Reach never get fixed

[zhangkaibest]

Backgrouds :

HK Government provide free VRS correction date and we have account , so we want to use one reach as RTK .
we can get the correction date from base , but reach never get fixed , wonder some setting of it that i have done is wrong
below are some pics to show the setting of reach . for your renference . oh our reach connect to pixhawk through UART 4/5

appreciate your help in advance !
is it because the single ,or something else ?

[TB_RTK]

There are couple of things i see:
1.Low amount of satellites with SNR above 45, and many of the sats visible to base is not visible to rover.
2.Do not use AR glonass with base other then Reach. For your case, turn it off
3.Disable additional correction, i can not see why you should use this as it is from the same service
4.Disable base mode

[zhangkaibest]

got it . thank you ~~so much

[thomas.purk]

@TB_RTK

2.Do not use AR glonass with base other then Reach. For your case, turn it off

Can you explain this? Why would you not use Ambiguity Resolution for GLONASS satellites from non-Reach correction sources?

[TB_RTK]

No.fix

Glonass is tricky for RTK because of the inter-channel biases. All satellites transmit on slightly different frequencies (compared to GPS or Beidou) and that results in receiver-specific delays in the front-end. Ability to get fixes with glonass might depend on the receivers in use. It is best to use Glonass ambiguity resolution when using a pair of Reaches. You can also use Glonass satellites, but not include them in AR by opening “advanced settings” and changing “GLONASS integer ambiguity reso…

[thomas.purk]

Thanks for the info, was not aware. Found a detailed explanation here.

http://www.insidegnss.com/node/1374

[Sylvain_POULAIN]

And so, what about gallileo ? Is this the same with glonass ?

[zhangkaibest]

hi， i do as you told me , but things dont get better . seems to be the gps single problem , is there any better RTK antenna that you recommand ?

[bide]

Could you show us a picture of your Reach and antenna?

Also, when you are using Reach, are you outside in a clear area, or are you beside trees or buildings? (you could also show us a picture of the area where you are using Reach.)

[zhangkaibest]

i have come back to office , i will show them to you next time , and at same time , i will try it again . thanks ~~~

[zhangkaibest]

hi bide , this is the environment of where we test the rech with free ntrip service.

[bide]

It is a very nice place you have there!

With my GNSS eyes, I see vegetation in one direction and a building in another. If the other two sides are similar, then your test area looks quite good. There are only a small amount of obstructions and they are very low on the horizon.

If you can share your log files from a test, please do. Include the raw log, the base corrections, and the solution file please.

If you do not wish to share your log files, then maybe you could analyze or present them in a similar way to this: reach-bouncing-rtk-float-to-fix-on-tractor

[zhangkaibest]

hi bide, thanks for you help!! here is the log file including everything i have for your reference ~~~
log.zip (686.3 KB)

[zhangkaibest]

oh by the way , bide , your hyperlink can t open . wonder it is correct

[bide]

Thanks. Hyperlink fixed now.

[TB_RTK]

Do you have a picture of the rover?

[zhangkaibest]

HI , TB, here you are .

[bide]

OK, my critique of these log files is:

• log duration is too short. only 4 minutes of usable data. 10 or 20 minutes or even longer would be better. although, in ideal conditions, 4 minutes should be enough time to acquire a fix.
• base is only using GPS and SBAS constellations. this is fine for acquiring fix and for troubleshooting, but if you add additional constellations (GLONASS) then it will hold the fix better.

In this pic, the rover is at top, and the base is at bottom.

• connection was lost to VRS after ~2 minutes, base coordinate changed location. ambiguity resolution must start over. this makes it take longer to achieve fix

• the only way to get a good amount of fixed points with these super short segments of log file was to use Fix-and-hold. The following plot is the result of post-processing 3 log segments. One for each of the VRS base coordinates:
  
  
  
  You can see three distinct groups of fixes and those are each from a portion of the log where the VRS base coordinate had not moved.

I think we had better wait for @zhangkaibest to record some longer log files.

[bide]

Hmmm! What do you think about that rover pic @TB_RTK?

[zhangkaibest]

ok got it . i will test it . thanks so much !!