1)RC Pro in Auto Mode does not use 5GHz band as far as I can tell and I get frequenbt disconnects in 2.5GHz
2)When I switch to Manual and select 5GHz the connection strength is lower (expected) but no disconnects. Switch back to Auto and only 2.4GHz and drops again
3)RC Pro (not connected to M3C) does not locate position utilizing internal GPS. BUT, when I use Find My Drone, the map comes up and is very close without M3's input
4)Once the RC Pro knew where I was I started M3C - no change in satelitte aquisition time
     - On RC Pro map, relitive position is correct within ~15 feet but continues to change within that radius even with 8 satelittes
     - Locks at 12

My findings are certainly different from others and the issues causing this can only be theroies on my part.

A) Connection issues between remote and M3C
     - Auto switching on bands does not appear to work. Significant nbumber of close field drops on 2.4 and none on 5.8GHz when on Manual

B) GPS in RC Pro is working but not communicating it's position to the M3C (if it is supposed to). Since the RC Pro can quickly locate position within a ~15 foot (+/-) radius very quickly when "Find My Drone" is used, if the units are supposed to communicate (which I believe they are), that aspect is not working

C) Satellite aquisition time to lock using all band antenna on tripod as a portable mountpoint - no internet
     - The M3C is showing similarities to how UBlock on my RTK device (satellite location over the horizon, signal strength, type of downlink info, etc.)
     - My RTK device is set for .5 meter accuracy with a time "window" of 30 seconds. I have it set now for a minimum of 12 Sats at 75% signal strenth holding for the full 30 seconds
     - Set-up is open sky with blocked horizon inclination of ~30 degrees
     - Generally, from a cold start, lock takes 10-12 minutes
     - When I change the parameters to 1 meter, time to lock is ~6-8 minutes
- When I change the parameters to 2 meters, time to lock is ~2 minutes

D) Potential (and only in my opinion) Conclusions - Kind thoughts and input appreciated ONLY
     - Latest RC Pro (and probably standard remote) firmware releases broke the connection for GPS data between the aircraft and the remote GPS devices (phone on the standard remote)
     - Disconnects (at least on the Pro as I don't have the standard controller) may be occuring due to firmware issues with signal level sensing and auto switching on the 2.4 and 5.8 GHz bands
     - I'm going out on a limb here but starting with the December update, the GPS accuracy requirement was increased and firmware changed in one or all of the following ways:
           - Signal strength minimum within the time window
           - Reduced the position parameter from 2 meters to 1 meter or less
           - Positional accuracy requirement based on the received satellites positions vs one another (wide spread map)
           - Broken data sharing between the GPS in the remote (Pro) and aircraft
           - Removal of the GLONASS constellation amd substitution of BeiDou (less satelittes at this time)

As the GPS time to lock issues started with the December firmware, it is a pretty strong indication that it is not a hardware issue as most here have said.  So, hopefully a firware fix in short order.

The disconnect issue on the RC Pro may be a hardware issue of some type that does not affect all units. I hope not as DJI will need to replace these units - likely early production (like mine). If that is the case..perhaps a published series of SN#s will be given to support to allow for replacement claims to be easier. Hopefully just a firmware tweak.

FYI - the more of us that send .dat logs to DJI the better for them to get a better data set on the issues. I have been using a OneDrive share link for them.

Ugh...no spell check - apologies for typos

Really appreciate the thorough testing.  It is not easy to take the time and fully document test measurements,  I agree with your statements about it being a firmware related issue that has potential to be resolved.

Nice Analysis, very well detailed. We all agree in one thing in every GPS thread I have read both here and in Mavic Pilots and that is that after the December update, GPS lock has got stuffed up. DJI, HEAR Your customers and fix the issue. We cannot be sitting around and waiting for over 6 minutes to be able to fly. Can I get a 6% discount on what I paid for the drone? I and every other pilot did not pay premium to wait..wait...wait for us to get our bird in the air. Even my Autel locks full SATS on cold start in 30 seconds and is ready to go. Fix this Dji, no more playing around.

Hello there EFRPIC. Good day and thank you for sharing these informative information with us and for giving out your valued insights with regards to these topics. Again, thank you for your understanding and for your valued support.

Thanks for the thorough test.   I think I will hold off upgrading for now.

I suspect we will see an improvement within the next few weeks.

So do you think the RC pro should be communicating with the drone for a faster connection?

I def agree that since the newer upgrade I am having bigger problems with my Mavic3 and RC pro with screen freeze and rc disconnects that I am not having on th air2s I am going to try the regular controller with the Mavic 3 and will report back.  my gut is saying it is the RC remote issue though.  Thanks for all of that info!

Hallmark, yes I believe it should be. But if it ever did - it is not now.

Have you checked to see if your set-up ever goes to 5.8GHz without being in manual?

EFRPIC Posted at 2-8 09:42
Hallmark, yes I believe it should be. But if it ever did - it is not now.

Have you checked to see if your set-up ever goes to 5.8GHz without being in manual?

I’m flying tomorrow I will check and come back

Very detailed analysis.  Thanks for sharing.  

EFRPIC Posted at 2-8 09:42
Hallmark, yes I believe it should be. But if it ever did - it is not now.

