Autopilot, Alt Hold

[Demiurgo]

Hi Everyone!

I've looked but I couldn't find anything. Sorry if this is double post.

SOMETIMES when i set up ALT HOLD the plane starts to dive and never recovers, even if the starting attitude was going UP.

I tried many different ways but i can't understand what triggers this behavior.  I saw it mainly in MP.

By the way, i always press NWS when vertical velocity es almost 0. In many cases I trim the plane to 0 to check and then the nose diving start.

Please let me know if this has being pointed out.

 

Thank you

[near_blind]

How fast were you going, and were all three SAS channels enabled?

[Demiurgo]

1 minute ago, near_blind said:

How fast were you going, and were all three SAS channels enabled?

Less than MACH 1 and 3 channels enabled.

[Rinz1er]

I've had issues engaging the altitude hold at anything higher than 35999ft on the tomcat altimeter

Edited September 19, 2022 by Rinz1er

[Demiurgo]

9 minutes ago, Rinz1er said:

I've had issues engaging the altitude hold at anything higher than 35999ft on the tomcat altimeter

 

Thanks. I will check again since I don't recall my altitude at those times.

[Demiurgo]

1 hour ago, Rinz1er said:

I've had issues engaging the altitude hold at anything higher than 35999ft on the tomcat altimeter

 

It seems you were onto something.....

AP + ALT HOLD seems to work below 36kft. If you go above that it starts oscillating (ie from 40kft to 36kft) and sometimes stabilizes in some other altitude; sometimes it does not.

I'll check some other source for limits on this AP but I don't recall this issue in the past.

I added a track if anyone is interested on this info.

 

Thanks

F14 AP TEST.trk

[IronMike]

Check what mach speed you are doing high up. Note that you can be significantly below 450IAS yet doing more than m0.9 high up, at which point alt hold will start failing, oscillating etc. That just in general. It will ofc also not be able to hold beyond a certain AOA, but would need to check at point.

Thank you for the track as well. To comment on that a bit:

The fail at 37k feet is expected, you are at m.9, which is where its limit is for alt hold, and engage out of a steep climb in terms of FPM, so it oscillates.

Likewise the fail at 45k feet seems ok, because you are getting so slow, you can barely keep the aircraft level yourself. So ALT HOLD now has to fight 3 conflicting things, from that particular negative FPM you engaged it:

1. It needs to null the rate, aka climb.
2. It needs to dive to lower the AOA, this in return makes the aircraft gain too much speed (to note that there was no speed management involved in your track), aka exceeds the limit again, which in return increases oscillation
3. And all that while trying to return to the captured altitude.

The result is this heavy mess. In other words, you made it maximum hard on the autopilot to succeed, or basically impossible. Which in itself is not a bug. ALT HOLD will disconnect only if stick force is exceeded or 30° pitch is exceeded, so it will keep the phugoid happily ever after if you do not disengage it. When exceed this limit, it should revert to ATT hold, in which case you would still have to save the plane manually from disaster, if that makes sense.

However, despite of the expected outcome seen in the track above, there is a bug, where above between 39 to 41k-ish - from personal experience - no matter how well within your limits you are, ALT HOLD engagement will always result in a dive first, and phugoid second, which seems to be wrong.

If you manage speed and FPM accordingly, it will work rock solid up until 39-41k-ish feet though. Just tried myself as well to confirm.

Flying with ALT HOLD on at 45k also may not have been a thing, but when within limits, the immidiate plunge to negative FPM still should not happen like that.

In other words, if you tried to do it correctly, it would not work either now, so definitely a bug, even if not clearly demonstated in the track (I also flew it myself a couple times now to verify).

Thank you again for the report!
 

Edited September 20, 2022 by IronMike

[Demiurgo]

34 minutes ago, IronMike said:

Check what mach speed you are doing high up. Note that you can be significantly below 450IAS yet doing more than m0.9 high up, at which point alt hold will start failing, oscillating etc. That just in general. It will ofc also not be able to hold beyond a certain AOA, but would need to check at point.

Thank you for the track as well. To comment on that a bit:

The fail at 37k feet is expected, you are at m.9, which is where its limit is for alt hold, and engage out of a steep climb in terms of FPM, so it oscillates.

Likewise the fail at 45k feet seems ok, because you are getting so slow, you can barely keep the aircraft level yourself. So ALT HOLD now has to fight 3 conflicting things, from that particular negative FPM you engaged it:

1. It needs to null the rate, aka climb.
2. It needs to dive to lower the AOA, this in return makes the aircraft gain too much speed (to note that there was no speed management involved in your track), aka exceeds the limit again, which in return increases oscillation
3. And all that while trying to return to the captured altitude.

