Magnetometer (original page)

The Magnetometer Project


Magnetometer is now up and running. Although it is in testing stage as we try to iron out some bugs. Here is the real time data (don’t use it, it’s not accurate). We have moved it (lost count) times now! Today (7th May 2014) we’ve had a break through and I think it is performing well. Scroll to the bottom of the page for full updates (and where we went wrong on the previous versions).

Latest data:

Aurora Service Magnetometer

There is some slight interference in the day time now because the house is being renovated (new roof) for the next 6 weeks and there are builders and or tools near the magnetometer.

1AM – 28 January
We plan to install a magnetometer at magnetic latitude 57°N (Geographic latitude 60°N) so we can monitor geomagnetic activity in real-time (every second) at a location that is very useful for a lot of people in the south of Scandinavia (and to a lesser extent Northern UK). Right now we use a magnetometer in Dombås in Norway at 62°N geomagnetic latitude. It has been quite useful for us in our aurora chasing, but we want to take it to next level and get real-time data from a site roughly in between the cities of Helsinki and Stockholm. This will give us, and more importantly, around 3 million people much more accurate idea of what’s happening directly above our heads. It is of course useful to people outside this area also as the auroral oval rotates west across Europe, anyone along this magnetic latitude will find this data useful before the peak of the oval reaches them in about 1 to 2 hours.

The magnetometer will be plugged into a live feed on the site and you will be able to see the data (with a k index) in a chart on the aurora forecast page.

What is a magnetometer?

Aurora Borealis is a visible manifestation of geomagnetism. If we can measure geomagnetic activity, we can to an extent, measure aurora activity (or at least know when it is quite likely). We monitor geomagnetic activity using a device called a magnetometer. Commercial magnetometers like the one below costs thousands of euros. But we plan to build our own…

magnetometer example

An example of a magnetometer (Not exactly like the one we are building)

We will site the magnetometer in a location far away from any sort of interference, along the south west coast of Finland. Here we can monitor the geomagnetic activity and we can see specifically, in real-time, when geomagnetic storms are occurring and are likely to produce auroras.

The data will be displayed in a chart similar to this:


Why is the South West Corner of Finland significant?

The auroral oval stays in the same place, but Earth rotates. Directly opposite the day side of the planet, the auroral oval is elongated. But for simplicities sake, we will refer to the oval as passing over countries than discussing planetary rotation.

When the auroral oval hits Europe, Finland is the first country it passes over. What this means is by using a magnetometer in Southern Finland, you have the first idea of geomagnetic activity and how it is looking before the peak of the oval reaches you in around 1 to 2 hours if you are a bit further West. Finland is only 2 hours ahead of Universal Time, so if you are in Sweden/Norway you will have around 1 hour to see how the geomagnetic activity is looking before the peak of the oval reaches you. If you are in UK you will have around 2 hours to see what is going on before the peak of the oval reaches you.


It is especially useful for Helsinki/Tampere/Turku in Finland and Stockholm and surrounding areas in Sweden. But anywhere within a couple degrees of magnetic latitude will find this data useful, such as Tallin, Oslo/Stavanger/Bergen and Northern Scotland (although Scotland has the excellent Mull magnetometer, which our system will be very similar to, but still useful to see what’s coming your way?)..

Get involved…

The magnetometer comes in kit form from Germany, compared to commercial magnetometers, it price is amazingly low, but the components still cost money. So the cost (in kit form) (inc DHL shipping) is €191 delivered to our door. So we are asking anyone that might find this useful, or even if you just want to donate towards a great science project, to get involved.


What do we need from you?


Yes, just 1 euro.

If 191 people donate 1 euro each, we will meet our goal of 191 euro’s and can order the magnetometer from Germany. Of course if you want to donate more than €1 feel free.

Amount raised so far: €222!!

WE MADE IT FOLKS! In less than 12 hours of asking for help on Facebook, the target was met. It is absolutely astonishing and overwhelming that we received that level of support, that fast. Thank you so much. If you still want to donate towards the project you can, because we still have to build it and bury it for which we need materials for and to buy cabling to and from it! But the costs of actually buying the magnetometer itself has been met, we can now confirm the order with the magnetometer suppliers in Germany! AWESOME! Guys. Everyone who donated/donates to this project pat yourself on the back. Just awesome. Scroll down for updates on the project we will continue to document it.

Where will the Magnetometer be placed?

It will be placed on an island just south of the famous Finnish observatory, Tuorla Observatory. Magnetometers are highly sensitive instruments, placing it near to a main road would produce false readings. So it will be sited off the beaten track on a very quiet island just south of Piikiö, Finland. The magnetometer should be able to be calibrated (in regards to K index) with help from Finnish Meteorological Institute’s Nurmijärvi magnetometer which is on a similar latitude further east, so the K index should be highly accurate.



So please help us get this project up and running. You can make a difference with just 1 euro!

Thanks to the following people who have donated so far (in no particular order..):

Amanda Mitchell, UK
Paul Telford, UK
Vikki Royal, UK
Garrie Powers, UK
Jon Austerheim, Norway
Lissa Wright, UK
Elizabeth MacDonald, USA
patricia courtney, UK
Anssi Toivanen, Finland
Christian Meier, Switzerland
Sam Hampson, UK
Ronny Årbek, Norway
Christina Jack, UK
Monika Mroczek, Austria
Aziz Errafay, Holland
Karri Pasanen, Finland
Carter Langley, UK
Jerome CANTALUPO, France
Rune Johan Engebø, Norway
Céline Simon, France
Martin Strauß, Austria
Caroline Bailey, UK
Teemu Saramäki, Finland
Carol White, UK
Cecile Planque, UK
Tami Owings, USA/Finland
Kari Kuninkaanniemi, Finland
Arun Narayanan, Finland
Paul Williams, UK
Peter Hulbert, UK
Martyn Clark, UK
Stuart Spencer, UK
Thomas Kast, Finland
Matthew O’Neill, Norway
Jukka Jokila, Finland

Thanks so much folks! We will keep updating this page as the project develops and we will show pictures all along the way right up to (and including) the point we install it. We will print the list of people above and bury it with the magnetometer.

