DCS: AH-64D Flight Model discussion

[GrEaSeLiTeNiN]

From the few attempts I’ve tried at OGE hovering, does seem to hold the hover now even with hands free. If it does oscillate for some pilot induced or SAS reason, I turn off ATT hold for a sec or two then re-engage it. 

[Amarok_73]

2 hours ago, Raptor9 said:

One final note based on some confusion I am seeing in this thread. (...)

Thank You for this clarification. At least now I know, that my way of using the Forces Trim is correct.
Apparently it seems, that for some users, the 2.25 % margin is way too small, especially, when the one is using the FFB stick, where usually the center is quite wobbly and loose. For me it's impossible to keep the stick steady within this limit as I am using rather small centering forces in FFB settings (about 15-30 percent) and the Logitech G940 is known of his relatively large area around the point of forces center, where no resistance is noticeable. For sure it's bigger than 2.25% of the axle range.
Because all above, and what You've said about the differences and specificity of the controllers, I'd suggest (not to mention really appreciate) to provide the parameter in Special parameters, that will let the user to set the satisfying margin instead of use the hardcoded 2.25%.

Edited May 19, 2023 by Amarok_73

[TheoreticalApex]

Maybe I just need to do some readjusting to my curves and get used to the changes, but I feel like before this patch I could get it into a rock solid hover without using any hold modes and only using trim. Now I am back to working around the weeble wobbles in a hover. Even in forward flight I feel like I am struggling to keep it at a stable position where as before I could be completely hands off with just using trim and no hold modes, and when I engaged the hold modes it would get even better. I just feel like that's not the case anymore. 

I also noticed that when you manipulate the cyclic while the aircraft is still cold and dark and watch the rotor blades, the don't smoothly move anymore, its like the jump as if I were giving some drastic cyclic input, when in reality I am barely moving my stick. Also the rotor blades reach their limit of travel while I still have quite a bit of cyclic movement available in my physical controls. So it is like having a dead zone on the outside of my controls rather than at center. 

I hope that makes sense, It is kind of tough to explain. 

[Raptor9]

56 minutes ago, Amarok_73 said:

Because all above, and what You've said about the differences and specificity of the controllers, I'd suggest (not to mention really appreciate) to provide the parameter in Special parameters, that will let the user to set the satisfying margin instead of use the hardcoded 2.25%.

Such options are being looked into for the very reasons you've mentioned. But as the case with all things, I cannot make any promises or provide any estimates, but just to assure you that we are already aware of how such situations can complicate the intended logic.

[Boogdud]

5 hours ago, Raptor9 said:

To get the Attitude Hold to work for you, fly the aircraft to the desired state while pressing the force trim, then let go of the force trim and do not move the cyclic. If you move the cyclic and exceed the "breakout" value, even if the force trim is not pressed, the Attitude Hold will disengage.

So how would this work with 'central position trimmer mode'? Wouldn't you have to release your stick (to center reset) at which point it would be pretty difficult to tell when you are reaching the ~2% breakout point.

I really don't envy you guys trying to find a compromise on so many control methods, especially when none of the consumer ones are designed to work like an actual helicopter control, with the exception of maybe ffb which is a niche of a niche.

[Raptor9]

5 minutes ago, Boogdud said:

So how would this work with 'central position trimmer mode'? Wouldn't you have to release your stick (to center reset) at which point it would be pretty difficult to tell when you are reaching the ~2% breakout point.

Sorry, what I should have said was "do not make any inputs from the force trim reference point using your stick."

In real-life, the cyclic would remain where it is. In DCS, when using a spring-centering stick, the assumption is you are letting it spring back to center when trimming, which means don't apply any inputs after that.

[Amarok_73]

2 hours ago, Raptor9 said:

(...) but just to assure you that we are already aware of how such situations can complicate the intended logic.

I think You're talking about harmonization of multiple parameters, as just increasing the rotor damping coefficient didn't gave the wanted effect, while noticeably decreased stability of the heli.
Now it's clear to me. Thank You for the information.

[Raptor9]

20 minutes ago, Amarok_73 said:

I think You're talking about harmonization of multiple parameters, as just increasing the rotor damping coefficient didn't gave the wanted effect, while noticeably decreased stability of the heli.

I was talking about how specific hardware devices may have mechanical play or inaccuracies, leading to instances where the user inputs may be violating the breakout values simply due to the fact the input from the device may be moving around due to loose springs or inaccurate input sensors. It isn't related to anything with the flight model, just hardware input limitations.

