dafcs is wip AFCS on flight far too erratic

[cw4ogden]

AFCS on flight is far too erratic, primarily in the pitch axis.  AFCS on should look more like how the aircraft flies with "Bob" turned on.  A trimmed out CH-47, with AFCS on, should require no significant inputs from the pilot in flight, even with significant changes to the thrust. 

Not trying to throw darts, just document what I see.  My initial impression of the flight model is that it is a very good starting point.  The flight modelling "feels" very close. 

Edited August 11, 2024 by cw4ogden

[TrigaNZ]

AFCS not yet modelled fully I believe.

Edited August 11, 2024 by TrigaNZ

[cw4ogden]

7 minutes ago, TrigaNZ said:

AFCS not yet modelled fully I believe.

 

Yes, I'm aware.  I took that to mean either more features possibly, or more tuning.  If it's just more features, this is a bug.  If it that this a work in progress, can't hurt to get it on the record.  Once trimmed, with the AFCS on, the pilot should able to take their hands off the controls for extended periods of time.  

 

[MPalmer]

Same, the pitch axis needs a little more fine tuning...No feeling greater then coming out of F10 Map view and not seeing any blue sky!

[II.JG1_Vonrd]

Seems like this might be related to my having difficulty achieving a steady hover. Pretty difficult to get external cargo hooked and delivered. Sure, practice will help but compared to the Hip I find the Chinook more difficult.

 

[Samythevilsnail]

Well spotted. ED did mention that the LCTs aren't working yet, so the pitch attitude will be incorrect.

The DASH isn't either, which means you'll get negative stick gradient on the longitudinal cyclic. In gist, the stick position to maintain a higher airspeed will actually be aft of the stick position for a lower airspeed. It's quite confusing to fly with this effect as you might imagine, and makes it difficult to maintain airspeed. 

Edit: @cw4ogden I see what you're saying now. The ILCAS should be doing pitch attitude hold during thrust changes, which requires neither the LCT nor DASH to be working.

Edited August 12, 2024 by Samythevilsnail
Accuracy

[Hawkeye91]

Seems setting the Cyclic trim switch from auto to manual makes a big difference in controlability. Try it out and let me know what you guys think.

[cw4ogden]

54 minutes ago, Hawkeye91 said:

Seems setting the Cyclic trim switch from auto to manual makes a big difference in controlability. Try it out and let me know what you guys think.

It may help as a work around while they iron out bugs, but shouldn't be necessary or even beneficial. 

 

LCT's just tip both rotor disks into the relative wind in forward flight.  It's the equivalent to pushing the cyclic forward in a traditional helo, but it happens on both heads, so you get a more level fuselage at high speeds, and reduced blade flapping.  Thereby Increasing the top speed you can attain before hitting retreating blade stall as well. 

[cw4ogden]

1 hour ago, Samythevilsnail said:

Well spotted. ED did mention that the LCTs aren't working yet, so the pitch attitude will be incorrect.

The DASH isn't either, which means you'll get negative stick gradient on the longitudinal cyclic. In gist, the stick position to maintain a higher airspeed will actually be aft of the stick position for a lower airspeed. It's quite confusing to fly with this effect as you might imagine, and makes it difficult to maintain airspeed. 

Edit: @cw4ogden I see what you're saying now. The ILCAS should be doing pitch attitude hold during thrust changes, which requires neither the LCT nor DASH to be working.

 

Yes the basic functions of the AFCS like pitch attitude hold, bank angle hold, heading hold (not heading select) are ILCA functions.  So can't blame dash or LCTs.

I've got another thread on what the 47 should do AFCS off, specifically how it rotates about the pitch axis with changes in thrust.  It's caused, best explanation I ever came up with is from the unequal tilt of the fore and aft heads.  The aft head sits a few degrees more vertical, but it's enough that equal "collective" amounts of pitch cause the aft head to generate more lift than the front.  So pull thrust, nose down.  Reduce thrust nose goes up.  

That phenomenon is or was rarely understood and it manifested with the AFCS on too, just much more subtle.  Most people had any number of explanations, but chinooks pre-digital AFCS so D not F model, had a tendency to drift forward when adding power at a hover and drift back when reducing.  Can't say if that drift AFCS on persisted to the F, but the AFCS off characteristics definitely didn't change.

Edited August 12, 2024 by cw4ogden

[Samythevilsnail]

3 hours ago, cw4ogden said:

Yes the basic functions of the AFCS like pitch attitude hold, bank angle hold, heading hold (not heading select) are ILCA functions.  So can't blame dash or LCTs.

I've got another thread on what the 47 should do AFCS off, specifically how it rotates about the pitch axis with changes in thrust.  It's caused, best explanation I ever came up with is from the unequal tilt of the fore and aft heads.  The aft head sits a few degrees more vertical, but it's enough that equal "collective" amounts of pitch cause the aft head to generate more lift than the front.  So pull thrust, nose down.  Reduce thrust nose goes up.  

That phenomenon is or was rarely understood and it manifested with the AFCS on too, just much more subtle.  Most people had any number of explanations, but chinooks pre-digital AFCS so D not F model, had a tendency to drift forward when adding power at a hover and drift back when reducing.  Can't say if that drift AFCS on persisted to the F, but the AFCS off characteristics definitely didn't change.

 

Yep definitely still happens AFCS on in the F model. You'll also speed up a couple of knots when raising thrust in forward flight, and vice versa. 

ED did mention that coordinated turns wasn't implemented yet, so we might also be missing a couple of other ILCA functions. 

[cw4ogden]

11 hours ago, Samythevilsnail said:

Yep definitely still happens AFCS on in the F model. You'll also speed up a couple of knots when raising thrust in forward flight, and vice versa. 

ED did mention that coordinated turns wasn't implemented yet, so we might also be missing a couple of other ILCA functions. 

Hey I posted a long thread in the Wishlist, most of it for DCS but my last point is something every 47 driver needs to know regarding a little known pitfall of the FADEC system being put into dual reversionary by tailwinds, or tailwind component as a result of rearward flight / drift.  It may no longer be a problem but if this sounds like something you've never heard of you should read the part I put a bunch of asterisks by.