameter of the scintillation crystal should be less than the outer diameter of the photomultiplier, because the outer rim of the PM is a less sensitive area.

Looking at our NaI crystal of image 1 we may conclude that with an outer diamter of 30 mm this crystal has obviously been designed for a PM with an outer diamter of 1.5 inch (38 mm) and that the thickness of the crystal of 90 mm destines it for detecting high gamma energies.

Photomultiplier

Light (photons) are being converted into photoelectrons by absorption in the cathode in a photomultiplier and pulled by an electric field to a dynode and amplified. Usually ten to twelve dynode steps are performed, with an amplification in the order of 10⁶.

The electrical field on the dynodes is build up by means of a high voltage that is applied to the dynodes by means of a resistor network acting as a voltage divider.
The change in charge of the anode in the PM is proportional with the emission of photoelectrons from the cathode, multiplied by the secundary emissions from the dynodes which is proportional to the applied high voltage. This implies that the amplification of the PM can be regulated with the (high) voltage applied. As a rule of thumb a raise of 85V yields about a factor 2 in puls height, though within certain limits.

In a simplified imaginary model the detector could be considered as a capacitor with a constant charge to form the electrostatic field, for separation of electrons and ions, and a variable charge due to the intensity of the (gamma) radiation. The height of the pulse, variably dependent from the energy of the absorbed radiation, is for a large part dependent from the capacitance of the detector because with a low capacitance of the detector with an unchanged charge the highest possible potential will be obtained, as shown in the equation

V = Q/C.

Knowing the importance of striving for a low capacitance of the detector (and a high impedance) we should have a look at the particulars for the transport of the pulse to the amplifier, which is usually done by a coax cable. Connecting a coax cable to a detector means that we connect the capacitance of the cable parallel to the capacitance of the detector, which means that the capacitances are summed. Taking a RG-58 cable as an example we will find that this cable has a capacitance of 100 pF per meter and an impedance of 50 Ohm.
Needless to conclude that the lenght of the cable between detector and preamp should be as low as possible. For that purpose sometimes the preamp is placed directly near the PM, as can be seen in image 3.

Our Burle (formerly RCA) photomultiplier (PM) in image 2 comes from a gamma camera as used in nuclear medicine.

Image 2: Burle Photomultiplier (© FRS 2017)

The hexagonal head shape of the Burle PM permits to place a large number of the PM's very close together to obtain a very large surface of PM's.
These Burle PM's are specially made to order and therefore it will be difficult to find a datasheet.
It is believed that this PM is identical to the Burle S83020F, for which a datasheet can be found in reference 6.

Reference 7 shows in a simplified calculation the performance of the scintillation crystal and the photocathode and dynodes of a photomultiplier:
when
e = the electric charge carried by a single electron, 1.6x10^-19 C
m = number of light photons produced in the crystal
k = optical efficiency of the crystal, i.e. the efficiency with which the crystal transmits light
l = quantum efficiency of the photocathode, i.e. the efficiency with which the photocathode converts light photons to electrons
n = number of dynodes
R = dynode multiplication factor, i.e. the number of secondary electrons emitted by a dynode per primary electron absorbed,
then the charge collected at the anode is given by the following equation:

Q = mklRⁿe

In a NaI crystal, depositing 1 MeV of energy will cause about 38,000 scintillation photons to be emitted. Assuming that the energy transfer is linear, a 100 keV signal absorbed in the crystal will generate about m = 3,800 photons. Assuming also that this crystal has an optical efficiency of k = 0.5 (or: only 50% of all photons reaches the cathode) where the photocathode has an assumed quantum efficiency of l = 0.15. Furthermore, we assume that our photomultiplier has n = 10 dynodes and an assumed multiplication factor of 4.5 then substituting these values in our equation we find a charge of Q =3800(0.5)(0.15)(4.5¹⁰)(1.6x10^-19) or Q = 155x10^-12 C, which equals a charge of approx. 155 pC.
Such a small electrical charge requires a pre-amplifier for reasons described below in the next chapter.

