Optimal GPS settings

Turn on GLONASS. Every Second recording will not make any difference.

IME Glonass doesn't help, and sometimes makes things worse. Experiment for yourself and see which seems to be better.
Turn off 3D speed and distance. They will only make a difference for things as steep as ski slopes (and for skiing I would use them); even for very hilly runs and hikes, the improvement in the very steep bits will be outweighed by the cumulative errors from extra noise in distance. (It needs to be a 1 in 4 slope before the difference in distance reaches 3%, and there are very few routes which are that steep on average).
One second or smart makes no difference, as Trippy says, the only difference is that it records more of the gps sampled points to the file.

Thanks guys for your reply. So If i wait for some few Seconds the GPS Signal is quiet good

Wait until you get a green ring and then wait a bit more (as in atleast 10-30 seconds more)... That at least kind of guarantees it's not the gps fix. Might be the F5 accuracy that;s bothering you though.

I read somewhere that adding GLONASS doesnt help because if you choose GPS + GLONASS on the F5, signals are not merged ! If GLONASS signal is better than GPS so the watch will use only GLONASS and if GPS is better the watch will use only GPS
In which country does GLONASS signal be better than GPS ???

After a lot of try, I believe that GPS default settings are the best for 90% runs and you'll never know what are the other 10% a modification could be better. So my believe is modify nothing or restore defaults ;)

I read somewhere that adding GLONASS doesnt help because if you choose GPS + GLONASS on the F5, signals are not merged ! If GLONASS signal is better than GPS so the watch will use only GLONASS and if GPS is better the watch will use only GPS
In which country does GLONASS signal be better than GPS ???

Uh huh. What's your source for that? It isn't generally how GNSS chips work, and it wouldn't be very efficient as you'd need to do two parallel position calculations then select one, rather than running all satellites into the same calculation and sacrificing, effectively, one satellite's worth of data to correct clock differences between the systems.

Uh huh. What's your source for that? It isn't generally how GNSS chips work, and it wouldn't be very efficient as you'd need to do two parallel position calculations then select one, rather than running all satellites into the same calculation and sacrificing, effectively, one satellite's worth of data to correct clock differences between the systems.

Sorry it's a french article :
http://www.montre-cardio-gps.fr/quest-ce-qui-est-mieux-avec-le-glonass-quavec-le-gps/

Another one here in french too :
http://www.my-trail.fr/gps-glonass-cest-cest-quoi/

Without translation :

Mais alors ça sert à quoi un récepteur GPS / GLONASS ?
S'il n'est pas possible de cumuler les données des deux systèmes, en revanche nos petites puces électroniques sont capables de choisir, dans une situation donnée, le meilleur signal (ou le moins mauvais) entre les deux.
Si par exemple dans votre position, seuls 3 satellites GPS sont visibles et qu'il y en a 4 du système GLONASS, alors votre puce calculera votre position à partir du signal GLONASS... et réciproquement.

With Google translation :

But what's the use of a GPS / GLONASS receiver?
If it is not possible to combine the data of both systems, however our small chips are able to choose, in a given situation, the best signal (or the least bad) between the two.
If for example in your position, only 3 GPS satellites are visible and there are 4 of the GLONASS system, then your chip will calculate your position from the GLONASS signal ... and vice versa.

But maybe are there not right...

Thank you.

I got as far as "The only way to explain this is that the Fenix ??3 always uses 4 satellites from the same system (either GPS or GLONASS), regardless of the number of available satellites. That is, there is no complementarity between the two systems, but a choice between one or the other" in the translation before I decided not to waste any more time reading :) Those both look like speculation without any basis in knowing how the chips actually implement location, or any information about best practice in doing that.

Modern chips have a lot of channels, enough to lock everything that's up - I think the 5 has the same MT3333 chip as the 3, and that has 66 channels. It's possible that it's deliberately only using the four strongest signals (actually it would be 5 if they're mixed across constellations), but unlikely; back in the days when Garmin sold watches that showed the GPS satellite bars, it was clear that the position improved as the number of satellites locked improved. Different chip, sure, but things like the performance of the Spartan Ultra, especially in the early days, show that a different chip is not the whole story and antenna location is very important.

This is useful for understanding what's going on https://electronics.stackexchange.com/questions/11884/how-many-gps-channels-make-sense and note that it does say that the location calculation is generally done from more than four channels for better precision.

I don't know what the explanation is for the shabby performance of Glonass+GPS - most likely that the watches are susceptible to multipath signals, and adding GLONASS into that mix just means you get even more multipath signals screwing up your calculations, in my view, which explains why the performance deteriorates so quickly around tall buildings even when the sky view is still reasonable.