Papa3 wrote:
All of the recreational
Garmin units (at least through early this year) logged a fairly limited
subset of the available NMEA data (understandably, since the average
recreational user needs only time, position, and altitude).

- As to which altitude it is storing, that depends. The basic
"altitude" in the 76S is the barometric altitude. The auto-calibrate
feature compares the barometric altitude to the GPS altitude and
periodically adjusts the barometric altitude.

Exactly how it does
that (ie. average delta over n number of fixes, how it determines
whether to discard a fix, etc.) is not public domain.

But, I have had
correspondence with Garmin engineers who pointed out that the algorithm
means that you essentially end up with GPS altitude but smoothed by the
barometric altitude to take care of any individual GPS altitude fixes
that might have been degraded (at least that's my current
understanding).

If you turn off auto-calibrate, then only the
barometric altitude is stored. There is no way to tell from the
output record whether auto-calibrate was on or off.

From Ian Strachan, the other side of "the Pond":
Very interesting, Papa 3. Further questions a to what zero-altitude
datum are the altitude figures:
(1) displayed on screen and
(2) recorded for later download.

I doubt whether recreational GPS units will be set to the ICAO ISA MSL
datum of 1013.2 mb (29.92 inches of Hg) that is the worldwide aviation
pressure setting for all "Flight Levels".

So, is it supposed to be a "mean sea level" datum? If so, how is it
achieved? (another proprietory algorithm, I suppose). The "64kdollar
question" is, how does it relate to spot heights and contours on maps?

The initial datum for GPS altitude is the selected ellipsoid, WGS84 and
local variants. But this is nowhere near MSL except in small areas by
sheer co-incidence. I understand that differences between the WGS84
ellipsoid and local MSL vary by some +65m S of Iceland to -102m S of
India, a total of 167m (548ft).

I suspect that these recreational GPS units have an electronic look-up
table that changes altitudes above the ellipsoid to what the WGS84
source document calls "an equipotential surface". As I understand it,
this worldwide equipotential surface is approximately, but not exactly,
equivalent to local mean sea levels. For instance, in UK mapping, Mean
Sea Level is taken as the average tide at a place in the SW of England
called Newlyn, I guess because it has regular and reliable tides. I
imagine that the mapping authorities in other countries have a similar
system, they have to have an MSL "zero datum" on which to base heights.

These are tricky issues if altitude figures downloaded after flight are
to be accurate and meaningful. At least you know where you are with
altitudes based on the aviation-standard 1013/29.9" sea level datum.
These 1013-based altitudes can then be converted after download into
others using public-domain formulas (the balloonists do this, I
understand, to get a sort of true height for the day rather than just
using the ICAO ISA as IGC does).

But with unknown algorithms applied before download, you do not know
exactly what you are dealing with. This is not a problem in the
recreational area, but may be in ours if accurate altitude is to be
used in a performance claim or to defend a pilot against a possible
airspace violation or other situation where a flight record wiuld be
useful. Then there is the issue of the general reliability of pressure
altitude compared to the sometimes erratic figures for GPS altitude
recorded in a proportion of IGC flight data files ...

Ian Strachan
Lasham Gliding Centre, UK