A portable, sensitive, low power, analog Geiger counter       

You may already have seen my previous page about a basic Geiger counter. As described there, building and playing with Geiger counters has accompanied me throughout my life as an electronician. I first read about Geiger-Müller counter tubes in an old German book that introduced all sorts of electronic devices, when I was 14 years old or so. When the Chernobyl disaster happened, I bought a small old Geiger tube and built my first counter. Now, quite a few decades older, I'm sort of closing this chapter, by building a highly sensitive and fast-reacting, yet extremely simple and power-efficient analog portable Geiger counter.

When it comes to Geiger tubes, size does matter. The larger the tube is, the more counts per minute it will give for a given radiation intensity, simply because of its larger capture area. At normal background radiation levels, a very small Geiger tube might produce only a very few counts per minute, requiring averaging over several minutes to get a meaningful measurement. A large tube will give many more counts per minute, allowing to average it over a much shorter time to get the same quality of measurement.

Instead at high radiation levels, smaller tubes are better, because a large tube would give so many counts that it will lose linearity, since too often a count would be missed because the tube is still recovering from the previous count.

Since I'm interested only in measuring very low radiation levels, like the background level and weak radiation sources that are at most a few times stronger than the background, I needed a large Geiger tube. I bought a Chinese J306β tube, the glass version, which is rated to give 88 counts per minute at average background radiation, or 8 counts per second at a radiation intensity of 1µSv/h. This tube is nearly 20cm long, and about 18mm in diameter.

I took quite some time to decide whether to use a microcontroller to count the pulses and display various processed measurements on an LCD, or choose the classical way and use dumb circuitry and an analog meter. Each method has its own specific advantages: A microcontroller can integrate the counts over a long time, achieving greater accuracy when measuring stable radiation levels, and also measuring the total dose over a long time. Instead an analog meter is far more convenient when scanning an area for radiation sources, because it's so much easier to watch a needle move up, than to read and interpret dancing numbers on a display. It was sort of a tie between the two approaches. The decisive factor, in the end, was that with this large tube I could easily do enough averaging by simply using a basic analog low-pass circuit, a feat that is not really practical when using small tubes. So the rule is: Small tubes used to measure low-level radiation need to be used with digital counters, while large tubes can use analog circuits even at low radiation levels. At high levels, any size of tube can work with analog circuitry. My decision to built an analog Geiger counter was further reaffirmed by the fact that many Chinese digital Geiger counters are widely available. Instead of building my own, it would be more practical and even less expensive to just buy a ready-made one from China. So, to bring any sense into building a Geiger counter at home these days, I went analog.
 
My counter has two ranges. In the low range, full scale is 1µSv/h, while the high range goes up to 10µSv/h. If anything I find pegs that scale, I prefer to run, instead of measuring exactly how much radiation there is! So I don't need a higher scale than 10µSv/h.

The averaging time is 4.4 seconds in the low range, and 0.44s in the high range. This change of averaging time combines with the higher count rate obtained at higher radiation intensity, to produce the same degree of needle stability in both ranges. In the high range, the reaction time is almost real-time, allowing very quick scanning of areas, while in the low range it's still fast enough to allow scanning, even if at a slower pace.
 
I included a speaker, so that I can hear a click for each count. But my practical experience is that hearing the clicks is a bit of an overrated feature. When scanning an area, it's hard to clearly notice by ear a moderate increase of the count rate, given the irregular spacing that the clicks of a Geiger counter always have. One really has to look at the meter. Anyway, the clicking sounds cool...
 