[Amarok_73]

4 minutes ago, Raptor9 said:

It isn't related to anything with the flight model, just hardware input limitations.

Roger that.

[M1Combat]

Thanks for the clarifications Raptor.

 

To note...  It's my humble opinion that there should be a master setting for FFB availability and the default should be "off"...  which would then preclude any aircraft from having their defaults set to use a FFB stick type.  I only suggest this because the Apache (at least when it was originally early access released) was set to use the FFB based control modes in DCS.  Many folks didn't and maybe still don't know this, and it took quite some "shouting it from the rooftops" for people to catch on.  I'm fairly certain the FFB sticks are very much the minority even in this community.

Anyway...  I know you're not the person that makes that decision...  but I do feel like this style of decision shoots ED in the foot all too often :).

[NeedzWD40]

I tested the FM and SAS changes today in a variety of trials and here are my current observations:

FMC off is markedly improved, with far more stability in hover, flight, and aerobatics. Pretty much translates 1:1 with the FTR depressed. It feels possible to get into a hover without lots of oscillations now. Aerobatics are markedly improved, with a lot of the troublesome characteristics (excessive roll tendency, yaw oscillations) gone now. Rolls and loops can be executed well and with precision. Hovering feels more like it does with other helicopters, with only minor cyclic changes to keep it steady. I was able to take my hands off the collective and pedals, using only the cyclic to maintain position. If a heading change was required, it wasn't nearly the same fight as it was in the past, with either FTR or pedal.

SAS behavior with roll and yaw is markedly improved, with the tendency to slip into snap rolls significantly reduced. I tried to induce the previous snap roll behavior and was unable to get an unwanted roll to occur, but someone with more issues with it will need to test as I was careful to not get into it (pressing FTR for evasive usually kept me safe).

Taxi behavior with tailwheel unlocked is still something of a challenge, with oscillations still occurring on the pedals. It's not impossible to dial in, but I suspect that most will still have major problems with taxiing and rolling takeoffs.

Takeoff and landing is no longer a "jump" up or down to stay safe and my general feeling is that I can get light on the wheels without risking a rollover. There's still a tendency for the aircraft to want to tip, but it's not nearly as pronounced as it used to be.

Overall, I find it an incremental improvement, but if you were struggling in the past I don't think the recent patch will change perceptions. It hasn't gone to "easy," but it has tamed some of the more annoying quirks.

[jubuttib]

8 hours ago, Raptor9 said:

Having said that, due to the varied experiences that people are having, we can only state when changes are made to the flight model, and when they are not.

Entirely fair. I work with racing sims, and sometimes weird things just frigging happen. You don't touch something but it suddenly breaks, an sometimes even inconsistently between different computers...

[corbu1]

vor 7 Stunden schrieb NeedzWD40:

I tested the FM and SAS changes today in a variety of trials and here are my current observations:

FMC off is markedly improved, with far more stability in hover, flight, and aerobatics. Pretty much translates 1:1 with the FTR depressed. It feels possible to get into a hover without lots of oscillations now. Aerobatics are markedly improved, with a lot of the troublesome characteristics (excessive roll tendency, yaw oscillations) gone now. Rolls and loops can be executed well and with precision. Hovering feels more like it does with other helicopters, with only minor cyclic changes to keep it steady. I was able to take my hands off the collective and pedals, using only the cyclic to maintain position. If a heading change was required, it wasn't nearly the same fight as it was in the past, with either FTR or pedal.

SAS behavior with roll and yaw is markedly improved, with the tendency to slip into snap rolls significantly reduced. I tried to induce the previous snap roll behavior and was unable to get an unwanted roll to occur, but someone with more issues with it will need to test as I was careful to not get into it (pressing FTR for evasive usually kept me safe).

Taxi behavior with tailwheel unlocked is still something of a challenge, with oscillations still occurring on the pedals. It's not impossible to dial in, but I suspect that most will still have major problems with taxiing and rolling takeoffs.

Takeoff and landing is no longer a "jump" up or down to stay safe and my general feeling is that I can get light on the wheels without risking a rollover. There's still a tendency for the aircraft to want to tip, but it's not nearly as pronounced as it used to be.

Overall, I find it an incremental improvement, but if you were struggling in the past I don't think the recent patch will change perceptions. It hasn't gone to "easy," but it has tamed some of the more annoying quirks.

Thanks for your detailed feedback o. The FM changes. Sounds very promising for me and I‘m looking forward try the recent Patch.

just a question, what is your controls setup?

