Thread: About 37xx converted to 34xx satellites reception accuracy

Is this even possible ?

[Only registered and activated users can see links. ]

Real 3490

Originally Posted by Kanopus

There is a difference in GPS chipset between 37xx and 34xx. You can expect, that different hardware will lead to different results, which can't be corrected by firmware.

I don't understand this obsession on accuracy value. This is some kind of synthetic number, for which we even don't know if values from different firmware are directly comparable. I think there is accuracy value in signal logs. As it happened, the best recorded accuracy was for firmware 37xx v3.60, which measured the worst signal quality.

I absolutely do not share the idea that our quest for accuracy is an obsession, but a legitimate challenge for our intelligences, for anyone who wants to understand, learn and maybe solve the problem. From our "obsession" for accuracy, intriguing new questions were raised that open new doors of knowledge to anyone who is interested.

The hardware and firmware and their controllers are different, designed for each device type, and the results may, and appear to be, different, but that does not mean that we can not fix this. The fact is, if we believe that we can not, we will never succeed. Fortunately there are those who believe and so many changes in firmwares were possible increasing the number of functions and benefits of the many Garmin gps..

I believe we should do comparative tests, side by side, with tracklogs to check the accuracy "real" between them, as suggested by the X-trail. Other possibilities would be to record specific 3790 original firmware regions on the 3490 fw. I know this is possible with some regions, but Loly is the expert on the subject.. (139 rgn?)

Originally Posted by angelozip

I absolutely do not share the idea that our quest for accuracy is an obsession

I didn't mean GPS accuracy in general but specifically accuracy value displayed on satellite screen. In my opinion it is totally useless as an criteria to estimate general GPS accuracy for the reasons I stated above.

Edit: maybe you know saying: garbage in, garbage out. If you use garbage as a test criteria then you will get garbage as a result of your research.

I completely agree, but we have to prove it. I understand that the only truly reliable way would be to have the raw data from satellites, not processed (static and dynamic) turning in RINEX format (for pos processing) and compare, but this is a complicated task in Garmin GPS (but not impossible). Therefore I believe that a simple and reliable method to compare the differences in accuracy between the two firmwares is simultaneously record tracklogs of some way (in off-road mode, Dem and 3D off) and thus to be able to clarify whether there are significant discrepancies and confirm or not if the values ​​displayed on the screen can serve as a comparative parameter, what has been done so far.

Originally Posted by giuseppe66

photoshop or what?

No Photoshop it's called Hotfix

If I delete the /.System/GPX Folder at this location it never picks up satellites even after waiting 20 minutes.

Then I go outside to the Garden and it picks up sats First 27m after a minute 14 meters another minute 7 meters another minute 3 meters, Then stays 3 meters.
Then I come back to this location which is under a steel roofed carport and it stays 3 meters for about 5 minutes and then it starts to drop gradually.
I think that this screen is indicative but it has a lot of "smoke and mirrors" created magically by the firmware.
The /.System/GPX folder holds the files needed for the hotfix. It already knows which birds are flying overhead at any time of the day. (It learns all the time)
It actually takes a couple of days to learn all 32 Sats. After deleting the folders in the evening the next morning it takes a long time to get a 3Meter Fix.
I believe that deleting the GPX folder is the same as cold start. (Starts without knowledge of sats)

Originally Posted by dasilvarsa

Is this even possible ?

[Only registered and activated users can see links. ]

Real 3490

Originally Posted by giuseppe66

photoshop or what?

I don't believe for i second that dasilvarsa would do such a thing as photoshop an image before posting it. I've seen that 3 metre so-called 'Accuracy' with similar low sat. lock numbers and signal strength on original 34/35xx units. I've stated before and will again categorically that i do not believe it is possible for a non-professional gps device without SBAS enhancement to consistently maintain that accuracy under a wide range of conditions. I assume the 'photoshop or what?' comment was a joke. Kanopus is right in that many of the tests (mine included) are not very scientific (garbage in, garbage out) but more anecdotal. Even the side-by-side tests rely on the units accurately reporting their true accuracy are suspect. Consistent 3-4 metre accuracy shown by 34/35xx is garbage imo. Angelo has done static tests over a permanent mark, can't get more reliable than that. But the big problem is moving accuracy for some units. I still think that the only real answer is by modding 37xx gps fw into the 34xx to overcome the hardware incompatibilities which are obviously worse in some 37xx coverted to 34xx. I can't do it, can any of you other blokes without 'Navigation software expert' behind your nicks do it? If not, we should all shut up and let the experts like Loly & Kanopus find the solution because i don't know if we can add anything more that's constructive unless the further tests are totally scientific and tailored, but really what will they prove? That some units have 'bad' accuracy, which their owners already know? We can do all the further testing we want, but more tests won't solve the problem ....