Some readers will recall the saga of Max’s non-counting Geiger counter (Click Here
to read that sad tale), to which I ministered some time ago (Click Here
to see that article). That was built from a kit and just didn’t work properly. I made a few design changes to improve it and since then it has been clicking away on his desk (assuming he hasn’t tired of it yet and let the battery run down).
Well, Max recently sent me his latest (non-working) project, an Atmospherics Monitor
. This is basically an audio amplifier connected to a ferrite rod coil antenna. So instead of picking up and demodulating radio frequency carriers, it picks up audio-frequency electromagnetic waves in the atmosphere and amplifies them for direct listening. It is based on an old Everyday Practical Electronics
design from the April 2003 issue (Click Here
to see the original article). The waves are produced by storm and auroral activity, sometimes at great distances, reflected around the earth by the ionosphere. The article goes into considerable detail as to the mechanisms of the sounds that can be heard on the monitor.
The nice thing about the stuff Max sends me is that I cannot, as with normal fault-finding, assume that they ever worked. So you really have to go back to basics and start from scratch. When I got the unit from Max, I could see it was drawing some – not excessive – current from the supply, but was not producing any sound at all.
In this case the design is a fairly simple audio amplifier, with a grounded-base transistor as the first stage (giving a very low impedance input for the ferrite antenna). The remaining amplification is by op-amps (LM 358s). The unit is operated from a 9V battery, so most of the stages have a divide-by-2 resistor chain to supply V/2 bias. So my first test was to measure the output pins of each op-amp. All were around 4.5 volts except the first one, which was at around 0.9 V. Now this stage gets its bias from the transistor input stage, so it was there I turned my attention.
Figure 1. Schematic of the Atmospherics monitor
to see a larger, more detailed image)
The transistor input stage would have a gain of roughly 10 (R3 / R2). It has a bias pot VR1 which will set the operating point of the transistor TR1. (This is necessary because variations in transistor gain will result in different collector voltages.) The object is to get a voltage of around the rail voltage divided by 2 at pin 5 (and hence at the output pin 7) of the op-amp IC1a. I tried adjusting VR1 but it had hardly any travel – perhaps 30 degrees instead of the usual 270 degrees or so. I couldn’t see if it was catching on anything so I removed it. Once off the circuit board it did seem freer, but I replaced it with one from my stock. Bingo, I could now get IC1a pin 7 to 4.5 Volts. AND I could hear a faint noise on the output piezo sounder. But it was very faint, even with the volume control turned up full. I had a square wave generator which I fed to a small coil I had lying around, and when I put 2KHz into the coil I could hear it faintly on the output of the monitor. So I was getting somewhere.
Max had a small piezo sounder on the output and I tried a larger one I had, about 2 inches diameter. This changed the tone of the output a bit, but it was still very faint.
I then turned my attention to the other stages of the monitor. My 2 KHz generator was putting out a 5V square wave (it was part of my breadboard test gear intended for a TTL test signal). So I put it through a 100K resistor and poked it around the circuit as a crude signal generator. It worked a treat, and I got loud tones when I touched it to the inputs of IC2b and IC2a (pins 5 and 3 respectively). Touching it on the output of IC1B still gave a loud tone but the input of IC1b (pin 3) gave only a faint output. So I looked at that part of the circuit a little more closely, bearing in mind that as this never worked, almost anything could be wrong, including the layout or connections.
Figure 2. The component layout of the monitor.
to see a larger, more detailed image)
It’s made up on stripboard (often called Veroboard, which is what the company that invented it called it). It’s a great way of making small projects like this where you’re not using high frequencies or powers. You get all the advantages of a PC board without having to make one. However it is difficult to achieve a high component density (this is pretty good) and it can be fiendishly difficult to trace. (For the uninitiated, the light strips are copper tracks with holes in them, and the dark circles across the tracks in places are cuts made to disconnect one part of a track from another. A 4mm or 3/16 drill is a great tool to do this.)
