Sim lockup when when NOPOL is loaded as a way point

[sbski]

Aircraft Version

Development

Build info

{
    "built": "2022-09-02T10:36:05+00:00",
    "ref": "refs/heads/master",
    "sha": "e0a8528902881dc66ec8a94646c7ac356d4de0eb",
    "actor": "2hwk",
    "event_name": "manual",
    "pretty_release_name": "master:e0a85289"
}

Describe the bug

The sim becomes unresponsive when way point NOPOL is loaded into the MCDU with a waypoint after without a discontinuity. As far as I know, NOPOL is the only waypoint that does this. In addition, autopilot will disconnect if you do manage to get to NOPOL. Feel free to reach out to me on Discord if you would like more info.

Expected behavior

Standard operation.

Steps to reproduce

Attempting to do any of the following in the MCDU:

1. Direct to NOPOL
2. Load a simbrief flight plan that contains NOPOL
3. Manually enter NOPOL into the MCDU without a discontinuity after.
4. Manually remove a discontinuity after NOPOL.

References (optional)

No response

Additional info (optional)

No response

Discord Username (optional)

sbski#6362

[CBGonzalez]

NOPOL being the north pole, right? 0 degrees north, zero degrees west? I wonder if the sim freezes also at the south pole?

But is this a sim problem or do you only see that in the FBW 320?

[tracernz]

Note: the polar navigation option is not simulated on A32NX. The ADIRUs simulated are not certified for navigation in polar regions.

[juliansebline]

We should set this at won't fix and close it

[tracernz]

But... we probably want to fix for the A380X if it's on our side..

[frankkopp]

But... we probably want to fix for the A380X if it's on our side..

I wonder - see discussion on Discord - will these issues here be combined A32NX and A380X? Right now it is only A32NX (see repo name). So I also would vote to close it.

[tracernz]

Well, the code is in this repo, and will remain in this repo... the A380 will move in. 😛

[Flaxx-Rat]

Just a comment, recognizing that the A320 is not certified for polar routes. Nonetheless I used today the experimental version 8f90853 (MSFS is up to date) in order to see how it behaves when comparing it to PMDG's 737. I used this flight plan (copy from SimBrief): PAFA DRRLL4 DRRLL BTT J515 BRW 7500N15700W 8000N15700W 8500N15700W 8900N15700W 9000N01000E 8900N01000E 8500N01000E 8000N01000E 7500N01000E KOMUX LOMV3M ENTC Around the pole it looks like this, even if the pole itself is not part of the plan:

Actually, the frame rate dropped down to ~2.7 FPS (only sometimes), the flight became unstable from time to time, specifically while the bird tried to follow the dog leg and back to course again (I5 9600k, Nvidia RTX 2070, no o/c). However, no crash at all. I think it's caused by big value changes close to (mathematical) poles that only can prevented when changing to a "flattened", linear mathematical calculation e.g. based on a grid based navigation.

[tracernz]

I think the sim is the thing causing the framerate issue. It's geo related calculations like VOR bearings are totally borked near the poles too.

[Flaxx-Rat]

It seems that your are right regarding the causing sim as the frames dropped with PMDG's bird as well. As far as the course calculation is concerned it behaves slightly different if not worse. Even when leaving the Pole away it lost completely the control, kicked out VNAV (why ever), and didn't show any path at all until I introduced a manually defined waypoint at N85E010. However, it didn't recover without switching on HGD as leading lateral control. So far, there is no real difference. All right, FPS recovers a bit better (closer to the Pole).

In my next test I'll return to FBW, after shifting the problematic waypoints close to the pole one degree farther South at both sides.

EDIT (hours later): I started the flight again, now using FBW's A320, and changed the previous path as mentioned: just ...85N157W 89N157W 89N010E 85N010E... The MFD looked (quite) good, but when looking closer to it in PLAN mode it seemed like WP 89N010E would be bypassed. Even if not as much magnified, a small disruption of the drawn path was visible.

Just when passing the 89N-border, things went strange. The frame rate again dropped rapidly and the A/P kicked out.

I started hand flying, but even holding the flight level (FL360) was quite impossible - the bird wanted to climb. After passing the Pole and close to 89N010E it became easier, nonetheless it took a few minutes, after passing the 1° "inner circle" towards South before the autopilot could be used again. I'm not sure whether flying without the "real" horizon would have been a success.

I'm going to try some other things, but today I'm done.

[Flaxx-Rat]

All right, I did my last attempt and the FBW A32NX made a quite good figure! I manipulated the flight plan in a way that I shifted the most critical points both to 88°N (just remember, each plane began to circle when passing 89°N), and I switched the weather to CLEAR SKY. The plan now looked this way (taken from LNM, as the following screenshot): PAFA N0455F360 6451N14747W 6457N14757W DRRLL BTT J515 BRW 7500N15700W 8000N15700W 8500N15700W 8800N15700W 8800N01700E 8300N01700E 7800N01700E 7300N01700E VAVAD KOMUX LOMV3M ENTC

You may see a previous attempt as well...

As you can see, the bird flies further straight North, begins to turn close to the Pole (following the turning MFD) until the following course line can be met when flying straight South. Furthermore the Autopilot worked all the time even if the calculated course between 8800N15700W and 8800N01700E should be > 3° (I rather think even > 6°), and it definitely refused to follow it from a certain point. The MFD even lost the course line the moment when the bird began to follow the turning MFD compass indication.

That's it from my side. Just an additional remark: A similar flight with PMDG's 737 in FSX caused trouble as well, but there was no serious frame hit - for me nailed at ~25 fps.