This is going to be a problem with L1 (single-path) hardware. I barely trust maintaining a constant fix on a UAV with L1/L2 (multi-path) gear so PPK is my recommendation. You can play with frequencies and satellites to see if you can get something acceptable, but it will end up being more trouble than just post-processing.
I have spent the last few days on an off trying to tweak the setup and at no point could get a fix long enough with the drone unit (rover) stationary to assume that it would achieve a fix in the air. PPK will be a solution but it does leave me just slightly irritated with Emlid’s highly “optimistic” sales material.
I think knowing the capabilities of single vs multi-path before you jump in would have clarified what you are experiencing. We need to greatly differentiate the expectations between RS, RS+ and RS2. Totally different animals. What you are trying to do may be an option with an RS2/M2 setup, but I still won’t trust it for what we are doing. If our Topcon gear can’t hold an infinite fix on a motor grader going 10mph then there is no chance that I am putting my drone imagery in that possibility. Even slight blips when you are collecting data every 1.5-2 seconds is asking for it. On top of that when people try to do this without GCP’s I wouldn’t be comfortable handing that off to a client.
Thanks for the response. I have the orginal Reach RS and module with the Edison board. With regards the problem I will generate some logs but I have a feeling that the issue is related to the inability of the radio link to maintain a constant correction signal from the base. With regards to vibrations this should not be an issue yet as we are “bench” testing with the drone one ground. With regards to rf the only emitter that could interfere with the GPS is a 900mhz rf link and that is positioned well away from the GPS which is on a mast with a back plane
Today we had better performance from the test system with longer fix on the Rover (drone). We reduced data to the minimum to obtain a fix and we also tried not to interact with the units via the web interface during the test we also increased the height of the Rover gps antenna above the unit.
During the log recording we changed the base from “continuous” to “fix and hold”
We will perform some more testing in the next week to validate if this has resolved the issue.