mag field sensors

Den fredag den 14. februar 2014 21.15.21 UTC+1 skrev John Larkin:

think this could work?

-Lasse

We put this is a hall sensor.

I used the above to map out the field in our lab space. It's a little drifty, but good enough for large changes.

George H.

Yeah, that looks too drifty for my app. Straight Hall effect is probably not good enough. Some GMR thing might work.

The sample rate is low, but it looks like the physics is there. Price is great.

The compass-type chips seem to have low readout rates, which is perfectly reasonable for a compass.

I'm tempted to design my own fluxgate, but I really shouldn't.

Den fredag den 14. februar 2014 22.11.32 UTC+1 skrev John Larkin:

they are cheap, maybe just use more of them?

I know it, one more for pile of interesting projects

-Lasse

I wonder what sort of sensitivity you could get using a stock unshielded ferrite-rod-core inductor? Surface-mount, even.

You can buy an official fluxgate core pretty cheap...

A colleague made a flux gate from two stock inductors, with a coil wrapped around both. I don't know other flux gates, but for this one it's a saturation effect and you've got to run the stock inductors kinda hot..

If you tune it up right you can wave it around and see the Earths field.

That's pretty cool.

Hey I know the scope trace at the bottom of this,

Well not the two little bumps on the ends. But two big peaks, in our two inductor one there is a larger asymmetry from side to side. (one slow and short, the other fast and tall.) This was driven by a voltage source, and I figured that beside the L*dI/dt term there was a I*dL/dt... So where the "inductance" is changing as it goes into or comes out of saturation there's a different "load".

Is that close to being correct?

George H.

Cool!

On a sunny day (Fri, 14 Feb 2014 17:14:13 -0800 (PST)) it happened George Herold wrote in :

My home made fluxgate, scope traces at bottom page:

The fluxgate is wound on the cut of top of a plastic bottle IIRC, filed some grooves in it. It is indeed simpler to buy one of them chips tha thave everything in it. ebay is full of them, I have one SPC01 module with temperature, air pressure, altitude, and magnetometer on it for the drone, it has a PIC on board to do the math, was 10$ or so. Not sure that one is still sold, i2c bus interface.

Hall Effect is a current hog. Try these people, they're experts in GMR

NVE Corporation (800) GMR-7141 (800) 467-7141

I don't know the area size you're thinking of but we've done a large VRS loop that worked very nicely for detecting magnetic or large metal objects disrupting the area.

It was a two coil system where the center coil was energized via a small DC current to generate a DC field and the outer coil was the variable reluctance pick up. This all works fine because you're only looking for a movement effect which this does.

Jamie

Interesting, but the output tracks the absolute value of the field. I'd have to bias it up, with an external field coil or PM, and operate on one of the side slopes.

We used some of the NVE digital isolators once, but they were buggy. They can get into a state where the output doesn't match the input.

The AMR (like Honeywell) sensors are bipolar and very sensitive, but need to be reset periodically with a roughly 1 amp pulse, to realign the domains or something. How often isn't clear. When something isn't clear on a datasheet, it's usually because they want to hide it.

Your description of how you plan to use them sounds rudimentary. Indeed that is one way to use them but not the most 'effective' way when planning to remove artifacts basic to all GMR sensors. Sounds like Honeywell already applied better ways to use the basic GMR structure.

Not familiar with NVE's isolators. What did NVE say about being buggy?

For what it's worth, GMR sensor's noise can't get below the 'johnson's noise' of the resistance, at 9k ohms, can be too much depending on what you're doing.

Take a four inch diameter coil with 100-400 turns on it, run the resulting current out to a 'rotating' transformer, then mount the coil on an old fan base, sit in the corner of the lab and measure field to your heart's content.

No drift, no BH curve characteristics, and can pretty much calibrate by using calculations.

Cool. I can get the bandwidth I need by spinning it about 200,000 RPM. Should be dramatic.

I want to measure ambient mag field along one axis. Is that rudimentary?

Indeed

The HMC1001 data sheet says that you have set and reset them regularly, with ampere-level pulses.

We didn't ask; we dropped in some Analog Devices parts.

At millivolts per gauss output levels, the Johnson noise won't be significant. But offsets and drift are high for most IC sensors. A fluxgate would be more expensive but low drift and low risk.

Huh?! You said 'maybe a few hundred Hertz' ? Obviously, you are proposing an extremely simple minded receiver. But, even with that why would you need to be so far above your Nyquist rate? 200 Hz times 2 [ok use 3] times

But this is a linear process. You can use your soundcard as [free] data acquisition and multiply the received signal by 1/cos(wt), since you KNOW the position of the coil, and block out a small bit in the 'dead zone' and get what you need at only 1000 rpm.

Or, use the rotating coil at 600 rpm [synchronous motor] and an adjacent fixed coil whose low frequency cutoff is set to coincide so over the whole bandwidth you get FLAT spectrum, then you can get that 1kHz you would have gotten using the 200,000 rpm fan.

more than one way to skin a cat.

I'm designing a product, not a one-off lab experiment. The idea is to place one to four field sensors in a lab, around an analytical instrument, measure ambient fields, and apply a compensation to the instrument to correct for its field sensitivity. The sensors need to be small, cheap, low power, and quiet. And preferably be made of all nonmagnetic parts, which things like motors won't be.

For a one-off experiment, I'd just buy a commercial gaussmeter of some sort.

ure, altitude, and magnetometer on it for the drone,

Jan That's quite an impressive device. I have several points about this area - the first relates to your fluxgate. The question is whether a "lock-in" approach can increase SNR. In this sce nario, the fluxgate is rotated periodically to induce a sinusoidal fluctuat ion in the signal. And this brings up an area of some mixed opinion. Some will say that you ha ve to 'intrinsically' vary the signal source. But others say not. Please al low to give some examples. The Foner magnetometer (c. 1955) used a loudspeaker or tuning fork to wiggl e a sample in a field. Then, another researcher would move a coil to and fr o a test object. Finally, the MRI wiggles the magnetic field to fluctuate t he NMR effect. True, these all modulate the source of the signal, but I cling to the idea that wiggling your fluxgate can have benefit. This is because you get away from 1/F noise, have a chopper effect on noise. It may facilitate also a si gnal averaging approach. Opinions?

1. The second point is I have a written DIY project for building a proton p recession device. I can't comment on it, since I don't remember it well, bu t somebody did it.

jb