|
Whatever You're occupied in, a question on measurement or at least
indication of ionizing radiations rises from time to time. For example
while buying alimentary products on some markets at regions known to
be not completely safe when concerning radiation. One should not stake
his life onto a trust in some seller. It means one needs a device, that
can feel this radiation...
I can not say that such a device is difficult to buy. On the contrary,
the market is literally full of dosimeters of "any shape and color".
A good digest of available models is given, for example here. However
after viewing of the available assortment, one gets... kinda strange
impression. The devices are either completely miserable, or absurdly
expensive. And those, that are miserable, are also absurdly expensive
for their miserability. They batten on our radiophobia.
Its a sin for a DIYer to droop in this situation. The mind immediately
thinks: "Phoey, it's simplier to make my own device! With current prices
of those arduinos and displays it won't be expensive."
And not. Suddenly it appears that the sensitive element itself (Geiger-Muller tube, photomultiplier with scintillator or semiconductor detector)
cost almost the same sum as the ready device does.
Next step: "Can one make the sensitive element all by oneself? And once
made, to calibrate it somehow? And if it was impossible to calibrate
reasonably well, it maybe possible to use is just as indicator, which shows,
where is more and where is less."
Brief surfing over the vastitude of the net, in search of what "DIY-industry"
can offer, has brought the next results:
♦ For the first glance it looks like usage of the photodiode is most
promising. Let's start with it.
The schematics of the device, given in instructions (see reference above)
is shown on fig 1. (Redrawn by me using LTSpice):
Fig 1. The offered here schematics of radioactivity
indicator based on BPW34 diode.
Originally a pot was used instead of R2 R3
resistors.
The first thing, that alerts, is usage of a MOSFET, having gate open voltage
as large as 4 Volts, as a device that should read a signal from the PIN-diode.
If this signal is large enough, no problem here. But is it really so large? -
It was to be checked out.
The next moment is that one should be pretty accurate, while assembling the
stage, containing this MOSFET. The currents there are minuscule (look at that
10 MegaOhm resistor) and any remnants of flux, grease contaminations
(fingerprints) and other leakages may be fatal for signal. One can thank
God, that voltages here are low enough - one can never mind about corona
discharge. Otherwise there could be lot more problems.
The circuit was assembled and tested. The stage with BPW34 and 2N7000 was
(as it was recommended) made by air wiring, and the rest of the circuitry
was made on a breadboard. The results were the next:
- The circuit behaves as if it worked. The LED at its output flashes from
time to time. And these flashes are irregular - very similar to how particle
count should look like.
- Potentiometer (on figure 1 it was depicted as a pair of resistors R2, R3)
is able to set the "count rate" to any necessary value.
- The "particle count rate" does not react to the proximity of source of
radiation. Even when an old aviation gauge, having its clock-face been
covered with radioactive paint, was placed in touch to the sensor the "count rate" was unchanged.
Generally this behavior is understandable. The words that this device
is sensitive to alpha radiation, one should take literally. Not as we usually
do, in sense that if a device can sense alpha, then it surely can sense beta,
and even more surely - gamma. Not. The device is sensitive ONLY to alpha
radiation. Only when heavy alpha particle hits the diode BPW34, one can hope
that the signal would be enough to open 2N7000. Maybe (with low probability)
some low energy beta particles (electrons) can do this.
However I can not test this statement. I have no alpha and beta sources,
The dial of the mentioned aviation gauge is under a thick glass and I have
no intention to break the device that has historical (and technical) value.
Smoke sensors, that they here and there recommend to use as a test source,
have happily vanished from the shelves and have been replaced by photoelectric
devices, - ROHS vigilantly watches for us for not to do anything wrong.
However If we are interested in making a homemade dosimeter rather than
clarification the reasons, why one or another circuit does not work, then
the absence of suitable source does not matter in this case. The deal is
that a device, sensible to alpha and beta, and being insensible to gamma,
is generally of no use as a dosimeter. When a device is sensible to gamma,
its ability to feel beta and alpha is of course a pleasant addition. However
if the devise does not feel gamma...
Imagine a case: a seller gives You a fish in a thick polyethylene bag.
