[BUG] Autopilot Altitude Hold

[mattag08]

I searched before posting, but I couldn't find anywhere where someone from Heatblur actually replied to a thread about this:

 

When engaging altitude hold with the NWS button, the autopilot immediately commands a rapid pitch down to 0° pitch and then steadily raises the nose to return the aircraft to the appropriate altitude. This occurs regardless of the original vertical speed rate of the aircraft. Even if you are perfectly level at 0 fpm vertically, the aircraft will command a pitch down and then slowly climb back to exactly the same altitude you were at when you engaged the ALT hold with NWS. This behavior runs counter to every autopilot I've ever encountered in my real world professional career in at least 7 types of light and commercial aircraft. Not saying that alone makes it inaccurate, I just find it hard to believe that autopilot laws and servo commands/gains are so differently programmed in the F-14.

 

An altitude hold autopilot mode generally does not primarily reference the pitch attitude for servo commands, but the altitude delta, pitch rates, and vertical speed. This is fed to the pitch trim which is adjusted to drive the elevator into the correct position. I know the F-14 does not have trim or elevators, but a similar concept of adjusting the trim position of the horizontal stabilizers to compensate for altitude/VS deviations would apply. The F-14B's manual states in the autopilot/altitude hold section: "Altitude hold should not be engaged during any maneuvers requiring large, rapid, pitch trim changes because of limited servo authority and slow automatic trim rate." In the attached image, you will notice that Pitch Channel B controls the autopilot/ACL functions. Pitch Channel A does not reference Altitude Error or Altitude Rate, while Pitch Channel B does, which I believe supports my theory that altitude hold references those parameters as opposed to pitch attitude. These pieces of flight manual information seem to imply behavior much more similar to what I've experienced in my real life flying.

 

If the current implementation is an accurate representation of how the ALT hold actually functions then that's acceptable, but absent any documentation on the autopilot programming it is difficult to believe this is correct behavior and I cannot find any public documentation that indicates that it is correct. As a pilot, the behavior seems unnatural and potentially dangerous. If someone from HB could weigh in and explain if this is a bug, work-in-progress, or accurate behavior as indicated by SMEs/manuals, I (and some of my F-14 squadron mates) would be very appreciative.

Edited November 1, 2019 by IronMike

[TOViper]

If someone from HB could weigh in and explain if this is a bug, work-in-progress, or accurate behavior as indicated by SMEs/manuals, I (and some of my F-14 squadron mates) would be very appreciative.

 

Me too!

Never in my life I heard of loosing altitude initially.

If this is really F-14's correct behaviour, I would accept it, but would still find it strange (in fact I would bet two coffees that this behaviour is not correctly modelled).

Edited October 31, 2019 by TOViper

[mattag08]

Can someone from HB comment on this please?

[IronMike]

I know exactly what you mean, and it comes from a too large offset from level flight. IIRC according to our SMEs this is appropriate, one had to be level and stable to engage the AP alt hold so it came in nice and steady. In time I had less and less issues with it, knowing when to engage it and when not, especially as lead in formations. But nonetheless you would get a small dip.

 

That said, I will re-ask, because it has been a while that we worked on that and discussed it. Marked as No Bug for now. Thank you for bringing it up nonetheless.

[mattag08]

This is not what's occurring. For reference, I am a real life commercial pilot with extensive experience and training. I've flown hundreds of hours in the F-14 in DCS at this point. I can fly the aircraft and hold an altitude. When I engage the autopilot I am perfectly level with a zero'd VSI and altitude holding steady.

 

However, your pitch angle will always be greater than zero (some small amount above the horizon) at normal cruise speeds. The current autopilot behavior is to always zero out the pitch (commands an immediate nose down) even if you are in level flight and then climb back up to altitude after the resulting altitude loss. This is counter to every autopilot I've ever used (including older autopilots that were designed and manufactured in the same time frame as the F-14).

Edited October 29, 2019 by mattag08

[IronMike]

This is not what's occurring. For reference, I am a real life commercial pilot with extensive experience and training. I've flown hundreds of hours in the F-14 in DCS at this point. I can fly the aircraft and hold an altitude. When I engage the autopilot I am perfectly level with a zero'd VSI and altitude holding steady.

 

However, your pitch angle will always be greater than zero (some small amount above the horizon) at normal cruise speeds. The current autopilot behavior is to always zero out the pitch (commands an immediate nose down) even if you are in level flight and then climb back up to altitude after the resulting altitude loss. This is counter to every autopilot I've ever used (including older autopilots that were designed and manufactured in the same time frame as the F-14).

 

Ok, I'll check. Thanks.

 

Btw, the drop I experience is tiny to say the least, but I engage the AP engage switch (and alt hold) first, then level, then egage AP ref button.

[mattag08]

I do the same. The altitude loss is on the order of about 50-100' depending on airspeed.

[IronMike]

Hey Mattag,

 

I talked to our SMEs again, and it seems I remembered their feedback wrong. The pitch down is indeed not supposed to happen, according to the feedback I got now, "The autopilot would enagage (...) without a bump or any pitch deviation. (...) the aircraft would simply level off and maintain altitude smoothly (...) The technique was to manually level off, null the rates, then engage."

 

We will thus take a second look at it. It's been so long that we implemented it that I can't fully remember the initial discussion we had as to why like that, but I guess it doesn't matter too much now anyway. It's in general something we use rarely, both devs and SMEs, since we predominantly test the manual flight behavior, so thank you again for bringing it to our attention. Personally I grew so accustomed to leveling and minor pitch deviation that I did not second guess it anymore.

Edited November 1, 2019 by IronMike

[Shaman]

Encountered similar problem on VPForce Rhino FFB stick.

[Dentedend10]

Having problem with VPForce RHino as well. Engaging autopilot the plane drops the nose forcefully. It does not hold altitude. I had no issues with these modes on spring stick.

Edit: found the solution. I did not have the FFB implementation enabled under F14 special options. Now it all works  great

Edited May 29, 2024 by Dentedend10