While developing the circuit I was in a power-saving mood. I started basing the circuit on the use of a 9V battery, and after some tweaking brought the current consumption down to 5mA. Commenting about this to an electronician friend, he dryly replied that 0.5mA would be better. He was right, of course, so I began working seriously to reduce power consumption. After a few days the consumption was down to 0.21mA at 9V. But then the question popped up: Why use a 9V battery at all? Those batteries are rather expensive for the amount of energy they contain, and are even becoming harder to find! It would be better to use something more cost-effective, such as AA cells. But... cells? In plural?  Why on earth? The Geiger tube needs 400V anyway, and converting from 1.5V to 400V is no harder than converting from 9V to 400V. And the rest of the circuit works from 5V, and converting from 9V down to 5V, in an efficient switching regulator, is also no different in complexity than converting up, from 1.5 to 5V!  So I modified my circuit to run from a single AA cell, and after just a little basic tweaking the consumption was down to 1.1mA at 1.5V, increasing proportionally as the voltage of the battery goes down during its life span. Both voltage converters stay in regulation until the battery falls below 0.8V. This gives a calculated operation time of well over 2000 hours for a single AA cell, and I think that this is good enough to leave it at that!

As usual on my website, you can click this schematic to get a high-resolution version that is good for reading and printing.

As you can see, the circuit is really simple! But there is quite a bit of out-of-the-box thinking in it, so a thorough description of its operation is in order.

Let's start with the left half of it, which is just the power supply, that creates 5V and 400V from the 1.5V cell. Both voltages are regulated, of course.

I cheated a little by using a tiny ready-made step-up converter to create the 5V. This module sells for a fraction of one dollar on AliExpress, but I provided its internal circuit here, so that if you want to copy my Geiger counter but without any ready-made modules, you can use the bare parts instead. The converter is a simple, bare-bones, classical boost circuit, using one IC that does almost the entire job. Only the storage inductor, filter capacitor, and the diode are external. In some other such chips the diode is internal too, often implemented as active rectifier! If you buy the ready-made module, be sure to get the 5V version, because it comes in 5V and 3.3V flavors.

The 400V converter contains some of my own creativity. Despite its simplicity, it's a quite efficient pulse frequency modulated flyback converter. High voltage sensing is provided by using common switching diodes in reverse, working as 130V Zener diodes (well, purists will claim that any so-called Zener diode above about 7V is actually an avalanche diode, rather than a true Zener diode, but that's another matter). Pressing common diodes into this use is not my own invention. In fact it's quite common in homebrew Geiger counters, because it provides good voltage stabilization at very low current in the feedback path (1µA in my design), and without requiring another voltage reference nor an analog comparator nor operational amplifier. Despite the fact that the 1µA in the feedback circuit, at 400V, causes a power waste of 400µW, which is a considerable part of the total power consumed by this Geiger counter, it's still more power-efficient than some other methods I tried, that give the same voltage stability. On the other hand, there is some room for improvement there, if I ever decide to take power saving to the extreme.
       
The operation of this 400V supply is the following: IC1 is a six-section CMOS Schmitt trigger inverter, one section of which is used as modulated oscillator and MOSFET driver. When the 5V come up after switching on the circuit, C3 pulls pin 13 of IC1 up, keeping it at logic high, and thus keeping the MOSFET off, hopefully until the voltage has reached 5V. Then R2 slowly pulls pin 13 down, eventually crossing the lower trigger point, making pin 12 driving the MOSFET on. A ramp of battery current will flow through the 60-turn primary winding of the autotransformer, while R2 is gradually pulling pin 13 up, discharging C3. When pin 13's voltage  crosses the upper trigger point, the MOSFET is turned off, and the energy accumulated in the airgapped ferrite core discharges through D1 into the high-voltage filter capacitor C2. Then the cycle repeats. We have a basic CMOS inverter oscillator, driving a MOSFET, which drives the flyback autotransformer.

When the voltage on C2 reaches the breakdown voltage of the three backwards-connected series 1N4148 diodes, roughly 390V, they begin pulling up pin 13, through R1. So it begins to take more time to pull pin 13 down, and less time to pull it up. As a result the pauses between MOSFET conduction pulses become longer, and the pulse duration becomes shorter. If the voltage on C2 is enough so that D2-4 and R1 can keep the voltage at pin 13 above the upper threshold while R2 is trying to pull it down, pulsing ceases completely, until the high voltage comes down a little.

The operation mode that results is that after turn-on the circuit pulses continuously, at about 8kHz, until C2 reaches 400V, and then reverts to a pulse rate of just a few tens of Hz, with a pulse duration of around 50µs, keeping the voltage in regulation. After every ionization event of the Geiger tube, the pulse rate goes up briefly, to replace the charge consumed by the tube, then reverts down to the base level required to replace the charge consumed by the 1µA flowing in the feedback circuit.

By the way, the feedback current is set by R2. The threshold values of the Schmitt trigger inverter gate operating from 5V are roughly 2 and 3V. Since most of the time the voltage is very close to the upper threshold, we can take the average voltage at pin 13 as 2.9V. This voltage, applied to 2.7MΩ,  makes a tad over 1µA, which must come through the diodes.  It would be possible to reduce the current in the feedback circuit, and thus the power consumption, by increasing the value of R2, and adjusting C3 accordingly. But running the 1N4148 diodes in breakdown at less than 1µA makes the breakdown voltage far less stable, so I settled for 1µA.

R1 serves as a "softener", or loop gain reducer, to keep the circuit stable. The roughly 1V hysteresis of the Schmitt trigger gets amplified by the ratio between R1 and R2, making several volts of hysteresis at the diodes, and thus at the 400V level. R1 also isolates the CMOS input from the capacitive coupling to the 400V bus, through the diodes.

Although the value of R1 can be tweaked to set the high voltage, it's better to select the 1N4148 diodes. I had a lot of them in my parts stock, some new, some used, some old, even very old. I took a random bunch of them and measured their breakdown voltage at 1µA. The 1N4148 is rated at 100V, so its breakdown voltage must be higher than that, but it's not something the manufacturer will precisely control. I got breakdown voltages between 115 and 178V, but most fell in the range of 125 to 138V, with old diodes generally having higher breakdown voltages than new ones, and diodes from different manufacturers having more difference between them than those from a given manufacturer. I had to test only about a dozen diodes until I found three that added up nicely to the required 390V.

Anyway the exact value of high voltage is not terribly critical. The J306 tube is rated to have a plateau going from 360 to 440V. So, to leave some margins, getting the voltage anywhere into the range of 380 to 420V is good enough. I got it to 402V, which is luxury!
 
To measure this voltage, it's best to use a multimeter with a high value series resistor. I used a 500MΩ one. Although this flyback supply can deliver a lot more current than the voltage multipliers used by many other Geiger counters, it struggles to stay in regulation when loaded by the 10MΩ of a typical digital multimeter, specially when the AA cell is weak. The value of the series resistor doesn't need to be ultra precise. One can simply use any high value resistor, and check with a suitable high voltage source and another multimeter in parallel, to find out the resulting scaling factor. Or if only one multimeter is available, a reasonably stable high voltage source can be measured with and without the series resistor, to calculate the scaling factor.

 
Now let's go to the right side of the schematic, which contains the measuring circuit. The Geiger tube gets fed through a 10MΩ resistor, which needs to be somewhat larger than the typical 1/4 watt ones, which are rated at only 250V and might flash over. The pulses are sensed on the negative side of the tube. When the circuit is at rest, R5 pulls pin 1 of the IC down to ground. As a result pin 2 is high. R8 pulls the right side of C5 down, and so C5 charges to 5V. The inputs of the four parallel-connected gates are low, so their outputs are high, and D5 is blocked.

When the tube ionizes, it pulls pin 1 high. Due to the voltage divider formed by R3, the tube, R4 and R5, pin 1 could go up to something close to 8V, but the input protection diode of the IC will clamp this to about 5.6V. The high level on pin 1 makes pin 2 go down, so that at pin 2 we get a nicely shaped down-going pulse for each ionization event of the tube. When pin 2 goes down, it pushes the right side of C5 below ground potential. The input protection diodes at pins 3, 5, 9 and 11 of the IC will conduct, and so the capacitor discharges to a residual charge of about 0.6V. During this process, the output of the four parallel-connected gates does not change, because their inputs have only gone from zero to -0.6V, which are both logic 0.

