INAV VTOL improvements required for safer operation

[Sibre3]

I have rewritten this post so it now aligns with the topic. And places all the shortfalls and issues INAV VTOL suffers from in the easiest to find location of this report.

I changed the topic name to better suit the improvements this feature badly needs to make it more reliable under all conditions.

INAV VTOL handling is rudimentary to say the least. With much guess work involved.
It requires many setting compromises to be made because some settings that should be individually defined to either VTOL or fixedwing, not both.. Are not found in the mixer or control profiles.

I'm doing the best I can to keep persevering with it. When it would be so much easier to use the mature Arduipilot quadplane support, which I have in the past.
It isn't very reassuring to watch every second iNAV VTOL Youtube video crashing; even the better known influencers.

I'm hoping by having this issue report open. And raising the many short comings this software has. Better support maybe implemented. I have a soft-spot for iNAV as the middle child in flight software.

Here are some of those settings that should be in the control_profile or mixer_profile for VTOL operation. Not in the battery_profile or master settings.
Or as @MrD-RC states in later comments. Have their own independent setting. One for fixedwing and one for multicopter.

• failsafe_throttle : FS GPS loss landing fallback throttle must be 1000us for a fw so it glides in an emergency landing. And just below hover for a mc, so it slowly descends to a touchdown.
• airmode_type : stick_center_once is best suited to a fw and throttle_threshold to mc.
• nav_rth_alt_mode : EXTRA or MAX for a fw. CURRENT or FIXED for a mc VTOL. The rational being stated in this post.
• nav_rth_altitude : Extension of the above mode. A fw may require a climb. A mc VTOL will require much less when near the home location.
• nav_rth_home_altitude : An airplane requires a loiter altitude when at home. A VTOL in mc form does not. It can land when home is reached
• throttle_idle : The fw motor does not require idle throttle. The mc motors do. The motor_stop function only half does the job. If you turn off motor_stop in the mc mixer profile. The fw pusher motor still idles at minimum throttle, even if the fw mixer profile has motor_stop enabled.
• nav_rth_climb_first : Having a VTOL travel across ground forward tilted while climbing if the feature is set to OFF, is not good. But that can be desirable for fixedwing RTH operation when an immediate turn for home is required to save battery power.
• nav_rth_allow_landing : ALWAYS is more suited a mc. And FS_ONLY or NEVER is more suited to a fw, so the motor doesn't shut off above the home location if the automated landing procedure is not setup at your flight location.

Other features that would make iNAV VTOL more reliable :

• Speed and timer based transition settings. In place of the user having to guess when the airframe has enough airspeed to switch to fixedwing profile.

• Motor rpm speed ramp-up settings. In place of a hard startup when the mixer profile switches, creating unnecessary torque reaction.

• A fix for the mixer issue Jetrell found. #11292 (comment)

• Another issue I've found that was also reported by you Jetrell in response to a bug found by the developer who implemented iNAV VTOL support @shota3527 Landing Detector VTOL support (multicopter phase) #9323

• tpa_on_yaw : is not implemented correctly. It is based on user throttle stick input. And doesn't achieve any purpose in preventing tilt-rotor yaw waggle during transition. Because the throttle isn't altered much during a typical flight transition. It should be based on forward flight transition airspeed now iNAV has that ability.

• Fixedwing and multicopter stabilization can fight one another in transition in Althold. See more.

• Support "weather vaning" for iNAV VTOL as with Ardupilot. To prevent yaw i-term windup and major hover instability when the VTOL is operating in wind. This ensures the aircrafts nose will always face into the wind, as dictated by the fixedwing vertical stabilizer.

• Add separate pusher throttle control. As requested in VTOL: Separate pusher throttle control during transition #11426
  This could also include pusher motor reverse thrust to slow the VTOL to a hover after transition.

[Jetrell]

@Sibre3 Your comments reminded me of these bugs. And why I switched my quadplane to AP.. It's funny it didn't even occur to me when I run the test for you the other day, and replied on Discord.
I experienced these problem back in early 2024. The feature was implemented in late 2023.
At the time I tried to contact the collaborator who wrote this feature, in a DM. However he wasn't around at that time. And I completely forgot of the issue until now.