We also acquired two Russian made photomultipliers, FEU-28 and FEU-35A, both with sockets included. The characteristics are compared in table 2:

Table 2: Russian PMT characteristics
Characteristiscs	FEU-28	FEU-35A
Typical use	photoelectron multiplier for indication and measurement of weak light fluxes in the red and infrared regions of the spectrum	photoelectron multiplier for detection of weak X-ray radiation in wavelength range 0.05-0.25 nm
Photocathode	silver oxide-caesium, spectral characteristic No. 1	antimony-caesium, spectral characteristic No. 6
Optical window	top side, flat surface	top side, flat surface
Diameter of operational area of cathode	25 mm	25 mm
Dynodes	8 amplification cascades	8 amplification cascades
Form factor	glass mounting with base	glass mounting with base
Mass	50 g	60 g
Principal data
Range of spectral sensitivity, nm	400 - 1100	300 - 600
Range of maximal spectral sensitivity, nm	650 - 850	380 - 420
Photocathode sensitivity, microA/lm when U_pt ≤ 1.3 kV)	≥ 15 -	- ≥ 45
Anode sensitivity, A/lm when U_pt ≤ 1.2 kV when U_pt≤ 1.6 kV	1 - -	- 10 30
Dark current, A with anode sensitivity 10 A/lm	≤ 3.10^-7 -	- ≤ 10^-8
Anode current, microA	100	≤ 50
Supply voltage, kV	≤ 1.6	≤ 1.6
Threshold sensitivity, lm/Hz^1/2	≤ 1.1 . 10^-10	-
Lifetime, hours with feed voltage 1.25 kV	≥ 1000 hrs -	- ≥ 3000 hrs
Evaluation Criteria
Anode sensitivity, A/lm with feed voltage = 1.6kV Photocathode sensitivity, microA/lm Threshold sensitivity, lm/Hz^1/2	1 1.10^-5≤ 1.10^-9	10 - -
Outline of the circuit of voltage dividers	Split the voltage unevenly: R1=0.4R;R2=1.6R R3-R12=R ≤ 0.3 MOhms. I - to load; II - to anode; III - 0 to feed source; IV - photocathode; V - to dynodes.	Split the voltage unevenly: R1=2R; R8=R9=1.5R; R2-R7=R ≤ 100 kOhms. Capacitance C ≤ 0.05 microF. I - to photocathode; II - to dynodes; III - to anode; IV - to load; V - 0 to feed source.

According to the characteristics only the FEU-35A is probably suitable for combining it with a NaI(Tl) crystal because its spectral range has an optimum between 380 and 420 nm, near optimal for the NaI(Tl) crystal with its luninescence wavelength output at 420 nm. Apparently its predominated use as a PM for detection of weak X-rays will require a thin layer NaI(Tl) crystal. However, for that purpose nowadays usually Si(Li) detectors will be used, or, more specifically, XRD detectors.

Surprisingly, the FEU-35A was sold including a black rubber sleeve containing a NaI(Tl) crystal with dimensions outer diam. 31 mm (optical window diam. 26 mm) and thick 8.7mm (including aluminium pot housing). Apparently, this appears to be a 1/4" crystal, which has its optimum resolution and sensitivity for energies in the 60 - 80 keV range. This range is useful e.g. for nuclear medicine investigations with the radionuclide ²⁰¹Tl, in the past used for visualizing damage from myocard infarct. The element Thallium has the same biological pathway as Potassium and accumulates in living heart tissue. In necrotic heart tissue, caused by myocard infarct, no uptake of ²⁰¹Tl is seen, thus providing information about the amount of damage caused by the infarct.

The socket for the tube was also included with the resistors still in place measuring a total series resistance of all the resistors in the voltage divider of 1.884 MOhms (image 4).

Image 4: FEU-35A Photomultiplier with socket, NaI(Tl) crystal and sleeve (© FRS 2017)

A fourth photomultiplier that we acquired is the RCA 6199, a rather popular PMT that was manufactured by RCA, Burle and Hamamatsu starting in the 1960's (image 5).

Image 5: RCA 6199 (Source: Supplier)

It is a 38 mm (1.5") 10 stage end-window PMT with a minimum useful diameter of 31.5 mm and a maximum response at a wavelength of 400 ± 30 nm. The maximum supply voltage between anode and cathode is 1000 V.