But when the tube recovers from ionization, and pin 2 goes up again, it pulls the inputs of the four parallel-connected gates up, through C5. So their outputs will try to go low. D5 will conduct. This diode and the 32Ω speaker are basically the only load on those outputs, because C6 never charges beyond 0.25V or so, and thus we always have something close to 5V on the upper side of the speaker. The gates then operate at their current limit, sinking just a few mA each. The total current creates a reasonably loud pop in the speaker, and charges C6. Meanwhile R6 is recharging C5, gradually pulling down the inputs of the four gates, until they cross their thresholds, and the outputs of the gates switches back to logic 1, ceasing conduction of D5. In this manner, for each ionization event this simple circuit delivers a certain amount of charge to C6, consisting of a current defined by the IC's current limit, for a time that is adjustable by means of R6.

I did at least two things "academically wrong" in this circuit, but I did it knowingly and for good reasons. One thing is using CMOS gates in current-limiting mode, which introduces a large factor of uncertainty, because the current limit of such ICs is not well controlled by the manufacturer. But this isn't a problem, thanks to the provision of a large adjustment range for the pulse duration. Also the current delivered by the IC may be quite unstable with temperature. But since I don't plan to use this Geiger counter outdoors in Antarctica nor inside a furnace, and the self-heating of the chip is totally negligible, and anyway a simple Geiger counter like this one will never be expected to have laboratory-grade accuracy, I simply don't care about any small temperature dependence my circuit might have! Using a separate, stable current-source circuit is theoretically better, but not worth the additional parts.

The other thing I did "wrong" was connecting CMOS gates in parallel, driven from a slow-changing signal. The thresholds of the four gates can be somewhat different, so the gates will not switch all at the same time. Normally that would be bad, because it would cause cross-conduction between gate outputs that are high and low. But when the outputs anyway operate into a very low resistance load, going into their current limit, it really doesn't matter much if they spend a short time shorted between themselves! I tried using four separate diodes instead of the single D5, to eliminate the cross-conduction between the CMOS outputs, but it didn't produce any measurable improvement in current consumption. So I just kept the shorted outputs.
 
The important point here is to not replace the IC by a poorly chosen alternative. I tried the circuit with both a CD40106BE of relatively recent manufacture, and an SCL4584BE made in the late 1970s. Both worked fine, but the old IC has a significantly lower current limit than the new one, requiring a much longer pulse setting. The 1MΩ value chosen for the R6 trimmer accomodates both ICs. And a 74C14 should be in that same range. But don't use a 74HC14, let alone a 74AC14! Those have far higher current limits, to the point of being self-destructive, and are unsuitable for this circuit. Also those have a much higher current consumption when an input is away from ground of Vcc, which is bad news in this circuit, due to the way I use IC1f. So it must be an old-style metal-gate CMOS Schmitt trigger IC, even if the 5V supply I used might make some people think that a chip from a more modern logic family would be fine to use!

Since every ionization event of the Geiger tube puts a well-defined amount of charge into C6, the charge there builds up to a certain voltage, while also draining off through the meter. So the voltage across C6 and the meter will be proportional to the pulse rate, and thus to the radiation intensity. The time constant is given by the capacitance of C6 and the resistance of the meter, 4450Ω in the case of mine, a HC Minipa model MU38, made in Korea many years ago. If a meter of much different internal resistance is used, such as a modern Chinese 91C4, which has a lower resistance, the time constant will also be different, and might be too short. This can be corrected either by adding capacitance, or by adding a resistor in series with the meter. This will affect just the time constant, not the calibration, thanks to the operating principle of this circuit. If assing a series resistor, the total resistance, including the meter's, should not be much over 5kΩ, to keep D6 from beginning to affect linearity. D6 normally doesn't conduct a significant amount of current during normal operation, and exists only for meter protection in case of extremely strong radiation that puts the meter way beyond full scale.      

The range switching works by connecting a resistor in parallel with the meter. If this resistor has a value exactly one ninth of the meter's resistance (including any series resistance, if used), the high range will be exactly 10 times as much as the low range. That's what I did.