The result is this heavy mess. In other words, you made it maximum hard on the autopilot to succeed, or basically impossible. Which in itself is not a bug. ALT HOLD will disconnect only if stick force is exceeded or 30° pitch is exceeded, so it will keep the phugoid happily ever after if you do not disengage it. When exceed this limit, it should revert to ATT hold, in which case you would still have to save the plane manually from disaster, if that makes sense.

However, despite of the expected outcome seen in the track above, there is a bug, where above between 39 to 41k-ish - from personal experience - no matter how well within your limits you are, ALT HOLD engagement will always result in a dive first, and phugoid second, which seems to be wrong.

If you manage speed and FPM accordingly, it will work rock solid up until 39-41k-ish feet though. Just tried myself as well to confirm.

Flying with ALT HOLD on at 45k also may not have been a thing, but when within limits, the immidiate plunge to negative FPM still should not happen like that.

In other words, if you tried to do it correctly, it would not work either now, so definitely a bug, even if not clearly demonstated in the track (I also flew it myself a couple times now to verify).

Thank you again for the report!
 

 

First of all, thank you for taking the time to answer and also for the outstanding job you are doing!

In regards to the bug/issue and the track, was just fooling around trying different things. What I see is the following:

1) In the past you could engage practically at any speed at 40kFt and would keep the ALT (if you engaged it correctly in terms of vertical speed). This is what I typically did so I am sure of the change. It's not clear to me if this change is intentional or not.

2) If the change is intentional I suggest changing the manual because (to my recollection) AP parameters are not listed. I checked the manual yesterday for this and couldn't find it. In any case a list of accepted parameters could be ok when you update the manual down the line

3)  The only thing where I am not having the same experience as you is above 36kft and below 39kft. I've tried several speeds and oscillation appeared. I'll try again and send you back a track (i am using the standar free flight mission on Caucasus) if i see this happening again. I always engange it at 0 fpm almost.

 

Again, thank you for your help and please let me know if you want something else to be tested. 

 

 

 

 

 

 

[Spurts]

One more thing is to be trimmed, not holding level with stick force, because IRC the AP system only manipulated trim and can only do so at a limited speed.

[Gun Jam]

what about with alt + heading hold with FFB joystick

I might get it to hold 1 in 15 tries and its unreliable and will usually oscillate after a minute or two then just enter an ever increasing dive. Is this a known issue or more or just me?

Thanks

[IronMike]

1 hour ago, Gun Jam said:

what about with alt + heading hold with FFB joystick

I might get it to hold 1 in 15 tries and its unreliable and will usually oscillate after a minute or two then just enter an ever increasing dive. Is this a known issue or more or just me?

Thanks

FFB is still an ongoing process and a known issue, our apologies that it drags out.

[IronMike]

2 hours ago, Demiurgo said:

First of all, thank you for taking the time to answer and also for the outstanding job you are doing!

In regards to the bug/issue and the track, was just fooling around trying different things. What I see is the following:

1) In the past you could engage practically at any speed at 40kFt and would keep the ALT (if you engaged it correctly in terms of vertical speed). This is what I typically did so I am sure of the change. It's not clear to me if this change is intentional or not.

2) If the change is intentional I suggest changing the manual because (to my recollection) AP parameters are not listed. I checked the manual yesterday for this and couldn't find it. In any case a list of accepted parameters could be ok when you update the manual down the line

3)  The only thing where I am not having the same experience as you is above 36kft and below 39kft. I've tried several speeds and oscillation appeared. I'll try again and send you back a track (i am using the standar free flight mission on Caucasus) if i see this happening again. I always engange it at 0 fpm almost.

 

Again, thank you for your help and please let me know if you want something else to be tested. 

 

 

 

 

 

 

I just tried above 36k again, and indeed, unless one is precisely at null rate, and well below m 0.9, it just pulls straight down. Definitely a bug.

Thanks again for all the feedback!

[Demiurgo]

34 minutes ago, IronMike said:

I just tried above 36k again, and indeed, unless one is precisely at null rate, and well below m 0.9, it just pulls straight down. Definitely a bug.

Thanks again for all the feedback!

Thank you for your help! 

[PD919]

On 9/20/2022 at 4:17 PM, IronMike said:

I just tried above 36k again, and indeed, unless one is precisely at null rate, and well below m 0.9, it just pulls straight down. Definitely a bug.

Thanks again for all the feedback!

Can confirm I have the same issue.