Samakopita Posted at 4-4 14:50
I got it first! Post #95!

Are you kidding me?

Labroides Posted at 4-4 05:33
Perhaps if I said: a practical difference.
Or maybe I could have given a few sentences to indicate that I meant anything that will actually improve things for users.
You know ... something that makes a difference to your use of the M3.

Am I speaking a foreign language or something?


No, what you asked was perfectly clear to anyone who was not trying to ignore.

thank you for sharing... long footage, interesting info...

HedgeTrimmer Posted at 4-4 17:11
Am I speaking a foreign language or something?

I was definitely not trying to ignore.  Those who know me here on these forums know that I do not instigate or use passive aggressive techniques.  I made various good faith efforts to answer the question I thought was being asked, and this is demonstrated by my genuine answer when the adjective "practical" was used.

Regardless of the various personality conflicts that exist and manifest in these forums, I have respect for Labroides, and value his contributions.

Labroides Posted at 4-4 14:52
Dji started manufacturing the Mavic 3 with the Professional version of the UBLOX chip, they ran short of chips and used the automotive version in the rest of production hence the issue of some have very fast lock times and others not.
There should be no practical difference between the chips.
The only performace difference is that the automotive grade has a higher temperature tolerance.

Should be.  Problem is some chip makers will "Fudge" ever so slightly on their chips involving tolerances.  Not only operating temperature and voltage ranges, but clock speeds, length of time a state is held, and what voltage transition does a change of state occur.   Bragging to make a sale?

TonyPHX Posted at 4-4 15:27
They can fix it a variety of ways.  

1.  Firmware to resolve whatever they mucked up.

#1 a real possibility.
#2 not so much
#3 highly unlikely

Samakopita Posted at 4-4 16:24
Mad is currently perusing hacker boards asking for help because he can’t get anywhere with the ublox chip. That’s why there is nothing for him to “add in more detail” right now. Nice to see him making an effort though, despite not being the savior that Hallmark made him out to be.

Something to keep in mind or understand as to where things could go astray.
ublox operates as a fabless IC and design house.  Meaning, they outsource making of ICs used.  
More hands in the pot.

Charles Adams Posted at 4-4 14:52
I have suspected that the "compass calibration" solution was more a placebo, though I've not tested.  My suspicion is that if one is calibrating their compass, they are taking time to do an unrelated activity while giving GPS more time to lock, and then experiencing "fast" lock because they took time to spin, turn and rotate.

The test would be to calibrate the compass, and then wait sufficiently to do a cold start in some location and determine if the GPS Lock occurs faster.  I have not performed this test because I do not believe it will work.

Had similar thought when a video was posted of a guy claiming he had solution by flying his drone up/down close to ground to supposedly improve GPS lock time.

9 pages, of which 8 are eternal clarification between the local monteks and capulet. it's very hard to read this in almost every thread on this forum - unnecessary info ...

on the topic, well, this is positive, IMHO, we are again waiting for a new firmware ...

Colleagues, when do you have time to fly? )
These disputes must take a lot of time?

Tornado12 Posted at 4-4 10:36
You remind me so much of teenagers in a video game forum, after the game released with a bunch of bugs. So passionate, so much emotion, but zero substance, and complete immaturity in your perspective.
You are too emotional.

I still say the issue is firmware and will be fixed. There is literally zero evidence of hardware issues so far. Absolute zero. The more we learn actually points more and more toward firmware.

Perhaps.  Unusal for same firmware running on same hardware to produce such disparent results.   Granted, there are external issues such as GPS signal interference, system date/time being off from actual GPS satellite date/time, large change in distance in short period effecting starting location guess from previous lock.

Samakopita Posted at 4-4 16:24
Mad is currently perusing hacker boards asking for help because he can’t get anywhere with the ublox chip. That’s why there is nothing for him to “add in more detail” right now. Nice to see him making an effort though, despite not being the savior that Hallmark made him out to be.

Ok lets cut some of the nonsense shall we.

I have spent over 60 hours on this since receiving the module and have learnt a LOT about the module, so much more in the last 48 hours as well having spent a lot of that buried in Unlox manuals and then speaking  with some one who designs GPS modules for a living.

While im not going to go into full details now I will say the module I have has no faults, issues or undesirable behaviour that's not expected once you fully understand some of Ublox own limitations and quirks., its neither the best or worst from a performance point of view and any talk of the M8030 being a problems is complete and utter nonsense period. The Ublox core is fine.

Also the lack of SQI flash has little to no impact either and is normal for DJI. Cold boot times ever with out Almanac is negligible compared to what people are seeing.

On the live stream I shared a little look at the HW and I knew someone would spin that into more noise but I don't hide what I do and it never intended to be anything but a look. thats it.

The plan for the video is this week or early next where I will be covering the full module, its performance in comparison to others as well as specifics of the system and my overall thoughts on what's going on as I do have some.

As for the posts on slack ect, Yes I have asked for some info from the guys on the groups for one specific thing and that's the DJI config for the module as there is no way to get that from anywhere but the craft firmware.  Its not because im struggling or cant understand  its because I want to see exactly the Ublox CFG DJI are using.   As the Mavic 3 module has no SQI flash it does not store the Ublox config, a DJI are loading the Ublox config at boot on the Mavic3 every time.  There is a massive amount of varying performance on all GPS modules based on its config.  Power setting, rate, filtering and more play into how a GPS module will perform.  Positional accuracy can be affected dramatically as a result of these settings.

So the mythical video with all the answerers will come next week

Actually no it wont, If you want absolute answers your not going to get it anywhere, What I will be addressing is the accusations on the HW, SW, showing its performance comparted to other models, showing data on that and then educating and informing on some what I THINK may be going on here and what that means for the M3 and what next.

I will say now its far to simplistic to say its just one thing wrong as the change to Beido and Gal has implications in its self, its not apples for apples compared to GPS/Glonanss.

If your looking for definitive outcome your not going to get it, but you will learn something.

Samakopita Posted at 4-5 14:25
So....the short summary is you have no idea what's wrong.
Could be hardware, could be software.

Please really do not watch. For those who like to make up lies, nonsense and hearsay the actual data presented will not suite your agenda.

Mads Tech Posted at 4-5 14:41
Please really do not watch. For those who like to make up lies, nonsense and hearsay the actual data presented will not suite your agenda.

Looking forward to viewing the video. Thank you! You have taken a lot of your own time out on this

Something I just want to expand on a little is the DJI Ublox Config.

Ublox modules have various settings that will affect the overall positional accuracy and acquisition times of the module.

For instance the Power Settings and Rates especially have a dramatic affect on HDOP and actual satellite lock too.

In the last update DJI said they tweaked the settings. One of the reasons I have asked for info from the guys who look at the firmware is id love to know exactly what they changed and test the impact of that my self .  

Furthermore all my tests have been done with the module is power save, balanced and how power settings but again we don’t know what DJI have chosen.  I would love to run their specific config and while I can’t do that today if I have too I will try to buy a damaged M3 and poll the UART when active to see what it is.  

As I said earlier the testing I am doing is a process of understanding and then verification.  On the hardware side I have even looked at the clocks on the oscilloscope to ensure its all as its should be. It’s not about being able to say hay I did it, it’s not about being right it’s about facts and understanding. There is simply too much hyperbole on this issue and it’s also imo much less widespread than the forums will have you believe.  In the end DJI were never likely to have made major mistakes but they have changed enough that the mix of that change could be having an impact we are seeing with some users.  

What I will say is if they were running GPS/Glonass these issues likely would not exists but again there are many reasons for that being the case as I will explain later.  

Mad_Angler1 Posted at 4-5 14:13
Ok lets cut some of the nonsense shall we.

I have spent over 60 hours on this since receiving the module and have learnt a LOT about the module, so much more in the last 48 hours as well having spent a lot of that buried in Unlox manuals and then speaking  with some one who designs GPS modules for a living.

I thank you Mad_Angler1 for taking time to dig into this and for providing information that we would otherwise not have.  I enjoyed your video, I learned from it, and I look forward to your next one.

Mads Tech Posted at 4-5 15:11
Something I just want to expand on a little is the DJI Ublox Config.

Ublox modules have various settings that will affect the overall positional accuracy and acquisition times of the module.

Cheers MAD’S “It takes strength and courage to admit the truth.”  looking forward to what else you might find.

Mad_Angler1 Posted at 4-5 14:13
Ok lets cut some of the nonsense shall we.

I have spent over 60 hours on this since receiving the module and have learnt a LOT about the module, so much more in the last 48 hours as well having spent a lot of that buried in Unlox manuals and then speaking  with some one who designs GPS modules for a living.

Actually no it wont, If you want absolute answers your not going to get it anywhere, What I will be addressing is the accusations on the HW, SW, showing its performance comparted to other models, showing data on that and then educating and informing on some what I THINK may be going on here and what that means for the M3 and what next.
No-one really cares what tiny little tweak caused the issue.
Knowing it (or not)  makes no difference to users.
Obviously it's complicated - it's had DJI stumped for months.
The only thing that matters to users is: will it be fixed?

If your looking for definitive outcome your not going to get it, but you will learn something.
Something that makes any difference?
Ideally just a  few paragraphs long that just cuts to the point?

Mad - this is EXCELLENT work and appreciate you digging in.  Honestly, I will never get as deep into the hardware as you have, but love that you are the type to share your knowledge.  Thanks for what you aredoing and hopefully we will meet up at a drone meet or conference someday so I can treat for a meal.  This kind of community sharing is why I check in here.  It is great to learn new things!!!

Mads Tech Posted at 4-5 15:11
Something I just want to expand on a little is the DJI Ublox Config.

Ublox modules have various settings that will affect the overall positional accuracy and acquisition times of the module.

Mad, thank you for taking your personal time to find an answer, much appreciated

Mad_Angler1 Posted at 4-5 14:13
Ok lets cut some of the nonsense shall we.

I have spent over 60 hours on this since receiving the module and have learnt a LOT about the module, so much more in the last 48 hours as well having spent a lot of that buried in Unlox manuals and then speaking  with some one who designs GPS modules for a living.

We (I) really appreciate your efforts Mad_Angler and it's refreshing listening to a fellow like you that obviously is very tech orientated and knows his stuff. Questions for you: What has your career been most of your life where you pick up stuff like this? Just a hobby instead?
Is there any truth from what I read (could be fake) that DJI mentioned that some USAF Satellites were not positioned correctly and they were kind of blaming things on that?

Your efforts are greatly appreciated, by me anyway.

Tom

Labroides Posted at 4-5 16:37
Actually no it wont, If you want absolute answers your not going to get it anywhere, What I will be addressing is the accusations on the HW, SW, showing its performance comparted to other models, showing data on that and then educating and informing on some what I THINK may be going on here and what that means for the M3 and what next.
No-one really cares what tiny little tweak caused the issue.
Knowing it (or not)  makes no difference to users.

No-one really cares what tiny little tweak caused the issue.
Knowing it (or not)  makes no difference to users.
Obviously it's complicated - it's had DJI stumped for months.
The only thing that matters to users is: will it be fixed?

You'll be surprised how many people enjoy listening to how the OTTO cycle works, while their cars broke down on the side of the road...

Mads Tech Posted at 4-5 15:11
Something I just want to expand on a little is the DJI Ublox Config.

Ublox modules have various settings that will affect the overall positional accuracy and acquisition times of the module.

First off nice test probe jig.  We had to either solder on enameled (tiny coil) wire and clip too, or hold a scope probe steady bracing against module tester cooling bars.

With that said, I don't understand how you can test Ublox board without it running DJI's MP3 config setup and data being read at rate DJI MP3 does.  Along with issue of you possibly having a stellar Ublox board, while some others have gotten less than stellar.   

HedgeTrimmer Posted at 4-5 19:01
First off nice test probe jig.  We had to either solder on enameled (tiny coil) wire and clip too, or hold a scope probe steady bracing against module tester cooling bars.

With that said, I don't understand how you can test Ublox board without it running DJI's MP3 config setup and data being read at rate DJI MP3 does.  Along with issue of you possibly having a stellar Ublox board, while some others have gotten less than stellar.

Good points from you as well

HedgeTrimmer Posted at 4-5 19:01
First off nice test probe jig.  We had to either solder on enameled (tiny coil) wire and clip too, or hold a scope probe steady bracing against module tester cooling bars.

With that said, I don't understand how you can test Ublox board without it running DJI's MP3 config setup and data being read at rate DJI MP3 does.  Along with issue of you possibly having a stellar Ublox board, while some others have gotten less than stellar.

Actually there is a huge amount I have learnt and knowing DJIs config would be nice but it’s not the be all and end all.  We know what the configuration roughly will be based on Ublox recommendation. I will also say there is no reason to think DJI has made a massive error with the Ublox config, that just don’t make sense. But again that’s not to say there are not things at play here by moving to BeiDou and Gal on its self.  There are things I have seen and observed that need explaining but will be playing into it imo.  

