Autopilot goes from VOR to ROLL when tracking VOR

jmarkows

Are you running the latest version? There was a bug at launch that caused that.

Jouxeur33

oh i have 0.1.0 but latest is 0.1.2 u think that could be the cause?

EDIT : i've read the change log and there is something about it indeed so its on me thanks anyways!

MarkS

@Black-Square said in Autopilot goes from VOR to ROLL when tracking VOR:

I'm happy to help with that. I just wanted to give you a quick answer before I left my desk for a few minutes.

@Black-Square Nick, sorry, I know you're a busy guy, but I'm going to have to ask for some help with this. My first calculator was a slide rule way back last century, but this RPN stuff throws me once the nesting starts. I have to tease it out into a non-RPN form to try to get a handle on it, and even then I'm not sure I'm getting it right.

(In fact, I'm pretty sure I introduced an intermittent timing issue with my "personalization" mod, because I get spurious warning button activations from time to time, sometimes only transient, sometimes they cycle for a while. That's on me though, not asking for your help with that.)

To recap, what I am looking to do is to trigger the autopilot disconnect tone if the autopilot fails over from NAV to ROLL mode. And of course it should cease sounding if I press the autopilot disconnect button as usual.

If CurrMode = ROLL and LastMode = NAV then
    PlaySound(AutopilotDisconnect)
LastMode = CurrMode

The code itself likely isn't that hard, but I'm having trouble locating the relevent mode variables, and there's nested code in that whole Flight Director section that I can't quite get my head around, so I'm not sure where to safely put the code even if I had the vars on hand.

Black Square

@MarkS Give this a try I'm not sure exactly what this will do, but it should be a good place to start.

<UseTemplate Name="ASOBO_GT_Update">
	<FREQUENCY>1</FREQUENCY>
	<UPDATE_CODE>
		(A:AUTOPILOT MASTER, bool) (O:lastApNavMode) (A:AUTOPILOT BANK HOLD, Bool) (A:AUTOPILOT WING LEVELER, Bool) or and and if{
			1 (&gt;L:var_AutopilotDisconnectTone, bool)
		}
		(A:AUTOPILOT NAV1 LOCK, bool) (&gt;O:lastApNavMode)
	</UPDATE_CODE>
</UseTemplate>

MarkS

@Black-Square said in Autopilot goes from VOR to ROLL when tracking VOR:

@MarkS Give this a try I'm not sure exactly what this will do, but it should be a good place to start.

Thank you. Any particular location in the Flight Director code section you'd recommend to put it?

Black Square

You can drop that anywhere in the file between other <UseTemplate Name="ASOBO_GT_Update"> tags. For instance, right after the  comment.

MarkS

@Black-Square I'll give it a go. I suppose it could be dropped in anywhere but in good practice it ought to be at least near the other relevant functionality.

(P.S. as hard as it is for me to parse RPN, your code is a treat to work with for this old automation engineer).

MarkS

@Black-Square no joy.

According to an Axes and Ohs watch list I set up to monitor, it appears the O:lastApNavMode variable isn't getting updated,(what is an O variable anyway?), or perhaps Axes and Ohs isn't capable yet of seeing it. I can declare it in the watch list, and it appears, but that doesn't necessarily mean there's communication to it.

I did a test flight with two VORs and deliberately flew through their zone 'o' confusion to force a failover from NAV to ROLL. Which happened as expected, but no alarm.

In fact, the AUTOPILOT BANK HOLD and AUTOPILOT WING LEVELER are never '1' at any point in during flight testing, and I'm not sure if it's something I'm doing or not doing or its just one of those weird things that are out of our hands like state saving.

This tickles my old engineering brain a little and I'm tempted to take a crack at it. I'm wondering if I can just declare a var in the interior.xml that Axes and Ohs can (demonstrably) see, to replace the O var. Maybe Mr Occam had a nice shave and will smile kindly on my attempt.

Here's the sequence with the Axes and Ohs watch lists.

Black Square

O:Vars are local to the scope of that UseTemplate tag, but if you just replace the O with an L, that should make it globally visible.

Do I misunderstand your message, or is AUTOPILOT BANK HOLD true in your last screenshot, right when we expect it to be?

I don't see anything wrong with the code, but I can test it when I'm next working in the simulator. Can you show me where you inserted it into the file?

MarkS

@Black-Square said in Autopilot goes from VOR to ROLL when tracking VOR:

Do I misunderstand your message, or is AUTOPILOT BANK HOLD true in your last screenshot, right when we expect it to be?

No, that's on me, I missed that it was set. Nice catch. Interesting though, that Wing Leveler is never true.

I put the new code at the top of the Flight Director section (I was bold and already changed the O to an L but haven't had a chance to test it yet).

I sat with it for a bit afterwards and thought it might simplify to something like this (sorry for the infix pseudocode, can't "think" in RPN):

If
	A:AUTOPILOT_MASTER = True and L:lastApNavMode = True and A:AUTOPILOT_NAV1_LOCK = False
Then
	L:var_AutopilotDisconnectTone := 1
Endif
L:lastAPNavMode := A:AUTOPLOT_NAV1_LOCK

The RPN issue I have with the above is that the NAV1 LOCK comparator is (I think) an "and not" operation in RPN, and I don't know how to do that. I'm not entirely sure what BANK HOLD or WING LEVELER are bringing to the party, but you surely have a better handle on that than I.

MarkS

Against my better judgement, I came up with this refactoring of the infix pseudocode I proposed above:

<!-- Mark's Occam-Channeling Autopilot NAV to ROLL mode failover alarm (probably doesn't work) -->

<UseTemplate Name="ASOBO_GT_Update">
	<FREQUENCY>1</FREQUENCY>
	<UPDATE_CODE>
		(A:AUTOPILOT NAV1 LOCK, bool) (L:lastApNavMode) (A:AUTOPILOT MASTER, bool) not and and if {
			1 (&gt;L:var_AutopilotDisconnectTone, bool)
		}
		(A:AUTOPILOT NAV1 LOCK, bool) (&gt;L:lastApNavMode)
	</UPDATE_CODE>
</UseTemplate>

I looks almost right to my eye, but I still don't know if that 'and not' is legal. Guess I'll find out later. At least it's easy to remove if (when) it fails spectacularly.

Addendum: It didn't fail, but it didn't work either. @Black-Square shouldn't (L:lastApNavMode) be (L:lastApNavMode, bool) instead?