1pm – 28 January
Target has been met, order has been placed with Magnetometer suppliers. It will take a few days to send payment over to them, then a few weeks before we receive the magnetometer (it will take a week just to deliver it). But the wheels are in motion so to speak now. This time yesterday, we were debating whether to make this page thinking no-one would be too interested. Wow.

1pm – 30 January
Today we sent the money over to Germany for the Magnetometer kit. It cost a bit more than they first quoted because they forgot to add the price of extra sensors (we will be using three sensors, one each for bx, by and bz components). But thankfully, we had a few extra donations so we were only 1 euro short.

1. electronic components, PCB , 1x FGM3-sensor,LCD.. 125.00 Euro
2. two additional FGm3-Sensors 2×32 Euro 64.00 Euro
3. enclosure, not machined 17.00 Euro
4.shipping costs 17.00 Euro

Total 223.00 Euro

Here are some pictures of the actual magnetometer. This is from the Reeve observatory in Alaska who run the same magnetometer. This is their installation. So we will do something pretty similar.

magnetometer1Bare circuit board in kit form.

magnetometer2Built circuit with 100 kit parts fixed & soldered

sensorsFGM-3 Sensors

magnetometer3Magnetometer control unit.

magnetometer4How they buried it at Alaska Observatory (we’ll probably do something similar..)

magnetometer5The reason it is buried is to keep a very steady temperature.


So now we wait for it to to be sent over from Germany. In the mean time we will stock up on coffee inpreparation for the building process…

11 February
We got an email from the dudes in Germany saying they have despatched it. We should have it within a week.

21 February
Magnetometer parts arrived…need to stop referring to it as a magnetometer because it is just a load of resistors, capcitors, chips and circuit boards for the time being…




24 February
We have made a start on the soldering side, which is only about 25% of the project. It’s actually a lot more technical (and time consuming) than we anticipated. I am sure if we built many of them we could get the time down to a couple days building each one. Right now it looks like it will take a couple of weeks to get it up and running. Building some sort of frame to hold the sensors at perfect X, Y and Z angles to the magnetic field while keeping it safe from being moved (and thus knocking the readings out) seems the most challenging part. As well as arranging cabling to the sensors as they say using one cable can cause interference so need to have 3 cables running there. Also the sensors have to be a certain distance from each other to avoid interference. So really the physical side of building the thing is the most challenging part. This soldering/building part of the control unit seems actually a straight forward part. Although incredibly time consuming because we are new to PCB soldering having only done automotive soldering before. Also we are moving house this weekend, to the same island where the magnetometer will be placed, 90km away from our current location, so that is a big task. But once we are in, I can get some serious magnetometer time going (also nice being 20 min drive to DIY store to buy parts to build it, now we are 3 to 4 hour round trip to DIY store!). Anyway, that is where we are up to!

9 March
Made some really good progress on the magnetometer this week & weekend. The financial cost of this kit is tiny in relation to labour cost…This has taken up most of my evenings and now a full weekend dedicated to it to get this far. Mainly down to my slow soldering, but it is a very slow process. Anyway, it has been a learning experience, and I have enjoyed it (most of the time) up to now and I’ve now got pretty much all soldering work complete. There is still a bit of cutting and drilling of the plastic case to do so buttons and stuff can be acessed. Then I can (attempt to) power it up and see if all that soldering and time has been worthwhile. It is still a while away from having it set up, I’d guestimate a week or two. So far all the work has just been for the control unit. That is all the kit is, a control unit kit and 3 sensors. You have to build the control unit from scratch with the parts supplied, but most other things you have to build yourself such as sensor housings/enclosure and cabling to and from the sensors and the control unit, which could be up to 50 metres distance between each other. I have some good ideas on the sensor side of things. I got some electrical bits on the weekend that the sensors can sit in and I can adjust their angles pretty accurately. Many pictures will show that whole side of things, for now I have just been trying to get the control unit up and running and it has all been soldering, nothing really to look at. On that note, I really don’t think this magnetometer kit is suitable for novices, I have done automotive soldering before but this was a whole new beast. I have wanted to give up on a few occasions, I really hit a brick wall and just thought I couldn’t do it, but I didn’t want to let anyone down so I kept at it. But it is A LOT of soldering and unless you have done PCB soldering before, I really think you will struggle (as I have). The last few days of soldering have got much easier as my technique improved, so I wouldn’t hesitate to get another kit now I have had so much soldering practice! Reeve observatory in USA sell the same magnetometer kits pre-assembled, I am not sure how much extra it costs, but I would really recommend that if you have never soldered before. But here is where we are up to now. Today I will order hundreds of metres of cat5 cable, patch cables and RJ45 connectors to connect the control unit up to the sensors per haps next weekend! IMG_6847

9 March
It has been an hour or so since my last update, but I must have put about 50 hours into this project this week and I wanted to see something from all that effort. So despite being moaned at I have house hold chores to do, baby to change etc.. I continued on and I just finished soldering the power connectors (AC adaptor was not included but I had one spare I cut the end end off) so here goes, first time powering up the control unit..

…fingers crossed…



Now I really do have things to do, but hopefully tonight I may attempt to connect this up to the computer and see if they communicate with each other. Once that is done, 100% effort can go into the sensor side of things and finding a suitable place for it somewhere near the house.

10 March
Ok I had a really good momentum going so I kept working on it into the small hours and a few hours this morning. I got the control unit up and running and communicating with the PC via serial connection. So I am happy to announce we have a functional magnetometer CONTROL UNIT, but we still have to do the sensor installation. As can see on the control unit display it is ready for sensor hookup and is at present showing blanks values for X Y Z sensors as there aren’t any plugged in. Most people build a sensor enclosure (like the reeve one above) and bury it, we are first going to attempt to build a magnetometer enclosure in the attic of our old farm house as it’s almost entirely timber built and it should be reasonably regulated temperature wise up there. Instead of hard wiring everything I want it so the magnetometer sensors enclosure is easily movebale, so I am wiring everything with cat 5 cables, RJ 45 couplers/connectors etc. A bit more work, but it will make life much easier should we need to move it, or bury it if we have to. I am just about to place the order for all the cabling, I need around 100 metres, + all connectors and a crimping tool, so it looks like the bill for the cabling alone is over 100 euros bringing the cost so far up to about 350 euros, and I have still to build and insulate some sort of enclosure for the sensors to fit in. But, it’s looking good so far!



