NTRIP wifi on base and rover?

I need absolute accuracy for my projects. Setting the base manually will work when I have a known control point. For sites that I don’t have a known control point I want to try using the NTRIP fix method to get my base position. My question is do I need a wifi for the base and for the rover to use this? When I walk with my rover away from the base, the base wont be in range of my mifi anymore. From what I understand I can use the rover with NTRIP without a base. However, the unit that I attach my rover to needs the rover and base to be connected.

1 Like

Can you expand on this a little? We run a rover on NTRIP with FieldGenius and never need a base, even if a site is localized.

When you say absolute accuracy are you talking about something on global or a site coordinate system? It is possible to get a “known” point that is on a local grid and vertical.

Land Survey, Construction, GIS?


I think they have a BASE & ROVER? From their other post also.

I run an Impulse Radar Raptor GPR unit. It uses both the base and rover to align lines of data as they’re being collected. The data gets processed in a program called Condor, the resulting features are imported into a CAD. It’s all for utility investigating. We want to start using a known point as a control. We’re experimenting with using a manual point and when that isn’t an option, NTRIP for correction. From what I’ve read on the emlid site, the base and rover both need to be connected to NTRIP in correction output and correction input, respectively. NTRIP requires the wifi through my cell phone. If I have the wifi with me as I’m moving around with the rover, the base won’t stay in range of the wifi access.

1 Like

The rover is connected to the GPR unit and captures its track?

You must enable the ntrip caster in base mode and then connect the rover to the caster to get the corrections.
This way the rover and the base don’t have to be in the same network. The only downside is that both base and rover must be connected to the internet. I use 2 sim cards for this, one for the base and one for the rover.



Obviously I won’t understand the ins and outs of your solution but unless there is machine control involved I don’t understand why you would need a base. A rover getting corrections from a service via NTRIP should be as accurate of a track as anything that you’re going to get.


Michael is correct… if you’re using the rover with the GPR and accessing an RTN network, you’ already have an absolute position from the state or private RTN. There’s no need for a base receiver.

I’ve seen this before on as-built surveys locating utilities and also on pre-construction sites. It’s so cool when everyone is on the same system and EVERYTHING is on the same system. Makes everyone’s job so much easier.

1 Like

Welp… looks like you have (2) rovers now if you use a RTN… twice as much work done, or finished half the amount of time? ; )

If you don’t have access to a RTN or too expensive, would need the base then.

There may still be some reason why his equipment requires a base. I just can’t think of one myself.

I agree with folks that if you’re within the coverage of an RTN, it seems a that using the RTN and 1 rover would give you great, repeatable accuracy.

However, assuming you do want to run a local base, want to use NTRIP to establish it and can use Lora to feed your rover corrections, then you only need 1 mifi. When you are setting up the base, connect it it to the mifi and NTRIP. Choose “average fix” to get the position of the base averaged over however many minutes you chose (since it’s fixed, somewhere in the range of 1-5 min is probably sufficient). Once the position is averaged, base mode will begin broadcasting its position and will no longer continue updating the base position even if it’s still connected to the RTN via NTRIP. Screen shot the coordinates that were averaged in base mode for your records, make sure Lora is up and running and your rover is connected to the correction stream. Then you are good to go. Your base no longer needs the mifi.

Doing it this way will give you the standard RS2 “cm accuracy” for the base (assuming this is the unit you are using) and then again “cm accuracy” to your rover. If your requirements require higher order survey accuracy, you’ll have to establish your bases’ position with a more rigorous technique (least squares adjustments of multiple static baselines).


Well said!!! ; )

Curious, in RTK settings for BASE, do you set it as STATIC (since NOT moving in this case?) and KINEMATIC for the ROVER (since moving in this case)?

Or is it best BOTH set to KINEMATIC with a moving ROVER?

Also, if NOT using RTN via NTRIP, I know you can Average SINGLE (meters accuracy), but why not Average FLOAT (sub meter accuracy)? What is the drawback? Much longer observation time versus a couple minutes for SINGLE? Obviously, prefer to set on a KNOWN point, traverse one using traditional methods or PPP one (if cm accuracy not needed) if no RTN coverage.


Correct, setting the RTK settings to static for the BASE and kinematic for the rover is right. Setting the base as static will allow the RTK engine to converge faster and have less process noise in its ongoing solution. Because you are averaging over time the process noise should cancel out anyways. The “what is happening behind the scenes” when you set the RTK engine to static or kinematic are related to the Kalmann filter of the RTK solution process and and how it uses the rovers velocity to estimate its new position. If there is no velocity, it can be removed from the equation (static).

For averaging RTK float, I don’t believe this is a setting that would work practically in a real situation. To achieve a float solution, the rover must be receiving RTK corrections from a base and locked on to the “code” of the GPS signal. The next step is for the GPS to estimate the most likely carrier phase position to “fix the ambiguity.” The ambiguity is that the carrier phase is a sign wave carrier signal and you can’t recognize one cycle from the next unless you observe multiple code progressions to identify the most likely one (requires clean signals - strong SNR). I’m summarizing, and it’s interesting internet reading if you want more detail.

So ideally if the base you are trying to set using another correction source is able to receive corrections to achieve FLOAT, it should also be able to achieve FIX. The cases where it couldn’t achieve fix would be poor SNR or poor sky view which means it’s a bad location for a base anyways. It could also mean you are at the edge of feasibility for using RTK for too long a baseline between your new base and the correction source your trying to use to set it. If the baseline is too long for RTK, better to PPP the base location or PP the location to the CORS afterwards when the precise ephemeris are out. Or like you said, set the base by using a passive monument nearby with a pair of GNSS or a total station, etc.


Awesome. VERY knowledgeable. ; )

Hi there,

Thanks for the discussion!

@evar, to establish a base position using Averaged Fix option using NTRIP, you need to connect the base to the Internet only at the beginning:

  1. Connect your base to NTRIP and wait for a Fix
  2. Go to the Base mode tab, let the base average its position for several minutes
  3. Save these coordinates to Manual by clicking the “Save coordinates” icon at the top right corner

After that, you can disable the NTRIP and Internet connection. By this moment, you’ll have a point for your base with already known coordinates.

However, you have 2 ways of setting up the RTK link between base and rover:

For the first way, you don’t need an Internet connection at all, just a pair of LoRa radios that are comes with our receivers. Another way require an Internet connection during all your survey for both base and rover.


This topic was automatically closed 100 days after the last reply. New replies are no longer allowed.