Looking for standalone Mini 3 for waypoint (MAVLink) app tests

[TT3023]

Hi. As described pretty much starting at:

Litchi on Mini 3 Pro with new controller?

That might be why they want to ‘mainstream’ the RC controller with screen built in: no more phone updates breaking compatibility. Providing mission-critical drones even sub-1000 dollar

mavicpilots.com

... an Android app I work with brought waypoints to the Mini 2 long before it finally was supported in the DJI MSDK (accomplished via device spoofing). It provides a MAVLink interface. For my own builds I've only tested with QGroundControl (QGC), but in theory other MAVLink apps and even python-based systems (e.g. DroneKit) should work.

I recently brought the app up to the latest DJI MSDKs which support the Mini 2 natively, and I'd like to test to bring support to the Mini 3 using the techniques originally used for the Mini 2 pre-MSDK. Given that it took over a year for DJI to support the Mini 2 in the MSDK, who knows how long Mini 3 support will take? Could it be faster? Sure, but probably not good to hold one's breath. So this app may bring waypoints right away.

This work is at a standstill now until I have a Mini 3 in hand. I don't need any controllers -- this work requires use with an RC-N1 of course and obviously I already have one of those. But it appears that standalone Mini 3s are going to be the last ones shipped by DJI -- they currently quote end of August!

If anyone knows of a source for a standalone Mini 3 (FCC version) that could be obtained more quickly, it would be much appreciated, because I'm very interested to see if I can get QGC-based waypoints going with it now. Before anyone suggests just getting the other packages, money is an issue, so I really need to keep this as rock bottom as possible. Thanks very much.

 

[MaximusCZ]

I know you are working with N1 controller, any chance those hacks could translate to RC(non pro)?

 

[TT3023]

AFAIK, the new RC can't run *any* third party apps at all at this time. Also if that changes in the future, it apparently only has 8GB of memory. I thought I read (though I can't find the reference right now) that only 2GB of that is free. So it might be cramped even if the ability to run third party Android apps is enabled by DJI (or hacked by someone else). Of course I wouldn't mind have the RC sitting around for that possible future eventuality, but like I said I have to keep my spending on this effort to the absolute minimum.

 

[globetrotterdrone]

Of course I wouldn't mind have the RC sitting around for that possible future eventuality, but like I said I have to keep my spending on this effort to the absolute minimum.

Same for me ...
But let's face it, it's more or less another SC in disguise with O3 but this time at least no promise or premium prize from DJI. I am pretty confident, that the RC will most likely be Mini 3 Pro only or may a bit backward compatible with Air 2S, Mini 2 ... but let's see in 6 months or so.

They clearly designed it that even with jailbreaking it, you will somehow run out of ressources pretty quickly.

 

[LakesGuy]

Hi. As described pretty much starting at:

Litchi on Mini 3 Pro with new controller?

That might be why they want to ‘mainstream’ the RC controller with screen built in: no more phone updates breaking compatibility. Providing mission-critical drones even sub-1000 dollar

mavicpilots.com

... an Android app I work with brought waypoints to the Mini 2 long before it finally was supported in the DJI MSDK (accomplished via device spoofing). It provides a MAVLink interface. For my own builds I've only tested with QGroundControl (QGC), but in theory other MAVLink apps and even python-based systems (e.g. DroneKit) should work.

I recently brought the app up to the latest DJI MSDKs which support the Mini 2 natively, and I'd like to test to bring support to the Mini 3 using the techniques originally used for the Mini 2 pre-MSDK. Given that it took over a year for DJI to support the Mini 2 in the MSDK, who knows how long Mini 3 support will take? Could it be faster? Sure, but probably not good to hold one's breath. So this app may bring waypoints right away.

This work is at a standstill now until I have a Mini 3 in hand. I don't need any controllers -- this work requires use with an RC-N1 of course and obviously I already have one of those. But it appears that standalone Mini 3s are going to be the last ones shipped by DJI -- they currently quote end of August!

If anyone knows of a source for a standalone Mini 3 (FCC version) that could be obtained more quickly, it would be much appreciated, because I'm very interested to see if I can get QGC-based waypoints going with it now. Before anyone suggests just getting the other packages, money is an issue, so I really need to keep this as rock bottom as possible. Thanks very much.

Because of altitude restrictions in my area due to a Class D airport, coupled with urban density and mature trees, my drones have quite limited range. I often fly using waypoints through third-party apps such as Litchi, Dronelink, and my favorite, Maven. The virtual sticks limitations of the newer DJI drones has led me to purchase both a used Mavic 2 Pro and a Mavic 2 Zoom on account of their native waypoints capabilities (my Air 2 is for sale). I am sure someone on this forum can explain why an enterprising drone techie cannot design and build a chip to provide that missing capability as it seems as though something like that is going to be required for retrofitting older drones to comply with RID in the near term, if it can't be done via firmware.

 

[TT3023]

While modification of DJI firmware itself is not necessarily impossible, it is extremely difficult and impractical in a general sense. That means the only practical mechanism for added functions is via the MSDK, either with a supported drone or the spoofing of a supported drone. Adding chips to a system like this is really not practical.

As I understand the *current* U.S. regs (but check for yourself, this is just how I interpret the rules at the moment) RID will not be required for sub-250 gram drones that are not operated commercially and are only used for recreational flying. That appears to include recreational-only use of the Mini series (without attachments increasing the weight) using standard size batteries (not the new extended time battery for the Mini 3, however).