[jubuttib]

@Raptor9 Any thoughts on how different trim modes have different maximum speeds for the stick movement? When you're in Center Trim mode you can move the stick around pretty fast, but when you're in unsprung stick trim mode the responses are way slower.

I don't know which mode is supposed to be the proper one, but if at all possible I'd love to be able to test the center trim mode's stick movement speed with the unsprung stick trim mode, even if it breaks IC. I use an unsprung stick, so centering the stick isn't really a good idea for me (it's immensely hard to find the center), but even then hovering the Apache is way easier in center trim mode because the input response is so much better. I just can't really use it for flying around because I can't ever trim in this mode. =(

Is there a way to adjust this in the files for the helicopter somehow for local testing? Could you guys in a future update make it so that the center trim mode's faster input speed is used for all trim modes? It really makes a big difference for me with the Apache, I go from feeling like I'm constantly chasing it to being in control.

Edited May 20, 2023 by jubuttib

[NeedzWD40]

11 hours ago, corbu1 said:

just a question, what is your controls setup?

I use a full CH setup, Fighterstick, throttle, and pedals. A slight curve on pitch, bank, and yaw axises, with Y saturation at 90 and a deadzone of 10 on each axis.

[corbu1]

vor 7 Minuten schrieb NeedzWD40:

I use a full CH setup, Fighterstick, throttle, and pedals. A slight curve on pitch, bank, and yaw axises, with Y saturation at 90 and a deadzone of 10 on each axis.

Thanks!

[Moxica]

Well, all my former three sets of muscle memory are out the door. I'm almost at scratch. 
Must start at the beginning.
-Or do something radical.
-Which I did.
-Ordered a stick extention.
 

[Boogdud]

I did a little bit of experimenting with some old sticks I had laying around, an old CH Fighterstick as well as a TM 16000 I had laying around just to see if there was any difference in behavior from other sticks vs my VKB NXT. The difference was staggering, I could fly much easier and with much more stability on both the CH and TM, with the same curve setup and even with zero curves vs my NXT. It's like the NXT is 3x as sensitive as the others. One touch of the stick sends the apache barrel rolling or trying to kill me I'm not sure why this would be, but it's very evident. It's so severe I can get SAS Saturation warning at will just sitting on the runway by moving the stick along it's axis in any direction. All other modules seem to be unaffected.

Does anyone at ED have NXT gear to test with? At first I thought the slower return to center on the nxt (from the built in dampening) might be messing up central trim position settings causing instability after trimming, which it sill may be. But it seems to me that the hall effect sensors  on this model are so sensitive it really makes flying very difficult. 

[NeedzWD40]

39 minutes ago, Boogdud said:

It's like the NXT is 3x as sensitive as the others.

This is extremely curious. What would you rate the throw of the NXT vs the CH stick? With my CH stick, I rarely ever use the full range of motion outside of extreme maneuvers. Perhaps the NXT would work better with a more extreme curve and dead zone?

[Poptart]

1 hour ago, Boogdud said:

I did a little bit of experimenting with some old sticks I had laying around, an old CH Fighterstick as well as a TM 16000 I had laying around just to see if there was any difference in behavior from other sticks vs my VKB NXT. The difference was staggering, I could fly much easier and with much more stability on both the CH and TM, with the same curve setup and even with zero curves vs my NXT. It's like the NXT is 3x as sensitive as the others. One touch of the stick sends the apache barrel rolling or trying to kill me I'm not sure why this would be, but it's very evident. It's so severe I can get SAS Saturation warning at will just sitting on the runway by moving the stick along it's axis in any direction. All other modules seem to be unaffected.

Does anyone at ED have NXT gear to test with? At first I thought the slower return to center on the nxt (from the built in dampening) might be messing up central trim position settings causing instability after trimming, which it sill may be. But it seems to me that the hall effect sensors  on this model are so sensitive it really makes flying very difficult. 

I have a VKB NXT. Seems fine to me. I have a 20 curve on each axis. Is there some maneuver you are unable to accomplish? I don't have another stick to test with but none of the other modules are significantly different.

[admiki]

3 hours ago, Boogdud said:

I did a little bit of experimenting with some old sticks I had laying around, an old CH Fighterstick as well as a TM 16000 I had laying around just to see if there was any difference in behavior from other sticks vs my VKB NXT. The difference was staggering, I could fly much easier and with much more stability on both the CH and TM, with the same curve setup and even with zero curves vs my NXT. It's like the NXT is 3x as sensitive as the others. One touch of the stick sends the apache barrel rolling or trying to kill me I'm not sure why this would be, but it's very evident. It's so severe I can get SAS Saturation warning at will just sitting on the runway by moving the stick along it's axis in any direction. All other modules seem to be unaffected.

Does anyone at ED have NXT gear to test with? At first I thought the slower return to center on the nxt (from the built in dampening) might be messing up central trim position settings causing instability after trimming, which it sill may be. But it seems to me that the hall effect sensors  on this model are so sensitive it really makes flying very difficult. 

Check settings in DCS and see that you get to 100% in game at full throw. This seems like calibration issue, where stick sends full input when it still has some throw.

[EbonySeraphim]

I can understand complaints about the flight model with respect to the flight performance and behavior in certain conditions, but there's a lot of "I can't control it" issues which is not really the flight model. I don't know what there is to talk about beyond a) calibrate your device however the manufacturer is says it should and b) check that DCS sees the full range of movement. Only after those two are confirmed should you then move to consider a) hardware adjustments or b) software adjustments. Hardware means, maybe get another device, maybe change up spring tension, clutch/dampener friction, or use an extension. Software adjustments involves deadzones, curves, and even saturation.

EDIT: forgot to add one thing. Make sure you don't have any unwanted devices influencing the cyclic and collective axes. The latest patch seemed to change in nature slightly and I had plugged in devices that would be fine if untouched. Now, they seem to impact the axes behavior of the device I am using even if they're unmoving and centered (or in any position). I haven't tested in detail but the way it messes things up isn't obvious -- as in, it's not a clear override sticking the axis to the position of the unused device. So if you notice worse behavior with the latest patch, recheck that you only have one device with a binding for those main axes.

If none of that makes the AH-64D flyable for you, then it's purely a "git gud" issue. Anyone who uses the term "squirrely" to describe the AH-64D is unfortunately telling on themselves. Hovering a helicopter is constant corrections -- a motor skill. If you call it squirrely, that means you're overcorrecting and/or not reacting fast enough when the heli is clearly accelerating in a direction so it ends up going all over the place. The thing about motor skills is that you can listen to advice all day, every day, about how to do it better and you won't have a eureka moment where things just click and it's easier. It takes pratice. Hover and fly for 15-30 minutes in the morning, and 15-30 minutes before sleep and you'll notice a huge difference over weeks. What takes your full attention to accomplish in week 1, by the end of week 2 you'll be able to do with 50% effort, then 25%, and eventually it'll slip into near second nature. What you're able to do with the extra focus is observe and deal with other factors: weather and atmospheric conditions, threat conditions, and overall operate some avionics while flying.

Edited May 21, 2023 by EbonySeraphim

[pii]

On 5/19/2023 at 7:28 PM, NeedzWD40 said:

Overall, I find it an incremental improvement, but if you were struggling in the past I don't think the recent patch will change perceptions. It hasn't gone to "easy," but it has tamed some of the more annoying quirks.

This is my experience also. Its better but still a long way to go IMO

[Nereid]

I have just tried the AH-64D again after a few weeks of absence from DCS and with the new update.

And for me it seems easier to adapt than before the patch. Still not easy to fly, but that is to expect. It's a helicopter, so it shouldn't be easy. 

Without any practise for three weeks or more, I was quite surprised that I could takeoff and land this beast without too many issues. Takeoff and landing were far from perfect, but that was to expect too.

There were some strange behaviours as I was trying to land on a flat roof of a building, but hey I was probably a little cocky even to try this.

 

[Boogdud]

9 hours ago, NeedzWD40 said:

This is extremely curious.

8 hours ago, Poptart said:

Seems fine to me.

6 hours ago, admiki said:

This seems like calibration issue

Well, I can't explain exactly why this worked, but here's what happened. After testing with different sticks and different curves for hours, I ended up looking at the collective. For some reason my collective axis had the "slider" checked (I don't remember changing it, but who knows). So I unchecked that box bringing it back to 'normal' axis and applied a small curve (I like some granularity on the collective). Once I unchecked that, suddenly the aircraft became tame as a kitten in comparison to what I was experiencing before?! Small inputs actually corrected movement unlike before where they were over-correcting every single input which cascaded into a wild pony ride every time. Level flight was a breeze (I do now see the sas in action during level flight). It was literally night and day. 

What's really curious is I re-selected the 'slider' setting just to double check and again the cyclic would overcorrect and any collective input seemed to throw the cyclic into fits, working against each other, maybe not quite as bad because I knew what I was looking for, but turning tha setting off for me definitely fixed it. It's at very least back to 'normal'. Extremely weird.