Search This Blog

Wednesday, October 26, 2011

I2C Geiger Counter




With nuke plants operating under questionable technical safety, natural events threatening humans and their dangerous energy production artifacts, and other humans disseminating fear through nuclear sovereignty upon the rest of the world, we came a long way from the happy ignorance of distant decades.

As such, in an attempt to further increase the sense of paranoia we readily accept to live in, I took some time to design and build another sensor, in this case a low level radiation detector. Having modularity in mind, I though of creating the device in such way to serve its purpose in more than one context. In order to keep it modular, I have put all the components necessary for it to operate in a stand-alone fashion, and communicate through a standard bus with other modules. Given its adequation to the solution, I have selected the I2C bus. The necessary elements for autonomous operation would have to be:


  • Geiger Tube - bought a russian made SI-39G. Known to detect at least hard gamma radiation;

  • HV power supply - the tube requires about 400 V DC, and very small current supply. In this case I have used a modified camera flash circuit. Originally would output 300 V. I had to tweak it to generate an extra 100 V without exceding the component ratings;

  • Pulse stretching and amplification electronics - the pulse output by GM tube is too short for the human ear to identify (less than a microsecond). At the same time, the pulse signal must be conditioned to be properly handled by digital electronics;

  • Pulse counting and I2C interface - a simple 8-bit microcontroller such as the ATMEL ATTiny85 is enough to handle the job. Additionally it also outputs a short tone (about 4 ms long), every time a disintegration is detected (e.g. gamma photon). This tone can then be amplified and fed to a speaker through external electronics;

  • Linear regulator - Unlike the rest of the circuitry in the module, which is powered at 5 Volts, the HV power supply only requires 3 Volts. As such a linear regulator had to be added (in this case a LM317 adjusted to produce 3 Volts);



The big thing about this modular design is that I can either use the sensor in my other project, the AMIRO robot (link), or attach it to a still-to-be-implemented UI module, made up by a display, speaker, and microcontroller to receive the data, convert event counts to microsieverts (radiation dose unit), and possibly log the data and make it available via USB. Pretty good! The only issue, having time to materialize the ideas. In this case I can at least say almost half the implementation effort is done, 100 % of the hardware design is done, 30 % of the firmware (which accounts for the full ATTiny85 routines), and a few drafts of the mechanical layout are done in my sketchbook. One step at a time, as after all this is a hobby to be kept as a pleasing activity, out of the race track. And of course the longer the ignorance about the radiation, the better the bliss :)

About the testing of this module, it has only been done in respect to the background radiation (that is the radiation that is all around us, coming from several sources such as the trace amounts of radon gas that is in the air, Potassium 40 from our bodies, from the soil, and of course from the outer space). An average of 6 or 7 counts per minute is obtained, which converting to microsieverts (using the tube datasheet) accounts for about 0.100 microsieverts per hour, which is consistent with the expected background radiation for our location and altitude.

Tests with other sources have not yet been carried out, given its scarcity (fortunately :) ). Yet to be tested are common things known to emmit slightly more radiation than background, such as Brazil nuts, granite, bananas, thorium gas mantles (a bit hard to find these days), americium smoke detectors (practically discontinued in the EU), etc. Some of these materials do not release quite many gamma photons, making radioactivity harder to detect. A pancake detector would in this case be far more efficient at detecting a broader range of charged particles.

Sunday, October 9, 2011

Baby monitor



While preparing to be a father for the first time, I have found most baby products to be more expensive than would otherwise be desirable, given its ephemeral usefulness. Considering the full blown economical crisis that we face today, spending copious ammounts of money in things that soon become useless, seems a little too light headed and perhaps irresponsible.

Either essential or non essential products tend to be more expensive than it is reasonable to pay for. As such it is my belief that the consumer must be clever and refrain from indulging the will to buy every fancy baby product that sellers expose.

One thing I found is that quality products are incredibly overpriced if bought locally on a pharmacy or baby shop (e.g. a mere breast pump, sterilizer, pacifier, etc) when compared to large online sellers (but here it not specific to baby products).

