Labroides
Core User of DJI
Flight distance : 9991457 ft
Australia
Offline
|
EFRPIC Posted at 1-29 13:16
After some research, I built a portable RTK streamer with antenna that provides a steam directly to my RTK module on Enterprise Advanced. Having to program the options on the GPS receiving card, there are several ranges that must be set to define the accuracy of the location based on time, received signal strength, etc. I have it set for 1 meter accuracy in a window of 30 seconds. It takes about 10-12 minutes (with a reasonable "open" sky) to get a lock the 1st time form a cold start. On a warm start it drops to 3-4 minutes.
I believe the settings on the Mavic 3 were adjusted/refined in the December update to require better accuracy which would involve going to a smaller accuracy limit. The antenna I use for RTK and the F9 board use all GPS satellite co9nstillations and is far better than the one in the Mavic 3. The antenna is about $240 US$ alone.
I suggest that DJI wanted better accuracy to support the new RTH and other features as well as increased fly-away protection.
Some of the suggestions being put forward are just wrong and not consistent with the facts of GPS or observations of the problem.
There has never been an issue with DJI drone's GPS "accuracy" before, and there wasn't for the Mavic 3 before the December firmware update.
And ... the M3 currently doesn't wait for more sats to ensure higher precision, it is slow to get even a small number of sats.
Before the firmware update, you could have 20 sats in a short time.
Now people are waiting ~5 minutes and still only seeing 7-9 sats.
That's not an accuracy issue, it's a failure to acquire sats issue.
When a Pilot takes off without waiting for a "lock" they are individually taking on the risk and DJI is protected. If a Pilot waits for GPS "lock" and it is in the data, the user should be covered for a fly-away or other issues.
With any other DJI drone, if you were impatient and launched without waiting for GPS and for the homepoint to be recorded, there wasn't much risk as the home point would have been recorded as soon as the drone had good GPS location data.
And in most cases, that would have happened only a short distance from the launch point.
Launching early doesn't cause a "flyaway".
Not understanding how your drone works can contribute to an incident that an inexperienced flyer might describe as a "flyaway".
Flyers are much more likely to launch prematurely when it takes 5 minutes or more to record a home point.
The current problem is a fault, not a feature.
Every one of millions of other GPS receivers out there gets GPS relatively quickly (and the M3 did previously).
|
|