I used an old tablet to finally understand how GPS works

Last week, while visiting my favorite e-waste store, I found an old Motion Computing F5t tablet. A long time ago, I used to work with these chunky tablets, and I remembered a few models had the legendary u-blox 6 GPS receiver built in.

I paid the sticker price, a whopping $25, and took it home. I wanted to see whether this old tablet could provide more accurate location data than the software-based location sensors on modern tablets.

Soon, I had it sitting on a roof outside while I watched the u-center software over a Remote Desktop connection. Slowly but surely, satellite markers appeared, signal bars climbed, and I started poking around the data responsible for the position fix.

That left me with a question I finally got an answer to. GPS depends on precise timing, so how exactly can this old tablet figure out where it was just from listening to clocks in space?

I found a window into satellite navigation in the e-waste pile

The old u-blox receiver gave up the fascinating tech behind the dot on a map

The appeal of the F5t tablet wasn’t its ruggedness, the leaking LCD screen, or the ugly handle at the top. Rather, it was its GNSS sensor that I could investigate, which allowed me access to something navigation apps tend to hide.

Modern tablets can offer location services. Some even list a “location sensor” in Windows Device Manager, but they don’t have dedicated GPS hardware. Instead, these software-based location services can estimate its position from nearby Wi-Fi networks.

I wanted to see which satellites its receiver could pick up, and what information it produced before an app turned all that raw data into a boring but reassuring location marker.

I installed an app called VisualGPSView, which confirmed that the receiver’s NMEA data stream (a standard format for navigation data) was working. VisualGPSView is comprehensive enough, but I soon switched to u-blox’s own software, u-center, to get more detailed processed results like signal history and satellite positions to explore.

Tabs like Sky View immediately had me hooked. Individual satellites appeared at different elevations and directions, while other panels showed timing, signal, and maps, and even a camera view.

Now, there’s an important distinction to be made in the name before we go any further. GPS is the US satellite-navigation system, but it’s not the only one out there. GNSS, short for global navigation satellite system, is the broader term used to describe these constellations of navigation satellites:

A compatible receiver can combine multiple constellations, but my old u-blox wasn’t receiving from everything in the table above. Seeing the individual satellites started to make the system feel more real, but I still wanted to understand how their signals end up becoming coordinates.

A location fix starts with a receiver understanding how wrong its clock is

Time is a huge part of finding a precise position

While clicking through the u-center panels, I came across the clock-bias graph. It showed a green line sloping downward, jumping toward zero, and repeating over and over. Essentially, my tablet was showing me a timing error, and it’s something I’d never seen on Google Maps.

Understanding what the zigzag lines on the graph mean involves following how a satellite signal becomes a position.

Satellites broadcast signals

GPS satellites carry high-precision atomic clocks and broadcast patterns recognizable by receivers called ranging codes. Alongside these codes, they also broadcast navigation data that describes their orbits and clock corrections. The u-blox receiver on my tablet uses that information to calculate where the satellite was when it transmitted. It does this passively, meaning it just listens and doesn’t transmit anything back.

Receivers measure how late the signals arrive

The receiver generates matching codes and compares them with the incoming patterns. How far it has to shift its copy to line up with the incoming codes reveals a degree of delay. Since radio signals travel at the speed of light, that travel time is also translated into distance.

One microsecond corresponds to approximately 300 meters (~984 feet). Given that there are 1,000,000 microseconds in 1 second, it really gives some idea of the intense levels of precision involved.

Its own clock adds an error

The comparison then depends on the receiver’s clock, which isn’t perfectly synchronized with satellite time. This introduces the exact same timing error into every single distance measurement, making the satellites appear either nearer or farther away than they actually are.

It has four unknowns to solve

The receiver needs to determine its position in three dimensions (latitude, longitude, and altitude), and how far its clock differs from satellite time.

Each satellite supplies a single distance measurement, giving them one equation each. Because the receiver clock error is another unknown, at least four suitably placed satellites are required to get the three-dimensional fix.

The fourth satellite helps reveal the clock error

Here’s where it all comes together. With only three satellites and an incorrect receiver clock, the receiver can’t tell if the distance measurements mean it's in a different location, or if its time is just wrong.

That’s why a fourth satellite is required to give an extra distance measurement. It acts like a sort of tie-breaker because a fake location and time that seemed to work just fine for the first three satellites won’t work for the fourth.

This extra equation means the receiver can now pinpoint its exact position and fix its clock error using all four satellites at once.

Back on my graph, the bias approached minus one millisecond, and then jumped toward zero twice. That pattern showed me the internal timing adjustments by the u-blox receiver. Every recorded position remained valid through those jumps.

Essentially, the receiver had to account for its imperfect clock constantly, rather than just wait for each jump to correct its position.

The tablet stayed still, but its position wandered

Just because a fix is valid doesn’t mean it’s stable

Since I placed the tablet outside on a roof, I more or less expected its calculated positions to be accurate and tightly clustered together. They kind of did, relatively speaking, but they also spread into a surprisingly large cloud.

The hardware hadn’t moved, so why did its reported location move around so much?

My stationary recording captured:

690 valid position samples over 11 minutes and 29 seconds. Six to eight satellites used and no recording gaps. Between 49 feet of east-west spread and 32 feet north-south

Several things can affect measurement results such as these. Signals encounter delays when passing through the atmosphere, while reflections from surfaces like metal roofs and bodies of water can make them take a longer path to the receiver.

Satellite positions relative to each other can also have an effect. Measurements received from satellites spread across the sky typically constrain a better position than measurements from clustered satellites.

So, how does navigation software deal with this position uncertainty? Apps like Waze and Google Maps use context markers like map matching that fit a sequence of GPS points to the road you’re most likely traveling along.

That’s what keeps a driving marker on the road, even when individual position fixes are falling beside it. It’s a sort of informed estimate, or educated guess. And as anyone who’s ever used offline turn-by-turn navigation knows, it can sometimes pick the wrong road.

My own receiver’s scatter plot showed reported positions without that context-matching layer, and exposed navigation uncertainty that I’ve never seen in an app.

I came for coordinates and left thinking about clocks in space

In the world of satellite navigation, agreeing on time is everything

On my clunky tablet, I kept returning to the clock-status panel. Alongside the receiver’s clock bias and drift, it displayed an estimation of its timing accuracy. As it turned out, position was only a tiny part of what this little chip actually did.

That helps explain why satellite navigation is so important, even if the equipment never moves. Cellular networks like LTE and 5G use precise timing to coordinate transmissions. Electrical-grid monitoring systems need to synchronize measurements to compare events at a distance. Even financial systems require accurate timestamps to record transactions.

The very same infrastructure that helps me find my way back home on Google Maps also helps other systems when they need to agree on when something occurred.

Calling GPS “clocks in space” is an accurate take, but I’d now seen more of the machinery behind the use of that description. The u-blox receiver needed orbit information to locate the satellites, codes it recognized to measure delays, and calculations that could account for its own imperfect clock.

All the pieces go together to make that familiar dot appear on my navigation software.

Finding the old Motion Tablet had paid off exactly as I’d hoped. It has helped me learn about the complex mechanisms I take for granted every time I open a map on Android Auto or see how far away my Uber Eats driver is using my phone.

The next time I see that familiar location dot on a map, I’ll have a much better idea of just what it takes to put it there.