I started with very basic function, showing bearing and heading angle from a digital magnetometer on an LCD, but there are a number of digital and analog pins are left unused on the microcontroller (Arduino compatible, JeonLab mini) so I added a digital temperature sensor and a phototransistor to adjust the brightness of the backlight (single LED) of the LCD.
I still have more than a few pins available and have a plan to add more sensors such as an exterior thermometer, anti-theft device, etc. (any suggestion would be appreciated!)
The picture shown in this Intro page is the fully assembled and mounted on top of the interior mirror of my car showing the bearing, angle and current interior temperature. Unfortunately I didn't take a lot of pictures when I assembled the electronic parts, but I'm sure someone who are interested in making one of these will be able to follow easily the schematic diagram I will show in a couple of steps later.
Step 1: Part list
- LCD: 16x2 HD44780 LCD (white text on blue background): This is pretty popular LCD that you can find and buy easily.
- JeonLab mini v1.3: This is a minimalist Arduino board I made. There are a number of small Arduino compatible boards in the market so pick anything small that can be hidden behind the LCD.
- Digital compass: MAG3110 (bought an assembled one from ebay): I have tried other digital compass a couple years ago but couldn't make it work. But this one is relatively easy to program. I will explain this in detail later.
- Temperature sensor: TMP36: take a look at the datasheet from the Digikey link. It's easy to use.
- Phototransistor: LTR-4206E: I'm sure any other IR phototransitor will do the same.
- 7805 regulator: You know what this is. Car battery's charging and regulated voltage is 13.8-14.4V while we need 5V for the controller and the LCD.
- resistors and capacitors (see schematic diagram)
- Car battery jack
- prototyping board
- push button switch (normal open, N.O.) for calibration of the magnetometer
- solid copper wire (1mm in diameter) for bracket frame
- cable ties
Step 2: Schematic diagram
The LCD connection is well explained with a library in Arduino tutorial page. I just changed the pins for my convenience. Note that you need to change these pins accordingly in the sketch. One important thing here is the backlight LED Anode pin is not hooked up to the V but to the digital pin 5 which provides PWM with which you can control the brightness of the backlight of the LCD.
The temperature sensor, TMP36 has 3 legs, V , signal, and GND. I connected the signal leg to one of those analog input pins of the JeonLab mini. Refer to the Arduino sketch (program) in the next step for how to calculate the temperature from it.
The phototransistor has two legs, positive and negative. You can omit the resistor between the negative and GND which I did as long as you get the values correctly when it's dark (I tested in dawn) and bright (you don't have to measure this cause it's already maximum of the V ). The brightness measured by this phototransistor is used to adjust the brightness of the backlight LED of the LCD as explained above.
Step 3: Preliminary tests and heading calculation
The magnetic field strength and direction is differ from location to location. In order for accurate compass and map reading, you need to know the magnetic declination (difference between the compass north and the geographic true north). In my sketch, I didn't subtract or add the declination, but if your location has very large declination, you may want to add this to compensate the difference.
By the preliminary tests, I found the value shows the maximum and minimum values of X and Y axis represented north and south. But the values did not evenly distributed in angle. In other words, the central value didn't represent east nor west. This is common, as far as I know, for all the semiconductor type magnetometer as long as they don't have any built-in compensation algorithm in it. In order to read heading (roughly) correctly I figured that if you know the values for north and south per each axis, X and Y, simply calculating ATAN of differences of current reading and averages of X and Y will give you heading angle (see Arduino sketch in the next step), and it worked fine. This is not perfect method to calculate the heading but we are not talking about decimal number precision. I have driven my car with this compass for about a week now and the results are quite satisfactory.
Step 4: Arduino sketch (updated)
UPDATED (April 24, 2013):
1. I couldn't upload Arduino sketch directly (weird, I had uploaded before...), so uploaded a compressed file.
2. I added the temperature in Fahrenheit as well.
3. In order to test the brightness values, added the analogRead(1) value and the mapped PWM value in the 2nd row.