GPS Time and Altitude Definitions

[nigelspeedy]

G'day Dynon,

I am busy doing some testing in my RV which has the Dynon Skyview system. I love the user data log files for post flight analysis. Some testing I did recently was takeoff and landing and GPS accuracy (static & dynamic). I've used a DGPS as the truth source. I've been experiencing a little trouble merging the two data sets. Primarily I think it is an issue of definitions rather than accuracy. To help me understand the system a little better I have a few questions.

1. How many leap seconds do you add to GPS time to convert it to UTC?
2. Why is the 'system time' 1 second advanced from the 'GPS Time & Date'?
3. Is the top bar time displayed on the screen the same as the 'system time'?
4. Is the 'GPS Altitude' height above the WGS-84 Ellipsoid or is it corrected based on lat/long to represent height above the MSL Geoid?

Thanks for a super system.
Cheers
Nigel

 

[dynonsupport]

1) This is done inside the GPS module, and should match the currently broadcast offset. We don't have control or visibility into this. Right now that should be 17 seconds.

2) We set the "system time" when the system boots and first sees GPS lock on. This is an independent clock inside SkyView that can continue to work even if GPS has failed. However, it can only be set to full seconds when we synchronize. Thus, it's expected that it can be +/- 1 second. It does not constantly synchronize so it can also drift very slowly, but this should be way less than a second on any normal flight that doesn't include fuel loads greater than airframe weights.

"GPS time and date" comes right from the GPS NMEA parser and reports whatever the GPS last said before we wrote to the log.

Also important is that there is variable lag inside the system from getting a GPS position and time to when it is recorded. This could be up to 300mS, plus whatever your datalog recording rate is.

3) The top bar is system time.

4) The altitude is MSL Geoid.

 

[nigelspeedy]

Dynon,
Thanks for your quick and full answer. What I was seeing in flight was very small across track error but an along track error that was equivalent to about 1 second of flight at my current ground speed. If I allow for a 1 second error in along track I also get very good accuracy here as well. Same for the GPS altitude. So a great system all round.
Cheers
Nigel

 

[nigelspeedy]

Hi Dynon,
Here are the test results that prompted the questions. If I understand your explanation correctly at 16Hz I could have a time error of up to 1.3625 seconds when comparing the time stamp of the Skyview with another source (not counting the errors in the other source). If this is the case the large along track errors fall to <50' and put them inline with the small cross track and height errors. I did this test as an exercise in the test method and data reduction, I'm not trying to imply that the Skyview system is anything but really good, I am really impressed with it and love flying with it in my aircraft. Keep up the good work.
Cheers
Nigel

 

[nigelspeedy]

Second graph