PCool wrote:
Thank you Mark, now it is all very clear to me, finally.

I wish to do a summary of what I have understood, very simplified.

What is Pressure Altitude for IGC standards
It is the altitude calculated with an ICAO-ISA formula . You need a baro
which has been calibrated at the factory, because the calibration is fixed
at the time the recorder leaves the manufacturer (just like mechanical
baro). The calibration is an important issue (exactly just like for the old
good mechanical barographs). Once calibrated, you read a pressure value.
You pass this value to a formula and get the ICAO-ISA altitude. In
practice, QNE and ICAO-ISA may differ by some tens of meters, once
calculated on the same value!
You can revert the formula: starting from an ICAO-ISA altitude you can get
the pressure.

You sort of lost me here 8^)

What is Altitude for a Garmin and/or a GPS COTS
The altitude measured by a GPS could be the real altitude over ground, but
intrinsecally may contain geometric errors. Everyone agree on the fact that
the GPS Altitude is not accurate.

Actually, GPS altitude is quite accurate at measuring actual height (to
within +/- 10 meters or so) *most* of the time. That last part is
important. Depending on lots of things, like bad satellite positions, a
wing or a rock blocking view of a critical satellite, the phase of the
moon, etc., GPS altitude can occasionally be hundreds of meters off. If
you take a longer term average, accuracy will normally be to within less
than a meter.

By contrast, pressure sensors are quite accurate and reliable at
measuring ISA pressure height (to within +/- a few meters or better,
below the tropopause), but can't accurately measure actual height. They
do not suffer from short term fluctuations and occasional wild
excursions like GPS altitude does (with the exception that some flight
recorder sensors will show large errors at low battery voltages).

Some GPS like Garmin's use baro sensor to correct the GPS altitude, and
vice-versa, in order to achieve maximum precision and obtain possibly the
Real Altitude, above mean sea level. When we say "correct altitude" normally
we refer to this.

Yes, to be more precise, when a Garmin is in auto-calibrate mode, the
pressure sensor is used to compensate for short term fluctuations in GPS
altitude, so you get the long term accuracy of GPS altitude, with the
short term stability and resolution of a pressure sensor.

Why COTS' Altitude is not good for IGC badges
The answer has nothing to do with precision. IGC requires to read ICAO
Pressure-Altitude, not the real altitude.
It is exactly the same altitude you may read on a paper from an old
barograph.
There could be little difference among the two, but in principle we are not
talking about the same thing.

Correct, though some also argue that the short term accuracy of GPS
altitude is not good enough to allow verification of height gains and
loss of height. I believe there are ways to work around this, others don't.

Of course a COTS could easily output an ICAO-ISA altitude, it's just a
matter of using the formula and unselect any other corrections. The
manufacturer could thus implement this feature, it is much easier than
correcting and auto-calibrating GPS altitude.

This is true, but the pressure sensors need to have rather good
temperature compensation and long term stability, which may not be the
case with the sensors in consumer grade GPS receivers.

BUT, but, the manufacturer should also provide a calibrated sensor at the
factory.

They could, but remember, the market for glider pilots is insignificant
in comparison to the number of these units sold. It may simply not be
worth the added expense to the manufacturer.

In other words: if three devices are standing at the same height, they
should all read the same pressure value, say 747 mb.
The garmin with sensor may say you are at 4750m , another COTS basing only
on GPS may read 4680m, while the IGC may declare 4820m.

No quite, if the Garmin is auto-calibrating, it is reading actual
altitude (actually height above an ellipsoidal Earth model, but that is
another issue), just like the GPS-only unit (without the fluctuations).
So, the Garmin with sensor might read 5250m, GPS only might be 5230,
and IGC might read 4820m. They are all more or less correct, the IGC
unit is measuring something different. If the Garmin pressure sensor is
set to a proper fixed calibration, it will read the same as the IGC unit
(assuming adequate temperature compensation).

Conclusion: without a pressure sensor no COTS can be used today as an
alternative to IGC altitude loggers.
And in any case, calibration is an issue.

At the moment, that is correct. The rest is up to the IGC.

Marc