It's been mentioned already times by many reviewers and users: Mavic 3 is very slow on initial satellite acquisition.
Mavic 2 was pretty much up a running after no more than a minute of the power-up.
With Mavic 3, I dare say it's close to 5 minutes. And a lot of times, you won't get all satellites until you are at least 20..30 ft up.
This delay eats into the battery time.
I mean it's great to have a longer-lasting battery, but if you always have to wait that long, it kinda brings this advantage down to where we were with M2P.
I've read that M3 stopped using GLONASS (Russian constellation) in favor of BeiDou (China constellation) system, maybe that's what causes it?
Regardless, DJI dear,  can we have the startup time back to what it was? Please?

Are you starting your drone out in the open where it has a clear, unobstructed skyview?
Or are there obstacles blocking a significant part of the sky?

Dropping Glonass and adding Beidou sats shouldn't make any difference to GPS acquisition times.

Yes. I'm in suburbia, no tall buildings around. And again, in exactly the same environment (my driveway) my M2P is ready within a minute.

DarthSLR Posted at 12-11 14:59
Yes. I'm in suburbia, no tall buildings around. And again, in exactly the same environment (my driveway) my M2P is ready within a minute.

I am experiencing the same issue.  And many of my flights originate from my backyard, so the environment is exactly the same, only the craft differs.

Got to admit my Mavic 2 Pro was not the fastest at acquiring sats, normally could get 13-15 on average after around 2-3 minutes, my DJI FPV is very fast at doing the same in a lot less time, & my Mavic 3 is around the same speed as my Mavic 2 Pro - maybe a tad faster depending on cloud cover/weather.

I can get enough sats indoors in less than 30 seconds.

Well, I did another test and checked when I was in a wide-open space (did my evening hike in the hills). Got 10 satellites within 10-15 seconds, and 15+ within 30..40 seconds.
Still, I maintain the notion that at my driveway it takes several minutes, while M2P gets to GPS mode within 30..40 seconds, so THERE IS a difference.

This is called GPS time to first fix.  And the amount of time this takes is very much time and distance dependent from the last time it had a locked location.  

On a "hot" start where you've very recently had a GPS fix, it can connect in under 30 seconds, sometimes just a few seconds.  However, on a "warm" start where its been a little bit, it may take between 30 seconds to a few minutes to get a lock.  

So if you want to compare between 2 different drones, you will need to compare them in similar scenarios.

https://ozzmaker.com/take-long-get-gps-fix/

Also, for the Mavic 3, DJI has removed the Russian GLONASS system and replaced it with the Chinese BeiDou system.  As far as I'm aware, this is the first time DJI has used this combination of positioning systems.

I am also having the same issue after the update. Previously to get sats was seconds but now to get 12 or sats for the homeport to be recorded takes my battery from 100% to 98% sometimes 99%. Its way too long compared to before. Today my drone sat on the floor for almost a minute before it gave the all clear to take off.

KlooGee Posted at 12-11 17:59
This is called GPS time to first fix.  And the amount of time this takes is very much time and distance dependent from the last time it had a locked location.  

On a "hot" start where you've very recently had a GPS fix, it can connect in under 30 seconds, sometimes just a few seconds.  However, on a "warm" start where its been a little bit, it may take between 30 seconds to a few minutes to get a lock.  

Would an appropriate test be to turn on M3, wait for lock, turn off M3, turn back on, and record time for lock, and repeat for my MA2?  And the times to compare are the 2nd locks (discarding the initial lock times for both craft)?

Tony64 Posted at 12-11 15:50
Got to admit my Mavic 2 Pro was not the fastest at acquiring sats, normally could get 13-15 on average after around 2-3 minutes, my DJI FPV is very fast at doing the same in a lot less time, & my Mavic 3 is around the same speed as my Mavic 2 Pro - maybe a tad faster depending on cloud cover/weather.

maybe a tad faster depending on cloud cover/weather.
Clouds and weather conditions make zero difference to GPS reception.

Charles Adams Posted at 12-11 18:48
Would an appropriate test be to turn on M3, wait for lock, turn off M3, turn back on, and record time for lock, and repeat for my MA2?  And the times to compare are the 2nd locks (discarding the initial lock times for both craft)?

That would be a better way.
Because the satellites are orbiting in polar orbits and the earth is rotating on its axis inside teh cage created by the orbiting satellites, there will be different numbers of sats in your sky from one hour to the next.
Comparing one drone with the time taken by a different drone at a different time is like comparing apples to bananas.

