[wingspan33]

Golly, Bob works out the way to down load a GPS file and the ones I sent him were from a different flight than discussed here. Sorry about that. Those tracks were from a longer flight and were closer to a usual length flight track log - better for a test upload.

But, for anyone interested, here's a Garmin GPS track log file of the not so good flight involved in this thread.

(If you have Gamin's "Map Source" software you will be able to open the file and "see" every data point in the flight)

[Bob Kuczewski]

Sorry about that.

There's nothing to be sorry about Scott. It's just so good having you here on the forum ... it reminds me of the old times when we were working together to build the HGAA. Thanks very much for joining us!!!

[Bill Cummings]

It just occurred to me that tight turns will show a slower ground track than airspeed since the two second interval setting will be straight line point to point whereas the glider is flying a faster arc.

[Bob Kuczewski]

It just occurred to me that tight turns will show a slower ground track than airspeed since the two second interval setting will be straight line point to point whereas the glider is flying a faster arc.

Like this?

subsampling_turn_points.png (4.4 KiB) Viewed 4926 times

I guess if you were flying a fast 180 degree turn and the 2-second samples were taken at the entry and exit points, it might show your ground distance as 2 times the radius, but the actual distance travelled would be pi times the radius. The difference between 2 and 3.14159... is pretty significant.

Of course, they may be sophisticated and use a curve fit (Bezier, for example) to come up with locations between samples which would be much more accurate.

[wingspan33]

It just occurred to me that tight turns will show a slower ground track than airspeed since the two second interval setting will be straight line point to point whereas the glider is flying a faster arc.

Good point. I wonder how significant the difference is between the true ground track speed and the artificially segmented ground track speed? Clearly, the faster and sharper the turn, the more your true path will be shortened - and effectively slowed.

Sometimes I like to exaggerate a problem to demonstrate it better. Consider a pilot doing a flat spin in which the timing of the GPS sample rate coincides with the same heading, one revolution further on, as the flat spin continues.

The resulting track would show your glider descending (close to) vertically with no apparent rotation. And while you actually have air speed (at least on the outside wing) your ground speed may register as zero, or close to it. Imagine handing that track log to someone to analyze, without telling them about the flat spin. lol

There's also another track log anomaly I noticed during my accident. In track log entry 741, I am 2 feet lower than my obvious at-rest point(s) registered in subsequent entries. I wondered how it was that I penetrated two feet into the ground, then came to rest up ON the ground - two feet higher. I hit my head, but there wasn't a two foot hole left behind!

With a little thought I realized that that track log entry was at (or close to) the moment of impact. That the abrupt stop to which the GPS was subjected caused a higher air pressure within (or upon) the air pressure sensor/transducer than the actual surrounding air (at a true 1128 ft MSL). This caused the GPS to register a momentary, slightly lower altitude of 1126 ft MSL. So it seems I hit the ground at 4:53 and 53 seconds pm.

I'd like to see everyone with a GPS running on their glider, set at a fairly high sample rate. The resulting record of data could make certain accidents much more understandable.

[Bill Cummings]

1) “With a little thought I realized that that track log entry was at (or close to) the moment of impact. That the abrupt stop to which the GPS was subjected caused a higher air pressure within (or upon) the air pressure sensor/transducer than the actual surrounding air (at a true 1128 ft MSL). This caused the GPS to register a momentary, slightly lower altitude of 1126 ft MSL. So it seems I hit the ground at 4:53 and 53 seconds pm.”

I noticed that subterranean altitude on your first post and wondered about it.
My old Garmin 72 only gets it’s calculations from satellite signals and usually gives me a plus or minus accuracy of 12 to 18 feet.

My track log on launch looks like either a kid went crazy with an etch-o-sketch or that I was being blown all over the launch before taking off.
I don’t know anything about the latest total energy verio flight decks. I am wondering how they integrated pressure altitude with GPS altitude.
My first thought was that your position was doing the normal deviation like my Garmin 72 but then I was surprised that the following altitude fixes were not moving.