If the walls of the bag are only 130 mcm, all the alpha particles with
energies below 10 MeV appear to be cut off. (A good place to say, that if
You had wrapped the sensor by 16 mcm thick aluminum foil, You have already
deprived Yourself the possibility to sense alpha particles with energies
up to 4 MeV. In other words - all alpha particles from common sources.)
[this conclusion was made using the data from book: Brief Handbook of
Physical Engineer. Atomic physics. Under edition of N.D. Feodoroff,
M., Gosatomizdat, 1961, pg 311]. Even more problems will be with wines
and juices. Their envelope is usually
thick (bottle or box) and it leaves no hope not only for detecting alpha,
but also for outcome even of high energy beta.
On the other hand, gamma radiation of comparatively low energies penetrate
such obstacles easily. One could ask, how can a gamma detector be of any
help if a product was contaminated by, say, beta emitting isotope? The
answer is that: in practice it is very hard to find a pure alpha or beta
radiator. Almost always there is an accompanying gamma radiation. It happens
due to various reasons: as a bremsstrahlung radiation (as in an X-Ray tube,
when beta particle is being decelerated in material of the source self)
or as the result of nucleus excitation relaxation after the decay act,
or when electrons fall to the inner shells, when the latter became free
as the result of the decay an so on. For example, even such a pure alpha
radiator as Po-210 gives a gamma quantum per 1e4 decays. And exactly these
accompanying gamma quanta allow to detect the contagion even in case it was
hidden inside a tight package.
So as "men does not live by beta alone", a detector suitable for practical
life have to be sensitive to gamma. Happily BPW34 photodiode can do this too.
However the useful signal becomes much lower here and the circuit shown above
becomes useless. One needs to make a more sensitive amplifier.
Fig 2 shows the circuit that appeared to be able to achieve reliable count
of gamma quanta.
]Fig 2. Schematics that allows to use BPW34 diode as a sensor, capable
to detect gamma radiation. Shown with preamplifier.
The circuit represents a two-stage amplifier with direct coupling, having
the voltage gain of about a thousand. The entrance transistor of the amplifier
is of low noise type (kp303i). It may be replaced by kp307e or 2N3819. After
the replacement one might need to adjust the value of R7 resistor. KT3102e
in the preamplifier can be replaced by BC337S, 2N5088, 2N5089. As in previous
case, after the replacement one may need to adjust R1 or R5 resistors.
As against the original schematics, this one loads the diode by much smaller
resistor: 300 kOhm. It gives lower sensitivity for leakages and allows to
assemble the whole circuit on a PCB, without bothering with air wiring of
the first stage.
Amplitude of the useful signal at the exit of the circuit (point "out"
on figure) is -60..-120 mV. The noise is 10..20 mV. The out port of the circuit
can be connected directly to a microphone port of any audio amplifier (clicks
on the back of noise will be audible), or to oscilloscope, or to a blinking
circuit with a LED, similar to the one, that utilizes LM358 chip on figure 1.
The feeding of the circuit should better use a constant voltage driver
(like some 7812 or LM358 voltage regulator), otherwise the level of the output
signal lowers with the discharge of the battery and one needs to continuously
adjust the trigger level of the oscilloscope, or the threshold of the
LED blinker. It is understandable, that if this happens, there can never
be any accuracy in measurements.
For people skilled in computative modeling, I will also show a circuit
with equivalent signal source. In LTSpice I succeeded to adjust the parameters
of the source, that more or less adequately models the signal that comes from
BPW34 diode, when it detects a gamma quanta (see Fig. 3.).
Fig 3. The circuit from fig 2, but with signal source
emulator. The latter simulates the signal that comes out from BPW34 diode,
when it registers gamma radiation. Current source I1 models useful signal,
current source B1 models the noise. Inductor L1 models parasitic parameters
of the battery and is used for analysis of circuit stability with partially
discharged source of power.
With knowing the parameters of the signal, You can design Your own amplifier.
The sensitivity of the resulting sensor appears to be low. When the mentioned
above aviation gauge was put in the close proximity of the sensor, it gives
1 count per each 15 seconds at average. For comparison:
a common commercial
Geiger counter SBM-20 gives about 200 cpm (counts per minute) when put at the
same place.
One reasonable workaround here is a set of multiple circuits like shown
on fig. 2. with their joint output being connected to a counting device.
Even better is the case when each circuit is connected to an individual channel
of multi-channel counter (for example it may be different ports of Arduino).
For an acceptable efficiency one needs maybe ten or more of such circuits.
It is sorrowful, but it is useless to connect ten BPW34's in parallel and
attach this stack to an input of a preamplifier. In this case the noise
from all diodes will be summed, but the useful signal will not. The useful
signal could be summed, if the passing gamma quantum would cause signal
in each diode. However unlike the heavy particles, gamma quanta do not behave
this way. All three kinds of reactions between matter and gamma quanta
(photoeffect, Compton effect and pair formation) are of pinpoint type.
There was a gamma quantum, it entered the reaction, and there is no gamma
quantum. With Compton effect, however, a new quantum with lower energy
can be emitted. But its probability to enter reaction inside our stack
of diodes is virtually the same as for the first quantum - i.e. very
very low.
Resulting verdict is: the circuit is attractive and promises high
stability of sensitivity. Precision of measurements may be comparable
with one of professional devices. However the value of sensitivity itself
for sole PIN-diode is fully inacceptable for usage in a dosimeter.
Usage of multiple diodes makes the circuit to be complicated and expensive.
The expenses for its creation may exceed the ones of buying a commercial
dosimeter of medium price.
Another shortcoming worth to note is that amplifiers of microvolt level
of sensitivity are also sensitive to electromagnetic field interferences.
If designing a practical device, one should foresee a serious shielding.
It's just not enough just to wrap it into a foil.
♦ Next turn is for homemade ion chamber (ionization chamber).
Fig 4. The scheme of homemade ion chamber.
Its scheme is winningly simple (see fig 4.), and videos of its operation
from YouTube dazzle with the efficiency. Hands are itching to reproduce. And
here one finds out a tangy bit detail: the circuit uses rather rare Darlington
transistors MPSW45A, deceitfully named "affordable" by the author of the design.
OK since we have no Darlingtons, but still have "superbeta's" - transistors
with very high (400+) current gain. Trying to assemble ion chamber on superbeta.
Result - zero. Ion chamber with superbeta transistor as the main amplifying
element shoes not demonstrate any sensitivity to a radiation source (in the
shape of the mentioned aviation gauge), and it does not show any sensitivity to
other ambient factors (temperature, shocks, air flows) that was described in
any guide on this topic.
The same zero result was demonstrated by a variant of ion chamber having
two superbeta transistors connected in Darlington scheme (see Fig 5).
Fig 5. A variant of ion chamber with homemade
Darlington made of two superbeta transistors.
To order transistor over the internet is a long story. By the time they arrive,
the very relevancy of the circuit, they were needed for, does usually secede.
So, to begin with, a search in nearest electronic shops was performed. The
only Darlingtons, that could be found were MPSA64 having their gain coefficient
of 5000 and having pnp structure. Without any serious hope one of these was
installed into the circuit of ion chamber.
Result: for the first time the indicator had shown something different from
zero. The readings vary from air flows and mechanical shocks, as they should.
However the circuit was still insensitive to radiation of radioactive source.
The circuit did not also react on ultraviolet irradiation of the internals
of measuring jar (by a mercury lamp).
A reader could rebuke me that I waste the time. If it was said that one
needs to use MPSW45A then use only it and nothing else. However such a strict
requirement would affect the ability of the scheme to be reproduced by others.
In addition the reason for use only that type of transistor is not completely
understandable, since the expected level of signal is not clearly known, even
by an order of magnitude. I am not aware how many pairs of ions may appear in
the volume of the chamber's jar due to a gamma quanta from my source, and I
don't know how many of these quanta are going there in reality.
However we can make a bit of handwaving on the info about the expected
signal level. As it has already been mentioned, SBM-20 counter near my source
shows count rate of 200 cpm. The size of the counter: diameter is 10 mm,
length is 100 mm. It means that we have 200 reactions in minute per 7.8 cubic
centimeters of gas volume. If the measurement jar of the ion chamber has
volume of 200 ml, the the amount of reactions there should be 25 times
greater than in the volume of the counter. Lets also take into account that
the counter is filled by a reduced pressure gas. If the pressure is, say,
0.1 bar, the ratio of reaction rates will shift from 25 times to 250 times.
(In principle the linear extrapolation is not fully correct here. It is known,
for example, that the probability for a gamma quanta to undergo a photo effect
is proportional to Z^5, where Z is average atomic number of atoms of the medium.
Probability of pair production is proportional to Z^2, and probability of
Compton effect is ~Z. For air in ion chamber Z=7.2. The counter is filled
with neon, that has Z=10, so even at the same volume and pressure the rate
of count will be different for air and neon. The exact value of the ratio
is hard to determine, since it is unknown what part of the reactions goes
to photo effect and what part goes to Compton. Even more uncertainty is due
to the fact, that the whole volume of the counter does not work effectively.
Only the area close to its central wire does. The size of this area is also
unknown. So the estimations by means of linear extrapolation from gas pressure
and volume can give their result only with the precision to two or three
orders of magnitude.)
Let's multiply the count rate of SBM-20 counter by the estimated proportion
ratio: 250 x 200 - 50000 cpm = 833 reactions per second.
The next question: how many pairs of electrons and ions each reaction gives?
Unknown. If we assume that each reaction gives one pair, we can go from the
reaction rate to the expected value of electric current. One does only need
to multiply it by the charge of electron:
833 cps * 1.9e-19 Cl = 1.58e-16 Amp = 0.16 fA.
We can see that the expected current is extremely low. One can try to unwrap
the problem from the other end. To start from the doses. By the definition,
dose of 1 Roentgen corresponds to the total charge of the appeared ions being
equal to 2.579e-4 Cl per kilogram [Physical Encyclopedia, article "Roentgen".]
If taking the air density to be 1 gram/liter we will get that 1 Roentgen
corresponds to 2.579e-10 per cubic centimeter.
Typical natural radiation background is about 15 microRoentgen per
hour, i.e. 4.16e-9 Roentgen per second. The current, corresponding to the
natural background for a 250 ml jar will be:
4.16e-9 R/sec x 250 ml x 2.579e-10 Cl/(R ml) = 2.68e-16 Amp.
I.e. the natural background corresponds to 0.26 femtoamps.
Next, using the technical data on the SBM-20 counter one can discover that
its count rate, corresponding to the natural background, is 15 cpm. And
my radiation source causes 200 cpm in SBM-20. I.e. in order to get the
expected current value one needs to multiply 0.26 femtoAmps by 200/15.
Finally we get that the expected current is 3.6 femtoAmps.
With taking into account the above notes on the precision of the first
estimation, one may say that the second estimation well correlates with
the first one. Indeed on the background of the expected two or three
orders of magnitude, the difference of twenty times looks well. However
other fact is interesting here. If one takes any of the given estimations
and compares it with noise parameters taken from MPSWA45A transistor datasheet
(several tenth of picoamp) or with leak currents of electrometric FETs 2N4117,
the the expected currents turn to be totally immeasurable. And nevertheless,
as one can see from numerous instructions or video clips over the Net, all
works well. It means that the estimations are lower than the real current
by about three orders of magnitude. Exactly this situation I mean when
saying that I am not aware of the signal level even with the precision to
the order of magnitude.
After many monthes of waiting for completion of internet order, the MPSWA45A
transistors have finally arrived. ("Arrived" is too good word to say. "Toddle",
may be more correct.) One of them was installed into the model of ion chamber.
I think You already foresee the result. Yes, the chamber does react somehow
to approaching of things (including the radioactive aviation gauge). But it
is hard to bind this reaction unambiguously with the action of gamma rays.
The reaction to irradiation by mercury lamp is ambiguous too. In principle
the chamber clearly reacts on the light of a spark gap, but it may be due to
electromagnetic interferences.
 
The sole schematic of ion chamber, that have demonstrated more or less
positive results was the bridge amplifier circuit (taken from "Mad scientist
hut" site).
 
After having been turned on, the circuit settles for a long time (up to
a half of hour). The multimeter shows 2..2.5 mV to the end of this process.
When a radioactive source is put in the close proximity to the (foil screened)
sensitive end, there is no reaction for several minutes. After that the
readings begin to rise slowly, and after ten minutes reach the new stable
level. Of about 4..5 mV. Action is clear and reproducible. But with taking into
account that the readings do float by 1 mV around the stationary value, one
could hardly call it normal measurements.
