homemade gaussmeter

Sorry for asking to be emailed directly, guess that was a big faux pas! Consider my wrist officially slapped. I understand that detecting microvolt changes in DC is difficult but did not think it was impossible. The Hall device I have is rated at 2.5mV/G. I could go with a Hall Device that is rated at 5mV/G but I'm not sure that will help much.

Anyone with any ideas?

This may be a simple approach, but it isn't elegant.

What is really required is a simple Hall effect sensor with four electrodes, one on each of all four sides of a square plate of semiconductor material. You drive a current across the square between one pair of electrodes, and measure the Hall voltage generated at right angles to the current path across the second set of electrodes.

Drive an alternating current across the plate, and you can synchronously detect an alternating voltage across the sense electrodes, giving you your chopping at source.

The Hall effect is temperature dependent, so for added extra sex-appeal, measure the resistance in the direction of current flow to make the plate its own resistance thermometer. This wasn't patentable back in 1991, so it isn't patentable now, and any patent that might have existed back in 1991 will probably have expired by now.

Getting hold of such a Hall plate might not be easy - Siemens (now Infineon) used to sell a lovely - if rather expensive - part for a few hundred dollars each, but it had dropped out of their catalogue when I last looked.

------------- Bill Sloman, Nijmegen

Thanks everyone. Those are great ideas. I think I may have gotten in a bit over my head on this one. It sounds like a fun project to figure out but I think it's well above my level of knowledge right now. I'll probably play with it sometime in the near future when I think I can handle it! Thanks again all!

I read in sci.electronics.design that snipped-for-privacy@gmail.com wrote (in ) about 'homemade gaussmeter', on Thu, 10 Nov 2005:

No, you post here, you are answered here.

These Hall things have sensitivities of millivolts per gauss. good luck with detecting milligauss changes, giving microvolt changes of d.c.!

I read in sci.electronics.design that " snipped-for-privacy@gmail.com" wrote (in ) about 'homemade gaussmeter', on Thu, 10 Nov 2005:

The simplest way is probably to connect your sensor to a chopper op-amp. Chop at audio frequency. That gets you an a.c. signal, which is still microvolts but is much easier to amplify. Then amplify it to some useful level (200 mV?) to work your meter. Restrict the bandwidth of your amplifier to cover just a narrow band around the chop frequency.

Also consider using a pair of devices 'back to back' in a bridge arrangement to cancel out offset effects - IIRC these sensors are temp dependent. You need to avoid self-heating effects, and also need to watch out for 'thermocouple' effects at the joints. Microvolts are hard to measure !

Your life will be much easier if you can modulate the magnetic field you are sensing, by (for inst) rotating source or sensor assembly - then you can use AC gain and synchronous detection, as above.

Dave

On Thu, 10 Nov 2005 17:18:26 +0000, John Woodgate wroth:

The proper solution would be to concentrate the magnetic field *before* it gets to the sensor.

Imagine a pair of ordinary kitchen funnels, but instead of being hollow and made from plastic or glass, they're solid in cross section and made from iron or a ferrite material. The two funnels are placed with the small ends together and just far enough apart for the sensor to be placed between them. The large ends have an area much larger than the small ends, so the magnetic lines of force will be concentrated in proportion the ratio of the areas. You get amplification of the field that makes the sensor output its self greater with no electronics needed. A thousand to one should be easily obtained.

Jim

Why not use a magnetoresistive device? They don't cost much and are way more sensitive.

I read in sci.electronics.design that Ken Smith wrote (in ) about 'homemade gaussmeter', on Fri, 11 Nov 2005:

So is an inductor on a ferrite bobbin; the 33 mH and 47 mH types make excellent magnetic receiving antennas. But the OP wanted to use his Hall device.

selection.

Would this chopper technique really work given that these Hall devices have a noise specification typically 10 mV p-p or even more?

With some, like the Allegro A1301 the noise spec is 150 uV. With many others, the noise performance is not even mentioned on the datasheet.

Regards,

Euthymios Kappos

In article , John Woodgate wrote: [... me ...]

Yes, but people in hell *want* ice water.

If the Hall device won't do the job, he should look for something that will. I don't think he's going to get mG out of a hall device unless he averages for a long time.