The power supply to the PMT and the voltage dividers are shown in image 6:

Image 6: Power Supply Circuit and Voltage Dividers for RCA 6199 (Source: Supplier)

Pre-amp

A signal coming from the photomultiplier (PM) should normally not require amplification. However, the requirement that the impedance of the detector should be high, whereas the impedance of the cable is usually low, asks for a solution like a circuit which combines a low input capacitance and a high input impedance with low output impedance. The disadvantage of such a circuit is that it adds noise to the signal and the length of cable to the spectrometer also has a negative impact on the signal/noise ratio. Noise is always produced in all conductors and semi-conductors which conduct a current and noise appears as small random fluctuations of voltages in a frequency spectrum between f = 0 and f = > the operational bandwith of the amplifier, according to:

v_n = √(4k_BTRΔf)

where
k_B = Boltzmann's constant in Kelvin per joule
T = temperature of the resistor in absolute temperature (Kelvin)
R = the resistance in Ohm
Δf = the frequency of the bandwith in Herz
and v_n is the RMS noise voltage.

In order to increase the signal/noise ratio prior to the transport of the signal through the cable(s) usually a pre-amplification will be applied.

Image 7: Burle PM with Divider Resistor Board and Pre-amplifier Board (© FRS 2017)

Image 7 shows the Burle PM with a pcb containing the divider resistors and the capacitors for the high voltage near to the neck of the PM (the black ring) and at short distance connected to the divider pcb we find (on top) another pcb with the pre-amplifier. This placement of the preamp very near to the PM is fully in line with our conclusion above that the connecting lines between PM and preamp should be as short as possible.

For estimating if we can make use of these two pcb's we need to sort out the schematic of the boards and the values of the components (images 8 and 9).

Images 8 and 9: Preamp PCB on Burle PM (©FRS 2017)

When the preamp present with the Burle PM cannot be used than we may need to make use of another, suitable preamp. The choice of preamps consists of two types: the normal "voltage" amplifier and the "current" or "transimpedance" amplifier. A voltage amplifier transforms the voltage at the input with a high impedance into a higher voltage at the output. A current amplifier transforms the current at the input with a low impedance into a voltage at the output. Current amplifiers usually are used to get fast signals from detectors representing a current source with a high parallel capacitance.

AD8541AKSZ-R2

Another preamp that we can use is the charge sensitive preamp as shown in image 10.

Image 10: Charge Sensitive Preamp (© Supplier)

It can work in a positive or negative HV-voltage configuration and the pulse duration can be changed from ± 50 µs to ± 250 µs. A 24-bit audio card is required for best results, yielding 9% FWHM at > 1000 cps. The amplification factor can be adjusted by adjusting the high voltage output on the power supply. The output of the preamp is provided by an SMA connector.
Input voltage ranges from 3.5 - 12 VDC, the opamp is 10 MHz and the pcb dimensions are 29x24 mm. It can be fed from the USB connector on the computer and its dimensions will permit installation near the PMT.

The electronic circuit for connecting the preamp to a PMT and a HV power supply is shown in image 11:

Image 11: Shematic for connections of PMT, Preamp and HV Power Supply (© FRS 2017)

HV-Power Supply

The high voltage power supply (HVPS) that we intend to use is kept as simple as possible by using a Hamamatsu C4900 module, which uses 15V DC as input for generating up to minus 1250V DC output (image 12).

Image 12: Hamamatsu C4900 High Voltage Power Supply Module (© FRS 2017)

The Hamamatsu C4900 HVPS module is rather tiny, which can be seen in comparison with the 1 Euro coin. The total width of the C4900 module is 46 mm. The specifications can be found in reference 8.

By adding only thee more components we can reduce the ripple in the output with 50% and we will be able to set the output voltage at a chosen value between -200 and -1250V DC.
These components are an electrolytic capacitor of 47 µF at 25 V, a capacitor of 0.2 µF at 2 kV and a ten turns potentiometer of 50 kΩ (image 13).

Image 13: C4900 High Voltage Power Supply (© FRS 2017)

Because it is our intention to make the gamma spectrometer as a portable unit, the DC power supply for the C4900 module will consist of ten AA type batteries of 1.5 V each in series, contained in a battery holder.