I initially suspected that C6 and C7 might have been swapped but closer inspection revealed that they were right. (This would have put a 10nF to ground and passed the signal through a 270pF – not great for AF frequencies…). Then I noticed that C4 was not placed correctly. C4 is a 1uF capacitor effectively across VR1 – it is the “Ground” of the grounded base transistor input stage, at least at AC. C4 was connected at K-8 and I-8 on the board, not K-8 and J-8 as it should have been. Max can be forgiven for this I think, the + symbol above C4 is almost on the I-8 hole. Here’s an enlargement of this part of the board:
Figure 3. A close-up of part of the board.
I also noticed that the cut on the track at I-6 (shown in red here) did not seem to be present. However poking around with a meter revealed that there was no direct path between the VR1 pin at I-5, and the R8 pin at I-7, so I suspect that track was cut with a knife.
What would be the effect of this? Well C4, instead of being from the negative line to the top of VR1, would now be from the negative to the junction of R8 / R9 / IC1b pin 3 input. At 1 KHz, 1uF has an impedance of around 160 ohms. Less at higher frequencies. So it was doing a good job of shorting to earth any signal that got through the input stage.
For those of you who may think “How could Max make a mistake like that?”
below is a photo of the board. It has my new VR1, but C4 is still in its original position. Can you see whether the + lead of C4 is in I-6 or J-6?
Figure 4. Max’s board.
The big bulbous black thing at bottom right is my meter earth clip. One thing I can criticise Max for is the fact that the 220 kilo ohm resistors above and to the right of C4 are fairly large. But then, who wants to go and buy stuff when you have it already in your junk box?
Anyway, a bit of work with my cutter and soldering iron sorted this out. And then the unit was buzzing away nicely. My workshop is right behind the wall with my main air-conditioning unit on it, and you can take it from me that it produces a LOT of EMI. At one stage a motor started and I got a nice rising whine from the monitor. So I was now pretty sure that it was working properly. Max had the volume pot wired so that turning clockwise reduced the volume, so I corrected that (I have done that I don’t know how many times, so I’m not going to say anything!)
The article says that using the unit inside a house will usually get a lot of interference, so I took it outside. By aligning the antenna roughly East-West (as recommended in the article) I could null out most of the interference from my house (which is roughly North-South). But there were still none of the intriguing signals referred to in the article.
As a final test, I measured the inductance of Max’s ferrite rod coil. The article says it should be 24 mH. I got 20.4 mH, but inductance meters are notoriously inaccurate and mine does not have different measurement frequencies like some fancy ones I have seen. But it was in the right ballpark so I did not go any further there.
The next afternoon I took my dogs for their customary walk, in a graveyard nearby, which has no mains around, not even streetlights, and is well away from any houses. There was a lot less interference and background buzz, but not much of anything else. I did this for two evenings running and I did get some random clicks on the second evening (and the article does make reference to these). Our local weather had been exceptionally fine so I wouldn’t have picked up much local activity, but I had hoped for something interesting from further away. I also got some variable rhythmic clicks and bursts of noise sometimes, but I am not sure if this was interference from a radio station – I tried the local ones on the car radio but couldn’t get any correlation to any of them. And whenever I switched the car ignition on to make the radio work, I got a lot of whines and noise from the car electronics, which did not make things easier.
The fact that I was picking up all sorts of man-made noise with the unit satisfied me that it was working, but I wish I could have kept it a bit longer. On the day I sent it back to Max, I had it all packed in the box, addressed and sealed up with tape, and we had a great thunderstorm in the afternoon just before I left. The temptation to open it and see what I would have got was great…..but the post office was closing and Max was waiting with bated breath, so I resisted.
I’ll be keen to see what Max picks up with it back in the Northern Hemisphere – there being a fair bit more land there might make a difference. He told me he has had some good storms around recently and he lives near “Tornado Alley” – and who knows what signals they might generate…
If you found this article to be of interest, visit Programmable Logic Designline
where – in addition to my Max's Cool Beans
blogs – you will find the latest and greatest design, technology, product, and news articles with regard to programmable logic devices of every flavor and size (FPGAs, CPLDs, CSSPs, PSoCs...).
Also, you can obtain a highlights update delivered directly to your inbox by signing up for my weekly newsletter – just Click Here
to request this newsletter using the Manage Newsletters tab (if you aren't already a member you'll be asked to register, but it's free and painless so don't let that stop you [grin]).