That's because GPS is not directional. It's a positioning system. Direction can only be inferred as the difference between locations.
GPS fundamentally has no way of determining your direction. It was never designed for anything other than giving you a single fixed 3D location on the Earth's surface.
Rubbish: from first principles it is obvious that direction can be calculated. You are correct for phones, which have omnidirectional antennas that can't calculate direction (also built up areas with multipath would make direction finding difficult).
Marine navigation systems are the most familiar example. Many boat GPS units advertise “GPS compass” or “true heading at zero speed.” . The antennas contain two GNSS elements, fixed a known distance apart, and the direction is found using phase comparison: they look at the tiny difference in arrival time of the same satellite signal at the two separate antennas.
Either (a) don't answer if you don't know what you are talking about or better (b) double check your "facts" (using Google or AI) before answering.
You just basically confirmed what the previous commenter said. GPS is not directional. You need two of them placed some significant distance apart (think ship length, not smartphone length).
When you have two points determined with high enough precision you obviously have a direction and even orientation.
Your comment could have been just that. Useful information many people could find interesting. Instead you decided to sprinkle it with bile.
I do apologize for liking brutal facts - I am an engineer type.
I clicked the Furuno antenna since it looked sexy - blurb:
SC-130 features a Tri-sensor antenna that provides a high system accuracy for the heading of your vessel [for Autopilot, berthing etcetera.]. 0.25°rms
Fist principles: I think wavelength of L1 GPS frequency is about 0.2m - the same order as smartphone length. A ship length baseline using phase would need precise engineering and would not work for satellites absolutely perpendicular to the ship (making it less ideal).
Calculating phase difference only needs a single satellite visible, so I would guess directionality can be precise in built up areas even where location is harder to fix. although I'm assuming multipath can be filtered out.
SC-130 has distance between the antennas on the order of 1 meter and has three of them (which probably let's them be closer than in two antenna unit).
If you test phase shift to one satelite you might get good accurancy for some directions and no accurancy at all for the orthogonal directions.
So maybe not ship-length but still 1 meter sized device is way too large to put under the hood of a car let wlon in a smartphone.
Again, GPS has no provisions for orientation detection. You can hack it with few antennas significantly spaced apart (thank you again for pointing out those solutions). I think it might be easier to detect location with GPS and orientation by testing phase shift to the closest cell tower (when on land). There's nothing in GPS system thst helps with determining orientation.
And you still can't drop arrogance which gets tiresome.
Yes, I was being rude, but the OP and you are and choosing to write incorrect information. You in particular are redoubling your errors and arrogantly stating incorrect "facts". Here's a dual antenna direction finder smaller than a cellphone:
Vectornav VN-310 embedded. Size: 31 x 31 x 11 mm. Power:<1.6W Weight: <15g.
I had chosen the 1m Furuno because it was pretty, not because it was a great example.
I personally suspect that the people that designed the GPS system deeply cared about using the system to calculate heading information for missiles and fighters. Unfortunately it is hard to get good information about either the system design or missile antennas. I did a quick search for information about missile antennas and learnt some stuff but I didn't find anything juicy.
Some early and experimental GPS compasses used a rotating directional antenna. The idea was to use a patch or helix antenna that has a gain pattern. Spin it, watch signal strength from satellites rise and fall, and infer orientation. A few marine and military prototypes in the 1990s did this. Once dual-antenna phase comparison became practical, that approach died.
An Android App could combine body-blocking direction finding (also called human-body RF shadowing) with GPS satellite strength metering to create a crude GPS based direction finder. If you designed an RF shade on a rotating axis, it wouldn't be that difficult to get a mobile phone to accurately detect true direction from the satellite ephemeris data. A lovely hacking project. Would be more fun with a highly directional antenna and pointing in 3D to any part of the sky.
Disclosure: I'm no expert on any of this. I just like hacking.
Interesting, however I'm doubtful about how this unit achieves compass function. It's filled to the brim with other sensors. It has gyros, accelerometers and magnetometers. I suspect that GNSS-compass is more of a marketing term for this product that integrates input from all sensors and keeps it over time for resting direction estimate.
> however I'm doubtful about how this unit achieves compass function
There's your problem: why continue to deny simple facts when you're shown them so clearly? Why should anyone ever waste their time to learn you?
> It's filled to the brim with other sensors. It has gyros, accelerometers and magnetometers
I'm guessing you don't realize how small those surface mount components are. Yet you must know your phone also contains them. The first duckduckgo result for a gyroscope was the A3G4250D in an LGA-16 package 4x4x1.1 mm. I didn't look for smaller ;)
Perhaps your employment doesn't have much to do with engineering? Facts are so awkward. I despise them!
> What is (1m)?
I've found AI too be pretty good at answering similar questions - so long as you form your prompts well.
> There's your problem: why continue to deny simple facts when you're shown them so clearly?
Who's in denial? You just completely skipped commenting on clearly visible twin external antenna sockets in gnss-compass version of the unit and didn't correct your wild assumption that smartphone sized brick is all that is needed for this gnss-compass to run.
And what's that stuff about size of all other sensors? Everybody knows those don't take any space at all and that every smartphone has them. My point was that having these additional sensors in a unit that boasts inertial navigational capabilities might mean that direction is also mostly innertially tracked since the last time the sensor moved or even mostly faked with magnetometer.
I can make better guess what (1m) is than AI. It tells us that this an accurancy is achieved when you put the two external antennas 1 meter apart.
Read Wikipedia page on GPS system and try to find any single feature unique to this system (or even just not common to all radio signals) that aids with determining orientation of the reciever (while stationary). You'll find none.
GPS fundamentally has no way of determining your direction. It was never designed for anything other than giving you a single fixed 3D location on the Earth's surface.