How to set up your own NTRIP service to work with Reach RX?

In our previous post, we talked about choosing between Reach RS2+ and Reach RX. Let’s say, you chose Reach RX. How can you set it up for RTK?

Reach RX is a network rover, so to achieve a precise position, it requires corrections over NTRIP. Usually, you can get them from a local NTRIP service. It works like this: you get a subscription, which, in most cases, is available for a fee, configure the receiver, and then start surveying. It’s all good, but what if you don’t have any NTRIP services nearby? Are there any alternatives?

Yes! You can place your own Reach RS2+ base and pass corrections to Reach RX via free Emlid Caster—the ‘pipeline’ to supply corrections via the Internet. All this will work just like an NTRIP service, but your own.


  1. Place your Reach RS2+ base in an area with a clear sky view without any obstacles.
  2. Set the position of your Reach RS2+ manually or using averaging.
  3. Go to the Emlid Caster page, create an account, and get your NTRIP credentials for both base and rover.
  4. Configure Reach RS2+ to output corrections using credentials from Emlid Caster.
  5. Configure Reach RX to get corrections using credentials from Emlid Caster.
  6. Make sure that your Reach RX is receiving corrections—check the Status screen in Emlid Flow (formerly known as ReachView 3).

That’s it! Moreover, with Emlid Caster, you can connect up to 10 rovers to your base for free.

Placing your own base may be a real catch, as it goes with you anywhere. This way you won’t depend on the location of NTRIP service’s bases and subscription in general.

Have more questions? Feel free to share them in the comments.


This saved me today. Had an odd situation where the receiver would not connect to the network service so I tried a base I have setup at my house via Emlid Caster and it was within 0.05ft of site control vertically from 30km.


Hi Svetlana, great write-up on setting up a Reach RX with Emlid Caster as your own lower cost NTRIP service. Just wanted to point out a couple of things. That you will need two cell/internet connections for a local onsite setup, one at the Base and one at the Reach RX receiver for using Emlid’s Caster service. That’s not necessarily a bad thing, because of certain field conditions, Cell service has been available when LoRa communication between Base and Rover has been cutoff because of site conditions, like buildings or an hill obstruction for me in the past. Though I mostly use an active NTRIP correction service nowadays.
Also, without an active NTRIP correction service to establish the absolute position of the Base point, you will have to either do PPK or PPP correction afterwards to get that position. Again not a bad thing, and most of the time I would assume that most everyone is out there collecting data for an least an hour or more with their static Base. And if you only need relative positioning, then it’s even a easier process of course like you have mentioned.
I liked Michael’s and PotatoFarmer setup with a Base receiver unit located at a office/house and then transmitting corrections using your Emild Caster service. That definitely would be a cheaper way of getting corrections than paying for a subscription service if your within the proper range. I do like your RX form factor and weight, especially after lugging a Rover around all day sometimes and I’m still considering purchasing one because every ounce counts.

Hello everyone,

I’m just curious, Reach RX is a network receiver,
for projects in places without internet service we need to pair it with RS2+. How will we get corrections?
What will be the corrections, is it relative or absolute.

I hope my query is correct.

Thanks a lot.


Hi Mon, with no available cell/internet service you would need to use a second RS2/RS2+ as a base to send corrections over LoRa to your RS2 Rover. To get a absolute Base position, you would either need a known point or establish one yourself using either PPK or PPP methods describe in Emlid’s excellent docs.

Without doing the work above, it would be just an relative position using the Average Single as the base coordinates entry method.

The Reach RX needs cell/internet service in order to communicate with either an NTRIP correction service or as Svetlana explained above, using Emlids Caster service in conjunction with an additional RS2/RS2+ receiver as a Base.


Hi @mark1st.john,

A very comprehensive feedback, I like that! I think it’s a good point for those who’re choosing between RX and RS2+.

You’re right, Reach RX is indeed designed for a specific (but still very wide-spread) application. It’s perfect for working in urban areas or near them where you always have a cellular coverage. And often, NTRIP services. If there is no Internet, then yes, Reach RS2+ is a choice.

Also, RX is compact because it’s a “city-style”, and you can walk around with it light. To a remote place, you usually drive and can take heavier equipment.

Yes, it’s really convenient. We actually use the same setup in the office for tests.

P.S. Thanks for your answer to Mon!


On base and rover, what additional settings should be done?
What is the ideal setting on this screen?

Hi Mauricio,

Svetlana didn’t specify the additional settings as the default ones should work great, so our general recommendation is to go with the default configurations. This is a really rare situation when you need to tune them.

The ones on the screen seem good to me, so I’d keep them as is :slightly_smiling_face:

Good morning Tatiana.
I understood.

But, I had an experience here, where in a baseline of 20km I had significant variations (+/- 20cm).
So I thought there might be something in the settings that could improve this.

We leased some points, after a while we leased them again and there was this difference.
We were doing this for a while and noticed this variation.

Due to the nominal accuracy of the equipment (RS2), working with Ntrip, in a baseline of 20km this could not happen.

1 Like

Sounds like you didn’t really have a fix. What were your residuals and PDOP?

Always getting patches and with great PDOP parameters and patch submission rate (age)

I think you had a false fix then. We always occupy each point after the initial collection is done and I have seen in cases where a point doesn’t check within tolerance and it’s always in a spot where the solution is struggling to get/stay fixed. How many points in the session were out like this? Same day or returning stakeout?

It was like 10 points.
But we checked and waited until we were sure of the fixed correction.

What methods do you use to secure a real fixed?

We have the habit of collecting the point, forcing the receiver to lose the signal by covering it with the hand, fixing it again and checking the point.

They were all on the same day.

Hi Mauricio,

That sounds really weird. Can you possibly email us all the data to from this survey so that we could look into that?


Haven’t you sent the files yet? Couldn’t find your ticket.

I’m checking with the field team.

Got you! I’m waiting, then.