Magnetic Field Detector

[gkec]

For some reason I can't remember I thought it would be fun to build a magnetic field detector to detect changes in the Earth's magnetic field due to Solar storms. So effectively making an Aurora predictor, not that I have ever seen the Aurora, that was impervious to the usual problems experienced by other detectors, e.g lack of space and easily disrupted.

After a bit of research I found that I would need a resolution of about 50nTesla. I found a device that claimed 15nTesla resolution which could easily be interfaced to a USB port via a SPI to USB converter.

I manged to wreck the first device through incompetent soldering so had a PCB build for a second device.

The SPI interface had a lot of issues but I eventually got it working and hooked it up to an old laptop running Linux which would average readings over 5 seconds and write the results to an SQLite database.

I then used a libwebsockets server to serve HTML/Javascript pages and sedn and receive data using websockets. Coupled with JSON/SQL queries this works really well.

I wasn't quite ready but I heard that a Coronal Mass Ejection had occurred and was due to hit Earth on Sunday 17th.

So I set everything going. I had previously done a test run and had an idea of what data I could get.

On Sunday at 12.55 I got an email from Lancaster University Aurora watch indicating magnetic activity. I also observed a spike on my detector.

I left it running to see if I could correlate any more events.

Here is the output. Blue is the magnetic field intensity change (in uTesla) and green the magnetic deflection from "nominal" in degrees.

Obviously something happened on Sunday 17th but what were the big deflections on Monday 18th and Tuesday 19th. I had seen these previously but had put it down to the Solar pressure during daylight. But why hadn't it happened at the weekend?

Also what was the step in the descent and ascent?

I ruled out the central heating and the burglar alarm. Whilst I was pondering this I gave my wife a lift into town. When I got back there was a dip in the data.

It was detecting the cars on the drive from an upstairs bedroom. My wife usually goes to work first and returns first. Which explains the steps. So you can tell how many cars there are on the drive and probably which one.

Also I went to Tesco's on Sunday so although there may be some effects from the storm it wasn't as obvious as I first thought.

<Removed Word>

[Gina]

Magnificent attempt at a serious scientific project for which I congratulate you Seems that for a sensitive instrument like that you need a site well away from any road or anything containing iron or any other magnetic material and also anything electrical capable of generating a magnetic field. Unless you live in a rural area that is not likely to be possible. I sympathise that after all that work, and cost, your project is a non-starter - what a shame

[astronymonkey]

Hi Gkec,

Id have to disagree with Gina in that the detector is obviously detecting something which iooks more than just short tern magnetic disturbance from cars or cables. Magnetic fields are generated in electrical items and ferrous objects but the extent of their field is relatively small when compared with the earth's field.As long as they are 10-30 feet away then I dont think they would cause too much interference.

I built an aurora detector about 25 years ago but it used a bar magnet suspended in a 15 litre bell jar , which was "vaseline" sealed onto a glass sheet to prevent the ingress of drafts. Attached to the magnet was a mirror onto which was shone a slit lamp, and the reflection from the mirror was projected onto a scale about 20feet away. The idea was that the change in the magnetic field would move the magnet and move the reflected beam across the scale. As you can imagine it was a nightmare to manage and took about 2 hours to stabilise once set up !

In order to try to work out what you are seeing Id have a look at some of the online professional measurements of the earth magnetic field or have a look at something like this;

http://www.swpc.noaa.gov/ace/MAG_7d.html

Im not sure what detector you have or how it is orientated but as the earths field is in 3 dimensions it would be worth understanding which component it is measuring , only then would you be able to cross reference against the professional measurements to understand what it is measuring.

Nice project though, very interesting, took me back a few years though and your system is much more advanced than mine ever was

[gkec]

Thank you both Gina and astronymonkey, you have both given me things to think about.

Effectively what I am attempting is an electronic version on the bell jar, or as it is known these days the coke bottle experiment. The device measures the X and Y components from which I derive the magnitude and direction of the vector.

My next step is to effectively do what the suspended magnet does and introduce digital low pass frequency filters to filter out passing cars and other short lived events. Then move it further from the front of the house although I did notice that it was affected by the presence of next door's car parked on their drive as well.

I have been using the Aurora alerts and data from Aurora Watch at Lancaster Univesity as a guide http://aurorawatch.lancs.ac.uk/ but the NOAA link looks interesting.

[astronymonkey]

Hi Gkec,

Two other good liks below, o e is the wiki link which is pretty informative and the other is the intermagnet link but it doesnt show all stations unfortunately !

http://www.intermagnet.org/Welcom_e.php

http://en.wikipedia.org/wiki/Earth's_magnetic_field

cheers

[JamesF]

I'm actually quite impressed that you've built something sensitive enough to pick up a vehicle parked on the drive I thought the bell jar experiment was about vacuums and sound propagation though, so you've lost me there.

James

[AndyWB]

I seem to remember reading about the US Air Force fitting AC-130 gunships with magnetic anomaly detectors during the Vietnam war for detecting engine coils in trucks... ...so I reckon it might be the car

[IanL]

Interesting experiment. Since your measurements are affected by localised influences, would it make sense to have two independent sensors separated by as much distance as you can manage? I know nothing about measuring variations in the Earth's magnetic field, but it seems to me that the change should produce consistent measurements over a reasonably wide area, whereas any local effects would drop off reasonably quickly with distance.