When a positive output high voltage will be required, than we can make use of the HV power supply unit from image 14:

Image 14: HV Power Supply positive output (© Supplier)

This regulated, stabilized HV power supply has an input voltage of 4.8 - 5.5 V DC and will deliver an output of +550 up to +1000V DC. The current drawn at 1000V with a 100M load is 10 mA and therefore the power supply could be fed by the USB port on the computer. The dimensions of the pcb are 140x45 mm. Installing the pcb inside a metal housing is mandatory for suppression of noise on the photomultiplier. The circuit includes a RC filter for the photomultiplier.

Cables

It is imperative to connect the output signal from the preamp to the sound card with a shielded cable (image 15).

Image 15: Shielded Cable for Sound Card (Source: Supplier)

This means that we will either have to add a 3.5 mm jack connector (image 16) to the preamp (when it has no output connector) or to add a BNC or SMA connector to the cable (when the preamp has a BNC or SMA output connector like in image 10).

Image 16: Connector for Preamp (Source: Supplier)

Sound Card

The sound card is a rather critical part in the setup of the gamma spectrometer as it contains the ADC for signal processing the output from the photomultiplier/preamplifier chain. It could be tempting to use the sound card that is present in your desk top computer or your notebook but this is not a good idea. Apart from the possibility that a wrong action might blow up your sound card, which is a (financial) disaster when it is incorporated in your computer, the internal sound card is subject to frequency related interferences from nearby electronic circuits inside the computer housing and negatively impacting the ground noise in the sound card. Moreover, for complying with the specifications as laid down by Theremino we will have to adapt the sound card with some extra electronic components and circuit changes, which is a permanent change and therefore unwanted for the internal sound card.

Considering these reasons it is probably a good decision to use an external usb sound card, which in case of a 16-bit sound card has to be slightly modified in order to improve the noise level to a lower value. All this has been described in a document published by the Theremino website in reference 9.

Unfortunately buying the recommended USB sound card adapter could turn out to be a bad surprise. The reason is that this USB sound card, usually originating from China for a low price, can look on the outside similar to the one as recommended in the Theremino manual (dated 2013) but inside the housing a totally different looking circuit nowadays is present. With time progressing and with increasing sales of these usb sound cards, they have been modified for cheaper manufacture by integrating some of the passive components, changing the circuit, omitting components (down-engineering) and by potting the electronics inside. Such a "modern" usb sound card cannot be noise improved conform the Theremino recommendations. The main change is that the IC-chip now is the CM108B (formerly the CM108), which is without an external crystal but has an internal oscillator circuit and "which integrates the USB transceiver, PLL and regulator modules, meaning that only a few passive components are necessary for the USB interface connection. Default USB descriptors are embedded in the CM108B, so no additional design effort is needed for generic USB operation". This quote is from the manufacturer, CMEDIA.

USB sound cards with a CM108 IC-chip make use of a 16-bit ADC/DAC. For use in gammaspectrometry a 24-bit ADC probably is a better choice because of their 192 kHz sampling, yielding a better S/N ratio of 108 dB and a total harmonic distortion of -90 dB. An example of our 24-bit sound card, a HD 7.1 USB Sound Card A01NC, which has a CM6620 IC inside, is shown in image 17:

Image 17: AS372 or A01NC 24-bit USB Sound Card (Source: supplier)

A description of the AS372, a manual, drivers and firmware can be found on the manufacturer's website (reference 10).

Currently (2017) a suitable usb sound card with the original CM108 IC, that still can be noise improved, can be found on eBay as shown in images 18 and 19:

Images 18 and 19: USB Sound Card with CM108 and external X-tal (© Source: Supplier)

The general electronic circuit for this sound card is shown in image 20:

audio card schematic

Image 20: General Schematic Circuit for USB Sound card with CM108

The circuit as shown is a general circuit as it contains components that can't be found on our USB sound card because it has a far simpler set-up:

MOSFET Q1 is not present because it is only used in configurations with an SPDIF output;
Push buttons are not present because they are used only in versions with adjustable volume;
Two capacitors of 470 µF/6.3V connected to the headphone jack are sometimes missing because of cost reasons. If present they can be removed to gain space and we do not use them in our configuration.

The green speaker or head phone jack can also be removed from the circuit board to gain some space and we do not use it in our application.

It is imperative for our use of the sound card as a gammaspectrometry ADC to check that the following components are absolutely present on the sound card:

Microphone jack;
Capacitor C11;
Resistor R13;
Capacitor C15;
IC CM108;
LED
X-tal Y1 with resistor R16 and two capacitors C12 and C13;
Capacitors C3, C4, inductors L1, L2 and resistors R1, R2,R3.

