Software alignment of magnetometers in a Gradiometer.

[pito]

pito

any 'virtual alignment' of a twin-fluxgate gradiometer.

It's simple, it is known that both frequencies should be the same if the magnetic field is not disturbed by the metal = rotating the gradiometer by 180 degrees does not affect the result.
In this case, the frequency difference is the amount of imbalance, we add it to the counter with a smaller value and the correction is done.

[Carl-NC]

Have you tried digging into the patent database? That's where the meatiest tech info can be found, it's just a painful digestion process.

[I M]

Here's the Bartington sensor(s) I believe they use in their 601:

https://www.bartington.com/products/gradiometers/grad-01-1000l-single-axis/

What interests me is "Interface allows directional and offset errors of the sensing elements to be nulled electronically". Unfortunately this is not explained in any detail but is what I want to know about!

Just to allay any concerns, I don't want to copy Bartington's products, rather I want to understand what they and others are doing from a fundamental perspective. I remain surprised that even eight years on from my last foray into this field the subject doesn't appear to be discussed much in the scientific literature that I've been able to locate...

[pito]

0 = Fs, another numbers = Δ F,
As I mentioned in my first reply, alignment does not have be perfect, rotating gradiometer will not cause change the error value because both sensors are rotated the same angle.
So you need 2 frequncy meters or 2 DC voltmeters programs.

[I M]

pito Thanks, this is more or less what I did eight or nine years ago (using a mixer) except it didn't have audio output; mine is a data-logging gradiometer for archeological research purposes.

It's likely I'll opensource the code/project, and I suppose audio could be introduced if you wanted it, but first I wish to gather some knowledge around the initial setup process, and particularly any 'virtual alignment' of a twin-fluxgate gradiometer.

[pito]

or make it simple for metal detecting, Δ F of your sensors,

[I M]

Carl-NC, thank you, I guess we must have been typing at the same time, but you've got to the nub of my query exactly.

In terms of dealing with the issue practically I agree that a three-axis sensor would allow corrections to be made on the fly. It may be that some of the commercial units do this, and I had thought about using an HC5883 in conjunction with the fluxgate's to do this, but I'm not absolutely convinced of the need for this .. yet I'd also considered the use of a MEMS gyro to act as a 'virtual gimbal', and it may be that these things would enhance operation, but I'm still trying to get to the source of what these guys are doing presently, at least in theory. I'd be keen to see this out in the open if it isn't already, hopefully this would contribute to the general body of knowledge around gradiometers and enhance future designer's understanding (and outcome!).

While typing this I recalled that at one stage I had some info on the sensors that Bartington were using with their gradiometer and I didn't think they were three-axis, but I'll have another look for that just to be sure and will report back.

[Altra]

Carl you are right. I also did the mechanical in addition. Thee screws to adjust

I built your version from a kit. Don't remember if I purchased from you or a guy in California, last name Kerns? Many years ago

[I M]

Apologies for my earlier short post - I had typed a lengthy response only to have most of it disappear on submission

Anyway .. I think I may not have made my original query very clear, sorry.

Presently I have no issues reading the sensor outputs directly. The Pi Pico (which is my first trial of this platform) seems to be doing a fine job of this task and is an improvement over the Arduino I'd previously used which could not handle more than one input of that frequency range (hence using the mixer).

I suspect our nomenclature may be different but also I don't see zeroing the difference between the sensors as alignment per se.

My interest lays in the theory of what's going on during the setup process of magnetic gradiometers such as the Bartington unit I described. My simplistic take is that they're using software to deal with anomalies introduced through imperfect mechanical alignment of the sensors along their axis. IOW 'virtual' alignment of the axial plane. It may be they're also assessing the direction + max and min of the local magnetic field. Either way it's not something I've found a great deal written about but would like to understand more.