Would probably take a bit of coding and some reading up on statistics, but it might well be possible to identify and remove effects that vary between the two sensors to just leave you with the net measurement across the pair. That would be nearer to the results you are trying to get?

[gkec]

I'm actually quite impressed that you've built something sensitive enough to pick up a vehicle parked on the drive I thought the bell jar experiment was about vacuums and sound propagation though, so you've lost me there.

James

You suspend a magnet in a dynamically static environment -- i.e. no air currents, it's doesn't have to be a vacuum. The magnet then moves as the magnetic field varies. By mounting a mirror on the magnet and shining a laser at it you can amplify the signal by having the reflection shine on a distant (say a few meters) flat surface.

[JamesF]

You suspend a magnet in a dynamically static environment -- i.e. no air currents, it's doesn't have to be a vacuum. The magnet then moves as the magnetic field varies. By mounting a mirror on the magnet and shining a laser at it you can amplify the signal by having the reflection shine on a distant (say a few meters) flat surface.

Ah, that makes perfect sense. Thank you.

James

[gkec]

I seem to remember reading about the US Air Force fitting AC-130 gunships with magnetic anomaly detectors during the Vietnam war for detecting engine coils in trucks... ...so I reckon it might be the car

The car is not running. It is just a lump of metal that has a magnetic field. The changes in detected field match when I and my wife leave and arrive home from work.

Running cars probably account for some of the spikes. I suspect engine coils might emit electromagnetic signals at particular frequencies which can be detected.

[gkec]

Interesting experiment. Since your measurements are affected by localised influences, would it make sense to have two independent sensors separated by as much distance as you can manage? I know nothing about measuring variations in the Earth's magnetic field, but it seems to me that the change should produce consistent measurements over a reasonably wide area, whereas any local effects would drop off reasonably quickly with distance.

Would probably take a bit of coding and some reading up on statistics, but it might well be possible to identify and remove effects that vary between the two sensors to just leave you with the net measurement across the pair. That would be nearer to the results you are trying to get?

Interesting idea. There are correlation techniques that work by transforming the data using a FFT into the frequency domain and identifying common frequencies. A bit like echo cancelling in reverse but without the attenuation or delay. Very hard to do in real time though. I think convolution is the technique come to think of it. Need to ask my friendly DSP engineer.

[skybadger]

Hi, maybe I missed it but I haven't seen a statement anywhere of what your detector is.

Speake instruments of llannfapley sell their fluxgate magnetometers for 20 or so each and they are.pretty easy to hook up. They are what aurorawatch.use as part of the SAMs network and the system drown is on the web too. Is this what you are using?

Cheers

Mike

[gajjer]

Ok, I get the fact that you are deflecting a laser beam, but how do you capture that deflection? Or are you nulling using Helmholz coils?

Very interesting project. Its a bit of a problem with the cars though, I've heard people have similar problems with seismic measurements. There is no easy answer

to that since its not easy to discriminate between signal and noise. You could detect the presence of a vehicle by the ignition coil or injectors but when its parked

you are stuffed.

Please keep at it, the flash of brilliance will come to you - probably while peeling the spuds!

cheers

gaj

[gkec]

Hi, maybe I missed it but I haven't seen a statement anywhere of what your detector is.

Speake instruments of llannfapley sell their fluxgate magnetometers for 20 or so each and they are.pretty easy to hook up. They are what aurorawatch.use as part of the SAMs network and the system drown is on the web too. Is this what you are using?

Cheers

Mike

I did see that device but I was unsure of the magnetic operating envelope and the electronic design was probably pushing my skills a bit.

I use a PNI Corp V2Xe device which has a built in microprocessor, a reasonably easy to implement digital interface and two field detectors at right angles to each other.

[gkec]

Ok, I get the fact that you are deflecting a laser beam, but how do you capture that deflection? Or are you nulling using Helmholz coils?

Very interesting project. Its a bit of a problem with the cars though, I've heard people have similar problems with seismic measurements. There is no easy answer

to that since its not easy to discriminate between signal and noise. You could detect the presence of a vehicle by the ignition coil or injectors but when its parked

you are stuffed.

Please keep at it, the flash of brilliance will come to you - probably while peeling the spuds!

cheers

gaj

When I said it was analogous to the coke bottle experiment I just meant it terms of what I'm measuring rather than the actual setup. No flashy lasers or helmholtz coils just a small circuit board.

I will keep thinking about it. Noise can be sorted, but like you say it is the parked cars that are the issue. I can't ban my wife or the neighbours from moving their cars when they want to!

[gkec]

Still not solved the car problem completely but it is up and running again for a trial period. Although my new (ish) Honda seems to be a lot less magnetic than my old Skoda.  

You can have a play here:

http://magnetics.no-ip.org:7681/

[Silveradogold]

Still not solved the car problem completely but it is up and running again for a trial period. Although my new (ish) Honda seems to be a lot less magnetic than my old Skoda.  

You can have a play here:

http://magnetics.no-ip.org:7681/

Hello gkec,

I am thinking you have a similiar, if not identical problem as the mariners of old (and possibly new) in that they had to contend with

their compasses being affected by the iron-hulled ships they were sailing on.

If my memory serves they used large masses of iron to counter-act the magnetic interference affecting their equipment.

One might want to follow up on how they solved this problem.

I think that those resolutions to that kind of problem may interest you, if not lead to a answer to the interference your detector deals with.

Clear Skies ----- Silveradogold