Tutshill - Chepstow - Monmouthshire


Home Made Magnetometer





Main Menu


Home Brew

Radio Home 

Weather Home


Sub Menu


Lightning Detector





Whilst searching around on the internet for circuits relating to lightning detectors, I came across information on building a magnetometer to monitor variations in the earth's magnetic field.

Most Radio Hams are familiar with the 'Boulder K Index' which is transmitted by the WWV radio station. The 'K' index is a measurement of the amount of disturbance in the earth's geo-magnetic field.

 Geo-magnetic storms, are caused by solar flares, (eruptions) which can create a 'coronal mass ejection' on the surface of the sun. 

Active areas of the sun eject charged particles toward the earth, if the active area is facing in our direction. As these particles reach earth, they interact with the 'magnetosphere,' (an area of magnetic field surrounding the earth, extending outwards into space) and cause changes to radio wave propagation and also sometimes visible aurora. 



During these events, it is possible to rotate your VHF antennas toward the aurora and 'bounce' a radio signal from it over greater distances than would normally be the case.

Whilst the science surrounding aurora  is complex, David Olean, K1WHS, has produced an excellent paper on the subject, explaining the background in an easy to understand manner. David also covers the building of the magnetometer which I constructed.

 The paper in 'PDF' format can be reached here. 


The magnetometer is available as a 'kit' of parts, and can be obtained from Speake Sensors. Their web site is at:


Other items, not included in the kit, are a case, power supply, switches, cable and control knobs. I also added a meter output to the front panel. The remote sensor housing is covered further down this page.


The magnetometer circuit board was built up as the clear instructions included with the kit indicated. The choice of enclosure and layout is left up to the constructor. As mentioned, I decided to add a meter output, in this case a 1944 vintage 0 to 1 mA DC panel meter, from my 'spares' box.

 The sensor can be operated in four sensitivity positions, however for this purpose the minimum sensitivity is more than sufficient. On power up, the meter swings to the central position. The 'meter gain' control is used to center the needle accurately. The minimum sensitivity setting gives a full scale deflection of 0 to 2.56 volts DC, which equates to 1000 Nano Teslas i.e. -500 to +500nT  

Should the trace wander off scale, or if you happen to turn on during a peak or trough, the magnetometer is also equipped with a 'reset' function.  A simple non latching 'push to make' switch is used here.

In order to display the data from the magnetometer over time, a data logger is desirable. I use a Dataq Instruments DI-148U. An output is included on the magnetometer circuit board for connection to the logger. 


Dataq Instruments web site can be found at: 

The data logger connects via a USB lead to the computer, and is controlled by Windaq Lite software, which is supplied with the DI-148U. The software is set up to provide a trace which scrolls from right to left across the screen. Current data is on the right of the screen. Compression has been enabled to give a period of 30 minutes per vertical division, and therefore the previous 17 hours values can be viewed. 

As mentioned above, the magnetometer sensor requires its own case. The sensor needs to be remotely sited, well away from possible sources of interference and ferrous objects. Following David's approach, I mounted mine inside some plastic water pipe, which I sealed to keep the assembly watertight. It is strongly advised to keep in mind these sensors are temperature sensitive devices, and to stop the output trace drifting over the screen I decided to follow others and bury mine. A two foot deep hole was dug and the sensor housing assembly placed inside that. Waterproof connectors were used to take the three wire control cable back to the control unit in the shack.  


The magnetometer sensor FGM-3h is so sensitive that it must be very carefully aligned on an east/west axis. The magnetic field of the earth's north/south axis is too strong, and will swamp the sensor. I used a magnetic compass to get a rough alignment, then used a frequency counter to fine tune. The idea is to take a reading, then rotate the sensor through 180 degrees and read again. The frequency output from the sensor will be the same on east/west and west/east alignment when properly positioned.

The hole was filled in with builders sand to give good thermal insulation, and the gravel from the garden replaced to keep 'The Boss' happy!

At some stage I intend to take the sensor out again, as, following some helpful advice from David, I want to try the less sensitive FGM-3.



During the course of the project, I looked at various methods of sending the data to my web site, so the information gathered could be shared with other Radio Hams or interested parties. I came across some screen capture software which has proved ideal for this purpose. The 'Snagit' software can be obtained from:  

The software, which is capable of timed ftp upload intervals, can be configured to capture either all or part of the computer screen. 

There is a drawback to this approach, which occurs when a computer re-boot is required. The previous recorded hours data disappears from the screen and it obviously takes time to build a full screen again. In these circumstances the page will display 'Magnetometer Offline' until sufficient data has been gathered.

I would like to pass on my thanks to David, K1WHS and to Bill Speake, from Speake Sensors, for their help and advice.


Update June 18th 2010:

After monitoring the output data for a year now, it has become very obvious that the FGM-3h sensor is far to sensitive for this application. Apart from alignment difficulties, and therefore trace linearity problems which have become apparent over the course of the experiment, the sensor is easily swamped by the variations in the magnetic field.