10 March
Ok, I got stuck into the magnetometer project first thing this evening. I have ordered a ton of cabling and associated stuff which will take a week to get here (from UK), but in the mean time I had a spare cat5 cable hanging about I was able to chop up and connect one sensor to the control unit. It works, but obviously the sensor is just on my desk, the readings are all over the place, but but I will keep it running for a few days to see if there are any patterns that emerge that will help me when I come to place it, whether it be local interference, a road, tractor, whatever. But here you can now see how it will look (albeit with 3 coloured lines instead of just one). I have also just emailed a local observatory asking for offset and K data so I can can calibrate this one so it is uber close to the commercial magnetometers (and thus reliable). A magnetometer is only as good as it’s calibration.



The magnetometer is just using default setup data at the moment, so this graph will be all over the place, as I test the sensitivity of the sensor, such as moving ferrous (magnetic) things near it, check the temperature affects, etc..obviously you shouldn’t be using it yet. Although I really wish it was 3000nt :)

But woohoo, we have an almost working magnetometer!

Right now we are waiting on all the cabling to arrive from UK. It is expected to arrive around 19th March. So for now, just to test the system, we have only one sensor (Y) connected using 1 foot of cat5 cable. It is just sat on my desk next to my monitor so the readings are pretty crazy. Worth noting is the magnetic field of a human body, it is actually stronger than Earth’s magnetic field, and as you can see if you have been following the graph is that each morning when the graph rockets upwards a few hundred nt, it is nothing more than a case of me going to sit at the desk. The human heart gives off a lot of magnetisim. It is a fascinating subject infact, biomagnetism and many other fields relating to magnetic fields and living organisms. It is absolutely unrelated to anything we are doing, but if you are reading this (and you find it reasonably interesting) you will like these too:


Electromagnetic Fields and the Heart

Humans have a magnetic sensor in our eyes, but can we detect magnetic fields?

Why did the chicken cross the road? Because it was following the magnetic field.

It is just sensing me because I am so close to it. The other fluctuations prior to that is due to temperature changes because I have a heat pump and also a huge fire in my office which is affecting it at the moment. There doesn’t seem to be too much interference from other sources, it is just temperature and myself that seem to be affecting it right now. So once the cabling arrives we have a few different options of where to place it within 60 metres of the main building. The cat 5 cabling we have ordered is indoor spec, so hopefully the magnetometer can placed somewhere sheltered. Attic is our first choice, then perhaps the barn, but there are high voltage cables coming to the barn which might interefere with it we are likely going to try a few different places. We’ll try it in the attic and see

11 March
So we had a first night with a test sensor attached. It is not aligned perfectly to the Y axis, just in the general direction, but already it is providing interesting feedback. It doesn’t seem to be affected by much interference in the local area, which is great, I was worried about the powerline coming to house which is about 35 metres away, this doesn’t seem to affect it at all. I was also worried about the road which is about 50+ metres away (albeit a very quiet road with perhaps only 1 car every few minutes), this doesn’t seem to affect it at all. There is an electric fence near the house to keep the sheep in, but based on the above 2 usual sources of intereference being ok, I think the electric sheep fence will be ok when the sheep are released into the fields after winter and it gets turned on in a few weeks time.

Last night geomagnetic activity was very quiet, so any readings on our magnetometer were from non-geomagnetic sources. It is VERY sensitive to temperature. I knew it would be, but it is even more sensitive than I anticipated. The gradual drop on the graph over night down to about -40nt was due to night time temps cooling the house about 2 degrees. The sudden up burst to 300+ nt around 6am UTC was due to me getting up and turning the heat pump up a few degrees and this resulted in the the sensor warming up a couple degrees. Even a sudden 1°C change in temperature can make the readings fluctuate by 100nt or more. So it is good information I can take on board when I come to build an enclosure and either put it in the barn, attic or bury it. Many people choose to bury them, but most commercial magnetometers are not buried, so it must be possible not to bury them. Ok those commercial magnetometers are in precision temperature controlled rooms. But I am thinking some thermal mass inside the enclosure and then insulated to about R100 should provide steady temperature conditions. It doesn’t matter too much if the temp changes gradually very slowly (ie over weeks), it is more a problem of sudden changes, or daily changes. So I need to make it so the sensors keep a very steady temperature 24 hours a day not varying even 1°C. I think it is possible without burying it because that is a last resort as I want the sensor enclosure to be easily accessible so I can service it, add things to it, etc..

11 March, sensor enclosure experiment
Got about 1 week before all the cabling arrives, so I am using this time wisely to run some tests on the enclosure I could use for the sensors so they remain at a very steady temperature. As mentioned previously, the sensors must remain at a very constant temperature, they cannot change temperature even 1 degree or it will produce false readings and look like a geomagnetic storm on the magnetometer. The goal is for the enclosure to be free standing in the attic or barn and not buried in the earth (that’s a last resort). It’s a tough ask to keep the contents at a steady temperature, and a lot of work now to figure out a way to make it possible, but it will be worth it in the long run and I will have access to the sensor enclosure at any time I want this way. If I buried it, yes it’s much easier to do now, but a pain in the @ss in the long run if I need to access the sensors for any reason and it is 1.5 metres in the ground. So that really is a very last resort if all my testing of an above ground enclosure fails.

To keep the sensor enclosure at a steady temperature, it’s all about keeping the heat (or coolness in winter) in there. So we need to to stop ambient air temps getting in the box, and the box temps getting out. There is 3 main forms of heat loss, conductive, convective and radiation(infrared). So we need to make the sensor enclosure the best possible insulation for all 3 of them.

So on with the tests…

Test 1:
Polystyrene cool box test
I’ve put a temperature sensor in a bare polystyrene cool box to see how the air temperature in there changes and how quickly it changes. I have put the box out in the garden, it is currently 1.6 degrees °C outside. With any luck, it will remain at a constant temp (1.6°C) inside the box for a while.

Conclusion: Failure. The temp dropped to 0.1°C inside the box within 1 hour. Not only that, the external temperature only dropped about 2 degrees, which is only a slight change. It can drop (or rise) much faster than that, which would make the problem even worse. This is nowhere near good enough. Back to the drawing board.

