How Do Digital Compasses in Watches Measure Direction?

[ffp]

Hi. I am curious to know how a digital compass in wristwatches works. How do they get the direction of the Earth magnetic field? I know the coils can measure the value,but how does they get the direction?

Casio's site explain like this:

"The magnetic field of the Earth is a complex network of north-south radiation. The direction sensor uses this field to accurately determine direction.

The sensor consists of two rectangular coils that detect slight tensions created by the Earth's magnetic field. Depending on the position and direction, these tensions vary in intensity. There is a non-magnetic resistance between the two coils, which absorbs the subtle changes in tension.

With each new position, the values that are measured by the resistance change. A microprocessor converts the measurements so that they can be displayed clearly on your watch."

Here is the link:
https://www.casio-europe.com/euro/products/watches/technology/digital-compass/

 

[f95toli]

Well, there are two of them (I would actually suspect there are 3 of them). If you know the measured flux for each of the 3 axis (which in the simplest case you can measure with 3 orthogonal coils) you can get the direction of the field with some simple vector algebra.

 

[anorlunda]

Here is a screen shot from my phone's magnetometer.

Look at the Wikipedia article on fluxgate. https://en.wikipedia.org/wiki/Fluxgate_compass

 

[ffp]

Well, there are two of them (I would actually suspect there are 3 of them). If you know the measured flux for each of the 3 axis (which in the simplest case you can measure with 3 orthogonal coils) you can get the direction of the field with some simple vector algebra.

Does the coils only get one axis component? I don't get how this calculation is done.

 

[anorlunda]

Here is a 3 axis chip used in mobile phones. https://www.akm.com/akm/en/product/detail/0059/
For 3 axes, I presume there are 3 sets of sensors.

 

[ffp]

@anorlunda, looks like it doesn't use Hall Effect nor fluxgate, but only two coils. I might be wrong, tough.

What I meant is that, ignoring the Z axis, since the watch must be parallel to the ground, both coils would take components of both X and Y axis at the same time. I don't get how, with the module of the field in 2 different coils, the watch is able to tell the field direction...

 

[anorlunda]

This is my guess. I never designed a watch.

There is a sensor, (pair of coils, or hall effect, or whatever) in each of the 3 axes, X Y and Z. Each sensor gives a single-axis measurements of field strength and polarity. The vector sum of the three measurements gives a 3D vector.

There are also 3 accelerometers, so the phone always knows it's orientation, which can also be expressed as a 3D vector.

Now we can calculate the projection of the magnetic vector on the orientation vector, which is compass direction.

I admit, the interpretation of a pointer is easiest when the phone is horizontal. But even if the phone was vertical, we could still do the calculation, but what to display on the screen? The programmer could choose to show the same screen as if the phone was horizontal, or show nothing if the phone is vertical, or to show a perspective view of a compass card.

The compass on my boat has a card like that. In theory, the card keeps it's orientation in 3D space no matter what gyrations the boat does in the waves. It looks really cool at night when illuminated. If it is really dark, that illuminated card is the only thing you can see and it stays fixed in space as you and the boat tumble all over. It would be really cool to have the phone compass app show a perspective like in this picture, and to have that image remain stationary as you twist the phone in all axes.

 

[davenn]

Here is a 3 axis chip used in mobile phones. https://www.akm.com/akm/en/product/detail/0059/
For 3 axes, I presume there are 3 sets of sensors.

yeah, 3 in the one package

@anorlunda, looks like it doesn't use Hall Effect nor fluxgate, but only two coils. I might be wrong, tough.

you are ... read the article and look at the block diagram

 

[Tom.G]

The watch probably implements a flux gate compass. i.e. a compass based on Flux Gate Magnetometers. Here is a reference with probably more information than you want, but the basics are on the first 7 (of 23) pages.
http://web.mit.edu/6.101/www/s2016/projects/farita_Project_Final_Report.pdf

Cheers,
Tom

 

[jim hardy]

""""two rectangular coils that detect slight tensions"""

mechanical or electrical tension?

 

[Lord Crc]

Many sensors rely on the magnetoresistive effect.

Here's one overview:
https://www.te.com/usa-en/industries/sensor-solutions/insights/intro-into-mr-sensor-applications.html

Wikipedia: https://www.wikipedia.org/wiki/Magnetoresistance

 

[jim hardy]

Honeywell application note

https://aerospace.honeywell.com/en/~/media/aerospace/files/technical-articles/applicationsofmagnetoresistivesensorsinnavigationsystems_ta.pdf

 

[dlgoff]

Honeywell application note

https://aerospace.honeywell.com/en/~/media/aerospace/files/technical-articles/applicationsofmagnetoresistivesensorsinnavigationsystems_ta.pdf

Very nice Jim. Thanks. I've never really looked into magnetoresistive (MR) sensors before.

 

[ffp]

Ok, I got how fluxgate works. This site explains very well: https://www.imperial.ac.uk/space-an.../fluxgate-magnetometers/how-a-fluxgate-works/

@jim hardy i had the same question, is it mechanical or electrical? and why there are 2 rectangular coils? Maybe the drive and sense coils? And why they are separated in the picture?

Maybe they are not using fluxgate (and it seems that it isn't magnetoresistence as well), maybe is some other technology developed by themselves...

 

[anorlunda]

Maybe they are not using fluxgate (and it seems that it isn't magnetoresistence as well), maybe is some other technology developed by themselves...

In a phone, miniaturization and power consumption would be high ranked considerations. Accuracy might be the least important property because most people do not use phone compasses to navigate planes and ships.

The consumer appeal of having a compass or magnetometer function in a phone seems to me to be slight. Perhaps others can educate me. Other than a compass, what is the value of a magnetometer in a phone?

 

[jim hardy]

@jim hardy i had the same question, is it mechanical or electrical? and why there are 2 rectangular coils? Maybe the drive and sense coils? And why they are separated in the picture?

i too was confused,
because two coils suggests a fluxgate but i couldn't see the toroid core for one in that Casio picture.

and the Honeywell appnote doesn't need coils for a magetoresistive compass

so I'm not sure what Casio has done.

Is that watch a metal detector too?
A Magnetometer is great to have when hunting in underbrush and muddy autumn leaves for old rusty iron survey stakes...

 

[anorlunda]

Yep, in our smart phones thread https://www.physicsforums.com/threa...-partially-replace.932077/page-6#post-5889280 metal detector is #78.

I know that passing under old bridges screwed up the fluxgate in my autopilot.

I just tested it. My phone would not detect a single Canadian coin, but it did detect a handful of pennies and quarters at a range of 2 cm.

 

[ffp]

i too was confused,
because two coils suggests a fluxgate but i couldn't see the toroid core for one in that Casio picture.

View attachment 234732

and the Honeywell appnote doesn't need coils for a magetoresistive compass

View attachment 234734

so I'm not sure what Casio has done.

Is that watch a metal detector too?
A Magnetometer is great to have when hunting in underbrush and muddy autumn leaves for old rusty iron survey stakes...

No, the watch only use it in a compass, no metal detector.

And looks like the 2 coils are separated. Aren't they those two cooper colored rectangles in the picture?

 

[Tom.G]

Of course they are rotated 90° from each other. The direction of the magnetic field is found by both the polarity of the sensed voltage and their relative amplitudes. If there was only one coil you would get no signal at all when it is oriented East-West, and to find North you would have to rotate the watch while looking for maximum signal strength. This way the watch does all the work!

Cheers,
Tom

 

[ffp]

Of course they are rotated 90° from each other. The direction of the magnetic field is found by both the polarity of the sensed voltage and their relative amplitudes. If there was only one coil you would get no signal at all when it is oriented East-West, and to find North you would have to rotate the watch while looking for maximum signal strength. This way the watch does all the work!

Cheers,
Tom

In your explanation, are you considering they work as fluxgate? Because from what I got reading about it, it doesn't have to be 90 degrees from each other. The coils should be one on top of the other, I guess...

 

[Tom.G]

it doesn't have to be 90 degrees from each other.

True, they do not have to be mounted 90° from each other; the important thing is they are oriented 90° to each other. The fact that Casio mounted them 90° around the periphery of the watch was very likely a mechanical consideration. If they placed the coils on top of each other the watch would have to be thicker to accommodate them.

Cheers,
Tom