@P-I-Engineer @shannonbaker If my bad memory serves me correctly. You are the only two guys and one other, that posted of making an INAV 4 + 1 build.
Did either of you actually use RTH on those specific builds ? Or run fixedwing Autolaunch ?
Back then when I ran my tests. I didn't end up using the feature because Shota didn't back in contact with me about the problem. So I wasn't going to risk the model.

[P-I-Engineer]

I used fixed wing auto launch. Not sure I ever did full rth.

I did add a note I. The docs about 4 and 1 and motors spinning up.

[shannonbaker]

I have only used VTOL take off and VTOL transition in the manual landing sequence. I have not used autolaunch. I don't recall using RTH but I think I would have but unsure as I flew this model a long time ago.

@P-I-Engineer @shannonbaker If my bad memory serves me correctly. You are the only two guys and one other, that posted of making an INAV 4 + 1 build. Did either of you actually use RTH on those specific builds ? Or run fixedwing Autolaunch ? Back then when I ran my tests. I didn't end up using the feature because Shota didn't back in contact with me about the problem. So I wasn't going to risk the model.

[Mecitinceoglu]

Hello ,

Could you please share full inav dump for 4+1 Vtol.I can't quite figure out the configuration. Do you have any idea how to activate the push motor in the mixer transition? Because this feature is only available for servos. I would appreciate it if you could share your configurations so I can review them. Thank you.

Edit : I saw and reviewed your configuration file later, and I understood that you weren't using mixer transitions, is that correct? Is there a method for smooth transitions in INAV for 4+1 VTOL?

[P-I-Engineer]

https://github.com/iNavFlight/inav/blob/master/docs/MixerProfile.md#motor

read 3rd bullet, need value between -1 and -2 for active throttle while transition active.

[Jetrell]

I guess it seems simple to us @P-I-Engineer. But maybe not so much for others the first time they read it.
I had a guy ask me the same. I just told him it this way..

-1 is the fixedwing motor OFF.

While -1 to -2 is the throttle mixer weight range between 1000 to 2000uS. At which the fixedwing motor will spin at during the MC to FW Transition phase. Depending on how much forward thrust your airplane requires to get it up to speed.

-1.100 = 10%
-1.200 = 20%
-1.400 = 40%
-1.600 = 60%
-1.800 = 80%
-1.900 = 90%

@Sibre3 I found part of the issue you're encountering. Or at least a work-around.
In your fixedwing mixer profile 1. You have the 4 lift motors allocated. And even though their throttle weights are set to -1. You also have weights added to Pitch, Roll and Yaw.
I know technically this should not matter. But unfortunately the mixer code doesn't account for a negative value in the FW mixer profile.
So set Pitch, Roll and Yaw weight values to zero in the FW mixer profile. Or just don't add them at all. Only the fixedwing motor in your case, when it is placed before the lift motors in order of the mixer sequence. S1 to S10.

[Sibre3]

Thanks Jetrell. That issue has now been sorted. I expected the mixer code to be more flexible.
I'm still not sure about the second part of the problem.

[Jetrell]

You could also have added nav_rth_allow_landing to the list.
With a fixedwing it's more often not desirable to set it to ALWAYS unless you land at the same location all the time using the fixedwing auto landing procedure... Otherwise FS_ONLY or NEVER is a better option.

However on a copter, ALWAYS is the best option.. But if you do choose that option, and don't have a fixedwing landing procedure set. The old fixedwing landing procedure will execute by shutting down the motor above the home location. Which can be a hazard.

[MrD-RC]

I actually disagree with a lot of those parameters moving. As it only makes sense to move them for a VTOL, if at all. For regular fw or mr it would make more sense to leave them where they are. It actually highlights that some of these parameters should have fw_ and mr_ versions but stay in their current profile/general locations.

For example throttle_idle with mr, and a VTOL in quad mode, will depend on which battery is fitted. A larger, heavier battery will require a different throttle value than a lighter battery. So it makes sense to keep it in the battery profile. But also to switch to fw_throttle_idle and mr_throttle_idle as the different platforms could require different values.

