Dual XGPS 160 accuracy issues

Discussion related to external and internal GPS / GLONASS / Galileo / BeiDou sensors
User avatar
rbpeirce
20 or more Posts ★★★
20 or more Posts ★★★
Posts: 86
Joined: Fri Jul 08, 2016 8:21 pm

Re: Dual XGPS 160 accuracy issues

Post by rbpeirce »

Interesting. I was going to buy a 160 for exactly that reason, The Track Attack map traces are bizarre and the Track Attack people felt that was because of the 1Hz signal in the iPhone. Apparently not. It is uncorrectable drift, which is probably much worse over a month than an hour.

Apparently GPS readings are nowhere near as accurate as I thought they were. Maybe I should wait to upgrade my hardware or not do it at all. The ONLY reason I felt the need to do it was for Track Attack and Tripper's plot isn't really the much different from what I got using the iPhone for GPS.
2012 Porsche Cayman R
2016 Audi A7
User avatar
Harry
Site Admin
Site Admin
Posts: 10516
Joined: Sun Sep 12, 2010 10:32 am
Location: Siegum, Germany
Contact:

Re: Dual XGPS 160 accuracy issues

Post by Harry »

The biggest problem with smartphone‘s internal GPS is accuracy, not update rate. As external receivers have an optimized antenna design, they deliver far better her. The higher update rate is a nice addition for smoother lines too.

As outlined above, the drift is one component reducing accuracy, but as it has almost no influence on lap timing. The problem component is the noise, it is higher for a small. Umber of satellites locked and it is higher for internal GPS due to the compromises smartphone manufacturers needs to make.

Harry
Image Image Image Image
aMacUzr
20 or more Posts ★★★
20 or more Posts ★★★
Posts: 26
Joined: Mon Oct 15, 2018 1:46 am

Re: Dual XGPS 160 accuracy issues

Post by aMacUzr »

This "plotted position on the track" accuracy was what I was referring to in my post at
viewtopic.php?f=19&t=5452&p=26650#p26647

For me, the "track plot" in Circuit Tools was more accurate from lap-to-lap throughout the day when using the VBOX as opposed to using the XGPS-160 (actual position judged via integrated front-facing dashcam video). To be fair, however, the number of samples I have are not statistically significant/conclusive.

Harry: I, for one, really appreciate your detailed explanations (of course, I am so much a geek that I think the term is a compliment). #;-) Your support is superb (and, as a software guy, I know how difficult and valuable that is).
User avatar
rbpeirce
20 or more Posts ★★★
20 or more Posts ★★★
Posts: 86
Joined: Fri Jul 08, 2016 8:21 pm

Re: Dual XGPS 160 accuracy issues

Post by rbpeirce »

Tripper. I see the difference in the Track Attack track plot but what about the data traces? I get a lot of variation from the iPhone, and I was considering the add-on GPS to improve that as well as the plot. The traces sometimes overlay exactly and sometimes are off by quite a bit. I'm not talking amplitude, which is what I am trying to examine in many cases, but the tendency for one trace to lead or lag another. Of course, there is some of this in real life, maybe 5-10' or so, but I see one case in the speed trace where a hairpin turn is 200' different and I know that never happened. I also see another case in an acceleration trace where one trace is consistently about 10' sooner than the other. These sort of things greatly reduce the use of the app. I can live with track plot problems but I was hoping a better GPS unit might cure the trace problems.
2012 Porsche Cayman R
2016 Audi A7
User avatar
PitrsRS
20 or more Posts ★★★
20 or more Posts ★★★
Posts: 156
Joined: Wed Jun 25, 2014 9:43 am
Location: Prague, Czech Republic

Re: Dual XGPS 160 accuracy issues

Post by PitrsRS »

Hi, I have a strange problem. As you may see in the picture. Same track, same XGPS160 (same firmware) only "2" years of difference and weather. 2 years ago was around 15° and today almost 35°. It´s not a problem of the phone, because I had simultaneously connected second phone to the GPS with TrackAddict app, and there is the same result.

It´s the problem with GPS shift, or something else (f.e. 2 connected phones at the same time)? I´m just asking to prevent it in the future.
Attachments
GPS fix.png
GPS fix.png (2.5 MiB) Viewed 3805 times
2016 SEAT Leon CUPRA 290 ST

HLT Grand Prix (Android) + OBDLink LX BT + XGPS160 + GoPro 11 Black + GoPro 7 Black
User avatar
Harry
Site Admin
Site Admin
Posts: 10516
Joined: Sun Sep 12, 2010 10:32 am
Location: Siegum, Germany
Contact:

Re: Dual XGPS 160 accuracy issues

Post by Harry »

That’s the so called “GPS shift”. It varies slowly but all the time. Results for the next track day may be better again - or shifted to a different position. It depends on the current satellite constellation and atmospheric disturbances.

RTK / GBAS systems eliminate that by using a reference station.

LapTimer has another feature to optimize it: select the lap from Lap List, follow Data Recordings / Fixes, select Remove GPS Shift from the second icon in the bottom.

LapTimer will apply a translation to all positions recorded so the overall difference between line driven and POIs is minimized.

Harry
Image Image Image Image
User avatar
PitrsRS
20 or more Posts ★★★
20 or more Posts ★★★
Posts: 156
Joined: Wed Jun 25, 2014 9:43 am
Location: Prague, Czech Republic

Re: Dual XGPS 160 accuracy issues

Post by PitrsRS »

I´ve already tried that feature to fix GPS shift, and this is result. I think problem was the second phone connected to the GPS itself. Because first 2 laps are nice and clean, after second lap I connected second phone and then it becomes messy to all remaining laps. XGPS can connect 5 devices at the same time, but I never tried it on the track. Maybe this was the problem
2016 SEAT Leon CUPRA 290 ST

HLT Grand Prix (Android) + OBDLink LX BT + XGPS160 + GoPro 11 Black + GoPro 7 Black
User avatar
Harry
Site Admin
Site Admin
Posts: 10516
Joined: Sun Sep 12, 2010 10:32 am
Location: Siegum, Germany
Contact:

Re: Dual XGPS 160 accuracy issues

Post by Harry »

Positions will not be different when a second phone is connected. Definitively not. That's GPS Shift.
Image Image Image Image
User avatar
PitrsRS
20 or more Posts ★★★
20 or more Posts ★★★
Posts: 156
Joined: Wed Jun 25, 2014 9:43 am
Location: Prague, Czech Republic

Re: Dual XGPS 160 accuracy issues

Post by PitrsRS »

if its GPS shift how is possible to have first few laps in the session nice, and from the third it was like this. Only difference was the second phone. GPS was mounted on dashboard under front windscreen
2016 SEAT Leon CUPRA 290 ST

HLT Grand Prix (Android) + OBDLink LX BT + XGPS160 + GoPro 11 Black + GoPro 7 Black
User avatar
Harry
Site Admin
Site Admin
Posts: 10516
Joined: Sun Sep 12, 2010 10:32 am
Location: Siegum, Germany
Contact:

Re: Dual XGPS 160 accuracy issues

Post by Harry »

The other difference is the time elapsed between the sessions. The satellite constellation changes all the time and atmospheric disturbances too.

The other reason I'm sure is that the position is calculated by the GPS chip once. There is neither a reason or the processing capacity to do position calculation twice (or five times). Instead the initial processing step will create the full fix information and a later step will distribute "copies" of that fix to the various consumers / clients.

- Harry
Image Image Image Image
Post Reply