I’m also going to ad now I don’t there there is a “fix” as such of what I think is happening is and it’s just inherent behaviour, that may be able to be tweaked or optimised but I need to see the config as I have learnt the best overall now my self however of DJI have done that then it it what it is.  Is is not as simple as replacing Glonass with BeiDou and Gal will just be the same.  There is differences in many aspects of the config and behavior.  

Mads Tech Posted at 4-6 03:09
Actually there is a huge amount I have learnt and knowing DJIs config would be nice but it’s not the be all and end all.  We know what the configuration roughly will be based on Ublox recommendation. I will also say there is no reason to think DJI has made a massive error with the Ublox config, that just don’t make sense. But again that’s not to say there are not things at play here by moving to BeiDou and Gal on its self.  There are things I have seen and observed that need explaining but will be playing into it imo.  

I’m also going to ad now I don’t there there is a “fix” as such of what I think is happening is and it’s just inherent behaviour, that may be able to be tweaked or optimised but I need to see the config as I have learnt the best overall now my self however of DJI have done that then it it what it is.  Is is not as simple as replacing Glonass with BeiDou and Gal will just be the same.  There is differences in many aspects of the config and behavior.

I’m also going to ad now I don’t (think) there there is a “fix” as such of what I think is happening is and it’s just inherent behaviour.

Are you suggesting that the Mavic 3 is a lemon and that's all there is too it?
That those that bought the Mavic 3, just have to put up with unacceptable GPS performance?
How does this match with DJI's claim that the Mavic 3 comes with a powerful positioning algorithm that enables Mavic 3 to lock onto multiple satellite signals faster than ever?

StanMaster Posted at 4-6 04:07
So when all is said and done...no ultimate reason -  and it is a lemon.
That was said at the very beginning of this mess - they were only arguing of their opinions of "why" it was a lemon. So much for the White Knight coming on a horse to report his findings to his special friends at DJI....

And if it is a lemon ... DJI would have been aware of this for quite a while.

StanMaster Posted at 4-6 04:07
So when all is said and done...no ultimate reason -  and it is a lemon.
That was said at the very beginning of this mess - they were only arguing of their opinions of "why" it was a lemon. So much for the White Knight coming on a horse to report his findings to his special friends at DJI....

But people need a white knight to tell them the Mavic3 is NOT a lemon, and whoever keeps saying it's a lemon will keep getting banned from this forum...

Mads Tech Posted at 4-5 15:11
Something I just want to expand on a little is the DJI Ublox Config.

Ublox modules have various settings that will affect the overall positional accuracy and acquisition times of the module.

Hello and thank you for your efforts! As we well know, we do not receive any information from DJI, so like other colleagues, I welcome your contribution.

Personally, I am tormented by the following question: Why can't my MAVIC 3 lock any satellite under ideal conditions in the first 40 seconds after power on?
I will be happy if you can share the behavior of the test GPS module mounted on a real Mavic 3. I mean how long it will take to record a home point. I am interested because I suspect that you may have a much better copy of the GPS module, than we have affected customers.
Out of pure sporting interest, what do you think DJI has changed in the configuration in 0600 firmware?

I am waiting for your video and I hope you will post your findings in this forum as well ... just my English is not good.

Monkey007 Posted at 4-6 04:25
But people need a white knight to tell them the Mavic3 is NOT a lemon, and whoever keeps saying it's a lemon will keep getting banned from this forum...

Captain Monkey007, I'm aware of one person who has been banned (repeatedly), while others who have made similar assertions have not been banned.  The reason that this person was banned was for different behavior unrelated to the assertions made regarding the M3.

I am confident that if one's behavior was limited to harsh critique of the M3 and did not include other poor behavior, such a person would not be (and in fact has not been) banned.

KokoFresha Posted at 4-6 06:33
Hello and thank you for your efforts! As we well know, we do not receive any information from DJI, so like other colleagues, I welcome your contribution.

Personally, I am tormented by the following question: Why can't my MAVIC 3 lock any satellite under ideal conditions in the first 40 seconds after power on?

Just a question.  Why did you think it should lock in under 40s ?  Hot start yes but warm and cold start is more than reasonable for it to take 1-2 minutes for goodlow HDOP.

