Joreg Ewald at Volirium hecks out all the information that Wesley
and I have been sending to him.

Joerg Ewald <<joerg.ewald>>
writes:

I finally found the time to investigate how we handle 2D GPS fixes
in our various varios. In short, everything is fine on our end, but I made an
interesting discovery in FS (see further down). For the more detail-oriented:

1. In the Flytec 6020/6030 (aka Bräuniger Competino+/Compeo+), several GPS
modules were in use over the years. We started out with a Furuno GH-81, talking
to it through its proprietary binary protocol. We check a number of flags: "has
valid fix", "3D fix" and "speed and course valid", if those three are true, we
then check again whether the number of satellites is greater than 0 (looks like
a classic fix to work around some bad data from the module to me), and only then
do we consider the position as valid, mark its B record with an "A", and include
the altitude we received from the module. In all other cases, we create a B
record with the last known position, GPS altitude 0, and marked with "V".

2. Later on, we switched to SIRF modules, for the 6020/6030 as well as the 6015
(aka Bräuniger IQ Basic GPS), and all the Element varios. Fastrax IT321 at
first, IT430 later, and most recently Telit Jupiter SE880. We communicate with
these modules through the SIRF Binary Protocol. And in that protocol, you can
set the receiver to only report 3D fixes as valid - which is what we do. So we
only have to deal with invalid fixes (use last valid position, set altitude to
0, and mark the B record with "V") or valid 3D fixes (use position, altitude,
mark with "A").

3. For the Flytec Sensbox and the Volirium P1 (formerly Flytec Connect 1), we
use a variety of modules from u-blox, communicating through the u-blox
proprietary binary protocol. Again, we set those up to only report 3D fixes as
valid, everything else is invalid and acted on accordingly.

But the really interesting question is, of course, what happens on the consumer
side? I don't know about any other scoring and online contest systems out there,
but as far as FS goes, I just made a surprising discovery: The validity flag in
IGC B records is ignored, has always been! Which means that as long as varios
record their B records as outlined above (only record positions obtained from 3D
fixes), everything is fine. But if varios also record positions obtained by 2D
fixes, regardless whether they mark them as valid or not, those positions will
be used for track evaluation. Which will falsify the results, as we have learned
over the last few weeks.

So not only should all vario manufacturers ensure that 2D positions are
appropriately marked as "V" in IGC B records (or even better, ignore them
completely, as we do). In addition, all the programs processing IGC files should
also be checked and modified where necessary to make sure they throw away all
those B records marked with "V".

Thank you to everybody who contributed to this learning experience!