Test 2:
Polystyrene cool box test with sensor wrapped in bubble wrap
External temp: -1.6°C

I’ve wrapped the temp sensor in about 20 layers of bubble wrap. 1 layer of bubble wrap isn’t great at insulating in itself, I think it is R1 value (about the same as single glazed window), but wrapping something many times really increases it’s insulating value to I would think at least R15! We recommend wrapping your camera batteries in bubble wrap if you are going out into very cold weather. We won’t wrap the magnetometer sensors in bubble wrap of course, because the sensors could get too hot and get damaged, rather we want the enclosure they are in to remain constant temp, therefore the sensors will be a constant temp also. So we could wrap the box they are in with bubble wrap, or substitute the bubble wrap for some proper insulation with similar (or better) insulating properties.

We will also be adding thermal mass inside the box. What this will do is regulate heat loss/gain inside the enclosure to a much slower rate. But for the time being if we can get insulation to such a level that we are only losing a few fractions of a degree, the thermal mass will take care of the rest and also offer back up for major ambient air temperature changes such as very cold night (eg -10°C) into very warm day air temps (eg +10°C) .

I have made a new testing procedure. This will replicate extreme changes in outdoor temp which could seriously affect the magnetometer sensor readings. I will allow the enclosure (and contents) to be outside for however long it takes for inside the box to reach the outdoor temp (equilibrium). Then I bring the box in the house which is a constant 22 to 23 degrees. I will set a stop watch and time how long it takes for the temperature sensor inside the box to raise 1 degree. The goal is for the temperature inside the box to raise a maximum of only 1 degree in 1 hour using insulation only and no thermal mass (which will be added later). Once we have the insulation sufficient enough to keep the temp sensor only moving 1 degree over the course of an hour, we will then add thermal mass (likely a few litres of water in sealed containers treated with ethanol for freeze protection) to the box.

Water is an incredible heat storage medium. 1 litre of water will hold considerably more heat than 1 litre of soil, sand, or event cement. The reason the box is losing heat so fast now is because it is so empty, it is mostly air. Once we add something to “store” the heat (or cold) which is inside the box, it will act as a radiator inside the box, keeping the temperature at at an exact level for long period of time, regardless of the what the air temp is outside the box. But this can only be achieved if the insulation is considerable to stop this heat escaping outside the box too fast. If a temp sensor which is around 0 degrees doesn’t register more than 1 degree rise in 1 hour inside a room which is 23°C, it should be more than sufficient for putting the box in the attic or barn where the temperature changes will be much smaller, likely only a few degrees per day (at best) or only a few degrees per hour (at worst)..

Ok results are in, in a 22.4°C room, the temperature sensor inside the box:
Raised 1 degree (from -1.6°C to -0.6°C) in 14 mins. Fail.

I did want to see how long it takes for the box to reach the same temp as the room, but it’s taking too long and will only get longer the more insulation I add, so that isn’t feasible. I will just cancel tests after 1 hour in future regardless of temps. So this test also failed. But it was a significant improvement over the first test. But still a long way to go.

12 March, sensor enclosure experiment
I got some insulation and some other bits today. One of things was a very small “traditional” style cool box which will sit inside the polystyrene cool box. It will be interesting to see how this does as it will cut down conductive heat loss/gain a lot.

Test 3:
Small normal coolbox inside a polystyrene cool box.

External temp: 2.1°C

Ok results are in, in a 22°C room, the temperature sensor inside the box (inside the box):
Raised 1 degree (from 2.1°C to 3.1°C) in 22 mins. Fail, but the best yet.

Conclusion: With the thick polystyrene box and now a normal cool box which has a layer of air trapped in it’s walls, and also considering there is nothing actually in the boxes storing the coolness, it is proving remarkably effective. So I think for the time being we have got conductive heat loss pretty much to the bare minimum. I don’t think there is much convective heat loss, because that only really happens when you take the lid off and we aren’t doing that here. So I am thinking radiation heat loss might be a factor here. So I am going to run the same exact test again now, but with aluminium foil lining the boxes, shiny side in.

Test 4:
Small normal coolbox inside a polystyrene cool box. Both lined with aluminium foil.
External temp: 0°C

Ok results are in, in a 21°C room, the temperature sensor inside the box (inside the box) with aluminium foil lining:
Raised 1 degree (from 0°C to 1°C) in 25 minutes. Fail.

The foil didn’t improve things considerably, but it did improve things. Losing 1 degree in 25 mins in what is basically empty boxes is pretty good. But we are still some way from our goal, so in the next step, I will add some professional fibreglass insulation into the mix and see how that improves things.

Test 5:
Small normal coolbox inside a polystyrene cool box. Both lined with aluminium foil. Fibreglass insulation added to sides and top of polystyrene box, and a few small pieces of insulation added in the smaller coolbox just to keep the temp sensor raised (I thought perhaps that the temp sensor picking up heat from the floor of the house may be an issue).

External temp: -0.6°C

Ok, results are in and wow, what a difference, nearly doubled the time just by adding this fiberglass insulation. The temp sensor took 43 minutes to rise 1 degree from -0.6°C to 0.5°C. It’s still a fail because the target is 60 minutes.

We’re getting close! We haven’t reached our goal yet, but it feels within reach now. What this test has proven is that moving a box from freezing temps outside suddenly to indoors +22°C, it takes 43 minutes for whatever is inside the box to heat up 1 degree. Finnish weather is not that turbulent. There is no way the temperature can change so much, so fast, so we have been testing with around a sudden 23 temperature change as an absolute extreme case scenario. In reality, I doubt we will even see a 10 degree temperature change like that. It is always a gradual change, a few degrees per hour usually. So we are confident now that we can have the magnetometer sensors in an above ground semi-portable enclosure and still enable the temperature to remain very steady. Right now is perfect time for testing these things as nights are dipping into minus and day temps are anything up to 10°C. This is where the magnetometer sensors would struggle to remain constant. The sensors might be a little too close together in that small cool box, so we might have to make a box from scratch.