Some older drones that will require RID will apparently be able to use separate RID modules that will be mounted on the drones to meet the requirements.

Again, these are my interpretations, I urge checking for yourself before making any decisions.

 

[LakesGuy]

Thanks for the thoughtful reply. As I think is probably apparent, I unfortunately don't have any technical expertise with these splendid little machines. My comment was really a translation of my understanding that the Mavic 2 series uploads the Maven, Litchi, or Dronelink waypoints flight plan into the drone's internal memory. It employs the onboard data to control the drone in flight, irrespective of whether it is receiving virtual or operator inputs via a signal from the controller. It seemed to me that a third-party chip could serve that same function, thought I freely confess the idea is a simple one to those of us who really don't know what is involved in creating and adapting the existing technology.

My understanding of DJI's response to the upcoming RID requirement is that it will adapt its existing drones to comply via either firmware, or through an inexpensive external device. I haven't really been concerned with the sub-250g models since I don't fly one, though I have an original Mini. My thinking was that if DJI is going to manufacture an external device for that purpose, other such devices could add additional features.

At least I hope we are moving toward more clarity from the FAA concerning the RID requirement.

 

[TT3023]

Thanks for the thoughtful reply. As I think is probably apparent, I unfortunately don't have any technical expertise with these splendid little machines. My comment was really a translation of my understanding that the Mavic 2 series uploads the Maven, Litchi, or Dronelink waypoints flight plan into the drone's internal memory. It employs the onboard data to control the drone in flight, irrespective of whether it is receiving virtual or operator inputs via a signal from the controller. It seemed to me that a third-party chip could serve that same function, thought I freely confess the idea is a simple one to those of us who really don't know what is involved in creating and adapting the existing technology.

My understanding of DJI's response to the upcoming RID requirement is that it will adapt its existing drones to comply via either firmware, or through an inexpensive external device. I haven't really been concerned with the sub-250g models since I don't fly one, though I have an original Mini. My thinking was that if DJI is going to manufacture an external device for that purpose, other such devices could add additional features.

At least I hope we are moving toward more clarity from the FAA concerning the RID requirement.

Keep in mind that (again, as I understand this), an external RID module need not interface with the drone itself (and I assume typically would not). They would typically be standalone units simply attached to a drone, and would not need to be purchased from DJI.

 

[LakesGuy]

Thank you again for taking the time to reply to my thoughts on this important subject. I'll try to muddle along without getting in over my head discussing things I don't really know about.

 

[TT3023]

Thank you again for taking the time to reply to my thoughts on this important subject. I'll try to muddle along without getting in over my head discussing things I don't really know about.

Never feel bad about asking questions here! As far as I'm concerned, it's a major (perhaps THE major) reason for this forum's existence -- a learning resource!

 

[TT3023]

I just noticed that DJI apparently no longer has a "pro" developer level (up to now that pro level required fees), and that all developers reportedly now have unlimited device activations on their apps (it used to be limited to 5 or some such I think). This, combined with the fact that Google does not (at least currently) charge for Mobile Maps SDK access, means that distributing an app might actually be practical.

 

[LakesGuy]

There seems to be waypoints apps that focus on one or the other of the two operating system. Litchi seems to me to be more oriented toward Android, though it offers IOS.

My favorite, Maven, is strictly IOS (Apple Maps) but allows users to import Litchi missions which were designed on a Windows or Mac computer using Litchi’s free mission hub (different map system). The advantage of planning a mission on Litchi and executing it on Maven is that a pilot can fly the mission virtually on the free Virtual Litchi which uses Google Earth and is also free.

In flying with Dronelink, Litchi, Maven, and some others that have come and gone, various maps have been employed in the various apps, including Apple, Tom-Tom, and Google Maps. I haven’t found one or the other of the maps to be an issue, even when switching between them through mission planning and execution, as described above, but I fly in mostly well established suburban areas where map reliability is more or less a given.

While I can’t offer any technical help, I do offer my best wishes for your success.

 

[TT3023]

To be clear, the purpose of this app is (a) to provide a MAVLink interface, permitting mainstream PX4-style programs like QGroundControl (which is open source) to be used to create and fly missions (including waypoints), and to permit interfacing with any other MAVLink system (e.g., DroneKit-Python for AI and other automation functions). And (b) to provide a "wrapper" permitting such MAVLink programs to work with DJI drones that are not yet supported by their MSDK that is necessary for the programs you mentioned to function. Programs like Litchi, etc. won't support the Mini 3 until the MSDK does. It took DJI over a year to support the Mini 2 in the MSDK after the Mini 2 release (actually that support is very recent). Maybe DJI will be much faster with the Mini 3, but we can't assume that. The reason I'm trying to get my hands on a Mini 3 is to test the app framework to see if it can support the Mini 3 without implicit support in the MSDK, as it did successfully with the Mini 2. That would at least provide waypoint support for the Mini 3 immediately, if successful.

 

[TT3023]

I should add that the app does support simulation mode and that works much the same as with the commercial apps.

 

[TT3023]

I'll also note that QGroundControl has versions for Windows, OS X, Linux, iOS and Android. The QGroundControl instance can either run on the same device as the wrapper Android app, or on a different device (on the same local network). So, while the wrapper app must run on an Android device attached to the controller, the QGroundControl instance can be on a different device with a different OS entirely, including desktop computers.