A single switch is used to turn the Geiger counter on and off, and to select the range. This is a common 3-position DPDT toggle switch. The center position is OFF.


I went so far as to overcome my almost infinitely growing laziness, and design a printed circuit board for this project! This decision was triggered mainly by the recent implementation of a new process (for me) to make printed circuit boards at home. I needed some deserving project to make a PCB for, to test my system! So, if you are interested in how to make printed circuit boards using a laser engraver, visit that page too!

I built the Geiger counter purely with through-hole parts, except for the little boost converter module, which I bought ready-made. That one uses SMDs, but is mounted on my board as if it were a through-hole part.

You can click the PCB image to get a 1200 DPI version good for making your board, if you choose to copy my design, and if this PCB design is actually of use to you. This is a valid question, because I designed it specifically to fit the little project box into which I built my Geiger counter, and also to fit the specific battery holder, transformer and other parts I used.

The speaker, meter, switch, and the Geiger tube don't mount on the PCB. Also I free-wired the high-range resistor into the box, because that allowed saving one wire going into the PCB. Everything else, including the battery holder, mounts on this PCB.

Here is a view of the fully assembled PCB. The electrolytic capacitor are physically small ones, being very low voltage types. I pilfered them from a deceased portable DVD player. The transformer is mounted by means of a homemade clamp, which I fashioned out of tinned steel sheet coming from a coffee can. It keeps a soft constant pressure on the core halves, to keep them together, and has tabs that go through the PCB.

By the way, this is the same transformer from my previous Geiger counter article, sort of. Same core, same spacers, same high-voltage winding, although I removed the original feedback and primary windings, and wound a new primary using thicker wire. The very low input voltage requires a higher peak current that made the original super thin wire cause too much loss. The core, turns and wire size information is in the schematic. The high-voltage winding is made in three thick layers, separated by double layers of masking tape, to insure enough insulation. The core halfes are spaced with 0.05mm thick adhesive tape applied to all three legs of one of the E-shaped core halves, so that a total 0.1mm gap results in the magnetic path.

Note that there are two 10MΩ resistors in this circuit, which are different. R1 is a quarter-watt type, while R3 is a half-watter, necessary for its higher voltage rating, not for the power rating. If you don't have half-watt resistors at hand, a good alternative is using two 4.7MΩ quarter-watt resistors in series.

It seems a little crazy to use a 1A diode in the high voltage circuit, which only has to deliver microamperes. The reason is that high-voltage, low current, ultrafast rectifiers aren't common, while the UF4007 is widely available and inexpensive. Don't confuse it with a 1N4007! While the latter would most likely work, it's a standard-speed diode, so it causes higher losses, increasing the current drain of the circuit. 
 
C2 is a ceramic disc capacitor with wide pin spacing. Its value is quite uncritical, so anything grossly in the same order of magnitude should work fine. C3 is a little C0G ceramic, hidden in the shadow in this photo, while C5 is a polyester film type, of the typical green epoxy-dipped type. These two should have stable values, so don't use any class 2 ceramic capacitors for them.

I used a BS170 MOSFET, because I have a bag of them, donated by a friend. Nowadays there are many such small MOSFETs, having much better characteristics, although most of them are SMDs. Any of them should work, as long as the gate input capacitance is low, it switches fully on with less than 5V, and has a reasonable RdsON. A big power MOSFET is not a good option, because it will have way too much gate capacitance to drive it fast enough from a CMOS gate.
 
The battery holder I used is a high quality one, having no rivets. The ones using riveted springs often develop contact trouble in those aluminium rivets, so avoid them. If you happen to buy the battery holder from AliExpress, like I did, pay attention that you don't fall into the claws of a bad seller who offers these holders there. That seller offers packs of five pieces, charges the price of five pieces, and then sends just a single piece. And what's worse, AliExpress rejected my claim for compensation!
    
