PPK result more than 100 meters off

Dear all,

We record camera events on our drone with an M2 that receives corrections via NTRIP. The M2 logs the raw data + corrections and we do PPK later on. On one day with five consecutive flights (1 big log), the RTK solution (.LLH file) looks fine (99% fix, position plausible), but the PPK is horrible: The first 3.5 flights are shown as fix but the recorded positions are shifted by 136 meters as compared to LLH (not plausible at all); the last 1.5 flights are shown as float and are off by 5 meters (not very plausible).

I suspect that the NTRIP base (which is a VRS) was moved/redefined mid-flight to a new position, which would be fine for RTK, but is not supposed to work for PPK. Does this make sense?

Did anyone ever experience something similar and is there a way to recover the correct positions of the VRS in order to split the base file and redo the PPK for the two separate segments?


Hi Marcel,

Yes, your suggestion sounds right to me. VRS can change the base position during the survey, and it can hardly be taken into account in post-processing.

Do you have a physical base close to your work area? If so, I recommend you process the data using the logs from it.

1 Like

Thanks, Julia!

We asked a nearby base operator and they only keep full logs for 6 weeks. The flight in question took place 18 months ago. This far back they can only provide archived base data (1 observation every 30 seconds). But the effort to retrieve the data for them is substantial, so no guarantees that they’ll do it.

I still have these questions:

  1. Can I expect any good ppk results for a drone moving at 2-3m/s with corrections at only 1/30Hz?
  2. We are in central Germany. Does anyone know of any service providing base data this far back in the past?

Many thanks

Hi Marcel.

Unfortunately, not. I would expect you would not gain any more accuracy over simply using the uncorrected positions collected by the M2 during the flight.

1 Like

Hi Marcel,

Oh, I see.

Agree with Dave. The 1/30 Hz correction rate will result in the age of corrections rising to 30 seconds, whereas we recommend it to be 1–2 seconds only. This rate just isn’t enough to get accurate coordinates.

I know there are some EUREF bases in Germany. Maybe you can get data from them.

I’ve also heard of SAPOS and Geo++.

1 Like

Yes! Thanks a lot! They seem to have data stored for the past 3-4 years at 1Hz.

1 Like

Yay! Glad to help :slight_smile: