RS3 Base w/ RX Rover- Elevation ~10 feet to high

We recently purchased RS3 and RX to use as base/rover for basic surveying and with drones. Tried it out today with few shots around the office. The points surveyed come out about8-10 feet higher than what Google Earth, local GIS maps, etc show even when accounting for the different geoids. Any idea what could cause this? Link below is csv file from the survey. From what i can tell we have the antenna heights, coord system, geoids, etc all entered correctly. We are using the RS3->Emlid Caster-> RX for the RTK corrections. Any help would be appreciated. Thanks.

Survey CSV File

If u didn’t set the base on a known point, and used average single it will be off.

GoogleEarth uses EGM96 fyi.


What datum will you be working in? I am assuming NAD83 2011 and you want your current State Plane with NAVD88 using Geoid 18.

As stated, Google Maps/Earth uses WGS84 and EGM96 as its geoid.

Where did you get the Base coordinates for the Base/RS3? If you SINGLE averaged, you could have an Absolute inaccuracy measured in meters.

In the future how are you going to setup your Base/RS3? On an NGS monument? Receiving corrections from a RTN RTK subscription service? Without doing either, you will not be tieing into NAD83 2011.

But you do have options!

  1. Make your own point at the office. Place your Base in a suitable permanant/semipermanant mark at your office and log for 12 hours. Use Emlid’s PPK software (Download free NGS CORS logs) or use OPUS (Submit your logs to OPUS) to obtain your new NAD83 2011 coordinates. You now have accurate coordinates for your Base/RS3 to send corrections to your Rover (RX or Drone).

  2. See if your state has a free RTK network.

  3. Purchase an RTK subscription. Rock RTK is around $50 a month or is free if you host a Rock RTK base on your property (Buy it for $800, install it, give it power and Wifi).

There are more options, but these are pretty easy and straighforward.

Welcome to the forum!


Hi @cincydrones,

Welcome to our community!

The best way to check accuracy is to collect a benchmark with known coordinates and compare the known values with the collected values. Please make sure that the compared coordinates are in the same datum. I’d also like to note that open-source maps, such as Google Maps/Earth, usually give only an approximate position of the objects. You can find additional details in this thread.

How did you set up your base? Was it at a known point, or did you use averaging? As the guys also mentioned, using average SINGLE can affect absolute accuracy.


Also check your interval settings. I had a customer who recently had some bad elevations but he had his RX set to 30 second interval. So he basically took 1 single observation per control point because he was only doing a 30 second timed observation. So that equated to a single observation vs 1hz interval which would be 30 observations over the 30 second time frame.

1 Like

Hi Tim,

Do I understand correctly that your user collected data with Reach RX using a 30-second averaging? The update rate on Reach RX can’t be modified.

If the conditions are good and the receiver has a clear sky view, collecting points instantly should also provide accurate data. If any kind of elevation mismatch occurs, please feel free to bring it up, and we’ll help you figure it out.