There isn't a lot to say about the bottom side of the PCB. Please be kind and don't laugh about my soldering. It used to be better when I was younger, and could still see at short distances.

The battery holder is held in place by two screws, so it doesn't dangle just from little wires.

This is a very coarse board, with no traces narrower than 1mm, so it's very easy to make, by any method.


The project box is one I bought at least 20 years ago, so I don't think that it's still available. But many other boxes are on the market. It looks pretty nice, although its black color makes it hard to get decent photos of its inside.

The meter is mounted in the good old traditional way, with four bolts. The speaker instead is mounted in the modern way - just thrown in, and fixed in place with hot-melt glue!

This speaker is a 32Ω one. I chose that relatively high impedance, because my circuit delivers constant-current pulses, so that a higher load impedance will take more power from it, and sound louder. That's one trick to drive a speaker directly from CMOS gates, and make it sound pretty loud!

This is a 50mm, 0.5W model, bought on AliExpress. I bought this one, and a smaller one, 36mm diameter, intending to use the smaller one if it sounds loud enough. It didn't, so I used the bigger one. The size difference, even if not huge, makes a big difference in efficiency! In speakers, just as in Geiger tubes and a few other things, size does matter!


I chose to install my Geiger tube in a protrusion in front of the box. That way it's free to receive radiation from all sides. I crafted this structure from a piece of 32mm PVC water pipe, and two polyethylene parts machined on my lathe. One is a simple end cover, while the other is an interface piece that holds the pipe to the box.

The positive end of the Geiger tube goes towards the box. This keeps the positive lead short, which is a requirement for long lifespan of a Geiger tube, unless the 10MΩ resistor is placed on the tube side of the wire. Also by this orientation the wire that has to run along the tube is the negative one, which is good because the external electrode of the tube is the negative one, so that there is no charge between the blue wire and the tube. In circuits operating with hundreds of volts, but just a few tens of microamperes of pulsed currents, even a few pF of stray capacitance matter.

The Geiger tube is supported inside the PVC pipe by means of two plastic foam cushions, delicately carved from a piece of foam by means of a knife, scissors, and needle-nose pliers used to rip little bits out.

The foam for these cushions came from packing material. It was an insert in the box of some appliance I bought.

Only the three-dimensionally curved ends of the tube are supported, given that they are sturdy. The more fragile cylindrical part doesn't touch anything. The inner diameter of my PVC pipe is more than 28mm, giving a good amount of clearance between the Geiger tube and the PVC pipe.

Although PVC isn't the most beta-transparent of plastics, it's the one I can get most easily in pipe form. A polyethylene pipe would have been better.

Any radio noise picked up by the long negative wire is filtered out by the low-pass filter created by R4 and the CMOS gate's input capacitance.

I crafted contacts for the tube's terminals from the same coffee-can tinned steel sheet that I used to make the transformer holding clamp. Talk about a universally usable material, these coffee cans! If you drink tea rather than coffee, no problem, since good tea also often comes in tinned steel cans.

If necessary, some cookie boxes are also suitable. Don't despair. You can surely find such fine multi-use material, if you need any!

The wires connect to the PCB on the bottom side. I immobilized them with hot-melt glue, because this is a portable instrument, after all, and might get banged around a bit. Wires tend to break at the solder points, if unsupported.

The Geiger counter is ready to be closed up! Here you can see the reason for the strange shape of my PCB. It conforms to the rails this box has, intended to slide in boards in vertical position. They didn't suit my needs, so I mounted my board horizontally. I hot-glued small scraps of plastic in four places along the box walls, as stand-offs, to make the board sit at a well defined height above the speaker. Once everything was ready and tested, I fixed the board in place with four drops of the glue. If the board ever needs to be taken out, it's easy to rip out those blobs of hot-melt glue first.

Note that the tube's positive (darker orange) wire is routed well away from the other wires, and hot-glued in place, to further reduce its capacitance.

You will see two resistors in parallel, at the range switch. Since my meter happened to have an internal resistance of about 4450Ω, I needed a resistor of slightly less than 500Ω for the high range. 470Ω was too low, 510Ω was too high, and I had nothing in between. So I used two 1kΩ 1% resistors in parallel. Typically resistors have values closer to the lower end of their tolerance band than to the higher end, and these were no exception: The combination gave 496Ω, which is almost perfectly right.

Another option is to use a 1kΩ trimmer potentiometer, and set it to the exact right value.

I even installed a brand new, fresh alkaline battery in my meter, before closing the lid! All the circuit development was done with a lab power supply, and for later testing I used three different, very old and worn AA cells. Now, with the new cell installed, and my rather infrequent need to use a Geiger counter, this cell will probably outlive me. A sobering thought...

The finishing touches include making a proper meter scale. The meter came with a 50µA scale, of course, and I needed a µSv/h scale. Also it's nice to put one's own brand name on the meter scale, right?

I used the "Meter" program by James L. Tonne, to make this scale. It's an old program, but still works, and is very flexible, even if with modern operating systems and printers it can take some effort to get the dimensions calibrated.

Speaking about calibration: It's a problem for me to make an absolute calibration of my Geiger counter, because I don't have any radiation sources with precisely known activity. For the time being, I trust the J306β tube's specs published on the web, which claim that it will produce 8 counts per second at 1µSv/h, when using a Cobalt 60 source. So I disconnected the tube, and injected an 8Hz square wave at 5V CMOS levels into pin 1 of the IC, then set R6 for precisely full scale in the low range. So far, so good, but I don't know how accurately my particular specimen of the J306β obeys the rules set by its manufacturer, and far less do I know how many counts per second it would give for the same 1µSv/h level, when it comes from a different substance, that emits radiation of different energies than Cobalt 60.

So my meter is only good to be used for relative measurements, detecting "hot" rocks, etc. At least for now.

 
My first little survey of my property brought up nothing unexpected. I'm not living on top of an uranium deposit, and that's good to know... Or at least, if there is one, it's not very close to the surface! The radiation level measured outdoors aligns well with the published averages for areas not known for high radioactivity.

I didn't find any "hot" rocks on my property, not even along my driveway, which was stabilized with gravel brought from a nearby river, that contains a wide range of different rock types. 

Given my circuit's time constant of 4.4 seconds, and the fact that radiation tends to come in irregular bursts, the needle moves quite a bit. It's necessary to watch it for about 10 to 20 seconds, to make a meaningful visual average.

Things changed when I went to my garage. The culprit was not the car, but the fertilizer! I store some fertilizer there, of different kinds, and the bag of potassium nitrate caused a radiation reading well above the normal background. There must be roughly 10kg of potassium nitrate left in that bag. That should contain about 3.9kg of potassium, and about 450mg of that is radioactive potassium 40. This stuff is several times more radioactive than natural uranium, and generates 1.31MeV beta particles and 1.46MeV gamma rays, according to web sources. The 1.6mm thick walls of my PVC pipe block weak beta particles, but the relative strong ones from potassium 40 should go right through it, and probably all the way into my Geiger tube, to be detected. And the gamma rays, of course, won't even feel this little bit of a plastic obstacle...

At first I planned to drill a grid of holes into the PVC tube, to get higher beta sensitivity. Maybe also add a removable shield, maybe made from aluminium, to block all beta particles, when so desired. But then I thought that this would be overkill for my use of this meter, and didn't do it. I can still do it later, of course.

So, even if there's less than half a gram of actual radioactive material in my fertilizer, distributed among 10kg of inert stuff, it does cause much more intense radiation then anything else I have on my property. Good to know! 

A good friend mentioned that I should paint the blue PVC pipe yellow, because yellow/black are the colors usually associated with radioactivity. And I had to admit to him that I had almost done this! But then I decided otherwise, and as always, for a good reason: Stuff that emits radiation is marked with that yellow color, but a Geiger counter doesn't emit any radiation. Instead it receives radiation, which is the exact opposite action! And blue is the color opposite to yellow. So it's just the right color!

Back to homo ludens electronicus.