[GTC] Doppler for all alphas in 2010?

So we know where we all stand should the GTC be doppler for GT11 Alphas from 1/1/10?
I intend to unless there is a really good reason not to.

Results of sailing today with 3 GPS. One GT-11 (V1.62B1115), One GT-31 (V1.3B1014a), One GT-35 with both 1 hz and 5 hz data (V1.3M1208T).

The GT-11 and GT-35 were worn in my Helmet.
The GT-31 was worn on my right upper arm. Not ideal for comparison but at least the
latter is how most wear their GPS.

Results: (For what they are worth )

Alpha 500 (Doppler):
GT-11 #1 - 25.773 @ 15.53pm *(Note: see below)
GT-11 #2 - 25.737 @ 15.50pm

GT-31 #1 - 25.767 @ 15.50pm
GT-31 ** - 26.260 @ 15.53pm (Note: see comments below **)

GT-35 5 hz #1 - 25.760 @15.50pm
GT-35 5 hz #2 - 25.735 @15.53pm
GT-35 1 hz #1 - 25.730 @15.50pm
GT-35 1 hz #2 - 25.583 @15-53pm


*This run has an excessive speed burst(spike?) on entry to the gybe. It was not so
obvious that I would ordinarily notice and reject it without having the evidence from
the other GPS's. After studying the others (and the fact that the GT-31 data was
rejected by GPS-Results - see below**) I would be inclined to think it was not
trustworthy, but I would not nessasarily know that if I was using this GPS alone.

**This this run did not appear in the results of GPS-Results because the filter
invalidated it due to excessive accelleration. I had to turn off the filter to get a
result for comparison.

Comments:
I think the 15.50 run is the best valid run. A comparison between different GPS with
that run is:

GT-11: 25.737
GT-31: 25.767
GT-35: 25.760 (5hz)
GT-35: 25.730 (1hz)

Most users of the GT-11 on its own (me included) would proably have posted the 25.773 results.
Even if this happened I think it would be within an acceptable range.

The big question is: Will there be other greater errors/differences if we have more
examples of GT-11 v's GT-31?

I have four other comparisons, Steve's (1 session) and Windrider's (3 sessions) and
these suggest not.

In those sessions the software (GPS-Results) finds the same best runs and the results are very close in 3 of the 4 sessions (within 0.15 knots and as close as 0.001 in Steve's case where there was very good satellite numbers - 8 on both GPS!).
In the fourth session the software found two better runs from the GT-31 because
satellite reception numbers were low for the GT-11 (dropped to zero for a few seconds, invalidating the run).

So, from this brief study, I think we can say that the Doppler results from GT-11's
with V1.62B1115 firmware and GT-31's with V1.3B1014a or V1.2B1405x firmware are
close enough for direct comparison and a fair fun competition (within 0.15 knots in the examples I have, and often better).

It also suggests that users of GT-31's will sometimes have an advantage in poor
satellite conditions where more runs will be valid and there is a likelihood of getting a better Alpha 500 as a result.

It also must be said that the GT-11 is more prone to strange errors or spikes in the Doppler than the GT-31 and we may not, even with the best intentions, pick up every one of those errors. But in the example of my session from today, the error was subtle and only gave a 0.034 advantage which is less than the variation between the two GT-31's (0.037). I am reasonably confident that greater errors would be more obvious and would probably be rejected by the GPS-Results filters.

It seems that more satellites give a greater correlation between units. This may be
something we can use to help spot suspect runs. I would suggest that runs with less
than 6 satellites should recieve extra scrutiny for strange bumps.

The GT-31 SDOP data is very sobering from the point of view of Alpha 500's. In almost every case the SDOP goes through the roof during the Gybe, suggesting that the confidence in the accuracy during the sudden turn is low. I have always suspected this anyhow but we don't have this facility in the GT-11 firmware and even if we did, I am not sure we could use it to effect anyhow. (see illustrations).

In an ideal world, we would all use the exact same best device available, the GT-31
but I think we can live with some people using the GT-11 for a fun competition. I
would urge those aiming for the top 20 in the Alpha 500 division to change to the GT-
31 if they have not already, so we all can have maximum confidence in the results.

I also think it is past time that we have postings from the Garmin Foretrex units (and
other non doppler units) in a competition environment, even a fun competition. The
results from these units are just not comparable to a fair degree.(But if you want hard numbers, that is another study ).

Oh, and another thing: It is definitely better to wear your GPS in or on your helmet, especially with Alpha's. It avoids the largest sudden movements during the Gybe and give much better and consistent satellite reception. It is even better to have TWO GPS's, preferably GT-31's, for comparison. It makes it soooo much easier to pick up any irregular values and gives greater reliability and confidence. Who want to do the gybe and Alpha (or speed run) of their entire life and have a spike right in the middle of it or not be able to tell if it is true or an error?

Illustration of SDOP values during gybes:
Also note the spike error in the trackpoint value coming out of the gybe in the highlighted run. This would definitely be an invalid run using trackpoints but the Doppler values are fine.



Even with 10 satellites, the SDOP values still go through the roof!