♦ Homemade Geiger - Muller tube. Reverse Engineering.
It is clear that one can take a thin wall metal tube...
Press there two dielectric leads into the ends...
Stretch a thin metal wire along the axis...
Vacuum seal all this husbandry...
Add a thin capillary for filling and evacuating,
And then fill with something like welding argon with addition of vapours
of medical iodine tincture...
It will be left only to connect this into a classical Geiger-Muller
schematic, and to measure the count rate dependence to the applied voltage
(count rate characteristic). From third...fifth attempt You will get a
counter not worse than commercial tubes (fig 7).
Figure 7. Internal structure and a method of connection of Geiger-Muller tube.
1 - dielectric stopper, 2 - thin metal wire anode, 3 - metal tube with thin
walls - cathode. The internal volume of the tube is usually filled with noble
gas (neon, argon, helium) with addition of some vapours of a low active halogen
(bromine, iodine). The circuit is shown with its output connected to a phone.
In this variant it represents classic click beetle. Instead of speaker, the
output of the device may be connected to some electronic count device.
However to perform this successfully, one should have a fair technical
junkyard in one's garage: one should have argon and vacuum pump, one
should be skilled in vacuum tight sealing. In other words: one should
have a fair technological base. However in reality one can avoid all
these excessives. In guide of Y.Onodera a DIY Geiger-Muller tube filled
with a common air at atmospheric pressure is described. I.e. Yo won't
need neither exotic gases nor vacuum technique.
Test reproducing has discovered the next things: it is amazing, but the
counter does really count. If You are in silent place, the clicks of count
are audible without any speakers of amplifiers. The wide tube of the counter,
with end having been covered with a membrane, is actually operational as
a sound emitter. (Remembering that old phonographs, one feels it right to
add a horn there).
Then it was found out that the counter demonstrates a good count rate
in conditions of natural background. And this count rate does considerably
(by several times) increase when the test source (aviation gauge) is put
close to it.
When reproducing the device there are two difficulties. The first is rather
high needed voltages (4..5 kV). The second is the fact, that the design and
usage circuit, given in the origin <href3> are twisted and differ strongly
from the design and usage circuit of classic Geiger tube (fig 7.). It makes
it difficult to understand the principles and to select what elements of the
design are of principle importance, and what are not.
The difficulties with high voltage are easy to overcome if one uses a
ready high-voltage unit. One can use <a HV unit from a spare stun gun>,
One should use 1.5 V butteries to power the unit. Otherwise the voltage
will be too high.
When counter is fed by a stun gun GV unit there exists one nasty thing.
The unit gives interferences. And You should either deal with proper shielding
or with EMI compensation. There also exist a more elegant solution: one can
charge a capacitor by the HV unit and then turn the unit off. In this case
the capacitor will serve as a noiseless source of high voltage. This variant of
connection is shown on figure 8.
Figure 8. Usage of homemade Geiger tube together with a stun-gin HV unit
and with a radio-set as a click-beetle.
A high voltage diode before the capacitor on the figure above is needed
for the capacitor not to discharge through the internal leakage resistors of
the High voltage unit. Zener diode at the output protects the next circuit
against high voltage spikes. It is less effective than a diode fork, but
it does not require additional supply voltage.
Alternatively one can use a high voltage unit from a home oven lighter.
Choose a lighter type that uses batteries. In lighter without batteries
(piezo ones or 220/110 V ones) there is no high voltage unit. Unit from
lighter is much cheaper than one from a shocker, and its output voltage is
readily close to the needed one. However one should keep in mind that most
lighter high voltage units have no rectifier, so it would be a good idea
to have a high voltage diode and high voltage capacitor in addition.
Attempts to solve the next issue (the twisted design) had lead to the
conclusion, that air filled Geiger counter is not of self quenching type.
And it is despite the abundance of such an electronegative gas as oxygen.
