I’m impressed how fast this updates are comming - 0.2.3 that’s understatement 
1 Like
Hehe, good one 
Hey!
Yet another update is out, v2.2.4. This one introduces automatic update notifications to the app. It is now much easier to stay up to date without monitoring the forum.
1 Like
Try pressing the Shift button while clicking on the reload button or Shift+F5 or Ctrl+Shift+r to force a fresh reload of the page.
1 Like
Yup, clear browser or run in incongnito mode
1 Like
Updates are too fast. 
My test is outdated again.
- Flight (could not get a better result. Help is highly appreciated)
The files (for the flight only): emlid.zip (7.7 MB)
I am still getting “Error memory allocation”. Restarting the program helps. It does not occur when the output folder does not exist (and is created when processing).
Their are significant problems when processing the base in static mode and choosing filter type combined. But using backwards seems anyhow the better solution.
5 Likes
I have notice too, that the reach solution for the base station is remarkablly good. It is difficult to get the same results in post processing. Any pointers on RTKpost settings?
2 Likes
The uploaded observation of the base station is only 0.5 h of the 3 h depicted above.
@emlid should add some description how to analize the observations and then choose the settings for processing.
The not so nice thing is, that even the flight paths of different fix solutions (calculated more than 60) vary for more than one meter.
Difference between the total observation and the fligth. Another hint that there is something very strange going on.
Yuhuu, second flight ever (with our Hexacopter), first flight with EMLID Reach & Navio2.
The absolute result of the solution is obviously wrong. Got a lot of different fix-solutions running RTKLIb with different time frames and different configurations.
Here are my raw files: raw files
3 Likes
I also have strange behaviour with RTKPOST b27. If I process combined (forward + backward) in the backward leg the whole duration is Q=1 but when I look at it in RTKPLOT it looks like roughly 50% Q=1 and 50% Q=2. If I only do backward the results are fine!
cI also use combined, maybe that is a bad idea? But for me combined gives far more fix-results.
I also get strange results when I compare my Reach base station and our official one. Some how the position (also calculated with the official base station) of the reach base station is wrong.
I observed similar offset between RTK and PPK. For me sometimes doing only backward gives better result than doing combined. Moreover, when I process the pictures with pix4d with Ground Control Points used as checkpoints the original RTK based positions give better match than the one with precise ephemeris and clock. So there should be something with the post-processing. If I exclude the .sp3 and .clk (and .i) files than I could have almost identical results with in PPK than in RTK.
1 Like
@vbnhu thank you very much for the valuable hint!
@emlid did you implement this fix: fix rtklibexplorer Pos2-arlockcnt
1 Like
In my experience post processing with RTKLIB demo5 b26a http://rtkexplorer.com/downloads/rtklib-code/ gives significantly better solution, just lacks the event processing feature.
2 Likes
arrg, “just lacks the event processing feature” - events are what dreams are made of.
What do you do about that?
Both RTKLIB versions, same settings, the results of the fix solution are almost the same:
left rtklibexplorer, right emlid yellow, rtklibexplorer orange
right emlid green, rtklibexplorer blue
1 Like
Hey, I am running 2.2.4. It’s the best yet for fixes! Usually just a few seconds. I have run into an issue with NTRIP. Can anybody confirm that NTRIP is working properly on this release? I don’t seem to be connecting properly to the caster. I don’t get a dropdown menu on rtk2go.com(although I can’t confirm that there are other users on, but other times there have always been multiple streams available). Also, I can’t seem to connect with my own caster either. I will try on different network but I was just wondering if anybody else has run into this.
1 Like
Hi Brent,
Yes there’re some bugs in NTRIP now, we’re working on fix.
Usually this version of NTRIP works properly if you put correct data in ReachView fields.
But I can say that there’s no answer from rtk2go.com:2101, so maybe that could be your problem.
Try to find another caster and write here if everything is OK 
I can confirm that when loaded with the right info NTRIP works fine for me in 2.2.3, but not updated to 2.2.4
I also had minor problems with NTRIP in version 2.2.4. No mount points in the drop down. I added the mount point after the ip separated with a slash. That worked for me.
@Emlid: It was my first look at the new ReachViiew and it’s great!
1 Like
I did a custom code in Matlab, to do the interpolation. Given the raw data log, I can process it with both emlid and rtklibexplorer and can get the event times from the emlid processing (to be precise I did some reverse engineering to know the timestamps with more than 1ms precision - assuming the interpolation emlid uses is linear). After that I can apply the same interpolation routine on the rtklibexplorer solution. I am doing lever arm compensation for the antenna - camera misalignment so it was not particularly difficult to implement this additional feature. 
1 Like