AFCS/SAS Correcting pilots inputs in lower modes

[Rav3nX]

Hi,

Easily reproducible by moving the cyclic in the hover while watching the controls indicator; it seems that the AH-64s SAS/SCAS is correcting attitude in response to pilot cyclic input. For example a sudden movement to the left will produce a right AFCS input to counteract an input. 

This would not occur in an aircraft, (with or without force trim held) without complicating it by talking about attitude modes and force gradients; in normal hands on flying if a cyclic input is made position sensors on the cyclic system (LVDTs/transducers) indicate to the AFCS that the pilot has made the input and not an external disturbance therefore the AFCS should not produce an error and correct the attitude/control displacement. 

I understand this may be a desired function of producing a more stable flight model, although the effect results in a less than predictable 'feel' and the aircraft trying to return to the original attitude. 

Additionally and possibly related pressing the force trip completely disables the SAS. While depressed the SAS should be correcting for external disturbance and secondary effects of controls, (yaw etc) but not pilot input. It seems while depressed the AFCS is completely disabled.

The force trim implementation on springed joysticks is a tricky issue to deal with, but normally forcing a control even without force trim held (whine NOT in attitude hold anyway) should not result in a SAS correction regardless of force trim release position. (Forcing a cyclic against a mag brake will have the same result as holding force trim; though only while not in attitude hold)

Anytime the joystick is moved (or forced) from the trimmed position should cause the trimmed position to follow. The current behaviour with Attitude hold off is more like what I would expect with attitude hold on. 

Additionally turning the mag brake off will not disable SAS. Not in any helicopter. (So long as the cyclic is held, it should be modelled as if this is the case)

(B1/B2 Avionics/Airframe Part 66 maintenance licence and several hundred hours of twin turbine helicopter experience -B212/B412 (functionally similar AFCS system))

Offered as assistance as it appears some mode/functional confusion has occured. (Understandable). 

Sorry for the dot pointing breif summary; writing on a phone. 

[Sandman1330]

Ahaha, as I was reading your post, I was like "this guy is totally a B412 pilot."  Then I got to the end and and was like, oh yeah, figures!

While I can't speak for the Apache specifically, I concur with your assessment. I find the SCAS in the DCS Apache always seems to be working against me.

Though (a little off topic), I'm not sure your assessment of API movement in SAS mode on the 412 is entirely correct. What you've stated is consistent with how our flight manual says it SHOULD work, but if I push against the mag brake in the 412, I'm still getting API movement (and even autotrim fails), which would imply the APIs are still moving.

There is also a use case where the APIs should work against you a little bit (in SAS mode), and this is in dampening aggressive stick movement. If I slam the stick to the side, the APIs will actually move against me to dampen that movement somewhat and make it less jarring. But they should quickly catch back up.

[Rav3nX]

Close lol; a bit of a medical issue prevented an even worse condition; I can say proudly I'm not the holder of a H-CPL; So I get to enjoy the life with normal benefits such as a fixed abode and a wife and kids that reside with me; a functioning liver too!  (Said in jest). I did most of my license before this issue pulled up my medical; and in a almost 20 year career since in a large family business (helicopter training which became aerial fire-suppression) has afforded me much more flying experience than a normal engineer. I'm mainly an avionics tech, that flies at any and every excuse. However; possibly close to 200 hours in 412's, probably 10-20 in BK117's, S76's and 206L and maintaining AS332L1's and UH-60s has given me over the years a good exposure to all the bits and pieces. I'm routinely board-level repair of AFCS helipilots and couplers etc; so, these are systems I know backwards (born of sadness-level fascination)

I don't say this as a grandstanding; it’s just my career; I do just so that hopefully my input can be distinguished from the noise because...I really want this Ah-64 to be all it can be. 

It is sooo close, functionally in assessed in individual components the functionality is all there, the bits work the way they should its just the engagement level logic has become confused I think; and that's totally understandable; these are complicated systems. Unfortunately, too many times I have heard pilots on here and around the net explain them, although usually what they say is true; its often-missing fundamentals that if the receiving dev or individual was missing or had a misconception about a particular function its easy to confuse the picture. That’s not meant as a slight; there is a ceiling to expected understanding of these systems for the average Joe, and a lot of real basic concepts are usually missed or assumed and the misconceptions compounded. I have wanted to put the hydraulic rig on one of our 412’s many times to make the ‘helicopter autopilots 101’ video and I still might.

Obviously, I too am limited in real world Ah-64D experience (until I find one! -jk), however having flown enough different types and read the real dash one to know that; functionally a lot of these modes are very similar to everything else; some things work the way they work because that’s just how they all work. Same as right pedal for right turn; its that way because that’s how it is.

The flapping you describe is true; and that’s a function of the RC circuits rate limiting function of the helipilot; fringe cases that at this point I think would only confuse the issue and I suspect the description of cases like these is what’s lead to the confusion at times.  I really think that there has been confusion of modes, and possibly a background in non-western aircraft systems from the devs has resulted in what is currently the case; and obviously the lack of public extensive data and detailed functional descriptions isn’t helping.

It seems that the focus has been shifted to the high-level modes which has left the confusion with SAS to sneak through; and currently I feel that the basics need to be given the highest priority and the rest can follow; otherwise, there is just function adding to an unworkable system, and its unfortunately really destructive to the experience as is. Flying straight and level shouldn’t be the task it currently is.

 

To the Devs; I do have good non-classified data and plenty of very-high level experience with these systems and I am more than happy to volunteer my time to assist in the understanding and improve this; if the assistance is wanted;  and I would caution against taking advise on engagement logic from Pilots, with respect I know that the many that I deal with have ‘enough’ understanding but much of the functionality is transparent to the operator.

I do stress though you guys have absolutely nailed the ‘how’ for the individual components given its current level of development; just not quite the ‘when’, and I offer my assistance if wanted. I’m sure many of the issues I describe are known.

Cheers

[Raptor9]

First off, I want to stress this is not a hit on you or any other person that has experience with real-world helicopters or helicopter components, either as a pilot or as a maintainer. However, I want to respectfully identify the fact that flight experience in civilian helicopters or experience maintaining them does not necessarily translate into practical application of a military attack helicopter's flight controls or flight control computers. Especially one that is produced by an entirely different manufacturer. Yes, there are some control principles that are universal across most rotorcraft types and designs, but that is a limited resource.

Second (and I say this as someone that has experienced both sides of the coin as both a maintainer and a pilot), simulating a helicopter is not just about re-creating an identical set of data points and numbers, it's also about replicating the "feel" of the aircraft, which is often difficult to translate into a simulation game that lacks the physical force feedback into the body, and deals with a wide variety of input devices and hardware support. To that end, the value of pilot input should not be underestimated in making the sim feel believable. It doesn't mean that they are the golden standard for what goes and what doesn't, because they are all human with subjective assessments. But between just two members of the SME team are a combined count of approximately 4,700 flight hours across multiple decades in just the AH-64 alone, with intimate knowledge of the control logic.

Finally, the dev team does have members with extensive aeronautical expertise, and are still working to improve the behavior of the AH-64D. I know not everyone reads every individual thread regarding the flight model; I can assure you that the current behavior of the DCS: AH-64D is not finalized, and improvements are already showing in internal builds. However, I cannot provide timelines on when such fixes will make it into Open Beta.

Having said all that, community feedback is still considered a valuable resource. If a need arises that the dev team feels your expertise and input could be useful in the product, we will certainly keep that in mind.