Super fast acquisition times as NOT good or normal.  Ublox specs of time taken is not actually to extremely accurate time with in 2-3m.  GPS 3D lock can happen at HDOP as much as 5 and Accuracy of 5+ meters.  That can happen in 40s but getting down to much lower numbers takes longer on most modules.  That is normal.

I think people need to have a better understanding of what is normal and what is not.  


I also want to add I never told anyone I would be telling you what’s wrong with the Mavic 3 or I will be providing that to DJI. I said I would be addressing the questions on the module and then giving an opinion.  

Frankly anyone seeing GPS home point lock times below 2.5 minutes has no problem at all imo. Anyone over that is not as it could be or could be better.  

Mads Tech Posted at 4-6 08:40
Just a question.  Why did you think it should lock in under 40s ?  Hot start yes but warm and cold start is more than reasonable for it to take 1-2 minutes for goodlow HDOP.

Super fast acquisition times as NOT good or normal.  Ublox specs of time taken is not actually to extremely accurate time with in 2-3m.  GPS 3D lock can happen at HDOP as much as 5 and Accuracy of 5+ meters.  That can happen in 40s but getting down to much lower numbers takes longer on most modules.  That is normal.

I cannot answer for the 40s, and I am not smart enough to know what the optimal time is reasonable for the optimal lock.  I will concede that DJI has opened themselves up to criticism (justifiably so) for making a claim "This enables the Mavic 3 to lock onto multiple satellite signals faster than ever" and then not meeting that claim.  The M3 is by far the slowest "locking" drone I have ever owned.  Not a biggie to me, but it's a "huge" issue to others.

I also acknowledge that your efforts are not intended to "defend" DJI in any way, and that you are just using your time and expertise to understand the details of what is happening.

Mads Tech Posted at 4-6 08:40
Just a question.  Why did you think it should lock in under 40s ?  Hot start yes but warm and cold start is more than reasonable for it to take 1-2 minutes for goodlow HDOP.

Super fast acquisition times as NOT good or normal.  Ublox specs of time taken is not actually to extremely accurate time with in 2-3m.  GPS 3D lock can happen at HDOP as much as 5 and Accuracy of 5+ meters.  That can happen in 40s but getting down to much lower numbers takes longer on most modules.  That is normal.

ok no problem, mad.  just make sure when somebody posts that "mad's scientific research has conclusively show" such and such or "mad did what dji couldn't and therefore...." or "mad proved that the problem isn't with...." or "mad clearly said..." please jump in there and quickly correct them; thanks.  because that would be most helpful to keeping the conversation civil and on the right track.  appreciate it.

we all have dji drones, air 2, mavic 2 pros and zooms, air 2s.... i think we all know what a good quick gps lock looks and feels like.
if my air2 starts taking 3 minutes to lock gps, it's going in the trash.

Mads Tech Posted at 4-6 08:40
Just a question.  Why did you think it should lock in under 40s ?  Hot start yes but warm and cold start is more than reasonable for it to take 1-2 minutes for goodlow HDOP.

Super fast acquisition times as NOT good or normal.  Ublox specs of time taken is not actually to extremely accurate time with in 2-3m.  GPS 3D lock can happen at HDOP as much as 5 and Accuracy of 5+ meters.  That can happen in 40s but getting down to much lower numbers takes longer on most modules.  That is normal.

For 40 seconds my Mavic 3 doesn't catch a single satellite, let alone record a home point. By this time, the other drones had already recorded a starting point. I think the inconvenience is obvious.
If you claim that everything is great with this module, just DJI have configured more accuracy, I will ask you why DJI did not do it when the engineering samples were tested by bloggers. I will answer you: they couldn't help but mention it in their reviews.

KokoFresha Posted at 4-6 09:56
For 40 seconds my Mavic 3 doesn't catch a single satellite, let alone record a home point. By this time, the other drones had already recorded a starting point. I think the inconvenience is obvious.
If you claim that everything is great with this module, just DJI have configured more accuracy, I will ask you why DJI did not do it when the engineering samples were tested by bloggers. I will answer you: they couldn't help but mention it in their reviews.

I think if that’s the case then you may have a more serious problem . Unless it’s not shown actual numbers it might be shown as in the reception guide . Look at the aircraft Icon see if it lights up that will give you an idea of the GPS health. If there are no lights then it could be worth getting it checked.