Using NTRIP station 178 km away?

Hi all,
I am having trouble justifying paying 1600 eur / year for a local NTRIP subscription. So I would like to hear your experiences with using the free euref network.

I have to place some permanent markers for stockpile volume measurements (so I can do the job faster, without having to measure out markers each time), so the absolute position is, as such, irrelevant. However, comparability between the jobs, I would like to have some sort of absolute precision (mostly for my self, not a requirement for the customer (not yet at least, maybe after seeing that they can actually compare, they’ll change their minds).

Would it be better/good enough to:

  • Use euref NTRIP using a point 178 km away (should give me 17,8 cm+ 7 mm of inaccuracy?)?
  • or collect 8-24 hours of raw-data and compile it using PPP afterwards, then feed the manual coordinate to my base station?
1 Like

I would suggest making your own reference point by using the free NTRIP service from 178 km away.

  1. Place Rover over new reference point.
  2. Connect to NTRIP service.
  3. Log ntrip raw data and rover raw data on your rover. (I don’t know how long to suggest.) Why not 24 hours?
  4. Post process using rtklib.
  5. Use your new reference point as your base station.
  6. This would be better then just using PPP with Raw data from your Rover.

I wouldn’t suggest using the station located 178 km away for real time corrections.

Thank you very much for your input.

What would you say is the main difference between using PPK, like you suggested and PPP, when talking a static position ?
As I understand PPP it also takes atmospherics etc into account, like PPK with NTRIP?

Since Brent seems to be asleep, i give a shot at it.
PPP is not possible with current RTKlib version from Emlid, you have to use 2.4.2 which is somewhat buggy in some areas.
PPS, not K, its static. With reachview version 2.13.0 you can now use Glonass AR against different brands, not possible in RTKlib. Crossing my fingers i got the facts right :relaxed:

1 Like

So even if I collected hours of rawdata, I could use it for PPP, even if uploaded to a service like Precise Point Positioning ?

I get better result with PPP in RTKlib 2.4.2 and process ionos/tropo/clock/epherm finals then nrcan.
I used CDDIS but investigating differente option now.
You should get better result with PPS long range ( 178km) then PPP in either way. You also get a position faster with PPS, since you dont have to wait for final products.
There are other ways but i havent fully tested it yet.

Here is an example from my known point (hjemmesoyle) vs long range post processing.
143km baseline, 50min processing. 4cm off from known position.

1 Like

For how long did you collect data?

50 minutes
Sorry, i did 5 hours, this is just er portion of it.

How much would the precision increase If you used all 5 hours worth? Given the already obtained precision, I guess it is almost not worth it?

Here is the entire run (5 hours ish). Not sure it would do me any good to record more then two hours.


I’m surprised by that accuracy for 143km baseline, do you know if that’s consistent or could it get worse with poorer weather conditions?

I guess another option, if your client insists on more accurate absolute accuracy and you have the time, could be to set up your own Base within a reasonable distance from the NTRIP station, say 90km, and find its location using PPS. Then use PPS again with rover data over your site reference point and logs from your new Base to find your know site reference point. You could shorten the baseline distance from 90km but you may have to repeat the process and step stone known points to your site. I haven’t done this myself to compare how effective it is but it should improve accuracy if you log plenty of data for each PPS.

1 Like

Nice results @TB_RTK
Mind you share your reach settings during acquisition ?

Single. Gps, glonass, Galileo and sbas at 15degree mask @1hz. But recorded raw.ubx.

1 Like

Ok nothing exceptionnal !
I will try in Indian Ocean from +/- same distance

1 Like

Bumping this one to see if any one els is getting some usable result with long range.
@Sylvain_POULAIN ?

I made a quick test of 20minutes log recording over know point. Results are good :

But point is 1.36m different from known point coordinates. Not the best know point. Need to find time !

1 Like

1.36m sounds a bit wrong yes, should get better result.

May be this know point from government is not good… Not the first time I saw that here! I will try with another and more time

1 Like

Trying with example data. From MUEJ to GOET (GREF Network) (both in southern Germany), baseline 390 km, GPS+GLO. 24h and 15 min obs time, both with 20 cm in XY. 24h obs is 4 cm away in Z…

1 Like