14 March, yet more sensor enclosure experiments
We thought we were doing pretty well with the double cool boxes and thought we were pretty close to making a temperature stabl eenvironment for the magnetometer sensors. We were wrong! By chance we left them out during the day time yesterday, and the temperature inside the box raised 7 degrees throughout the course of the day. That is a catestrophic failure. If the magnetometer sensors were in there, that sort of temperature rise would show kp9 on the chart despite it being only kp0! It would have been catestrophically bad. Yes we hadn’t added the thermal mass (water) inside which would have slowed down the temperature rise a few degrees, but it’s still bad, no matter how you look at it.

Test 6
So it’s back to the drawing board. We have ditched the cool boxes and will make a whole new super insulated enclosure. I had a look at what kind of plastic tubs I have already and found a good candidate for this project. So we immediately set to work on it. New box size comparison to cool box:

We picked up some crazy thick polystyrene we will cut up and make and line the box with it. We don’t need a great deal of space inside the box, the sensors are very small, they have to be about 10 centimetres away from each other to avoid interference, but we should have this space in this box even with mega thick insulation.



Ok results are in, it took 40 minutes to raise 1 degree. Fail.

We thought it would be better than that, but it’s not as well sealed as the cool boxes, so air can easily escape or enter through the lid. Either way, with just a single layer of polytyrene, albeit uber thick, there is certainly room for improvement here. So we done another all-day test with it. We left it out all night and all day, the temperature inside the box actually only raised 1 degree the whole day, despite the ambien t temperature rising about 12 degrees. The temp inside the box never rose above 3.5 degrees. So that is very positive. It should be noted on this all-day test we added 5 litres of water as thermal mass. It really shows how well it works.

But 1 degree rise throughout the day is still unacceptable. So we need to super-insulate the box further still. There is no more room to add any more insulation inside it, so we will do a similar thing where we put a cool box within a cool box, we will put this box inside and even bigger box.

Test 7
We had a bigger box in the barn that the grey box above could fit in. We added fibreglass insulation on the bottom top and all around it. The spare wheel is just to keep the lid on for the time being. I need to make some sort of way to lock down the lid and also seal the it airtight.

The results are in, when I brought the box inside the house from outside (about a 20 degree temperature difference) it took 1 hour and 10 minutes for the temperature to raise 1°C! That’s a PASS. The lid weren’t sealed properly and it was without any foil so this will be improved further still when those 2 factors are done! Pictures will follow tomorrow. The box is pretty big now, but still portable, and it is incredibly lightweight as it is just filled with insulation, I doubt it even weighs 3kg (without thermal mass added). I have just put the 5 litres of water inside it and put it in the yard and will do an all night/day test on it now to see how it performs. I don’t really intend for the sensor enclosure to be sited outdoors (it was going to go in attic or barn or somewhere sheltered at least), but on the all-day test yesterday it was windy as hell and the box was just sat in the garden being buffeted by wind and even a bit of sun hitting the box and still the temp only moved 1°C all day. So it is very interesting to see how this new box performs tomorrow, it may actually have the potential to be sited outdoors in a field or in the woods right amongst the weather elements and still be temperature stable, which will make it very very, very flexible indeed! Fingers crossed the temp inside doesn’t budge!

The new box is black, which will absorb serious heat from the sun (1000 watts per metre squared!), so I need to move it into a shaded area by about 11 am before the sun hits it. I also need to paint it white which will reflect most of the suns rays and not heat up anywhere near as much. So tomorrow I will buy white spray paint.

Update will follow tomorrow (15th)

15th March
Ok just a quick note, it is performing awesome so far today, the temperature inside the box has been a solid 2.2°C for a few hours. Ok it’s snowing so the ambient temperature is very steady which makes it much easier for the box to do it’s work. The past few days temperature have been going from -2 at night to +13 in day time. It is those sort of days we are going to so much trouble to super insulate the box to avoid the magnetometer sensors changing even half a degree in 24 hours. Ideally, we don’t want to see any temperature change at all on a daily basis, but we expect to see very slight gradual changes due to seasonal changes. We will perhaps put a temp logger in the box when it’s permanently set up so we can monitor the exact temps on a graph all temp changes, whether daily, monthly, seasonal or whatever.

The reason we are going to so much trouble to super insulate a box so we can keep the sensors above ground (when everyone else buries them a least 1 metre in the ground), is because we want to do a lot of testing with the magnetometer in different locations around Aurora Service HQ for interference, performance, testing, etc. It would be a real pain if we had to dig a new metre deep hole each time we wanted to move the magnetometer! All the testing is for a reason, it would nice to build a few more magnetometers and perhaps put them in remote places of Scandinavia and UK and perhaps even write our own kp system exclusively for Europe (instead of the global system we currently use from NOAA which sometimes misses localised activity to Europe). So burying them is not a feasible option. For instance, if the site we place them turns out bad and we have remove the magnetometer after a few weeks etc. They really need to be semi-portable magnetometers that can be installed/uninstalled at locations in under an hour. Anyway that’s a long way down the road…back to our first magnetometer…

15 March
Ok I had the big box out in the yard for nearly 24 hours. It wasn’t sheltered at all, it was just next to a spruce tree in the middle of the yard. It was snowed on and was being blasted by serious arctic cold wind all day. Despite all that, the temperature inside the box only changed 0.7°C in 24 hours. And still this is without foil or even the lid on properly. So I think we have a winner. Tomorrow I will refine the box in preparation for the sensors to go in (probably Wednesday), paint it and try and lock down and seal the lid somehow. Then I will put it in the attic for 24 hours, where the temperature should be more stable (and no wind chill) and see how the temp goes. I am hoping for 0.0°C change in temp over 24 hours. So this is the big test. There are no bigger boxes if this test fails!

18 March
Weather has slowed progress on the sensor enclosure the past couple days. The weather the past week had been very warm, but then winter decided to return with plenty of snow and minus conditions. There isn’t a workshop/garage here so everything must be done in the yard. I tried painting the box yesterday and the paint froze on the box the instant it landed. So, I had to put it off until the mercury rises above 3 or 4 degrees.

All the cabling (about 100 metres of) should be arriving tomorrow, so then we will have everything we need to to hook up the sensor enclosure. We just need the weather to warm up so I can paint the box, foil line it with glue, and do a few modifications to it for the sensors to fit and the cables to attach. If all goes well, I am hoping to have it up and running in a couple of days time.