To perform an external quenching the resistor R1 on figure 7 must be
very very high: 1..100 GigaOhms. On the contrary C1 capacity should be
very low (a few picofarads) to keep the time constant of the counter
to be reasonably low. It is disappointing, but usage of commercial high
voltage capacitors (KVI-1, KVI-2) has shown that their intrinsic leak
currents are too high for this circuit. Even after cleaning with ethyl
alcohol the circuit was unable to develop the necessary voltage already
when R1 was 10 GOhm. The problem can certainly be solved with a help of
a homemade capacitor like some Leyden jar, but this solution is ugly and
increases the size of the device. Apparently the twisted design of the
original device (look at href) was born from hybridization between
classic Geiger-Muller tube and Leyden jar of the signal capacitor.
Happily the connection scheme given on fig 7 is not the only possible.
For example there exists variant with picking the signal from a load
resistor (that forms a resistive divider with the ballast resistor),
see fig 9.
Figure 9. A scheme of connection of Geiger-Muller
tube with picking the signal via a resistive divider. Shown with an entrance
stage of some further device (amplifying or counting, dashed line). Note the
protection of this entrance stage against excessive voltage, Here a diode
fork D1+D2 was used, but it is not unique solution.
It appears that when connected as shown on fig 9, with proper selection of
ballast resistor R4, classic design of the counter is very operational.
The classic design consists of a metal (or metal coated) tube - cathode and
central wire - anode. When designing a classic type air filled counter, one
should pay attention for:
- Quality of insulation. The central wire must be insulated very good. Since
the current through the ballast resistor is minuscule, even a small leakage
can suppress the operation of the counter. The insulators should be dry
and degreased. The dielectrics, suitable for making these insulators, should
have minimal intrinsic transconduction. Good ones are: plexiglas, perspex,
ebonite, polyethylene... all those dielectrics, that You could choose for
experiments with static electricity. Materials with significant conductance
(e,g, paper) are not to be used here.
- Absence of sharp ends. Any sharp pint means a corona discharge, and the
corona discharge means inadmissible leaks. The edges of foils should be
rounded or rolled. Tin snotters of the solder places must be blunted with
filer. Cut ends of wires should be either rolled into tubs, where the sharp
ends being hidden inside, or covered with solder until a round shape was
obtained.
- Proper sizes. And in the first order: for wire diameter. When the diameter
of the wire is much less than the diameter of the cathode tube, exactly the
central wire diameter sets the operation voltage of the counter and the size
of its sensitive area. After having changed the diameter of the wire, get
ready to become involves into the search for the proper supply voltage
and ballast resistor value. At the supply voltage of 4 kv, the best diameter
of wire is 0.8 mm. A wire having its diameter of 1.0 mm requires 5 kV of
feeding voltage. Too small wire diameter tends to produce a continuous
corona rather than lower the working voltage. So I can not recommend to
use 0.3 mm wire or thinner.
As a multi gigaohm ballast resistor (R4 on fig 9) one can use a piece of
common paper having its length of 1..2 centimeters. It is robust to clamp a
stripe of paper between two alligator clips. By changing the distance between
the alligator clips it is possible to vary the value of the ballast resistor
(figure 10).
Figure 10. The internal structure and connection scheme
of homemade air filled Geiger counter. Between the alligator clips a stripe
of common usual paper is clamped. It serves as a 1e9 - 1e10 Ohm resistor.
A tap at the end of wire anode is for suppression of corona discharge.An
example of usage of such a circuit was shown on fig 8.
The main and unbeat advantage of the air filled Geiger counter is its high
sensitivity to gamma rays. And if its casing was equipped with a window (open
one, or one covered with thin film) then sensitivity to alpha and beta
particles is added too. The next advantage is rather high level of the output
signal. It simplifies usage and allows to work under rather high electromagnetic
interferences.
And now about the shortcomings.
The main and first one is its too short plateau. (Plateau is a part of
the count characteristics, where the count rate is practically independent
of supply voltage.) The data of Yasuyuki Onodera (the author of the original
design of the counter) shows that length of the plateau is only 80 volts.
And 80 volts are as low as 2% of the supply voltage. The value is comparable
with the precision of not bad voltage regulators. Even worse is that position
of the plateau can be shifted in dependence on atmospheric pressure,
temperature, and even humidity of the air. I.e. when we have a constant
intensity of radiation, the count rate will wander with external conditions
changings. If one wants to make a measurer based on this counter, one should
equip it with sensors of pressure temperature and humidity. And using the
data collected from these sensors, the device should make corrections to
the value of the count rate. And the correction of the supply voltage without
other measures is simply not sufficient. With changes of temperature and
pressure the very amount of substance inside the counter does change (remember
the ideal gas state equation), and it means that the efficiency of particle
count itself does change.