I had thought to joke with you about the subterranean altitude data point but since we don’t know each other that well I thought it might be in poor taste.
(Something like --crashing the hell out of yourself and your equipment and going straight on to your reward equipment and all.---And then THEY realized that you weren’t finished with your mayhem on the surface yet and you were sent back.--
Which reminds me that I should mount a fire extinguisher on both of my down tubes and have my passport printed on asbestos paper. I don‘t know if that would help but where I‘m headed anything would be worth a try.)

2) “I'd like to see everyone with a GPS running on their glider, set at a fairly high sample rate. The resulting record of data could make certain accidents much more understandable.”

I think I have my GPS set at two or three seconds. The memory in the Garmin 72 is short and if I have a long flight I start to loose the first part of the flight when the memory is filled and I’m still flying.

I use my GPS in Track Up mode so that I can easily rejoin a lost thermal. The circles on the screen aren’t. They look more like stop signs or triangles.

When thermaling in a strong wind the pattern on the screen looks like a heart beat pattern on an oscilloscope more so than it does circles, stop signs or triangles.

Having an in flight GoPro camera for your forgotten event may have been a lot of help analyzing what went wrong.

[Bill Cummings]

Yes Bob, Exactly what my minds eye was seeing. Also found this"
http://www.math.ucla.edu/~baker/java/hoefer/Bezier.htm
Also this:
http://www.math.ucla.edu/~baker/java/hoefer/Lagrange.htm

[Bob Kuczewski]

Yes Bob, Exactly what my minds eye was seeing. Also found this"
http://www.math.ucla.edu/~baker/java/hoefer/Bezier.htm
Also this:
http://www.math.ucla.edu/~baker/java/hoefer/Lagrange.htm

That's an amazing coincidence ... Here's my Bezier Curve program ... also written in Java:

Screenshot_Bezier_Curves.png (3.39 KiB) Viewed 4909 times

I used that basic code years ago to develop a landing approach system for a simulator I was building at work.

[wingspan33]

Hmmmmm. Being the practical type, what I think I'll do is walk a "lap" of a routine hike, which I've already got logged at the 2 second sample rate, but in this case set the GPS for a 5 second sample rate first. I'll then be able to compare the mileage readings from both, otherwise identical hikes, and see how close or far off they are from each other.

That should tell us if (what I'll call) the curve calculation is being done "automatically" - in my particular GPS (a Gamin GPSmap 60CSx). It could be that a different or newer GPS may include curve calculation.

I've got 3 other models of Garmin GPS unit so I might just check more than just the 60CSx.

Also, if I'd had one attached, a GoPro or something like it would have been very useful indeed. I actually have a fairly small MiniDV video camera with a setup to connect it to my glider's rear keel. But I haven't been flying with it recently. With my luck, if I'd had it attached, I would have busted it during that not so good landing.

And here's another fun fact - Two spectators (husband and wife) were at the launch while I setup and after I launched. They used a digital camera they had to take video of some of my flight. But they didn't want to run out of space on their memory card so they didn't video my landing - but they saw it. However, they did resume the video as the emergency response people showed up. They were very nice people and may have been the ones that first called for help.

[wingspan33]

So, I did a GPS test this evening.

Rather than hiking, I figured that driving my car the same way around the neighborhood twice would more closely simulate a glider's path. I drove between 20 and 30 mph roughly and had to make many tight turns as I went.

For the first run I set my GPS to an every 10 second sample rate and drove the course. I got .5 miles. For the second run I set my GPS to an every 1 second sample rate. That resulted in a .6 mile reading for the exact same route. The Garmin GPSmap 60CSx does not seem to do what I'm calling "curve calculations".

The view of the less defined track log in Map Source looks obviously simpler with a few corners cut due to when one or another position reading was taken. The 1 second sample rate track is much more clearly defined with no cut corners.

Considering the above differences, I would guess that even a 2 second sample rate could "cut" some corners, at times, in an average hang gliding flight. Particularly in sharp, high speed turns. It also makes sense that the registered speeds in those cut corners may be slightly slower than in reality. My guess is that at a fairly high position sample rate the error probably isn't that great. With a 5 or 10 second sample rate (or even longer), the errors will be more significant. But it needs to be understood that during certain quick (flying) transitions even a couple second sample rate could miss critical or revealing information.