SparkFun will be closed Monday 7/4 for the holiday. All orders placed after 2pm MT on Friday 7/1 will be shipped out next week.
The goal of this tutorial is to show you how to use GPS NMEA data and demonstrate some of the features and hardware one finds in embedded GPS modules/receivers. In addition, I will show how the features and hardware can affect tracking accuracy. This is not a buying guide per se, but it will help you get an idea of the differences, shown in Google Earth, between chipsets, antennas, and update rates for the six SparkFun GPS modules seen below.
I choose modules with different chipsets, antenna/hardware configurations, and update rates in order to get a general idea of the wide variety of features. All modules were left in their default states (except for the SUP500F which I had to configure for 10Hz update rate). The Copernicus and uMini use SMD GPS modules and are assembled by SparkFun. The other four are from outside suppliers.
|GPS Module||Chipset||Antenna||Update Rate|
|Copernicus||Trimble Trim Core||ceramic patch||1Hz|
|SUP500F||Venus634||small ceramic patch||10Hz|
|MN5010 (uMini)||SiRF Star III||chip||1Hz|
|LS23060||MediaTek MT3318||ceramic patch||5Hz|
|EM-406A||SiRF Star III||ceramic patch||
Refer to individual product pages and datasheets for the detailed technical information for each module.
Here is the bill of materials (BOM) for testing each GPS module:
The following are some issues that I had to consider when designing the test rig.
GPS modules are radio frequency receivers.
For reliable performance, your GPS module needs to receive good unobstructed signals from the satellite constellation. GPS modules have relatively tiny antennas to receive data from satellites moving at relativistic orbital speeds, over 10,000 miles away, and so they need somewhat of a clear view of the sky. I know this might sound really obvious, but I cannot stress this enough, because...
GPS sensitivity and accuracy is highly dependent on environmental conditions.
Not only does this include physical objects, like trees, earth, and structures, but also includes atmospheric effects, space weather, the overall health of the GPS constellation, and multipath interactions, all of which contribute to constant variability in receiving the GPS signal. To minimize errors due to these environmental effects, I wanted to test and track all six modules at the same time with varying views of the sky.
What I needed was a mobile logger for each GPS module. SparkFun stocks more than enough supplies to accomplish this.
Tracking Setup. Note: The uMini was oriented vertically for the test and is not shown that way in the picture
Connection Block Diagram
Each GPS module should have its own clean (regulated and decoupled) DC power supply.
The GPS power supply feeds into the antenna power supply on some of the modules, so any excessive noise or unexpected loading on the circuit can affect performance. Attaching a high capacity LiPo battery to a regulated Breadboard Power Supply Stick for the power supply for each module is sufficient.
Next, I capture the data coming out of the GPS module with the openlog.
All that needs to be done is to connect the transmit pin (TX) from the GPS module to the receive (RX) pin on the openlog. The openlog will log anything it sees on that pin, just be sure to configure the baud rate on the openlog to match the baud rate coming out of the GPS module.
A brief word or two on the VBATT pin and the TTFF (time to first fix).
What is the VBATT pin used for? Some embedded GPS modules contain volatile RAM that is used to store the almanac and ephemeris data of the satellites, which is globally updated on a regular basis. Think of the almanac data as a "bus schedule" of when and where each satellite will be in the sky. The ephemeris data contains the errors in the schedule. If your GPS receiver does not have the schedule or the errors of the satellites, it will have to download that information when there are updates. Since the RAM stores this data, the VBATT pin should be powered or connected to some kind of energy supply (i.e. supply voltage, battery, or supercapcitor), so that your GPS module can quickly start-up. If you don't power the RAM, all of the almanac and ephemeris data is lost with each power cycle.
The TTFF (time to first fix) is determined by how quickly the GPS module can access the almanac and ephemeris data. If the module doesn't have any of this data (some modules don't have this data out-of-the-box) the module might take a bit of time to power up. Same goes for if you don't connect power to the VBATT (RAM power) pin, you will lose all of the data upon a power cycle. If the almanac and ephemeris data can be permanently stored in flash, like the SUP500F can, then you should get a really fast TTFF. Some modules, like the EM406, have a supercapacitor attached to the RAM. This means you will get really quick TTFF for only a period of time after you power the module down. Wait too long, the capacitor will discharge and you will lose your information. The reason for the choice in using a supercap over a battery or flash chip, is that it is in expensive, small, and the almanac and ephemeris data needs to be updated regularly anyway, so if your position is changing a great deal over the globe, you don't need to keep all of the same almanac data.
In terms of the TTFF performance for each of these modules, all have similar stated fix times that are under 1 minute, but will differ with environmental conditions and antenna hardware.
Download Google Earth to view the KMZ files and zoom around my track in 3D (be sure to turn on 3D buildings). Just click the link and Google Earth should open and zoom to the the SparkFun building.
For my track, I decided to walk the GPS modules around the SparkFun building, instead of testing them in a car. GPS modules, to an extent, seem to work better the faster they move, so walking the units around the building should create a test condition where the GPS modules will have a greater chance to show errors.
Actual track walked (top is northish).
Here is a brief summary of my track: I turned all of the module's power ON next to a window on the second floor of the SparkFun building. I waited a few minutes, then made my way outside to the north most point of the parking lot, where I waited to make sure all of the modules had a lock before I proceeded south around the building and back to where I started. For more info on the outside track refer to the 2010 Autonmouns Vehicle Course Preview.
What does it mean when the receiver has a lock or a fix?
When the receiver has a lock, you get accurate position (3D) and time data. This is accomplished when the receiver sees at least four satellites in good view. Why four? Well, I'll give you hand waving explanation, but the full explanation is well beyond the scope of this tutorial and most users. For a thorough explanation, there is a two volume text book written on the subject. Here is the first volume, free to view and download. Thanks Dan!
You need 4 satellites for 3D trilateration and 3 for 2D trilateration, no need to consider time [see wikipedia]. GPS uses time to measure distance to satellites though.
The 2D case is easiest: if you know the distance to a satellite (of known position) you know your position is on a circle with the satellite in the center and the radius equal to the distance. Two satellites gives two circles, your position will be on the intersection which is usually 2 points. A third satellite will give you a single point. In the 3D case one satellite gives a sphere, two a circle, three two points and four a single point."
Overall, getting accurate time is not an easy task for satellites and receivers alike. The GPS satellites need really, I mean really, accurate timing. So accurate, that there are actually atomic clocks on board the satellites in order to keep them in sync. Since the speeds and distances are so great between satellites and receivers, relativistic offsets are programmed into the satellites position and time data. This information, along with additional correction data the satellites receive from permanent ground based stations (this is how WAAS works), turns your GPS receiver into an amazingly accurate and powerful little device with, on average, five meter accuracy.
RED = Copernicus, BLUE = uMini, GREEN = EM406, ORANGE = D2523, LIGHT BLUE = SUP500, GREEN-BLUE = LS23060, YELLOW = actual path
*North is at the top*
This view only shows latitude and longitude data (no elevation) and I only am showing the tracks after I waited in clear view of the sky to make sure all of the modules had a lock. Notice that all of the tracks in clear view of the sky (to the north and south) are decently accurate; within the claimed 1-5 meter accuracy for C/A code receivers.
There is about a 150 foot span in between the ~25 foot tall SparkFun building and our neighbors ~25 foot tall building to the west that creates a small 'urban canyon' where most of the modules had some expected trouble. In addition, there is tree cover in part of the SparkFun urban canyon, which also hindered the accuracy of the GPS modules.
Notice how the SUP500F's (light blue) track is not a straight as the others, this is because the position is being updated at 10 times a second compared to one time a second for most of the other receivers. Modules with faster update rates usually work better within faster moving objects.
This is a 3D view looking at the west side of the building. Looks like the D2523 (orange) had a hiccup in the urban canyon. Also notice the uMini (blue) is not performing as well as the others. This is expected since the uMini is using a passive chip antenna. And wait...I only see five tracks, where is the sixth? The LS23060 module's elevation data is below what Google Earth calls ground height. It is there, just not visible in this view.
Don't try to use GPS inside!
Wow, what is that? Those are the tracks of the GPS modules trying to find their location inside of the building when I first turned the units on. This is why you don't rely on GPS inside of a structure.
How to create files in Google Earth
Overlaying your tracks in Google Earth is really easy. After you have logged your positions to the openlog, remove the SD card and plug it into a SD card reader on your computer. Look at the contents on the card. You should see a config file and at least one log file (if you cycled power during logging, new log files would have been created). Open the log file with a text editor, like Notepad in Windows. You should see a bunch of text in the form of NMEA sentences or something in a similar format. The NMEA sentences contain, among other data, your position and time.
Here are my raw text files:
All you need to do now is convert the file to a Google Earth file with a kml or kmz file extension. The website I use to convert the text file to a Google Earth file is GPSVisualizer. If you click on the Google Earth KML link in the middle of the homepage you can configure settings.
The settings I used are shown above. Once you have loaded your data into this page, hit the Create KML file. You will then be presented with a link. You can either click it and, if you have Google Earth installed, it will zoom to your tracks or you can right click and save as to save for future use or to show your friends.
If your txt files are over 3MB, you will not be able to use GPSVisualizer. GPSBabel is another site you can try if you have really big log files.
This is by no means an "end-all-be-all" summarization, it is merely a comparison on how well 6 very different GPS modules performed at the SparkFun location (not in any kind of order).
MN5010 (uMini) Pros
MN5010 (uMini) Cons
We now have a pretty good idea of the general performance characteristics for some of SparkFun's GPS modules, however this is obviously not all inclusive, in terms of the performance capabilities for other GPS modules with similar features out on the market today. GPS modules with combinations of the same chipset, hardware, antennas, and configurations all will have different performance characteristics in different environmental situations. All have their benefits and costs, but it will be up to you to take the data points I have found to help choose the correct GPS module for your project.
If you are interested, here are some random tracks I took preparing for this tutorial.