The first thing I noticed when I received my M3 was the slow GPS position lock. I have launched all drones I own many times in my front yard and this was one of the slowest. I will delay judgement until after using the M3 for longer. Too busy with real life activities right now ;-)
Also I have not used it since the recent update.
And..... no I did not have the water rescue buoy attached those times ;-)

Labroides Posted at 12-11 20:00
That would be a better way.
Because the satellites are orbiting in polar orbits and the earth is rotating on its axis inside teh cage created by the orbiting satellites, there will be different numbers of sats in your sky from one hour to the next.
Comparing one drone with the time taken by a different drone at a different time is like comparing apples to bananas.

That makes sense, but assuming that each sat system source has a different cluster in the sky at different times, I would anticipate a variance in the duration to acquire lock on both craft, with a variance in which is faster.  If we do this test say 10 times for each craft, I would expect a fairly random distribution.  If one craft uniformly under performs the other, that would be an interesting data point.  The test would need to span some time just to assure a variance in the available cluster.  Say...  10 minute intervals?

Do we know how much coverage the Chinese BeiDou system has?

DAFlys Posted at 12-12 02:37
Do we know how much coverage the Chinese BeiDou system has?

Those of us that know how to use Google do.

DAFlys Posted at 12-12 02:37
Do we know how much coverage the Chinese BeiDou system has?

Apparently 35 satellites, in geosynchronous orbit (so the perceivable constellation does not change over time).  

I have not repeated this test sufficiently to be worthy of a "scientific" result, but my limited testing results support KlooGee's observations.  The first start from a location took a significant period to acquire lock (over a minute), but subsequent sat locks were dramatically faster.  I tried to time them with a stop watch, but I obtained lock before I could activate the clock.

I would say the proper test would be starting two different units simultaneously a few feet apart and comparing the results.

DarthSLR Posted at 12-12 07:50
I would say the proper test would be starting two different units simultaneously a few feet apart and comparing the results.

I might do that (M3 vs MA2).  However I am less motivated to test when subsequent time-to-locks after the initial are near instantaneous.  If the problem repeats, I will do so and report the results!

Charles Adams Posted at 12-12 08:27
I might do that (M3 vs MA2).  However I am less motivated to test when subsequent time-to-locks after the initial are near instantaneous.  If the problem repeats, I will do so and report the results!

Yeah, it's interesting how subsequent acquisitions are so much faster...

Yes, it now takes much longer to aquire enough satellites to record the homepoint as it did prior to the latest firmware update. I just performed a test this morning using first my Mavic 3 and it took 5:23:58 minutes to record the homepoint. I never test the Mavic 3 prior to the update but have noticed it being much longer compared to any other of my drones.

Flormo2002 Posted at 12-12 09:26
Yes, it now takes much longer to aquire enough satellites to record the homepoint as it did prior to the latest firmware update. I just performed a test this morning using first my Mavic 3 and it took 5:23:58 minutes to record the homepoint. I never test the Mavic 3 prior to the update but have noticed it being much longer compared to any other of my drones.

Wanted to follow up by saying I did the same testing using my Air 2S immediately following the Mavic 3 test and it only took 38 seconds to acquire enough satellites to record the homepoint. Hopefully further firmware updates will correct this issue

I just started seeing this problem recently. At first I didn't have the issue.  

I had a chance about 45 minutes ago to try this out from a different location.  I only tested with the M3 (I did not include the MA2).  On first startup I obtained satellite lock in 32 seconds.  I subsequently tested the new features provided by the latest firmware, and then landed and powered down (everything).  I then re-powered back up, and all systems started immediately with lock.

Charles Adams Posted at 12-12 09:56
I had a chance about 45 minutes ago to try this out from a different location.  I only tested with the M3 (I did not include the MA2).  On first startup I obtained satellite lock in 32 seconds.  I subsequently tested the new features provided by the latest firmware, and then landed and powered down (everything).  I then re-powered back up, and all systems started immediately with lock.

I wonder if there is an internal cache of sorts...

DarthSLR Posted at 12-12 10:41
I wonder if there is an internal cache of sorts...

During a cold start the GPS system in the drone also downloads the ‘almanac’ which is basically a time table of which satellites are going to expected be at any given time. This gives you the short warm start later.

This almanac changes over time and has to be updated periodically, which may give you the variations in start up time.

