Prototype GPS Speedometer DYI for around 35$

A quick project for tracking basic data during flight. This is a garage project I made out of own curiosity. Not tested in flight yet, if you decide to do it you're doing it at your own risk.

There are advanced solutions on the market like Dekunu One or Aon X2 and if you can afford them I encourage you to use professionally made equipment.

You'd need:

* Car GPS speedometer (I used this one because its shape and size are already almost snag-free: https://www.aliexpress.com/item/4000265066382.html

* Powerbank (with least capacity and smallest size possible - in the prototype this one was used: https://www.aliexpress.com/item/32945597181.html )

* 850 mAh battery (ideally 2000 mAh, but unlikely they'll fit inside the case)

* Small switch

Assembly is pretty straightforward: unsolder USB ports on both GPS and powerbank, solder switch and battery. Additionally I flipped the capacitor (it's under white tape on the first photo) - put it on the opposite site of the board to give some room for powerbank's charger.

The device is mounted on a rig using velcro strap with a loop.

So far I tested it on the ground on a bike and the speed is pretty accurate - +/- 1 km/h to what smartphone GPS shows

Even with 850 mAh battery the device help up for around 8-10 hours (screen constantly on) which is quite impressed actually.

Talking about the battery - attentive reader might notice that it got swollen - probably because powerbank's changer is for 2000 mAh, while battery is 850 mAh. That's a critical flaw because swollen battery has high risk of explosion as pressure changes during climb to the altitude. If I ever decide to use it for actual tests, the batteries will be replaced with non-chargeable lithium batteries just like in audibles (powerbank's module would still nicely convert that into a constant 5V supply). Meanwhile I'm doing some more research to find safe chargeable batteries.

Now, what this thing can do:

* Display current ground speed (both metric and those other, inferior units). Speed alarm can be set at as high as 200 km/h, so I'd assume device can measure at lease that much

* Display one of 8 directions ( N/NE/E/SE and so on) - yeah, it's pretty basic, but better than nothing

* Display distance covered, accuracy 100m

* Display few other stats, but those are useless for WS flying

* Oh, yeah, you can change colors

Again, this is a garage project. I'm only sharing the fruits of my curiosity and taking no responsibility for what you might do with it.

Quick update: battery has been replaced with other, more reliable and I was able to test it today.

Long story short - the unit didn't work as well as I'd expected: it was perfectly fine during the climb, showing both ground speed and distance, but during the jump when altitude drops quickly it just stops processing all the readings. Under the canopy it works again, but the moment you do any kind of dive it stops reading again for a second.

Well, it wasn't designed for skydiving after all, so nothing to complain about.

At least I was able to measure my ground speed under the canopy, so it could be used for some basic canopy piloting drills, but nothing more than that.

P.S. Turns out the plane travels in circles roughly 70 km at ~190km/h to climb to a 4k

GPS units generally have a physics model of what is "realistic" speeds and accelerations. These are used in kalman filters to clean up noisy GPS data. It probably thinks that falling at 120mph is not realistic.

Also the satellites are mostly near the horizon, so GPS is inherently less accurate on altitude than position. 

And finally -- check where it's mounted on your body. If its chest-mounted, you're going to be blocking its view of most of sky. Do you backfly? 

Cool project! 

you'd be better off using an barometer along with the gps.  i made one once using a gps module and a barometer, a small display screen, and a raspberry pi zero, but couldn't get them both to read at the same time.  the forums i was looking at when i put it together suggested using gps also to check the accuracy.  getting the raw gps data works fine, and it works with just the barometer, but not both together yet.  i had to stop when shit went bad a few months ago.  when i get some time, i will post some links that i used to help you with yours.  i used the same battery you have and it seemed to do fine.  when i get it working this winter, i plan on testing it in the air as a wrist mounted altimeter.  as a chest mount it would be great.  i was going for a sub $75 altimeter/gps/logbook.  if that works out, i was going to add bluetooth and mate it with the pi i am putting in a helmet over the winter, but that may take a backseat also, a little ambitious anyway.

Interesting. How about replacing the LCD with a monocular then? https://www.aliexpress.com/item/1005001355384065.html

Very cool projects!

