Anyone done any research on high-accuracy magnetometers?

[alanambrose]

Hi,

Has anyone done any research they can share on high-accuracy magnetometers? Leads on specific devices, technologies, algorithms, geometry techniques etc would be great. This is for a position sensing application. Tests I've done so far on consumer grade magnetometers suggest a lot of adc noise...

Hints on some go-to devices would be great.

TIA, Alan

p.s. I guess I'm looking for single / multiple chip solutions rather than big physics e.g. SQUID set-ups.

[FriedLogic]

When I looked last, the most sensitive chip sized one that I could find was the TI DRV425, which is a single axis fluxgate sensor.

Bigger but more sensitive are the ones from stefan-mayer.com.
The last prices (Euros, ex VAT) that I have for them are over 2 years old, but were:
FLC 100 - 79,  FL1-100 - 360,  FLC3-70 - 499

The old Speake & Co fluxgate sensors seem to live on at fgsensors.com. I've not tried these, but I found that the original Speake ones were quite sensitive to things like temperature, supply voltage and loading, so they tended to be a bit of a project in themselves.

[alanambrose]

Thanks, that's very helpful and interesting.

Alan

[maxwell3e10]

You can look at GMR sensors from Honeywell, HMC1002. They can get performance better than 1 nT/sqrt(Hz).

[intabits]

This may be of interest:-
http://www.siliconchip.com.au/Issue/2018/December/An+incredibly+sensitive+Magnetometer+to+build

[AzzyAsi]

I have relatives that are into geophysics reasearch at a national institute. They use instruments from Danish Instruments and Bartington.
There are two types mainly used for the magnetic field of Earth:
Fluxgate type that are variomagnetometers (they measure acurately the variable field and delta but must be first calibrated to the constant field around usually by an absolute magnetometer nearby) https://www.bartington.com/catalogs/three-axis-fluxgate-magnetometers and https://www.space.dtu.dk/english/research/instruments_systems_methods/3-axis_fluxgate_magnetometer_model_fgm-fge

Proton precession magnetometers that measure the absolute field. www.geomag.bgs.ac.uk/documents/gdas2002complete.pdf

Usually you need a PPM to take an accurate reading of the absolute field and then you calibrate the FGE fluxgate vario-magnetometer to account for any minor change in the field with ultimate precission.

Equipment range from a few thousands $ to a few tens thousands $. The exact FGE they use is about 30k $

These are only for measuring constant fields of low value like the Earth with roughly 40 to 70 microTesla, and not for measuring a magnet. In fact around those measuring devices at the observatory there is a room deep into a mine in a mountain where there are no ferrous objects neraby.. only wood and copper nails, and you need to leave outside (about 500m away at the enterance of the mine) any metal you have on you (belt, cellphone, even clothes that have metal zippers, shoes with metal gromets, etc.. is that sensitive!)

For measuring large magnetic fields (magnets, coils, etc) there are other types that i havn't worked with or seen.

[alanambrose]

Just to report back for anyone in future going down this path:

>>> http://www.siliconchip.com.au/Issue/2018/December/An+incredibly+sensitive+Magnetometer+to+build

This is a slightly wacky design using mains transformers as pick-up coils and 4096UB hex inverters as amplifiers. Interesting though.

>>> Re: interesting parts

Here are the headline specs of some interesting parts (the first two suggested above) :

HMC1002 - 1 nT/sqrt(Hz) / +-2G full scale / 1 and 2-axis / magnetoresistive analogue

DRV425 - 1.5 nT/sqrt(Hz) / +-20G full scale / 1 axis / fluxgate analogue

RM3100 - 15 nT total rms noise / +-8G full scale / 1-2-3 axis / 'magneto-inductive' hall effect with packaged asic supporting I2C/SPI

MMC5883MA - 40 nT total rms noise / +-8G full scale / 3 axis / magnetoresistive supporting I2C

New (as yet unnamed) TDK part - 40nT total noise / ?G full scale / 3 axis / magnetoresistive with asic supporting ? protocols

https://www.tdk.com/corp/en/news_center/press/20180109_01.htm

>>> They use instruments from Danish Instruments and Bartington. / Equipment range from a few thousands $ to a few tens thousands $. The exact FGE they use is about 30k $

Interesting snippet thanks, but atm I'm looking for components atm to build into some prototypes.

- Looks like I need to dig into the datasheets to unravel the noise numbers properly...

Alan

p.s. the HMC1002 seems the most honest / forthcoming re its noise characteristics.

[guenthert]

Just to report back for anyone in future going down this path:

>>> http://www.siliconchip.com.au/Issue/2018/December/An+incredibly+sensitive+Magnetometer+to+build

This is a slightly wacky design using mains transformers as pick-up coils and 4096UB hex inverters as amplifiers. Interesting though.

[..]

Not to mention the *relais* (in a highly sensitive MAGNETometer) !?! 
"The advantage of a relay switch is that the magnetometer might be put to good use by switching things."  Riiight.

[ykurban]

According the datasheet HMC1002 has 27 micro gauss resolution, that's 2,7 nT

[alanambrose]

FYI the "new (as yet unnamed) TDK part - 40nT total noise / ?G full scale / 3 axis / magnetoresistive with asic supporting ? protocols" I think is the AKM AK09940 chip:

https://www.akm.com/akm/en/file/datasheet/AK09940.pdf

https://www.akm.com/akm/en/product/detail/0059/

- so +-12G full scale and 10nT resolution with i2c/spi.

Unlike the press release, the 'product brief' doesn't have noise data. I can see digikey have loaded it into their database so presumably coming in the next month or two.