20 March
171.5 metres of cat5 cabling arrived today. the 50 metres of black cable is for the sky cam project, that is outdoor spec cat5 cable. I have 120 metres of standard cat5 cable for the magnetometer, hoping I only need half of that, but I though I best buy twice as much as I think I’ll need just incase. I have to run 1 cable to each sensor, so if the sensor enclosure is 20 metres away, I will need 3 x 20 metre cables.
I also bought shielded patch cables to use inside the enclosure to cut down on any interference. I thought I would buy coloured cables in the colours of the X, Y and Z colours on the graph :)

Basically what I have in my mind is the logn cables running to the sensor enclosure, where I will fix network sockets (RJ45) on the outside of the sensor where I can just plug the long cables into without even opening the box.On the inside of the box the sensors will have the coloured patch cables going to the socket, perhaps spaced 10 cm apart. This will make the sensor essentially plug n play, so it is very easily moved. I am still worried about temperature affecting the sensors, but if it does affect it, I have some great plans to fix that. I will be making the sensor enclosure thermostatically controlled heated. It is just find a way to heat it that won’t affect the sensors. But I will cross that bridge when I come to it, I will see how it behaves for a week first once it is up and running. So now all the cabling is here (although one part was missing from the order that I really need to connect a sensor) I can really make a lot of progress on this now and get it pretty much up and running.

27 March
Been waiting on parts to arrive, today the last piece of the jigsaw arrived. I have put the laptop under the house already and the magnetometer and the sensors are all but ready to connect up, which I should be able to do tomorrow. The temperature fluctuations we will have to deal with for now, but I am planning to add a very basic heating system to the sensor enclosure to keep it at a super steady temp all year round.

I actually resorted to back the small blue cool box because when I came to attach the sensors to the polystyrene boxes the screws wouldn’t hold tight in them and the sensors would easily move out of position. The plastic cool box the sensors seem to be staying put at their correct angles, and I will address the issue again in summer perhaps and make any improvements. But for now I just want this thing up and running it’s taken too long! So tomorrow I’ll put magnetometer control unnit and sensor enclosure under the house. It’s all timber construction using dovetail joints, no metal parts. So it should be ok down there. Also the temperature is pretty steady (though not steady enough I think).

28 March! It is completed!
Ok, the last few parts to arrive was really holding the project up, but once they came (yesterday) I got the sensors installed and everything up and running in the cave beneath the house.



























It will take a couple of days for the sensors to settle after all the fiddling and moving today. So the graph will take a day or two to start being steady. I can then monitor it and fine tune it, make changes to the sensor enclosure as needs be. Once it is pretty reliable I will put it on the Aurora Forecast page :)

Total costs:
Magnetometer: 222 Euros
Cabling & connectors: 50
Enclosure hardware, screws, sensor holders: 20 Euros
Compass, spirit level, multimeter, soldering stuff: 40 euros
AC Adaptor: FREE (I found one in the garage that had the right voltage/amperage)
Some small cabling for the circuit board: Free (Had some lying around)
Labour costs: 5000 Euros. Seriously time consuming project this. Basically because every aspect of it was new to me so I had to learn everything as I went, especially soldering PCB boards, very tricky (or this one at least, it’s very small and quite intricate soldering).
Total costs (excluding labour costs): 332 euros

I also had to donate an old laptop to the project so I didn’t have to put cables everywhere in the house. I put a laptop in the basement and that wirelessly sends data to this website. Saved a lot of work.

Post install issues/updates
It was not doing too well underneath the house. Although it was under the floor in very empty and incactive part of the house, there was still something that was affecting the readings really quite majorly. I am not sure what it was, but there were a large few spikes on the graph. It was even happening at very quiet times, like 3 am. There was no traffic, no human activity in the house or whatever. It was a bit strange, I wondered if it was an animal living under the house walking past the sensors…

Anyway, all that time testing the sensor enclosures was not wasted because I have now put the sensor enclosure outside about 20 metres from the house. I might still get some temperate fluctuation affects, but they are pretty easy to spot and I can deal with them. For now I am trying to find a spot where the magnetometer will be unaffected by anything human related or manmade.









1 April
Well it looks like we solved most of the interference issues. The By and Bx sensors are producing perfectly smooth lines on the graph now (see top of page). But what the heck is going on with the Bz sensor! It is all over the place, jumping 10,000nt in the blink of an eye. That is strange because the Bz was actually reliable with little interference when the sensor enclosure was under the house. The By is the most important sensor, with the Bz the least important, but we would like to get the Bz sensor working as it should be. So more troubleshooting ahead.

2 April
It looks like the Bz sensor was pointing directly at it’s own cable which may have been the issue with it acting strangely. I thought the cables being shielded would prevent that from happening, but evidently not, so something learned there. I also made some changes to the sensor enclosure to try to prevent the temperature fluctuations so much. I had it stuffed with insulation, but it wasn’t proving effective enough. So I removed the insulation, and put some bricks on top of the small blue cool box (to stop it floating) and filled the bigger box it sits in with about 30 litres of water, which is now surrounding the blue cool box. I also put 1 litre bottle of water inside the cool box too. It takes a heck of a lot of energy to raise water temperature even 1 degree, so I am hoping this water cooling will keep the sensors at a very steady temperature. If it works reasonably well, I might put a fish tank heater in an set it to 20°C all year round. Anyway, now we wait for another day for everything to settle down and see how it looks tomorrow.

3 April
Ok I give in, I buried it today. It was just so temperature sensitive everything I tried above ground failed. I suppose a perfectly heated box or something would work, but the heater would emit a magnetic field which would interfere with the sensors no doubt. So for now, I just buried it. Not especially deep, so I will monitor how the temperature fluctuates now and take it deeper if I have to. The Bz sensor is still going crazy, and now it is buried. So I will see how the temperature goes, and if I need to dig deeper I will fix (or at least attempt to fix) the Bz sensor then. I’m really not sure what could be causing it to go crazy like that. Perhaps there is old iron plumbing or electrical cable under the ground or something.

