Qstarz 818XT Nightmare

Hi guys,

After a while using RaceChrono Pro + cheap 1 Hz AD-350 GPS with good overall results, we decided to upgrade to a 10 Hz device in order to improve the race line resolution. So, we bought 3 Qstarz 818XT 10 Hz for go-karting race analysis, which has become our nightmare. After a number of runs I realized 10 Hz is unusable due a very strange effect:



Or this below:


I'm aware RaceChrono might experience problems running @ 10 Hz, but I couldn't find a reasonable explanation for the effect presented in the linked pictures above. Lap after lap, the race line was going displaced to northeast in a semi-regular pattern and the correspondence from real race line and the represented race line was being degraded. So, I decided to lower the frequency to 5 Hz and 1 Hz and got some driving around a block trying to keep the car on the middle of the road. But the results are also quite bad:

5 Hz> http://i.cubeupload.com/HLqkLV.jpg

5 Hz> http://i.cubeupload.com/5pLo1p.jpg

1 Hz> http://i.cubeupload.com/h9toUN.jpg

I also noticed there's a zoned effect on overall error:

Zonal Error> http://i.cubeupload.com/XkmX3f.jpg

When travelling through a rough E-W direction the error significantly increases in comparison to N-S directions which make no sense to me if you ask me. The expected difference from real life and registered lines was 1 meter (the informed Qstarz error during the run) but it delivers differences about 10 meters mostly running on the yellow zones of the last linked picture. So, under 1 or 5 Hz there is no more progressive displacing effect of the race line, but the accuracy is still poor, even more if you bare in mind the cheap 1 Hz device has been presenting better results.

When testing under 1 or 5 Hz I made sure Differential GPS was enabled in all tries. 10 Hz forces DGPS off by default. Also tried to play around with NMEA settings, trying to force all information in several rates, but it has no impact on the GPS performance. I tested all 3 devices as well to make sure isn't a fault unity but the results were the same.

Tips and advice are welcome. I ran out of ideas here.



  • aolaol
    edited February 2017
    Hi Jorge,

    Sorry to hear about your bad experience. I think this post sums up why in the last couple of years I've been recommending receivers with GLONASS support (such as Garmin GLO and Dual XGPS160). The main issue with bad quality data is lock to too few satellites. When you got both GLONASS and GPS satellites available, chance of getting good data is better.

    1) Have the GPS receiver in a position with as clear view to sky as possible. Chances are most of the satellites aren't directly up, but somewhere between up and horizon. This means that if there's anything blocking the device's view to certain direction, fix to some satellites will be lost in turns. Not sure what your setup looked like, but having different accuracy to different directions make perfect sense to me. Something in your setup blocked some of the satellites at the turns. So optimal place on a kart would be maybe drivers helmet. But as that might not be safe, then somewhere in the front.

    2) What comes to QSTARZ Q818XT settings, have it 5 Hz as you did. The 10 Hz mode is impaired due to lack of DGPS. I think the first screenshots is what can happen without the DGPS. There's probably some atmospheric anomalies going on, and the location keeps shifting due to lack of DGPS correction.

    The 5 Hz screenshots do not look hopeless to me. Yes, they're not perfect, but you're not going to get perfect with a cheap GPS. I'd like to see how the logs are from the kart track in 5 Hz mode. I'm sure it will be usable.

  • aolaol
    edited February 2017
    Also city or a park maybe not best way to test a GPS, as there's trees, houses and structures that distort the signal, but also your speeds are lower. Better to test on track.
  • Thank you for your support Aol.

    I'm testing with 7 - 10 satellites average. I'm not very familiar with the subject but is not a bad amount I guess? Maybe there're too much of them lower on horizon so you'd be right.

    Now I'm also considering to upgrade to a GLO device at once if things remain the same.

    What bothers (and intrigues) me is the previous device (much cheaper, not Glonass support as well and only 1 Hz) used to provide much better lines in RaceChrono Pro at the exactly same conditions (gps positioning, number of satelites, buildings and trees blocking, etc). It's also intriguing the estimate error DOP is around 1.00 most part of time if not always.

    I'll provide a full comparison between devices and frequencies and make it available here ASAP.

  • Weather also affects GPS. You may have different data quality. And if you do buy Garmin GLO, buy just one for comparison and testing at first.
  • I use the same device, and when I use it on kart, I got the same problem like you, shifting out of places. But I got perfect result when I use it on car racing track.
Sign In or Register to comment.