Hidden Flight Model Changes 2.8

[Poptart]

Hey all,

Even though the patch notes don't say it, it's clear that there have been some flight model changes with DCS 2.8. Here's a breakdown of what I am seeing:

• Effective Translational Lift is much more gradual now, both speeding up and slowing down. Previously crossing the 30 knot threshold would result in a significant change in lift that you needed to take account of. The effect is now more gradual and seems to scale with speed. As a result it's much harder to put the aircraft into VRS when transitioning to a hover from high speed.
• In forward flight there is a wicked oscillation of the nose up and down that makes altitude control much harder especially with the collective FCS on. This oscillation cannot be trimmed out and needs to be counteracted with cyclic. This is pretty noticeable when exiting a hard coordinated turn. 
• Crabbing in level coordinated flight with no wind has been reduced, especially with no stores. 
• Tail rotor thrust has been changed again. In 2.7, I could take off with virtually no left counter-torque. I now need to add a touch of left pedal to remain facing forward.
• FCS changes: The collective FCS no longer seems to be trying to keep you at the same altitude. After a few minutes flying it seems to wonder aimlessly. It is now much easier to fly collective FCS off. Attitude hold will now bring you wings level in forward flight if you are close. However those discussed oscillations will overpower the attitude hold at some speeds. The collective FCS still exceeds its bounds. 
• Climbing seems to be much easier. Especially at speed. I was able to reach 4,000 feet quicker than I ever have in the apache.
• Uncommanded roll while at speed seems to have been entirely eliminated. I was able to reach 170+ knots with no ill effect to roll. I also could not detect any retreating blade stall. 
• Gun rounds seem to de-spawn after about 1.5km. They do not complete the arc to the ground. The ballistic computer does not give any out of range warnings.
• The ballistic calculator for rockets is still off. It will show a good solution at about 3.8km @ 1,000 AGL. Actual range is closer to 2.2km @ 1,000 AGL.
• George now has a setting to WAS no weapons. 

Edited November 1, 2022 by Poptart

[bart]

See Raptor's reply here

Quote "This has not occurred. From my understanding, the last 2.7 patch in September was the only instance when the flight model was updated and not in the changelog. I do remember making a comment in a separate thread in Bugs & Problems that sometimes things are missed in the changelog, but that it isn't intentional; not a deliberate effort to conceal changes.

Further, no flight model changes were made in last week's patch to 2.8, so I think there may be some placibo effect going on. Or maybe you all are just that good."

 

Edited November 1, 2022 by bart

[Floyd1212]

Interesting observations, but most of them are rather subjective, and are hard to quantify as definitive changes.  I'll pay closer attention to some of these the next time I fly, but I would say I disagree with a few off the bat.

• I haven't seen any nose up/down oscillations during forward flight.  At what speed were you seeing this?  With or without ATT Hold?
• The amount of left pedal needed during lift-off was reduced when they made the changes in the previous patch, but I wouldn't say it was "virtually no left pedal" in 2.7.  This seems about the same to me prior to 2.8.
• Uncommanded roll to the left is still a thing for me when pushing the nose down at 100+ knots.

[admiki]

One thing that seems noticeable to me is change in required pedal when accelerating from hover, but this seems to be across all helicopters, not just Apache.

Holding FTR, Apache seems a lot more stable than before.

While it is still somewhat twitchy in hover, it does seem subdued in response to inputs (note: I use 20 cm extension, with no curves and 100% saturation).

 

[Poptart]

3 hours ago, bart said:

See Raptor's reply here

Quote "This has not occurred. From my understanding, the last 2.7 patch in September was the only instance when the flight model was updated and not in the changelog. I do remember making a comment in a separate thread in Bugs & Problems that sometimes things are missed in the changelog, but that it isn't intentional; not a deliberate effort to conceal changes.

Further, no flight model changes were made in last week's patch to 2.8, so I think there may be some placibo effect going on. Or maybe you all are just that good."

 

 

I read that after my post but it is clearly not the case. I flew the same profile back to back after updating and the changes are fairly obvious and repeatable. Maybe the changes are caused by modifications to the background sim but there are changes.

 

3 hours ago, Floyd1212 said:

Interesting observations, but most of them are rather subjective, and are hard to quantify as definitive changes.  I'll pay closer attention to some of these the next time I fly, but I would say I disagree with a few off the bat.

• I haven't seen any nose up/down oscillations during forward flight.  At what speed were you seeing this?  With or without ATT Hold?
• The amount of left pedal needed during lift-off was reduced when they made the changes in the previous patch, but I wouldn't say it was "virtually no left pedal" in 2.7.  This seems about the same to me prior to 2.8.
• Uncommanded roll to the left is still a thing for me when pushing the nose down at 100+ knots.

The oscillations occur between 40 to 90 knots. There is no one speed because the oscillations also cause changes in speed. I've attached a photo. In the photo the helicopter was trimmed but otherwise had no other inputs. The oscillations occur both with and without ATT HOLD. 

Using very gentle application of collective no pedal takeoffs were possible. I have another photo of a 5-foot hover with only aft and left cyclic. Doing that now would result in a spin.

I was unable to replicate any kind of adverse roll behavior at various high speeds. 

 

Edited November 1, 2022 by Poptart

[Poptart]

Just checked again. The only way I can do a no pedal takeoff in 2.8 is if the FCS adds left pedal. No roll (left or right) at 120 knots. Oscillations seem to start after going more than 90 knots. The oscillations are impossible to correct with collective only.  

This is an apache on the test fire range with 65% fuel and all standard rockets.

[Raptor9]

I just spoke to the dev team about this, and they confirmed that no flight model changes have occurred since the tail rotor VRS behavior was modified in the Open Beta patch on September 30th.

4 hours ago, Poptart said:

• Gun rounds seem to de-spawn after about 1.5km. They do not complete the arc to the ground. The ballistic computer does not give any out of range warnings.
• The ballistic calculator for rockets is still off. It will show a good solution at about 3.8km @ 1,000 AGL. Actual range is closer to 2.2km @ 1,000 AGL.
• George now has a setting to WAS no weapons.

As for these items... The below screenshots are from the current OpenBeta, and rockets are clearly reaching beyond 2.2 km, even at a lower altitude and a hover; and the gun is also clearly landing rounds beyond 1.5 km. And the new George setting was also clearly identified in the changelog:

[Floyd1212]

22 minutes ago, Poptart said:

Just checked again. The only way I can do a no pedal takeoff in 2.8 is if the FCS adds left pedal. No roll (left or right) at 120 knots. Oscillations seem to start after going more than 90 knots. The oscillations are impossible to correct with collective only.  

Interesting.  I just flew the Live Fire Range instant action mission on Caucasus with 70% fuel, 70% gun rounds, and 16 Hellfires.

I was able to lift to a 5 foot hover very slowly without any pedal input, but the SCAS added some input to keep the heading for me, like you observed.  I am normally more aggressive with my lift, and apply the left pedal myself.

I was able to cruise at 95 knots at 300 feet level without any oscillations in any axis.

I increased my cruising speed to 120 knots and got the bird straight and level at around 300 feet, then performed and aggressive nose-down maneuver like perhaps I was trying to dive down to the ground to avoid a missile lock, and the helo rolled to the left (maybe to about 40 degrees?) before I pulled out of it.  I tried it both with nose-to-tail trim, and with aerodynamic trim, and got the same results. To be clear, I can cruise at 120 knots straight and level all day long, it's only when nosing down for the dive that I get the roll to the left.

Have you modified any portions of the flight model .lua?  I can't really find an explanation for the difference in the aircraft behavior for you while performing these tests.  Very strange.

Edited November 1, 2022 by Floyd1212

[Poptart]

1 hour ago, Raptor9 said:

I just spoke to the dev team about this, and they confirmed that no flight model changes have occurred since the tail rotor VRS behavior was modified in the Open Beta patch on September 30th.

As for these items... The below screenshots are from the current OpenBeta, and rockets are clearly reaching beyond 2.2 km, even at a lower altitude and a hover; and the gun is also clearly landing rounds beyond 1.5 km. And the new George setting was also clearly identified in the changelog:

Interesting. 

My rocket steering cursor is solid at 6k. If I do running fire at 5k the rockets fall well short.

Hover fire from the CPG at 4.7k give me the same result. Rockets fall about 1km short.

I wasn't able to reproduce the gun issue so maybe a glitch on my end. But I do have the tacview which is what I saw in game [i.e. gun shoots and no ground impacts]. [Edit: I had my units mixed up. I was at 1.5nm, not km. I'll try again at 3km.]

Edited November 2, 2022 by Poptart

[Poptart]

And here's the tacview from those two shots. Green line to target.

Hover fire:

22 knot running fire:

Edited November 2, 2022 by Poptart

[NeedzWD40]

1 hour ago, Poptart said:

My rocket steering cursor is solid at 6k. If I do running fire at 5k the rockets fall well short.

Hover fire from the CPG at 4.7k give me the same result. Rockets fall about 1km short.

Ballistic calculations for rockets are quite a bit off and the TSE is still not where it should be. Ballistic calculations only accurate at about 3km or less. Beyond 3km you'll have to manually elevate your sight to gain the necessary pylon elevation. Generally speaking, with M151 warheads, you should be able to hit targets with pylon elevation alone out to about 5km; beyond that you'll have to manually pitch up to get the necessary elevation.

[heloguy]

4 hours ago, Poptart said:

And here's the tacview from those two shots. Green line to target.

Hover fire:

22 knot running fire:

 

22 knot running fire?  That's a stretch...

[pii]

On 11/1/2022 at 3:57 PM, admiki said:

While it is still somewhat twitchy in hover,

yeah it seems like mine has a dead zone where even in a hover it says I'm gong Zero when it is clearly floating all over. It used to hover within the box you can put down in HB mode but now it floats all over. Oh well I wait for next patch