10 April
I dug it up again. The Bz sensor was going crazy for some reason and all the sensors were still being affected by temperature affects (I didn’t bury it very deep though). I think I fixed the Bz sensor, it was two pins touching each other shorting it out. I think I fixed it, but then I was fiddling with the sensors again so much the pins might have touched again. Time will tell. But the sensors are reasonably accessible now they are suspended directly in 50 litres of water. I can get photos next time I got the box. It’s pretty cool now. I hope it works. If the temperature still rises a few degrees, I will put a fish tank heater in the box of water to keep it at a steady temp. This will also freeze protect it in the winter months. The only thing I am concerned about is the fish tank heater might interfere with the sensors. The sensors are now quite far apart from each other, so that should totally rule out any interference they can get from each other. I filled it up with water from our well, it was very cold water, so it will take some time for this water to reach similar temps to ambient temps. Which is a good thing and exactly why I done it, because it will take a long time for the temperature to change if the ambient temp changes fast. It’s all about keeping the sensors at a very steady temp.

13 April
The sensors submersed in a box of water seems to be working beautifully. The first experiment actually failed because I put each sensor in a sandwich bag thinking that would keep them water proof, but all 3 sensors got wet within 1 hour which caused a whole host of problems, from minor corrosion, to shorting the pins and crazy readings coming back. So I had to do it again. But how the heck could I water proof them sufficiently was the challenge. But using some outside the box thinking, I found a solution. You can imagine the look on my partners face when I ran into the house asking for 3 condoms urgently!





But, they are working a charm. And so far so good, the sensors have been submerged in water for a couple days now and remain waterproof.

The Bz sensor is STILL giving me problems, I have done everything to ensure it is getting clean connection and still the readings are slightly erratic (but MUCH better than they were). The Bz measures vertically, and as such it is pointing directly into the ground. So I can only assume there is ground currents or something that is causing it to behave like that. When the sensors were under the house in the first test, the Bz was behaving normally, so today I am going to move the Bz sensor to a new location and see how it behaves. As the other 2 sensors are performing perfectly (they just need a little fine tuning) I will leave them in the water. I might make a new water box just for the bz sensor if I can find a place where it is not behaving strangely.


16 April


Back to the drawing board with this idea…

Ok, where to begin. I started the whole sensor thing again from scratch. I found out those copper connectors and brass screws I was using are not 100% copper or 100% brass. Perhaps the copper was, but the brass screw started to rust submerged in the water, which tells me the manufacturer of those screws has not been entirely honest and they were a mixture of brass and iron I think. So they are more like brass coated. I originally believed them to be non-ferrous (ie 100% brass), which should not interfere with a magnetometer (which is basically a super sensitive metal detector), but I am of the firm believe they were interfering somewhat.

Secondly, although the box of water was superb for keeping everything temperature stable, the condom trick didn’t work as well as expected. I don’t know how it is possible, but there was moisture getting into the sensors, they were not wet, water was not getting in, but there was moisture/condensation on the sensor pins which were causing some issues. So I basically give up on the water idea (for now anyway, I still think it’s a good system).

So I started the whole sensor enclosure again, using nothing metal whatsoever, non-ferrous or ferrous, because manufacturers cannot be trusted.

So back to the drawing board. What can we use to hold the sensors instead of the copper clips. Well, we have these orange plastic pipes you stick in the side of the roads in winter, so the snow plow driver knows where the edge of the road is.


It just so happens they are almost the perfect fit for the speake fgm-3 sensors we use for the magnetometer to fit inside. So I set about setting up the new sensor enclosure using them.





I put a few five litre water bottles in to act as thermal mass. Five of them are filled with water, making 25 litres of water. The 6th one was filled with sand and the Bz sensor (which has to be vertical) pushed in.

IMG_0837 IMG_0838





















I put the white cover on to stop the sun heating up the soil. In winter (which is when we need the magnetometer most due to it being aurora season) It will likely be under snow, so we don’t have to worry about the sun so much then. But for this summer when we will be seriously testing it and scrutinising the results against proper observatory magnetometer traces, the white cover will do. If this is much soil is not enough to keep the temperature stable, I will put another half ton of soil on. Note I removed the bricks from the top of the lid before I started throwing soil on because the bricks + soil probably would be too heavy for the plastic box. But soil on it’s own should be fine. I filled the box full of insulation too, the lid could barely shut, that is why I had the bricks on it in the first place.

You might be wondering where I am getting all this soil from?


Well there is now a crater in the garden. Someone is not going to be too please when they see that….but it is after all, in the name of science and that takes precedence :)

17 April
I put a couple more wheel barrow’s full of soil on top of the mound today. I really think the temperature should be more steady that it is, so I think the exposed cables could be to blame as copper is great conductor, the heat could be travelling up the data cables to the sensors. So I buried the cables for a metre or so leading up the the enclosure too as well as the extra soil on top of the enclosure box. If that doesn’t do it..I will put more soil on…It’s getting pretty big this mound now though….

19 April
Ok, so there were still some temperature fluctuations affecting the magnetometer so today I buried it. I didn’t intend to bury it too deep, just so the lid was above the soil line. It was just aswell, because I hit bedrock at about just this exact level! But anyway, it should improve things, how much will remain to be seen. The graph at the top of the page will be going crazy for a while until the temperature stabilises.















23 April
Ok, now I am really pulling my hair out. I have done everything possible. This should not be behaving like this anymore. There must be interference, there is no way this is down to temperature. Take this image from today for example:

In the middle of the night, it appears to be behaving reasonably ok, all the traces are steady, this is good, this is exactly what I want to see. Then look at 3am UTC, which is 6am here, things start going crazy, then at 6am UTC which is 9am here, it goes nuts, the Bx especially. There are 3 issues which can be causing it.

1. The sun comes up about that time, could it be the direct sunlight on the mound of soil? I don’t think so. Too much soil and thermal mass for temperature affects to change so rapidly.

2. There are electric power lines about 20m to the left of the magnetometer.

3. The road is about 60m-70m to the right.

Could it be I am seeing the powerlines becoming active at 6am when everyone on the island is putting their kettles on? Or could it be the traffic on the road is none existent at night time then it slowly starts picking up in the morning. Could it be this?