Incidentally I'm not that enamoured of the FGM-3, something I gather from comments others here would agree with. I liked Bill, and he produced an inexpensive product that 'worked', but they have some issues, however before addressing those shortcomings it's this initial setup process I want to deal with, and I see it as something that is probably independent of the manufacturer of the (fluxgate) sensor?

Thus what I'm looking for is some [theoretical] detail on what each step of this initial setup process is actually doing. Some example code would be great as it would no doubt assist understanding, but perhaps it's all a big secret

Thanks again for responding, such interest in these things is good to see!

[Carl-NC]

The FGM sensors output a frequency proportional to the magnetic field. What I did was measured the frequency of each sensor. Then with a push button command subtract the two frequencies. The difference became the offset value subtracted from the sum of frequencies. This removes the alignment error between the sensors

This doesn't really remove the alignment error, it only removes the static error for the particular orientation and field you are in. Rotate the gradiometer 180° and there will likely be a new error if the sensors are not aligned.

I M, the only way I know to do software calibration is with 3-axis sensors. Suppose you have only the Speake sensors and they are slightly mis-aligned. So as you rotate the grad axially you will get a difference response. The question then becomes, how do you distinguish a rotational-misalignment response with a target response? The answer is, you have to be able to tell when the mag is being rotated. A 3-axis sensor will tell you that. You would then do a calibration procedure on all 3 axes to get them "software aligned." If the axes are equally sensitive, then theoretically this means that the sensors don't have to be aligned in any particular way, as you can software-correct for just about anything. However, it is easier to start with them closely aligned, and then do minor correction in software. And, often, one axis is designed to be far more sensitive than the other two so a decent static alignment is needed.

I suppose that single-axis sensors could be software-aligned by using a 3-axis accelerometer, but I expect the 3-axis mag sensor is a better way.

[I M]

Morning all, appreciate the responses

[Krzysztof]

When changing columns, I always recommend the "parallel" method, not the "zigzag" method.
Precisely because imperfect FGM sensors are very dependent and unstable in orientation/rotation/position.
The introduction of a constantly available "degaussing" solves all of these problems more effectively.

In addition, you should make sure to always hold the magnetometer with the same side to you and not to wear unnecessary iron.
I saw positive and negative (inverse) pulse degaussing before each measurement in integrated magnetometers with dimensions of 3mm x 3mm, used explicitly or embedded in the internal program - they used an additional coil above each measuring coil.
At this point, Ivica, an idea came to my mind: could an additional coil in the FGM led out of the factory be used for a similar reset (setting) of the sensor characteristics?
I'm attaching photos of the use of the Helmholtz coil.

[ivconic]

I did what you described at my magnetometer. And on real terrain for the purpose of amateur archeology, that solution proved to be more than good enough.
Because we, as "archaeologists" amateurs, are mainly looking for large anomalies, whose recognizable contours we can see on the recordings.
So the non-linearity and instability of the FGM sensor is not a big problem.
Of course; for professionals in geology, that measurement quality is not good enough.
But here we will limit ourselves only to amateur purposes. And for that the described method is good enough.
Specifically, in my area, we mostly use such devices to locate walls, entire houses, paved paths... covered with layers of soil that have been washed over by rivers over the centuries.
All of the above was built and made of ancient ceramics, baked clay, which has a large percentage of ferrooxide.
So the FGM-based magnetometer gives very clear and recognizable contours of artificial origin.
Another mass application is in the search for military, where it is mainly steel and iron, materials with extraordinary magnetic properties.
These are anomalies mostly measured in the range of several tens to hundreds of nT.
It is clear that for such purposes this approach is more than good.
...
I always tend to refer to a very useful book "Seeing beneath the soil" by Anthony Clark.
Everything I would say further on these topics is already well described in the book.
...
But Krzysztof is also right. Excessive manipulations with measured values in the code; just introduce bigger mistakes.

[Altra]

ivconic, I agree totally! The op asked for ideas. I never did any field test, but ended up buying a surveyors mag for field use.

[ivconic]