Alright then, maybe the default value of 0.5 m for EK3_POSNE_M_NSE makes sense. I would be surprised if a quad-constellation dual-band receiver will have a worse accuracy than the M9N.
Yeah, if GPA.HAcc, GPA.VAcc GPA.YAcc are all 0 all the time, it means the LC29HEA is not sending the NMEA messages corresponding to the estimated accuracies that Ardupilot expects to receive. It’s also possible that the NMEA GPS driver doesn’t even read accuracies.
Maybe it’s a lag issue? What does the position look like when you do not move your drone? Does it wander a lot?
When in Loiter mode with no Stick change it well start drifting until i correct it. and then it drifts the other way and so on.
Like the control loop for Loiter mode has too much gain of something. But the only think i do is change out the GPS and change the Ardupilot for NMEA input.
Dennis
My understanding was that the parameter was the minimum accuracy value to use. If the GPS would report 1 cm accuracy, the EKF3 would take the value of EK3_POSNE_M_NSE (0.5 m). But if the GPS reported 1 m, then the EKF3 would use 1 m. But you know probably more than me.
But anyways, we’re diverging slightly from the conversation, unless the real accuracy of the LC29HEA is over 1 m. Which is why, @dennis_bowman1, I’d be interested in a log where the drone isn’t moving (ie, on a table, without props, without flying). You can set LOG_DISARMED to 1 for Ardupilot to create a log without needing to arm. And maybe log the GPS position for 5 to 10 minutes.
Sorry, but it is certainly not wise to try to judge the accuracy of a GPS by adding the control deviation (no matter how precisely someone has followed your instructions)
I can get an indication of the accuracy of the GPS by leaving it still and recording the measurement points long enough to assess their scatter. I don’t need an Ardupilot or UAW for this, it just gets in the way.
What exactly do you mean by GPS responsiveness? If you mean the reception strength, the same applies
I have yet to hear from anyone that has done a flight with the LC29HEA on a Drone running Ardupilot. So, i take it that I am the first? I have ublox M8, M9, F9P on other Drones and this one and have not seen any problems like this with no answers. So, i guess i well not try to use the LC29HEA until i see some that has done a good test flight.
Thanks for the Help
Dennis
I think the point he was trying to make was there was a drastic difference between the 2 GPS receivers with everything else the same. even if he had ardupilot totally untuned, as long as the compass is calibrated it will hold position as long as there is no wind, his multirotor was flying fine with the m9n so there is obviously something wrong with the new GPS configuration if it’s wandering around.
@dennis_bowman1, with the LC29HEA (and without the M9N), were you only using an internal compass? That could be why the drone was unstable, if you hadn’t performed a full calibration of the compass.
Evaluating the accuracy of a GPS unit can only be done if you have a ground truth, which you don’t obtain by looking at a flight log. And the easiest ground truth is if you don’t move the drone at all (leave it still on a table).
Yes @Rolf you are correct, that is the proper way to determine accuracy.
But it looked to me as the user is looking at the drone movements to infer GPS receiver quality.
My point was : if you really want to do that, then at least tune the drone first.
I was talking about complete system response not just GNSS signal response/quality/accuracy
If you want apples to apples comparisons, then do apples to apples tests. Keep all variables equal except the GPS module (ideally even using the same antenna, assuming both modules support the same antenna hardware and range). Testing should ideally be done as close in time and location as possible in order to limit environmental variables.
Static testing is good but imperfect. I think it will likely showcase the major differences.
However, an on-vehicle/in-flight test should not be overlooked, as receiver sensitivity to environmental interference or susceptibility to signal loss at various attitudes could be make or break items for making a decision on a module’s practical use. Again, limit variables. Fly a similar (or identical, if possible) profile. Use a well proven platform that has been properly set up and tuned.
Yes, that is what i did as i said i only changed out the GPS receiver and just changed ardupilot for NMEA input and not ubx. and Same Antenna Mounted in the same position. The only thing different is the GPS receiver. The same compass too. Same location for flight.
can you change the GPS baud rate to something higher? im wondering if you are saturating the serial connection with all those constellations at a high update rate and that’s causing the update instability.
or try limiting it to just the GPS constellation on its own.
Ah OK. All is enabled in his GNSS configuration which can be problematic with other modules. But I don’t know if this rate instability is the culprit or not.