Emlid Caster wrong FIX?

I’m confused about the documentation of how the emlid caster works. From the docs one can think that it is possible to connect only a rover without a base. But it looks like this is not the case. Here is what I did:

  1. If I connect only one RS2 as rover (no base) with the given credentials for NTRIP I get a ‘connected’ status in ReachPanel. In ReachView3 (Version 26.5) the status is ‘Waiting for correction data’. The solution is always single.
  2. The situation changes if I configure a base on the very same RS2 by using the same mountpoint. The correction data is received and the RS2 gets a FIX.
  3. If I then measure a well know fix point fixpoint LFP2 the measuring is totally wrong even tough the RS2 got a fix. Easting Error: 6.69meters / Norting Error: 0.959meters / Elevation Error: 22.095 meters.


  • Why do I get this huge difference? Is it because I configured the same RS2 as rover and base? Yes, sounds stupid but after configuring a mountpoint for the base I got a fix.
  • Is it really not possible to use emlid caster with only one rover and no base?
  • Any thing else I did wrong?
  • What can you recommed I should do to improve the results?


This is expected behavior. For RTK you need a correction source.

This is you root problem, yep.

The point of of the Emlid caster is to provide correction from your own base over a longer distance than LoRa can offer.
If you don’t have a base of your own, you need an NTRIP subscription from a state of commercial supplier.


Thanks. I’ll organise a second RS2 to setup a base on it.

Hi Roland,

Besides what Christian said, I want to share the post from our blog about the difference between NTRIP caster and NTRIP service. In a few words, a caster allows you to connect your own base and rover using NTRIP protocol while a service provides corrections itself.


Hi Svetlana,

Yes, I completely misunderstood the concept of the emlid caster and thought it is a service.
Only after reading the answer from Christian I got the message.
Thanks for sharing the blog post anyway.

1 Like

Hi Roland. Your not the only one that has misunderstood NTRIP caster and NTRIP service. I’ve tried to help a local surveyor doing this same thing and he couldn’t understand either. I explained to him to visualize a “bridge” (NTRIP caster) between two receivers along a road. The “bridge” allows the “car or truck” (correction data) to deliver the information to the rover unit. He was complaining about the limited radio RTK range and I suggested him to try this instead. I also suggested that he get a subscription to our state RTN VRS service. He finally understood and currently he uses both for his field crews. I told him the only disadvantage to this is additional costs for the services and possible lack of cellular coverage. I got a free lunch after he started using both services !!

Also, the accuracy of your Caster is dependent on two things, is the base feeding it surveyed in and how far are you physically from the base.

If you just setup a base using single average, it can be off by 3-6 feet on LAT/LON and ALT can be off as much as 30 feet. So if you are verifying your system with your rover on a known (surveyed) point, and it is off in LAT/LON the most likely culprit is your base LAT/LON. If it is off by ALT (or height) it could be your base station ALT or you have not put in the height of your survey pole.

All points collected with a rover are relative to the base data used. If you manually put the same information in your base and physically put it in the same location, you should get the same rover data. If you put your base in the same location and do a single average then survey the same point, you will find a difference as the LATLON/ALT have changed for the base location. Single average is just using WAAS data to get the base location, hence the difference in location data.

1 Like

Yep… all is true. However, even if you did a “here” position at the base (without known coordinates, just an averaged position for a few minutes) you can translate the base/rover data to whatever passive control marks you may have located or just PP the base and translate rover data to the new base position. Of course you could enter known coordinates too for the base. I’ve used an NTRIP service a few times with our Javad receivers just out of curiosity, but using the RTN service is cheaper (one service to pay for) and I don’t have to worry about a base receiver unattended. I’ve used a base here at the office and also one at my house (about 20 miles west to east direction) for a few projects as a baseline for static use when doing projects in the southern part of the county just for experimentation. Of course, long observation times (minimum 1 hour) are necessary due to baseline lengths. Post processing accuracy is usually less than 2-3 cm accuracy in horizontal/vertical. PP data also compared the same using OPUS, DPOS and using RTN service. It just amazes me every time I occupy some of the old NGS triangulation stations when I have free time also. The accuracy that was accomplished back then in the 30’-50’s was phenomenal when compared to what we can do now. Even after multiple adjustments, different geoids and different reference frames down through the years. NGS knows what they are doing !

In my use, I really do not have to be true with the world. It is nice if I am so if someone wants to locate the 4" tile which is 2 - 3 foot deep they can.

It is amazing what the surveyors of old could and did do. Where I live almost all of our roads are magnetic North/South or East/West. Pilots are taught to fly by them. Occasionally there is a small 5 - 10 foot offset which creates a small S curve in the the road. I have been told this is where the surveyor crews met and the offset was the difference between them, not sure how true it is though. The S curve is in about every road parallel to it at the same point N/S/E/W of the other.


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