Once large unnecessary components have been removed and we have checked that the minimally required components are present, we can start changing the circuitry by adding components as described in reference 8.

The purpose of changing the circuit is noise reduction and current limiting in case of shortcutting by inserting or removing the mic jack plug from the mic jack connector.

Because our USB sound card is different from the originally advised (apparently obsolete) adapter, we will have to do some additional work, which is not easy to perform because of the tiny size of the USB sound card and because of the printed circuit copper traces being difficult to reach.

In general for any USB sound card the following changes will be required:

A high-pass filter needs to be added to the microfone input by adding a capacitor;
5 V DC from the USB card needs to be connected to the central connection of the mic jack for powering the HV module and the preamp;
5 V DC from the computer through the USB card needs to be filtered from noise by adding a low ESR capacitor and a current limiting circuit;
For realizing these modifications a pcb track needs to be cut as well as a soldering leg from the mic jack connector on the sound card;
Two capacitors of 470 µF (if present), connected to the headphone jack, should be removed from the USB sound card;
The headphone jack should be removed from the USB sound card.

Despite the fact that this procedure has been described extensively in reference 8, we shall show the modification in detail in the following steps:

Step 1:

Opening the housing of the USB sound card and checking if the card is suitable for modification:

The curent limiting circuit (source: Theremino) is shown in image :

Miscellaneous

Calibration

For the purpose of identifying unknown samples it will be required to calibrate the gamma spectrometer with sources of known radionuclides, preferably with a known strength.

Nowadays, it is a bit difficult as a private person easily to obtain calibration standards. However, it will be possible to find objects which contain a low amount of radionuclide(s).

Spark Gap Tube

Gas tube arresters (or spark gap tubes) are devices that can be used to protect circuits from high voltages by conducting such a high voltage to ground, but also as triggers for pulsed high voltages and high currents. Once the nominal high voltage has been reached, the spark jumps over the gap into the follower circuit. For better functioning, the tube is filled with a small amount of an ionizing substance (a radionuclide), such as Cesium-137.

An example is our C.P. Clare TG-36 HV spark gap tube (image ):

This particular TG-36 spark gap tube was manufactured around 1975 - 1985 and it contained at that time < 1 µCi ¹³⁷Cs,equivalent to < 37 kBq or < 37000 desintegrations per second. Cesium-137 has a half-life of 30.17 years and therefore less than half the amount of radioactivity can be expected to be present.
The radionuclide emits a single γ-line at 662 keV, yielded by ^137mBa, a radionuclide in equilibrium with ¹³⁷Cs.

Thorium Lantern Mantle

Thorium lantern mantles, manufactured before the year 2000, contain Thorium-232. It will not be easy to find these Thorium lantern mantles, because they are obsolete now.
This radionuclide can easier be found in some (colour coded with a red tip) thoriated tungsten welding rods for TIG welders (containing about 2% ²³²Th).

Because the welding rods consist of Tungsten and Thorium only, Thorium can be separated by putting the rod(s) in hot concentrated Nitric Acid. Thorium will go into solution as Thorium Nitrate, but Tungsten will not as it is not reacting with acids. Heat the Thorium Nitrate solution to 450 - 500°C to convert the nitrate into Thorium Oxide. Wash with demineralised water and damp dry until solid Thorium Oxide is obtained.

A different approach is dissolving the welding rod(s) in 3% Hydrogen Peroxide (H₂O₂) until fully dissolved and crystallized. This will take quite some time. The yellowish product obtained is a mixture of Thorium Dioxide (ThO₂) and Tungstic Acid (H₂WO₄). Dissolve the crystals in a volume of 5 ml per single (one) welding rod of saturated Sodium Bicarbonate solution (Na₂CO₃) and heat for a few seconds. The yellowish crystals will dissolve, soluble Sodium Tungstate (Na₂WO₄) will remain in solution and an insoluble (yellowish white) Thorium Dioxide will precipitate. Add again 5 ml of saturated Sodium Bicarbonate solution (or multiple amounts of this volume when more than one rod is processed) and heat for a few seconds to react with possible remaining tungsten. Filter off the Thorium Dioxide over a paper filter and wash with demineralised water. Dry the precipitate at 105°C in an oven. A yellowish powder of Thorium Dioxide will result.
Reaction: H₂WO₄ + Na₂CO₃ ⇒ Na₂WO₄ + H₂O + CO₂î
Thorium Dioxide powder should be kept in a hermetically closed glass ampoule, which is placed in a polyethylene jar and stored in a safe way. Image shows a glass ampoule containing 1 (one) gram of Thorium Dioxide, 99% purity, processed from TIG welding rods:

Note: Dissolving TIG welding rods can be accelerated by using a stronger concentration of Hydrogen Peroxide, up to 50%. However, selling this high concentration of Hydrogen Peroxide is in the EU since 2016 permitted only to registered professional users because the substance has been listed now as a precursor for explosives. Usually, concentrations up to 12% are exempted from this regulation. The disadvantage of using higher concentrations peroxide is that gas production (CO₂ and water vapour) becomes more violent, with the risk that radioactive thorium particles may be ejected as an aerosol in the air, possibly endangering the operator's health.

Warning: this is a dangerous procedure because you a dealing with a very poisonous, radioactive substance (Thorium-232), a strong oxidizing substance (H₂O₂) or a concentrated strong acid (HNO₃). Do not try chemically processing welding rods, unless you are a skilled laboratory professional (chemist), working in an appropriate workplace (laboratory) equipped with appropriate materials such as eye- and respiratory protection, protective clothing, a hermetically closed glove box with fumes extraction, appropriate glassware, ceramic crucible, etc.

The gamma spectrum of Thorium-232 may show γ-radiation from the ²³²Th decay chain, such as ²¹²Pb at 239 keV.

Smoke Detector

Smoke detectors, manufactured before 2005, may contain Americium-241 in an amount of 0.9 µCi or 33.3 kBq of ²⁴¹Am with a half-life of 432.52 years. Americium-241 decays to Neptunium-237 with a half-life of 2.144 x 10⁶ years.

Our smoke detector (image ) was manufactured in the early 1990's and the relatively long half-life means that almost all ²⁴¹Am radioactivity is still present, but also that about 5% of ²³⁷Np has accumulated.

For use as a radiation source we will have to take the smoke detector (very carefully) apart by dismantling the ionization chamber (the black cilindrical container with the radioactive symbol on the circuit board) and by taking away the radiation source, which is placed on a small disc. Store the Americium-241 source (like all other sources) in a plastic bag in a closed and locked metal container out of the reach of children and away from unqualified adults.

The gamma spectrum may show γ-radiation from ²⁴¹Am at 59.54 keV (highest peak) and 26.34 keV (lower peak). Also a small röntgen peak at about 18 kV can be seen.

LYSO Crystal

A LYSO-crystal is a scintillation crystal that is slightly radioactive (image ). The scintillator consists of a Lutetium based and Cerium doped scitillation material. Lutetium is present as Lutetium-176 that decays with a half-life of 3.78 x 10¹⁰ years to Hafnium-176, in 99.66% of the time to the 579 keV excited state.

Image : A rectangular LYSO Crystal Scintillator (Source: Supplier)

¹⁷⁶Lu decaying to ¹⁷⁶Hf in the 579 keV excited state emits three gamma's at 307, 202 and 88 keV.

Image shows a gamma spectrum of our LYSO crystal with peaks (from left to right) on the shoulder of the first peak at 54 and 61 keV from internal β decays, a γ peak at 88 keV, a γ peak at 202 keV (the highest peak in the center) and a γ peak at 307 keV (the last large peak at the right):

Image : Gamma spectrum of LYSO Crystal

Uranium glass

Another source of known radiation can be found in uranium glass objects, manufactured prior to 1945, such as the pre-war Bohemian glass buttons shown in image :

Uranium glass objects are still being manufactured but since 1945 only depleted uranium (i.e. Uranium-238 with Uranium-235 separated from it) is used for the manufacture, which therefore has no significant radiation from Uranium-235.

In pre-war glassware, containing Uranium-238 and Uranium-235 we can see γ-radiation peaks from the decay chain of these radionuclides at 185 keV from ²³⁵U, at 92 and 63 keV from ²³⁴Th.

Radium