It is interesting that Onodera himself shows the results, that prove the
count rate instability with changes of atmospheric pressure. However he
interprets them inadequately. With lifting from the height if 2020 meters
to the height of 2305 meters his count rate changes by three times: from
1028 cpm to 3270 cpm. And when height was changed from 1596 to 2305 meters
the count rate changes by an order of magnitude: from 455 cpm to 3280 cpm.
Onodera explains it as: "increasement of the efficiency of the counter with
decreasement of the pressure in it." In reality with decreasement of the
pressure, the efficiency can only decrease - the number of gas molecula
in the volume decreases as so a gamma quanta has less chance to interact.
However at lower pressure the counter needs lower supply voltage. If one
keeps the voltage constant, the counter will easily escape from its
plateau into the area of false count. Exactly what had happened to Onodera.
The next drawback is... paper. Yes, exactly the paper, that makes the
whole design to be possible. The paper has a property to become wet and
it changes its resistance drastically. It can totally suppress the
count of the counter. It means that one can not take the device with
him/her outdoors during rainy weather. One can not take it to water
stroll. (Generally the same is true for the described above ion chambers,
which are intolerant even to picoamp leaks).
Here a thoughtful tight sealing could help. But it is at least complicated,
and also it is in conflict with a wish to provide a highest possible
sensitivity to different radiations. As an alternative to paper one could use
grandpa's method of making a gigaohm resistor: one should take a glass tube
filled with dry alcohol and insert two wires there. It can make possible to
avoid the usage of paper in the design of the counter. However it won't
allow to get rid of the humidity problem completely. The next problem place
will be the insulating stoppers (where the wire is being attached to the tube).
The third shortcoming is the limited count rate. "Dead time" of the counter
is determined by the time constant of RC-circuit, where R - is the high Ohm
ballast resistor an C - self capacity of the counter (and its connection wires).
When capacity C was charged through resistor R to the proper voltage, the
counter is ready to count the next particle. It looks like everything is in
our hands: we can reduce ballast resistor, or we can reduce the capacity of
the counter. However it is not true. The capacity of the counter depends
mainly to its geometry, and it is almost impossible to make it lower without
making the counter smaller. The latter, however, affects sensitivity. And
the ballast resistor was choosen so large to give the time for the discharge
in the counter to quench itself normally. If one steps towards ballast resistor
reduction (towards dead time shortening), at the beginning the counter will
give double pulses per one particle, further on it will give triple pulses
per particle, then multiple, and finally it will turn to the continuous
discharge mode. It is clearly audible. With proper resistor the counter
yields quiet but distinct single clicks. As far as resistor value becomes
smaller, the clicks begin to have "an echo" (they become double), then they
turn into warbles (naturally warble is multiple click), and then only a
quiet cracking noise of corona discharge remains. Besides, it is possible,
that those completely epic count rates, that Onodera gives in his tables,
are due to multiple and false pulses. (In the opposite case he may congratulate
himself with that during making and testing the counter he have got a good
strong dose of radiation.)
Practice shows, that without difficulties one can chose the charging
time constant so that the counter would not ring and would be able to trigger
once or twice per second (at some distance from radiation source). It is
three to six times higher than the background count rate. Need to say
that the original design of the counter, just like the simplified one, can
give multiple pulses, and to suppress this mode one needs to feed it through
a (paper) resistor. Only the adjustment of its value is complicated by
leaks from the wire brush and distributed resistance of paper cathode (both
hard to determine and to take into account).
From the above one can conclude that it is impossible to make a serious
device from this counter. But it is still too good to make a verdict that
it is useless. Such a counter is ideal as an indicator. Closer to the object -
the more count. Further from object - the lower count, One can easily see
whether the object is contaminated or not. A different matter is that such
a simple indicator does unnecessary have to be equipped with digital
controls and indication. A common "click-beetle" circuit will suit. And in this
case everything becomes peculiarly simple. Provided that one has a HV unit,
the device and its circuit can be assembled in less than an hour from readily
available and affordable parts.
Video of its operation is below.
<< HOME PAGE |
