Hoenywell's IC Comparison
This is a comparison of magnetic sensing ICs offered by Honeywell.
|HMC1042l||2||~1 degree||± 6 Gauss||Signed Analog voltage|
|HMC105x||1-3||~1 degree||± 6 Gauss||Signed Analog voltage|
|HMC6352||2||~2.5 degrees||0.10< x <0.75 Gauss||I²C|
 How AMR IC's Work
Since the project was proposed to us by Dr. Peczalski, there was no previous work for this specific project. To the best of the group’s knowledge and research, no one has done any research on using multiple nanowires to increase the accuracy over a single nanowire. Furthermore, to the best of the group’s knowledge no other devices current use multiple IC chips in various orientations to increase accuracy. However, within some IC sensors multiple magnetoresistive wheatstone bridges are used to broaden overall range and in some case increase accuracy. The extent of this while remaining coplanar has been limited to two bridges and therefore has not yet been fully explored as this project plans to do.
It is important, however, to have an understanding of how other digital compasses and similar navigation devices have been designed. Due to the fact that anisotropic magnetoresistive (AMR) sensors will be used for this test of concept the scope of this description will therefore, be limited to AMR sensors and those which use them.
AMR sensors measure the orientation of a magnetic field by means of a differential voltage that is produced by a wheatstone bridge configuration of four magnetoresistive elements. The wheatstone bridge is shown below in Figure 1.0, where the pink strips are the magnetoresistive elements, and the voltage measured across the center is the differential voltage produced.
Figure 1.0: AMR Wheatstone Bridge
This configuration works to create a differential voltage that varies with respect to the direction of the applied magnetic field due to that fact that the resistive properties of the individual magnetoresistive elements change resulting in different voltage drops around the bridge. The output differential voltage as a function of the applied magnetic field direction with respect to the bridge orientation is shown below in Figure 2.0.
Figure 2.0: AMR Wheatstone Bridge Voltage Output
From Figure 2.0 above there are a few points of interest. First, it is important to note that there is a window from -45 degrees to +45 degrees before all output values are repeated. This means there is only a 90 degree window for which a magnetic field direction can be determined using only one bridge. It is also important to note that this function is not linear. This means that at some positions a small change in magnetic field direction will cause a large voltage change than at others. Obviously, where the voltage change is greatest for the smallest direction change the more accurate the sensor will be, this happens during the linear region shown from roughly -35 degrees to +35 degrees. These items of interest need to be taken into account when formulating a working compass and compensated for.
Other designs have compensated for the limited 90 degree range by introducing multiple bridges and Hall Effect sensors to specify which quadrant or hemisphere the measured magnetic field is within. As a bonus, this also easily increases accuracy by creating more linear regions which can be staggered to create effectively one continual linear range. Figure 3.0 below shows the plot of a device using two bridges positioned 90 degrees apart with an addition of a Hall Effect sensor.
Figure 3.0: Dual Bridge with Hall Effect Sensor
Using Figure 3.0 above it can be seen that whenever sensor A or B is outside its linear region the other sensor has entered it linear region, always supplying a very accurate voltage reading. Furthermore, using the different signs from each of the bridge output and the hall effect sensor the resulting direction can be placed within the correct quadrant extending the range to a full 360 degrees.
To further increase the complexity of using these devices is the fact that the output voltages produced are fairly low, less than 1 volt. Therefore, amplification is generally required to allow for sufficient accuracy. This often posses an additional challenge because the amplifiers need to be tuned with great precision. Fortunately, this has been perfected and there are many devices available that will perform the amplification needed and many that also provide a digital output that can easily be interfaced to a microcontroller for processing.
The bulk of this design revolves around the AMR IC sensor selected. Andy has given us the choice to use any of three sensors made by Honeywell; the HMC1042l, the HMC105x series or the HMC6352. After many in-depth data sheet reviews and debates were placed over the conclusion was made to go with the HMC6352.
One of the main reasons the HMC6352 was selected by the team was its I2C capabilities. I2C is a communication protocol that allows you to communicate to multiple devices using their addresses as a way to distinguish one from another. In the case of the HMC6352, each chip can calculate its own heading, voltage readings, or (X Y) coordinate. This will allow us to greatly simplify the circuit. Without these capabilities the team would be forced to have many analog to digital conversions along with several operational amplifiers. The processing capabilities far outweigh its only foreseeable down side, its field range of measurement. It field range can measure B fields anywhere from .10 - .75 gauss. This is fine for applications such as navigation systems because the earth’s magnetic field around this region (upper Midwest) is around .6 gauss. The only foreseeable problem is getting a constant B field to test in. If the team were using another sensor that can measure fields at ± 6 gauss the team could cancel out any of the earth’s magnetic field with a stronger artificial B field created by winding coils.
The last reason that the HMC6352 was chosen was that its accuracy is a little less than the other available sensors. The HMC6352 is specs say that it is accurate to within 2.5 degrees. This is less accurate then the other sensors that are accurate within 1 degree. This is good because it will allow the team to see more definitive results. Although the team has a highly accurate rotational table that will allow for us to measure precision, other resources like a highly accurate directional gauss meter might be needed to see significant improvement in accuracy. With less accuracy the team will be able to answer the general question about using multiple sensors more definitively. With everything taken into account the team settled on picking the HMC6352 as the sensor that would be used in the final circuit.