I'd also disagree about the RTH parameters. They are more set for the area you're flying in than the craft (with the possible exception of if you land or not).

nav_rth_alt_mode should be decided on how and where you're flying and the likelihood of obstacles between you and the home point. For the same reason nav_rth_altitude will be set based on the location to ensure that obstacles are avoided. Not the craft type in use. Trees and mountains don't care if you're using a quad, fw, or VTOL. Possibly nav_rth_home_altitude could be different for different craft. But less important if the mr is just going to land anyway. It could be simply set to the fw value.

In all honesty, I think the best thing for VTOL would be a new nav_rth_vtol_transition_distance with maybe a default of 500. So if you're over 500m away from home, it transitions in to fw mode for a more efficient RTH. I know mixer_automated_switch exists. But it will always transition, no matter what. In cases where you're close to home. It would be safer and simpler to just stay in mr mode.

[Sibre3]

Thanks for your input @MrD-RC
I think it would be acceptable to have one setting for mc profile and the other for fw profile.

I see your point. Although the short falls of iNAV VTOL are due to it being cobbled together from two platforms. With lack of concern of what could and will happen under certain flight conditions.
After weeks of trial, error and learning the hard way. I assure you iNAV VTOL is a crash waiting to happen for newbies.
I have provided more valuable information in this report, than is contained in the documents. They don't tell you about the dozens of gotcha's waiting to destroy your expensive VTOL model. And I have many more I've not yet reported.

I see you mentioned mixer_automated_switch and its short falls. I also have another to report. If it transitions from fw to mc when at home. And you inadvertently switch out of RTH during the mc landing when you are close to the ground. It will revert back to fw mixer profile and try to launch again as a fw with no airspeed. It should have logic checks in place for this and many other conditions as well, accounting for the other problems I raised in earlier comments. Otherwise it's simply dangerous and not trustworthy as a true FPV aircraft.
I have raised all these matters so the logic short falls can all be found in one place, if a developer has the willingness to make it better one day.

[Jetrell]

In all honesty, I think the best thing for VTOL would be a new nav_rth_vtol_transition_distance with maybe a default of 500

@MrD-RC Something like that would be a good idea. Allowing transition to occur before the aircraft gets to the home location.
Instead of Yoyo-ing back to the home position. Which isn't as easy for a VTOL as it is for a copter.

For the same reason nav_rth_altitude will be set based on the location to ensure that obstacles are avoided. Not the craft type in use. Trees and mountains don't care if you're using a quad, fw, or VTOL

I do understand your point, when it comes to simplicity. However on dedicated platforms like an airplane or a long range quad, nav_rth_altitude modes like MAX or EXTRA are absolutely needed, for reasons you mentioned.
But when two platforms are combine. It's really only the fixedwing side that needs those rth climb modes. Because it's as a plane it will require those climb actions.
But when the aircraft is in MC VTOL use. That will only occur within a few hundred meters of the takeoff/landing location. No user in their right mind would fly a fixedwing airframe in VTOL mode for any distance. Or even take the chance of transitioning at a great distance away.
My point is. If you're flying around as a plane and then finally decide to transition to a VTOL for landing. But as many I have seen do, want to land automatically by RTH. The aircraft as a MC VTOL will have to climb to the nav_rth_altitude via EXTRA or MAX which you set for safe fixedwing operation. And this will happen within a stones throw of the home location.
I have encountered the same issue myself. And also thought it would be useful to have them different for each platform, for the reason mentioned.

nav_rth_alt_mode : EXTRA or MAX for a fw. CURRENT or FIXED for a mc VTOL
nav_rth_altitude : Extension of the above mode. A fw may require a climb. mc VTOL will require much less.

I don't mind coming along side @Sibre3 in this case. I have been waiting a while for someone to raise these VTOL support issues.
I have 3 VTOL's and only one uses INAV. Because I simply don't trust its feature set enough to use it on my more expensive VTOL aircraft. Or it isn't flexible enough as mentioned.

[sensei-hacker]