On 11/13/2020 at 6:33 PM, platypii said:

GPS units generally have a physics model of what is "realistic" speeds and accelerations. These are used in kalman filters to clean up noisy GPS data. It probably thinks that falling at 120mph is not realistic.

If this is the culprit, I bet it's the fact that the speed is so vertical.

On 11/13/2020 at 6:33 PM, platypii said:

Also the satellites are mostly near the horizon

I don't think that's true but maybe there's some geometry and/or statistics thing I'm missing

On 11/13/2020 at 6:33 PM, platypii said:

GPS is inherently less accurate on altitude than position. 

This is true though. I haven't found a great explanation for why, but I'm thinking it's because we get satellites in all directions horizontally but only across 180 degrees vertically? This answer from Garmin is pretty vague but it sort of hints at what I'm saying.

GPS satellites are in an almost circular orbit - they are not all at the horizon, but appear at the horizon, rise in the sky on their orbit and then set on the horizon. Altitude is less accurate because you're getting all of the fixes from above (I believe this is what they're trying to explain on the Garmin web page) and because of differences between the reference ellipsoid (GPS model of the earths surface) and the topographic surface (GPS users actual altitude on the surface of the earth)
 

GPS System

GPS Altitude
 

When I said "satellites are mostly near the horizon" I meant that statistically at any given time, you are more likely to have satellites off to the side, than overhead. And If you're trying to use triangulation to measure height and vertical speed... that's going to be less accurate without satellites overhead.

Geometrically it's a fact that they are more likely to be near the horizon than overhead. Here's why. If you draw a line extending out 45 degrees above the horizon, then the area of sky "above" you is much smaller than the area to the sides. I get a ratio of more than 2:1 "near the horizon" vs "above" at GPS altitudes.

Picture is drawn to scale. I will say though, I expected a bigger ratio. But because GPS orbits are actually quite high, this is less of a factor than if they were in lower orbit.

On 11/20/2020 at 4:52 PM, platypii said:

When I said "satellites are mostly near the horizon" I meant that statistically at any given time, you are more likely to have satellites off to the side, than overhead. And If you're trying to use triangulation to measure height and vertical speed... that's going to be less accurate without satellites overhead.

Geometrically it's a fact that they are more likely to be near the horizon than overhead. Here's why. If you draw a line extending out 45 degrees above the horizon, then the area of sky "above" you is much smaller than the area to the sides. I get a ratio of more than 2:1 "near the horizon" vs "above" at GPS altitudes.

Picture is drawn to scale. I will say though, I expected a bigger ratio. But because GPS orbits are actually quite high, this is less of a factor than if they were in lower orbit.

Interesting. So the 2:1 ratio you mention is the area of region B to area of region A in your diagram? Or the volumes?

Do you have a source for this actually being a contributing factor for the worse accuracy in vertical?

3 hours ago, nwt said:

Interesting. So the 2:1 ratio you mention is the area of region B to area of region A in your diagram? Or the volumes?

Do you have a source for this actually being a contributing factor for the worse accuracy in vertical?

i too wondered and looked it up.  i found this post on quorum to be quite informative.  the reason i looked was that i had heard the gps is an accurate altimeter.  it seems that it is, as long as it gets a fix on four satellites.  the fourth is the one that gives the time and allows the receiver to adjust depending on the differences in timing. 

of course i didn't stop looking and found a much better source of info here.

5 minutes ago, sfzombie13 said:

i too wondered and looked it up.  i found this post on quorum to be quite informative.  the reason i looked was that i had heard the gps is an accurate altimeter.  it seems that it is, as long as it gets a fix on four satellites.  the fourth is the one that gives the time and allows the receiver to adjust depending on the differences in timing.

I guess intuitively it makes sense that a satellite directly overhead would result in the largest difference in time measurement per difference in altitude.

If the 4 satellites happen to all be on the plane tangent to the surface at your location (might be impossible by design), I wonder if it would be impossible to calculate altitude. I understand how triangulation works, but how they get a distance from the timestamps without an atomic clock in the receiver is something I never wrapped my head around.

that's what the fourth satellite is for.  they take the timestamp and calculate the difference between known reference points of the others.  one of the links has a really in depth explanation i am halfway through.  i just saw this pop up and checked.  fascinating stuff and has motivated me to finish up my altimeter, although using just the barometer and the pi.