As a consequence, I have removed the sensor installation, am rebuilding the housing and replacing the sensor with the FGM-3 model. I hope to have this work complete by the end of July. 


Update July 21st 2010

The FGM-3 sensor is now waterproofed, boxed, and ready for burying underground. However, due to rainy weather here, I have not been able to start digging as yet. I am also taking the opportunity to move the final location of the sensor, as I found some local interference was being experienced in the original location. 

I have the unit working and sending data back, which can be viewed by following the link below. Please note, the steep inclines (twice per day) are probably temperature effects on the sensor enhancing the diurnal change, and please remember also I am working on improving the sensor, as time permits, so when I occasionally move it around there will be data swings which are down to me - not the magnetosphere! 

I have altered the software settings in the  Dataq DI-148U to give 34 hours data instead of the 17 hours previously displayed. I have also reduced the sample rate, as this seems to be effective in reducing 'local noise,' and gives a more representative trace.

Regarding temperature stability, I have also encased the sensor unit in a 24 inch polystyrene box, packed full with polystyrene sheet, to give even better insulation.

I have re-instated the internet upload for anyone who would like to see the 'above ground' output, but the unit will again have to be switched off when I can get its new home underway - hopefully soon!  


Update 2nd August 2010

The final stage of installing the magnetometer sensor unit has now started - at last! I have prepared the hole, and the sensor, inside it's polystyrene box, is sited, level, and aligned. There is a solid paving slab on which the box is sitting. I have added some sand to the base and sides of the box, to provide some stability. I want to run some more tests to ensure proper alignment prior to filling the hole back in.  

I have also set the sensitivity of the control unit to a higher level, and am using switch position '2' which makes the 0-2.5v swing equivalent to -275nT through to +275nT. This has also been reflected in the Windaq software settings, as can be seen from the output trace. 


Update 9th August 2010

The magnetometer hole is now being back filled, a bit at a time, to allow the soil to stabilize around the magnetometer box. 

Live data is being uploaded, and I have further modified the output, to display 24 hours recordings. The data now scrolls from left to right, as I am experimenting with the logger operating in oscilloscope mode. The white vertical bar on the output screen-shot represents the current  position of  the data trace at the displayed upload time. To the right of the time-line bar, the position of the trace 24 hours earlier can be seen.

I am also working on a small circuit which will apply a 'reset' function at 2359z each night to the control unit. I intend to set the Windaq software start time to synchronize with the 'hard' reset. This can be achieved by using the windows task manager, and will ensure data upload will always commence at 2359z in the event of a computer re-boot. I hope to have this modification tested and installed in the next couple of weeks, and also have our garden put back to normal to keep the 'Station Manager' (my XYL) happy!




Update 16th August 2010

Following some fine weather, the sensor installation is now complete, and the garden has been returned to normal.

In addition to my next task, ie. building the circuit mentioned above, I have also come across some 'strip chart' software which provides additional functions (data logging) and a much improved display at: 

On the surface, the Skypipe software looks more suited to my particular project. However, I am unsure as to whether this will work with the Dataq DI-148U, but an application has been written for the DI-194 RS. This leaves me with the option of either building an A/D converter suited to SkyPipe, (details on the web site) or purchasing a DI-194. These options are under consideration once I can determine if Skypipe will work with the DI-148U.




Update 11th December 2010

Since the last update, I experimented with the software configuration, and came to the conclusion that (for me anyway) Skypipe software would not work with the DI-148U. However, as the Skypipe screen chart out performs the Windaq/Snagit setup on a number of levels, I was still keen to use this program. I decided to follow the route towards a Maxim MAX-186 based data logger.  

As time had marched on, and I wanted the new system installed and working by the year end, I decided to 'cheat' a bit and purchase a new data logger. These are available as a ready built module, or as a kit, from:

Quazar Electronics

In addition, a circuit diagram is available at the Radio Skypipe website for those who wish to build a unit themselves from scratch.

Having obtained the new logger, which fits within a small DB-25 (25pin) plastic housing, and has male / female connectors at either end, I also purchased a DB-25 breakout board. This enabled me to take the power and data cables straight to the rear panel of a project box in no time at all. This box also holds the 'reset' circuit board, again available from Quazar. 


The board on the left is the reset unit, the data logger is contained within the black plastic housing on the right. A ribbon cable takes the signal from the data logger to a chassis mounted DB-25 serial connector on the rear panel. This connects to the parallel printer port (LTP1) on the computer. Signal 'In' from the magnetometer control box, and 'Reset' signal are routed through BNC connectors. I also added another BNC connector, looped to the 'Signal In' to provide an additional test point / output if needed. 

The new system is currently on test, it seems to be working well so far, and has now replaced the Windaq screen uploads on the main magnetometer page.


The magnetometer output can be found in the 'Ham Radio' Section of my web site:

Live Magnetometer Chart


Return to top of page