Have you checked to see if your set-up ever goes to 5.8GHz without being in manual?

I have checked today, in Auto it stays in 2.4ghz . Now where I live is predominantly 2.4 used, it didn’t jump to 5.4ghz. But I flew manually in 5.4ghz and reception was fine.

Starting the drone cold . I turned on the drone removed RC pro from my bag placed sticks in gimbals started Remote and had 10 sats and 5 seconds later I had gps lock and could take off. The maximum it took was 90 seconds

Hallmark,

I need to test a theory on powering up the aircraft for approx 2 min and then the controller. Then extend it to 3 minutes and see if any differences occur. Using the same location, of course.

I want to rule out that the RC Pro and the aircraft are putting out data that conflicts somehow until a benchmark is reached.

If there even is communications between the 2 GPS modules.

EFRPIC Posted at 2-9 10:53
Hallmark,

I need to test a theory on powering up the aircraft for approx 2 min and then the controller. Then extend it to 3 minutes and see if any differences occur. Using the same location, of course.

Not that this precisely follows your test, but I just tried letting my craft sit for 5 minutes powered up, then turned on the RC (attached to phone), and saw that I had 15 satellites immediately.  I'm going to decrease the time my craft is powered up before turning on the RC by 1 minute, but I have to wait for a while to let the cache "empty".

Interesting post. I note one inaccuracy that I keep seeing though regarding the GNSS. People keep saying that BeiDou has less satellites than Glonass. This isnt true. In fact, Beidou currently has more sattelites in orbit than all other GNSS systems at 35. The American GPS system has 31. The European Galileo has 24 and the Russian Glonass has 24. The Chinese BeiDou is the newest of these systems, and on paper claims one of the highest levels of accuracy. I just keep seeing people refer to BeiDou as if it has only a few sats up, or is unfinished or something. Im not sure why this is. Dropping Glonass for BeiDou should have actually improved the mavic 3's gps performance, not hurt it.

I will be testing with the RC Pro. I believe that when you pull the battery it will initiate a cold start. I'm not sure if the GPS chip has a small back-up CAP power supply.

Tornado - you are correct on the size of the constellations. It's their viewability above the horizon in a specific location I would suspect...if anything.

EFRPIC Posted at 2-9 12:02
Tornado - you are correct on the size of the constellations. It's their viewability above the horizon in a specific location I would suspect...if anything.

for sure. I really do not think though that the gps issues we are seeing are the result of them dropping Glonass for BeiDou however. I think for just about everyone around the world, this should have netted an increase in performance on the gps system rather than a decrease. The BeiDou is the better constellation at present.

Now....This doesnt mean that they dont have some bug that has resulted from the switch to BeiDou from Glonass, but if you just look at Glonass versus BeiDou on paper, there should be no loss of performance as a result of the change. I have seen in many places however people suggesting this could be the issue itself.

Thank you for the thorough testing and analysis!

We can surmise, postulate, think and rethink this. It really will not make any difference except for DJI knowing there is discontent with the current performance of the M3. The forum here is a small (hopefully valuable) source of data points for them.

What we don't know is how many units have been sold, how many complaints, returns, etc. they have had and what metrics they are using.

This forum is very vocal. I'm not sure how many other forums and direct communications they have had with dissatisfied customers and vendors (BestBuy, etc.).

If it was 1000s of responses about a specific issue, they would likely do a hot fix asap. If it is not setting off "disaster sirens" for them, they will take the time necessary to solve the firmware/software issue without, hopefully breaking some other feature(s).

Most likely the two firmware updates (December and January) were marketing driven. They had to get the features working to avoid a significant blowback. December was a sunset that was ready then and January was likely the roadmap release that was planned. What happened is a guessing game.

My guess, they did not catch some downstream issues (GPS, disconnects, etc.) that started in December when they released the January package. Now DJI needs time to fix, test and fix. They will not likely rush this next release. They are basically feature complete now and don't want to distribute anything that will not improve the situation.

My guess, 30-45 days from the end of the NY celebration. I'd say mid to late March (that's 2022 :-) )

EFRPIC Posted at 2-9 14:27
We can surmise, postulate, think and rethink this. It really will not make any difference except for DJI knowing there is discontent with the current performance of the M3. The forum here is a small (hopefully valuable) source of data points for them.

What we don't know is how many units have been sold, how many complaints, returns, etc. they have had and what metrics they are using.

I suppose it might become a real problem for them if sales targets are not met because of underlying problems still plaguing the M3. This could force the issue.

Agreed, that is a good distillation of what I took too long to say. :-)

Sales Quotas, unsold inventory on the books, returns would drive the speed of decisions.

With +/- 70% of the consumer market, they have far less risk as long as the issues get solved withing a "reasonable" time frame

Hi all, just purchased a brand new RC Pro here in Australia to use with my M3 and yeah, same problem,  good stable connection with 5.8ghz but terrible dropping/signal loss when using 2.4 and auto dual band. This is even when using the drone out in the western desert of Western Australia, 500kms from nearest town. Good connection for 5 min then drop out, then it drops out every 1-2 min consistently. Pretty much unusable on the 2.4 dual band mode but perfect on 5.8ghz....