Placement and Orientation
When determining an orientation and position for placing AMS chips on a board it is important to understand who the chip works and take into consideration where the field is that is trying to be measured. The reason this is so important is due to the fact that inside the chip there is a Wheatstone bridge (Figure 1) made up of magneto-resistive material, that is it changes resistivity as a function of applied magnetic field. Unfortunately, the properties of this material and bridge setup yield non-linear output as shown below in Figure 2.0.
From Figure 2.0 it can be determined that an IC chip setup in this configuration will be most accurate while operating within the linear region between -45 to 45 degrees, a full range of 90 degrees. Ideally we would like to cover a range of 180 degrees (note that 360 degrees is not required due to the fact that 180 to 360 is the same direction as 0 to 180 but negative). There are multiple ways this can be achieved, (1) two (or more) sensors could be used and placed at staggered orientations, or (2) multiple Wheatstone bridges could be incorporated into one IC chip. Both of these situations have been used and are widely available.
From the above description the assumption that for a single Wheatstone bridge IC the direction of the magnetic field can be most accurately measured when the direction is perpendicular to the Wheatstone bridge can be made and is illustrated below in Figure 3. The worst case scenario would then be when the magnetic field is parallel to the IC direction as shown in Figure 4.
With this in mind it is our task as group is to determine the best placement and orientation to place multiple sensors (9) such that any applied magnetic field at any angle can be accurately at consistently determined. Using the suggested number of 9 chips the following orientations seem most logical and are analyzed below with the assumption of a single Wheatstone bridge contained within the IC.
From these images it can be seen that in Figure 5 the magnetic field 1 will cut numerous IC axises near its optimal range (perpendicular) while magnetic field 2 does not cut any IC axises near optimal conditions. This has been fixed in Figure 6 by placing the chips from 0 to 180 degrees with a great spacing minimizing the number of IC axises optimally cut at any one position. Note that in Figure 6, any magnetic field will cut the same number of IC axises equally, a huge bonus in terms of consistency.
As mention previously, some chips have two Wheatstone bridges built in to improve accuracy. According to the data sheets these bridge are rotated 45 degree with respect to each other to improve the optimal range of -45 to +45 degrees to -90 to +90. Below Figure 7 trys to shows this but doesn't seem to make much sense as the two regions overlap and therefore do not double there range.
More research is required here to understand how a 45 degree shift in bridge orientation can double the optimal range for such a chip. From this point on let us assume the data sheet is correct and the new chips ranges is from -90 to +90 degrees, with one optimal axis at 0 degrees (like before) and a second a 90 degrees. as shown below in Figures 8 and 9.
Some of the chips we are investigating actually rotate the second bridge a full 90 degrees from the first making the two 90 degree regions cover a full 180 degree region. This also creates a chip with one optimal axis a 0 degree and a second at 90 degrees as shown above in Figures 8 and 9. The output of these bridges is shown in Figure 10. Note that in this setup the linear regions overlap and are not spaced out evenly as we would like. A better positioning scheme would be orientating the chips with a 45 degree offset between them or applying barber poles to effectively change the sensors optical angle. This will be discussed later.
The following is possible orientations and layouts for 9 chip combination. I have done this to show which layout places the chips in positions such that all magnetic fields will be nearly perpendicular to some chip under the assumption perpendicular magnetic fields yield the best result. Next to each layout I have also plotted the associated bridge voltage outputs that would result. This should help determine which layout is most uniform.
This is with each sensor rotated 10 degrees apart, spanning 90 degree section. All optimal axes are 10 degrees apart as well.
This is with each sensor rotated 20 degrees apart, spanning 180 degree section. All optimal axes are 10 degrees apart as well.
This is with each sensor rotated 30 degrees apart, spanning 270 degree section. All optimal axes are 30 degrees apart.
This is with each sensor rotated 40 degrees apart, spanning 90 degree section. All optimal axes are 10 degrees apart as well.