Sharp Shooters

On May 1, 2000, the President of the United States wrote, “Today, I am pleased to announce that the United States will stop the intentional degradation of the Global Positioning System (GPS) signals available to the public at midnight tonight. …This will mean that civilian users of GPS will be able to pinpoint locations up to ten times more accurately than they do now.” Now, far be it from me to question our Commander-in-Chief on military matters such as GPS accuracy, but just to be sure I decided this was a great opportunity to drag out the ol’ GPS-shootout bag and put a few marine GPS receivers to the test. What you’re about to read may be ground-breaking research, and it’s one of the first test of its kind to appear in online-boating press pages.

As you probably know, ever since GPS became fully operational we’ve had to deal with the scourge of Selective Availability (S/A), which was an error deliberately introduced into the pure GPS signals by the U.S. Government. S/A degraded GPS accuracy to a published accuracy of +/- 100 meters, but most of my testing showed regular GPS accuracy to be good to +/- 35 meters in the field. Regardless, in an effort to improve GPS accuracy in this country, several years ago the Government rolled out Differential GPS which is a system that sends GPS corrections to receivers from land-based radio towers, thereby taking out the error that S/A put in. About five years ago I conducted a DGPS vs. GPS shout and found that DGPS was, indeed, far more accurate than uncorrected GPS by a factor of about eight times. But now that S/A has been turned off, the big questions are how close are the two technologies to each other in terms of absolute accuracy, and is there any reason to buy a DGPS from here on out? Here’s what we found out.

For this test we pitted two DGPS units—Simrad’s CP32 and Northstar’s 951XD—against three regular GPS units: Garmin’s GPSMAP 230, Magellan’s Nav6510, Raytheon’s RC520. While all of these units have full charting capabilities, we concerned ourselves only with the unit’s latitude and longitude position readings, as well as its display of speed-over-ground both standing still and while in motion.

We took all five units out to a Class One U.S. Coast & Geodetic Survey monument named Station “ZIEGLER,” and its lat/lon coordinates are accurate to within +/- one inch! Once we were set up on location we fired up all the units and began recording each unit’s lat/lon readouts over a period of one hour. We conducted this test twice on two different days since weather may sometimes play a role in GPS accuracy. The first day was overcast and raining intermittently, while the second day came with fair weather and clear skies. Back in the office, we entered all of the results into the computer and averaged them, and the final results appear in this article. What we discovered was nothing short of amazing.

In contrast to some marine-electronics websites that are reporting accuracy of the “new” GPS to be +/- 30 meters, with only one exception all the units in our test were accurate to within +/- 5 meters (about 15 feet). And in the case of Simrad’s DGPS vs. the Garmin GPS, these two units finished in a dead-heat tie for first place with an average accuracy of +/- 2 meters (about 6 feet!). Now, I know what you math majors are thinking…if one reading was 100 feet to the left, and another reading was 100 feet to the right, the average would be a bull’s-eye right on the target. That’s true, but that also didn’t happen in this test, as you’ll see in the data ranges for each unit detailed below.

Garmin GPSMAP 230 GPS. Average error: 6 feet. Data Range: 12 feet. The GPSMAP 230 has an unusual large, vertically oriented monochrome screen that’s easy to read. From an accuracy point of view, this unit was the most impressive of the GPS group, with lat/lon readings centering around a 12-foot radius. It consistently showed a speed reading of 0.0 knots standing still, and its lag time during our acceleration tests was just 5.5 seconds from the time we hit the throttle to the time the unit was reading our speed accurately.

Magellan Nav6510 GPS. Average error: 30 feet. Data range: 120 feet. While these numbers seem unimpressive, they don’t tell the whole story. The 6510 was at a distinct disadvantage in this test since it only shows two decimal places for position, not three as in all the other units. That third decimal point can provide accuracy good to six feet, so without it the best the 6510 can ever do is round up or down to the closest .00 lat/lon for an accuracy of +/- 30 feet, and in fact, it did exactly that. Once it locked on, the 6510 provided rock-steady readings as close to the mark as it could. However, because of its daylight-readable, backlit color LCD display it turned out to be a power hog and failed on our portable battery pack while all the other units read fine. Against that, when wired into your electrical system, low power to the unit should not be a problem.

Northstar 951XD DGPS. Average error: 10 feet. Data range: 18 feet. Back in the loran days, Northstar made its name for precision accuracy, and it carries that reputation today in GPS. Surprisingly, the Northstar did not fare quite as well as some of the other units, and consistently gave us a latitude reading 12-18 feet north of our actual position. Plus, even with the differential beacon in use, the unit never “zeroed out,” but consistently showed a speed over ground of 0.1knots at rest while the rest of the field showed zero. Still, this is nit-picky in the grand scheme of things.

Simrad CP32 DGPS. Average error: 6 feet. Data range: 24 feet. Simrad’s color LCD DGPS unit automatically located and locked onto a differential GPS beacon in just under five minutes, and during that time it wavered quite a bit both in terms of position and speed. But once the unit settled down, it stayed down and its averaged readings matched Garmin’s exactly both in terms of latitude and longitude. Now that S/A is turned off, Simrad’s northeast representative Allen Schneider explained that most DGPS owners can turn their unit’s position-averaging features down to “0”, and when we did the unit had a speed lag time of just 5.5 seconds, so it proved to be highly accurate during both our static and dynamic tests.

Raytheon RC520 GPS. Average error: 10 feet. Data range: 18 feet. Raytheon also turned in a strong performance, with tight tolerances on latitude with a little wider variation on longitude. This unit acquired satellites quickly, but because it is designed to be interfaced with other units we had to wire its antenna into a junction box and power both the unit and antenna separately. In addition, the instructions included with the unit were inadequate and we had to make a call to tech support to wire the unit correctly. While this was a cumbersome process for testing, it should not be an issue aboard once the unit is installed, and with its big screen the RC520 was one of the easiest units to read.

So what did we learn? We learned that—at least during our tests—the “new” GPS is a heck of a lot better than the 10-meter accuracy the government describes. In fact, the accuracy is more like ten feet, not meters. So does this mean there’s no longer a need for differential GPS? It’s far too early to tell based on these tests, and what the government giveth, it can also take away. So my advice is that even with S/A turned off, if you’re running anything bigger than a 30-footer you should still consider DGPS as a very accurate and reliable source of positioning information. Yes, DGPS cost more, but my guess is that thanks to the removal of S/A the price of a DGPS receiver may come down in the short term. So in the end whether you go with regular or differential GPS, the demise of S/A is good news for all of us. And, oh by the way, thanks Mr. President. It looks like you were right.