The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Mixed-Technology Design >> Current spinning in integrated Hall effect elements
https://designers-guide.org/forum/YaBB.pl?num=1214621909

Message started by thechopper on Jun 27th, 2008, 7:58pm

Title: Current spinning in integrated Hall effect elements
Post by thechopper on Jun 27th, 2008, 7:58pm

Hi all,

Let me open the fire in this new section. :)

Does anyone has experience on Hall effect sensors, and in particular with current spinning hall plates?
I'm particularly interested in state-of the art number of current phases that are used for reducing the hall element offset down to acceptable levels.

Any comments are more than welcome.

Regards
Tosei

Title: Re: Current spinning in integrated Hall effect elements
Post by Christoph M on Jun 30th, 2008, 3:40am

Buenos Dias thechopper,

interesting coincidence that you ask this question right now ...

I have a little experience with spinning current Hall plates, but it's been quite a while ... these days I don't deal much with choppers, but fly fixed wing aircraft around San Diego, BARET.... intersection and such.
It would be interesting to stuff a CMOS Hall sensor compass with low enough offset into the wing of an airplane.
I wonder whether somebody tries to do this these days.

Anyway. I'll use your post as an excuse to look at some references, to find out about the state of the published art.

You eliminate most of the offset of a Hall device by building a device symmetrical under 90 degree rotation, and by measuring the Hall voltage in two perpendicular directions by interchanging drive and sense terminals.

From there, it's relatively straightforward (mostly, you need a few more switches) to reverse the drive direction also, and average over current measurements in all 4 directions of symmetry.

S. Bellekom et al., "Offset Reduction of Hall plates in three different crystal planes", Transducers '97 Digest, p. 233-236,
P.J.A. Munter, "A low-offset spinning-current Hall plate", Sensors and Actuators, vol A21-23, pp. 743-746 (1990),
may be a reasonable places to start looking at the basics.  

Udo Ausserlechner, "Limits of offset cancellation by the principle of spininng current Hall probe",
Proceedings of IEEE Sensors, 2004.,  24-27 Oct. 2004 Page(s): 1117 - 1120 vol.3
[may be categorized under author Udo, A.]
appears to have run some simulations assuming nonlinear Hall plate resistance
(e.g., due to variations in the width of integrated Hall plate structures that are isolated from the substrate by a depletion layer)
and to have come to the conclusion that averaging over 2 perpendicular current directions reduces the offset by some 4 orders of magnitude, while going from 2 to 4 directions buys you about a factor of 2.

J.C. van der Meer et al., "CMOS quad spinning-current Hall sensor system for compass application"
Proceedings of IEEE Sensors, 2004., 24-27 Oct. 2004 Page(s): 1434 - 1437 vol.3
is a paper from Delft (Huijsing's group) that reports an array of 4 Hall sensors with 8 current directions.

You call yourself "thechopper", so I would assume that you already know this one:
A. Bakker, et al. "A CMOS nested-chopper instrumentation amplifier with 100-nV offset.",
IEEE J. Solid-State Circ., vol. 35, pp. 1877-1883, Dec. 2000.

Another nice overview is http://www.xensor.nl/pdffiles/art/orlando.pdf.

In theory, you can increase the number of directions of measurement ad nauseam, with diminishing returns.

Under the buzzword "continuous spinning current" you'll find a proposal how to spin the current literally, by superimposing two sinusoidal drive currents in quadrature on a 4-terminal device.
Works like a charm with precision laboratory sources;
the peculiarities of making the idea work on monolithic silicon, where mismatch and error propagation issues become important, have been described extensively by Archibald Putt.

You can probably make some headways along the lines of "continuous spinning current", by selecting a discrete number of current superpositions that can be realized by a small number of matched (maybe dynamically matched) sources.
A conference paper by
J. Bartholomeyczik et al. "Cartesian lattice spinning current method for the simple extraction of σxx- σyy and σ xy stresses and the Hall voltage from four-contact elements",
Transducers '05 Digest, vol.2, p. 2127-2130,
sketches the method. There may be an earlier patent about this, too.


It seems to be "state of the [published] art" to superimpose a known, time varying reference magnetic field on the field to be measured, such that the signals can be separated by the chopping/spinning scheme.
M. Pastre et al., "A Hall Sensor Analog Front End for Current Measurement With Continuous Gain Calibration",
IEEE Ssnsors journal, vol. 7, no. 5, May 2007, p. 860-867
reports such a system.

For the commercial state of the art, you may find something at
http://sensors-transducers.globalspec.com/Industrial-Directory/hall_sensor_application.html
and at
http://news.thomasnet.com/fullstory/540251


I've found most of the articles I quote on the full-access IEEE Xplore at the local university.

Hope that helps ...   8-)
CM

Title: Re: Current spinning in integrated Hall effect elements
Post by thechopper on Jun 30th, 2008, 8:29pm

Hello Christoph,

Thanks a lot for your thorough and comprehensive answer on this subject.
I was aware of only some of the references you cited, but certainly there is one which caught my attention, which is the one from Udo's. I knew the reduction of offset magnitude when using 2 directions was at least 10x ore even more. From what you say this reference suggested we are talking much more than that, and - this is good piece of info for me - going from 2 to 4 buys an extra 2x.

I will try to cover most of your suggested references. This is great help

Thanks again
Tosei

PS: yes...I knew about Bakker´s paper on nested chopper  ;)

The Designer's Guide Community Forum » Powered by YaBB 2.2.2!
YaBB © 2000-2008. All Rights Reserved.