If the drone has moved a significant distance from when it was last powered it may download the whole almanac again. It's transmitted continually.

A cold start today for me at the coast in the UK took about 30 seconds, after a battery swap full GPS lock took about 5 or 6 seconds.

In the grand plan 30 seconds is nothing, in the early days of the GPS system you could wait a whole 90 seconds to get even a 3 satellite lock & due to Uncle Sam not wanting it to be used by the bad guys they introduced timing errors, which only they knew about so accuracy was rather poor!

Buzzyone Posted at 12-12 10:58
During a cold start the GPS system in the drone also downloads the ‘almanac’ which is basically a time table of which satellites are going to expected be at any given time. This gives you the short warm start later.

This almanac changes over time and has to be updated periodically, which may give you the variations in start up time.

Ah, that makes perfect sense. Thank you for this info!

DarthSLR Posted at 12-12 11:00
Ah, that makes perfect sense. Thank you for this info!

Since the last firmware update, I have also problems to get a good satellite position (flying always at the same spot!).

×



There must have been a software change in aquiring satellites!

Slow GPS Lock here too...After the FW update
Very sluggish inside my house..
Outside is better of course.
So I fire up my RC Pro Controller and start the Fly app.
Then start my M3. Go through pre-flight checks...
By then I usually have GPS lock....
As with any Radio Signals....  approach dusk... Improved reception..
The lock speed improves too!

Yes it takes forever to get satellites on the mavic 3 my air2s was faster.

NGC Posted at 1-24 12:04
Apparently the most recent fw update didn't fix it.

Not for me, at least, but I only ran the test ones, and not during my usual time. I'll check again later tonight, when I have much larger statistics flying

DarthSLR Posted at 1-24 12:10
Not for me, at least, but I only ran the test ones, and not during my usual time. I'll check again later tonight, when I have much larger statistics flying

Day or night makes no difference to GPS

Buzzyone Posted at 2021-12-12 10:58
During a cold start the GPS system in the drone also downloads the ‘almanac’ which is basically a time table of which satellites are going to expected be at any given time. This gives you the short warm start later.

This almanac changes over time and has to be updated periodically, which may give you the variations in start up time.

My times completely match what you’re experiencing, I’m northwest coast Ireland

Labroides Posted at 1-24 14:09
Day or night makes no difference to GPS

I was not referring to the visibility or otherwise optical quality of my surroundings.
But according to https://www.uavforecast.com/ there is a different number of satellites available for any given time, so comparing one time of day to another is not always apples to apples.

DarthSLR Posted at 1-24 14:14
I was not referring to the visibility or otherwise optical quality of my surroundings.
But according to https://www.uavforecast.com/ there is a different number of satellites available for any given time, so comparing one time of day to another is not always apples to apples.

Don't waste any time looking at satellite numbers in UAV forecast.
There are far more satellites than needed in your sky at any time of the day or night.
Even more so with the Mavic 3 accessing 3 different satellite systems.

UAV Forecast has lots of things that are irrelevant and sat numbers is one of them.

Labroides Posted at 1-24 15:00
Don't waste any time looking at satellite numbers in UAV forecast.
There are far more satellites than needed in your sky at any time of the day or night.
Even more so with the Mavic 3 accessing 3 different satellite systems.

Really? Interesting...
TBH, thus far I have observed a decent correlation between the forecast and my actual number of sats.
Not in pure scientific sense - as in, I didn't run tests specific to this issue, but in a general feeling I have noticed that different times of days act differently for me, and if the forecast says I'd have fewer sats in general, that would be my "gut feeling" as well.
I guess I need to be honest with myself and actually run the test and compare my findings with the forecast. Just need to find time to do this, usually too busy with other things...

DarthSLR Posted at 1-24 15:06
Really? Interesting...
TBH, thus far I have observed a decent correlation between the forecast and my actual number of sats.
Not in pure scientific sense - as in, I didn't run tests specific to this issue, but in a general feeling I have noticed that different times of days act differently for me, and if the forecast says I'd have fewer sats in general, that would be my "gut feeling" as well.

TBH, thus far I have observed a decent correlation between the forecast and my actual number of sats.
The number of sats will vary due to the orbital patterns and the earth's rotation inside the cage they form around the earth.
But you won't ever find a time when there aren't many more than you need.

The need to check numbers of sats goes back to the early days when there was only one system in place and the full number of sats hadn't been launched.