Another case are the products which are easily replaceable by general purpose alternatives. One such example is a non-contact thermometer. While being quite convenient and fast to use, baby non-contact thermometers are technically the same (but more limited in the temperature range) than a general purpose non-contact thermometer (these are sometimes even more precise). While first considering buying a baby thermometer, I found that besides being equally effective, a general purpose equivalent would cost me about half the money, and I could use it for a myriad of other applications than for checking the baby temperature. While having the look of a work tool, it didn't really bother me, as it does the job equally well.



But the case I want to focus on is about baby monitors. While technically simple devices, these tend to cost way more money than an intelligent human being is tipically willing to pay. Even CTCSS walkie talkies are usually more powerful than these devices, and sometimes provide the same function, for less money.

In a first approach, I checked the feasibility of building a baby monitor from scratch, but soon found it to be too laborious, even though I would probably not spend much money in parts. I also considered buying the CTCSS walkie talkies, but while being less expensive I considered that I would not give future use to these, so this option was also discarded.

I have taken a look at the stuff I had in lying around at home, and found that I had at least two of the three parts of the puzzle already. One thing is an all band communications scanner I bought a few years ago for listening to all sorts of analog transmissions:



Another thing was a pair of wireless headphones from Sony, which featured quite a reasonable quality, but which I rarely used:



The headphones transmitter had a pretty good range, and I found that I could tune the scanner to either of its two transmitting frequencies: 863.520 MHz or 864.520 MHz. I could listen in perfection selecting wideband FM. Only the stereo signal didn't seem to be encoded the same way as regular broadcasting stations. So I had to listen in mono, but no problem for monitoring the baby anyway.

The only part of the puzzle left was how to connect microphones, as the transmitter only had line level inputs (0.7 Vrms), and the electret microphones signal would be well below this level to be properly heard.

So I had to build a pre-amplifier. I first tested on a breadboard some single transistor approaches, but soon realized that the quality would be below what I would be interested in.

So I looked at the popular LM358 opamp chip, which seemed to potentially provide more than enough gain for this purpose. And besides this, each package contains two amplifiers, easily allowing a stereo pre-amplifier to be built, without having to add an extra opamp for the other channel.

The following schematic describes the device:



And after testing on breadboard the finished circuit, mounted on veroboard, and placed in a nice appropriate enclosure.



First I used random electret microphones I had in my "junk repository", but inevitably found that even the most similar parts, have quite distinguishable characteristics. So for a stereo setup this should be avoided. As such I went to the cheapest place I could eventually get microphones: a chinese store. After a bit of searching in the electronics section, I found PC microphones that seemed to be quite suitable:



Costing 2 euros each, I thought it would be worth it. Took two of these and went home to test against my circuit. One would provide excelent results and the other had lousy sensitivity. This really made me even more aware of what to expect from quality control in cheap chinese products. Curiously the package had the indication "We have passed ISO9001 certification". Whatever...
Anyway, I went back to the store and replaced the bad microphone with another one I selected randomly from the stack (crossing fingers to be more lucky this time). And in fact this time I was lucky (I dont know what the odds are, but perhaps 50/50 ?). Both microphones had apparently equal performance (as far as my hearing can tell), and the final result was very good (true immersive stereo sound when heard from the sony headphones).

All in all, ignoring that the first two parts of the puzzle I already had (which by the way were not all that cheap - 50 euros the sony headphones, and about 100 euros the all band scanner), mounting the pre-amplifier was cheap (around 15 euros and one day of labour). In my case what mathers is that I saved money, as otherwise would be throwing 50 to 70 euros to something that would later become useless. This however can also be used connected to a computer or other device with audio inputs, and provide excelent sound sensitivity (similar to human hearing) for recording a conversation, music performance, etc.



Now that we are talking about electronics engineering projects, I take the opportunity to refer to a company developing great devices, aimed not only to hobbyists but to integrators. Microsensus solutions, given its use of wireless technology fits greatly into applications such as people monitoring. One of the products (the Tiny family) even has the baby monitoring as one of the design motivations: