wip AH-64D Attitude Hold

[Skeeter1]

I have noticed that the attitude hold in the apache will constantly shift to pitch up or down. For example, I can trim my heli for perfectly level flight turn on att hold, and watch my heli slowly pitch forward or backward until it crashes. With alt hold on it slowly speeds up/down. I had the alt hold on, and was cruising for about 10 mins hands-off. I turned on the attitude hold and about 3 minutes later the aircraft could no longer maintain altitude for the nose position.

[Skeeter1]

Before anyone asks, no my stick is not putting in false inputs, I've checked, I have a deadzone, my stick is a warthog. Yes I am confident I am using the mode correctly and I am well trimmed.

[Grennymaster]

as far as i know, that is normal. the hold modes in the Helo are not as sturdy as in a plane. Should be caused by the forces that punches on the frame, disk, etc.

from my understanding you always have to retrimm a helo in flight.

 

[admiki]

2 hours ago, Skeeter1 said:

Before anyone asks, no my stick is not putting in false inputs, I've checked, I have a deadzone, my stick is a warthog. Yes I am confident I am using the mode correctly and I am well trimmed.

Are your servo sleeves centered?

[Lord Vader]

Hi @Skeeter1

The AH-64D auto-pilot is very unique. We would require a track of the behaviour in order to analyse your claims.

[JSpidey]

10 hours ago, Skeeter1 said:

I have noticed that the attitude hold in the apache will constantly shift to pitch up or down. For example, I can trim my heli for perfectly level flight turn on att hold, and watch my heli slowly pitch forward or backward until it crashes. With alt hold on it slowly speeds up/down. I had the alt hold on, and was cruising for about 10 mins hands-off. I turned on the attitude hold and about 3 minutes later the aircraft could no longer maintain altitude for the nose position.

Make sure you're in aerodynamic trim before engaging ATT hold. I find that it transitions easier that way.

[Skeeter1]

14 hours ago, admiki said:

Are your servo sleeves centered?

Yes, I have held down the trim to ensure that is the case

6 hours ago, JSpidey said:

Make sure you're in aerodynamic trim before engaging ATT hold. I find that it transitions easier that way.

I always trim for the ball in cruise

15 hours ago, Grennymaster said:

as far as i know, that is normal. the hold modes in the Helo are not as sturdy as in a plane. Should be caused by the forces that punches on the frame, disk, etc.

from my understanding you always have to retrimm a helo in flight.

 

Doesn't that make the attitude hold useless? It's actually less stable than nothing at all when well trimmed. I will try to pull a track, but the issue is easy to replicate by just turning on the hold modes in straight and level flight and waiting.

[Grennymaster]

well its not useless. you can fly cirles at relativ low speed with ATT and ALT mode on.

I think it will get better wenn the flight model will be more completet. The SCAS is also not complete atm. so that also may effekt the hold modes more than it should.

My last flight in a real helo was a decade ago, so i may also not remeber it correctly. Im not a lizenced pilot sady, had to give up to illnes .(

 

[admiki]

7 hours ago, Skeeter1 said:

Doesn't that make the attitude hold useless? It's actually less stable than nothing at all when well trimmed. I will try to pull a track, but the issue is easy to replicate by just turning on the hold modes in straight and level flight and waiting.

Track would be the best. TBH, I use holding modes only if I need to go head down for some reason. So, not much at all.

One of the reasons I am in DCS is because I like flying helicopters. If it was just button pushing and watching scenery go by I would have gone to train simulators.

[Skeeter1]

On 6/19/2023 at 11:02 PM, admiki said:

Track would be the best. TBH, I use holding modes only if I need to go head down for some reason. So, not much at all.

One of the reasons I am in DCS is because I like flying helicopters. If it was just button pushing and watching scenery go by I would have gone to train simulators.

I like trains . . . But seriously I enjoy a good flight as much as the next fellow. But a 30 minute flight from waypoint to waypoint can get repetitive for a video game. Otherwise it's basically just me sitting there making miniscule corrections every 30 seconds.

[Scaley]

The DCS Apache seems to have a maximum speed it can maintain with the combination of ALT and ATT hold. Even if perfectly in trim in level flight at moderate power settings, if you set an attitude that corresponds to much above 110kts the two systems will enter a little positive feedback loop that lead to gradual collective increases, followed by gradual application of forward cyclic to maintain attitude, the combination of which leads to increased airspeed, thus requiring increased collective to maintain level flight above MaxE. This cycle then repeats and eventually as you say the aircraft will enter a dive and start descending. I have no idea if the real aircraft does this, but the ED one has for at least the last 4 patches. The solution is to slow down a bit (generally below 110). I'd guess the hold mode logic will improve as the module evolves. 

[Skeeter1]

12 hours ago, Scaley said:

The DCS Apache seems to have a maximum speed it can maintain with the combination of ALT and ATT hold. Even if perfectly in trim in level flight at moderate power settings, if you set an attitude that corresponds to much above 110kts the two systems will enter a little positive feedback loop that lead to gradual collective increases, followed by gradual application of forward cyclic to maintain attitude, the combination of which leads to increased airspeed, thus requiring increased collective to maintain level flight above MaxE. This cycle then repeats and eventually as you say the aircraft will enter a dive and start descending. I have no idea if the real aircraft does this, but the ED one has for at least the last 4 patches. The solution is to slow down a bit (generally below 110). I'd guess the hold mode logic will improve as the module evolves. 

Happens regardless of airspeed, it will eventually speed up to that point even from a completely stable config at 100 kts

[Grennymaster]

i tested it yesterday ab intesivly. what i could find out is, speeds till 120 work good as long as you always keep the bird in trim. above the modes are not able to keep the bird.

A thing that many may not have get right. The hold modes are NOT an Autopilot system like in a jet.

[Skeeter1]

On 6/22/2023 at 2:35 AM, Grennymaster said:

i tested it yesterday ab intesivly. what i could find out is, speeds till 120 work good as long as you always keep the bird in trim. above the modes are not able to keep the bird.

A thing that many may not have get right. The hold modes are NOT an Autopilot system like in a jet.

Just go to 95 kts, turn on alt hold and att hold, and wait about 5 minutes

[admiki]

I have done a 10 minutes test.

In that time, I have gained exactly 2 knots from my start speed.

I think you might have wrong ideas of how hold modes work in concept of helicopter flight.

I think at this point, every further discussion is moot until you give us track file to check over.

 

[papaz]

Here's a track file. The attitude hold makes the pitch channel slowly drift forward, until it drives you to the ground, even if you're trimmed for stable flight.
This mostly happens above 50% torque cruise. The faster you fly the quicker this behaviour manifests.

AH-64_ATT_hold.trk

Edited September 13, 2023 by papaz

[Lord Vader]

Hi @papaz

I ran your track and I saw you only engage the attitude hold. I did observe the SAS pushing forward but you're not compensating the vertical axis by also engaging altitude hold while cruising. The pitch will be maintained, accelerating and adding negative VS. We need to think in three axis, not just two. Also, remember you trim to speed, not altitude.

That being said, the SAS and overall autopilot systems are currently being reviewed and fine tuned by the team. Look forward to further updates. 

[papaz]

2 hours ago, Lord Vader said:

Hi @papaz

I ran your track and I saw you only engage the attitude hold. I did observe the SAS pushing forward but you're not compensating the vertical axis by also engaging altitude hold while cruising. The pitch will be maintained, accelerating and adding negative VS. We need to think in three axis, not just two. Also, remember you trim to speed, not altitude.

That being said, the SAS and overall autopilot systems are currently being reviewed and fine tuned by the team. Look forward to further updates. 

Thanks for taking a look at it. As others have said it here before, engaging the ALT hold at the same time will compensate by adding collective, for a while, after that it just breaks.
Even with perfect stable flight trim (trim for speed) the problem is present. Strangely enough if you trim for maximum speed, the opposite drift will occur. (pitch up)

I can provide more tracks if needed. Maybe not the same topic, but the basic flight model is the most unstable out of all rotorcraft modules, which probably makes troubleshooting the SCAS difficult.

Edited September 13, 2023 by papaz

[Lord Vader]

Hi again.

Like I said, this is all being tweaked. I believe, at some point, you will see improvements on this system and overall dynamics. 

[papaz]

Does the dev team know that the SCAS and the ATT hold are still broken? DCS 2.9.1.48111 Open Beta - 16.11.2023
It's a PID controller, which works in all other rotor modules.

Edited November 16, 2023 by papaz

[NeedzWD40]

21 minutes ago, papaz said:

Does the dev team know that the SCAS and the ATT hold are still broken? DCS 2.9.1.48111 Open Beta - 16.11.2023
It's a PID controller, which works in all other rotor modules.

 

Tagged as WIP by the team, latest issues are still being addressed:

 

[papaz]

Much better in the latest OB (2.9.2.49629). Though the SAS pitch channel is still prone to drifting forward or backwards after a minute or so, just like when I first reported the issues.

Edited December 20, 2023 by papaz

[papaz]

DCS 2.9.11.4686 - 24.12.2024

The SCAS drift problem is still present, and so are the unexplainable pitch oscillations. The helicopter cannot be trimmed for stable flight and requires constant micromanagement.

[SloppyDog]

4 hours ago, papaz said:

DCS 2.9.11.4686 - 24.12.2024

The SCAS drift problem is still present, and so are the unexplainable pitch oscillations. The helicopter cannot be trimmed for stable flight and requires constant micromanagement.

Don't know about the drift, have to test it.

 

Regarding the pitch oscillations, they noted it on the changelog:

• Helicopter SCAS - known issues
• When turning ON or OFF the Position-Hold  mode, several damped oscillations in roll and/or pitch may occur

[papaz]

AH-64D_ATT_hold_test_Dec_25.trk
Uploading a track again, where you can clearly see the issue:

1. Trim for stable flight. 2. Activate ATT hold and ALT hold. 3. SCAS ATT hold commands forward pitch (as to why you can see later). 4. ALT hold tries to compensate with additional collective. 5. This speeds up the nose down command of the SCAS. 6. The feedback loop drives you into the ground.

Near the end of the track I deactivate the FMC and trim for stable flight. What happens is that you get an slight nose bob, up and down, for no reason (FM issue).
That's what I meant for oscillation, not the oscillation on ATT hold activation-deactivation (refered to in the changelog).

I believe the ATT hold tries to compensate for this pitch-up movement where the problem starts.

edit:
With the latest patch it's possible to deactivate only the Pitch channel. This makes the Hold functions actually useable (in cruise, not hover). Still I feel that the FM is the source of these problems.

Edited December 25, 2024 by papaz