I have ran out of patience now now, I am don’t have the time or the means to investigate what is causing it, so I am just going to move the magnetometer to a whole new location in the middle of the woods. It crossed my mind to put it there in the first place but it is SO far from the house. Also the laptop that the magnetometer is connected to is also connected to the sky camera, so it would mean moving the sky camera also. Both cause major issues, because cable length is the limitation. I already have issues with the camera using 30 metres of cable. And I Know the magnetometers start having major issues with long cable lengths, so I want to stick with the 20 metre cables I have now so I don’t open a whole new can of worms. So it basically means I will have to run a super long power lead into the woods and have the laptop there with the magnetometer and sky camera attached to it. But then I need internet connection to ftp the camera images and magnetometer data how am I going to have internet connection in the middle of the woods. What a headache.

Here is a quick image to show the layout here and the I am basically caught in the middle of a triangle of interference.


Ok, I just dug up the magnetometer, the bx sensor is actually measuring north to south and there is really nothing there which should should be causing issues. It’s just a field directly north and south is the house but it is timber built on stone foundation and there is practically nothing on in there generating currents (therefore magnetic fields). I did notice some copper corrosion on the connector which could be causing it, so I swapped the connector out and we’ll see tonight if that was the issue.

7th May


I have got the sensors very temperature stable and they are recording the magnetic field well with zero infterference from the house, power lines or the road.

I have actually buried them in the same place as above and didn’t need to move them at all.

What was going wrong?
Firstly, the sensors were interfering with each other causing some of the erratic readings. I changed the way the sensors went in the box so they were not close to each other and they were not pointing at each other (or each others cables).

I removed anything metallic, whether ferrous or non ferrous. Which may or may not have been causing issues in earlier versions. (Especially the brass screws).

I removed the water bottles acting as thermal mass as there could have been iron in the water.

But the biggest problem, was that the sensors were overheating. Also the sun heating the exposed cables was carrying heat to the sensors causing them to heat up in the daytime.

Wyy were the sensors overheating? Those orange pipes I was using actually turned out to be a bad idea. Because they were a tight fit around the sensor it wasn’t allowing air (or heat) to escape from the sensor. When I say they were overheating, they were not getting hot, but they were luke warm to the touch, which on a temperature sensitive device, was just sending the readings crazy. Like 5000nt crazy.

So how did I fix the overheating issue? I removed all the insulation, the water bottles etc so that big white box was just empty. I put it in the ground, directly on to the bedrock itself which was only half a meter deep. I put a bit of soil underneath the box to act as a thermal bridge from the bedrock to the box rather than just having pockets of air there. Then I started to shovel in soil straight into the box. After I put a few cm of soil in the box, I set the sensors directly on this soil, aligned them with N-S and E-W then piled soil directly on top of them and filled the box right up to the brim with soil. This should keep them water proof in the rain. Then I put another ton of soil on top of the box too for good measure. By insulating the box in previous versions, the very tiny heat output from sensors was building up inside the box allowing the sensors to get warm. Now there is zero insulation, the bedrock half a metre under the ground is at a very, very steady temperature, I actually want any heat in the box to leach out so the box stays the same temperature as below ground. So far so good, it is working perfect.

Once I figured out the overheating issue and the fact the sensors were also interfering with each other. I thought I didn’t need to move them to a new location like I mentioned above. The road is such a quiet road and at at least 80 metres away so I was quite confident that wasn’t an issue. I was monitoring the magnetometer values in real time as cars drove past, and it was not registering any activity.

Secondly, the power lines are countryside powerlines (or actually it is just an individual powerline). Like this: powerline

So I just don’t believe there is much current going through it enough to cause any issues. If it was one of those ugly big metal powerlines that you can actually hear buzzing and making weird sounds, then yes, it would be major headache. But these countryside powerlines don’t seem to be much of an issue. As a test I had one sensor directly under the powerline and even then it wasn’t registering a thing. So pretty safe at the distance it is from the sensors buried in the ground.

I also buried the cables going to the box up to about 2 to 3 metres before the cables enter the box, to stop the sun heating up the cables as was happening previously. But in winter, everything will be under snow and/or the sun is at such a low angle on the horizon the cables won’t be heated anyway, but I want to test the magnetometer this summer so I want it to be as good as can be.

So that’s that, all my problems were down to:
1) moisture
2) sensors interfering with each other (too close to each other and/or pointing at each other)
3) sensors overheating
4) heat transferring from cables to sensors
5) non-ferrous metals I was using in early versions probably DID have ferrous metal in them. Plus the sensors were too close together (in the first blue cool box version)

So now it is time to test the magnetometer. Now it is summer you might think is a bad time to test the magnetometer. But that isn’t the case. The fact that we can’t see auroras in summer is just that, we (humans) can’t see them. But the auroras are still there, and the magnetic field still behaves exactly the same as in winter. So the magnetometer will function exactly the same now as it will in winter. So despite not being able to see auroras personally, I am keeping a keen eye on geomagnetic activity this summer so I can see how the magnetometer behaves in a geomagnetic storm, which is really the purpose it was built for to record geomagnetic storms at this latitude. More specifically, I will be comparing our magnetometer to a commercial magnetometer on a similar latitiude and seeing if the peaks and drops of nt’s (nanoteslas) are of a similar range. I can adjust the sensitivity of the sensors if they are over sensitive.

Now, I can relax for a short while. But I have the next aurora service project starting soon (top secret at the moment) which is also electrical and then outdoor and is also susceptible to interference so another few sleepless nights beckons surely with that project :)

Thanks for reading. I documented it all in painful detail for people who want to install their own magnetometers. No detail is too small when you are building something and are looking at where other people went wrong. Hope all the above info helps.

HUGE Thanks goes to (in no particular order):

EVERYONE who donated. The donations covered probably 75% the cost of this project, without which this project would never have gotten off the ground.

Roger (GM4PMK) from who runs the same magnetometer and provided some very useful advice.

Bill Speake from Speake sensors who manufacture the fgm-3 sensors who gave us some useful advice (especially it’s ok to solder the pins on the sensors).

Dirk Langenbach and Karsten Hansky who provided the magnetometer kit and have provided some really great support since.


There is still interference. I will re-start from scratch! A new page will follow!

Add Comment Register

Leave a Reply