Previous Thread
Next Thread
Print Thread
Rate This Thread
Hop To
#4554183 - 01/30/21 02:41 AM Crosscoupling poll  
Joined: Jun 2005
Posts: 2,132
Reticuli Offline
Member
Reticuli  Offline
Member

Joined: Jun 2005
Posts: 2,132
Dayton, OH, USA
Considering crosscoupling OFF is not working correctly for FM 0 & 1 but (with the necessary tweaks) most everything else works better with FM 0 compared to FM 2, I thought I'd do a poll to get some feedback on its use in the community. Thanks.

Do fly with Cross Coupling turned OFF in EECH?
multiple choice, up to 1 choices
Votes accepted starting: 01/30/21 02:39 AM
You must vote before you can view the results of this poll.
Last edited by Reticuli; 02/03/21 03:38 PM.

The term "necroposting" was invented by a person with no social memory beyond a year. People with a similar hangup are those o.k. with the internet being transient vapor.

http://www.openfuelstandard.org/2011/12/methanol-wins-open-wager.html

Saitek X65 and X52, Glide, Winx3D, and GlovePIE Profiles http://library.avsim.net/search.php?SearchTerm=reticuli&CatID=miscmisc

http://library.avsim.net/register.php

X52 + Silicone Grease = JOY stick
#4554364 - 02/01/21 09:05 AM Re: Crosscoupling poll [Re: Reticuli]  
Joined: Jun 2005
Posts: 2,132
Reticuli Offline
Member
Reticuli  Offline
Member

Joined: Jun 2005
Posts: 2,132
Dayton, OH, USA
By the way, the real Apache and Viper have some tail-mixing compensation automatically as well as heading hold modes in their hybrid stability augmentation stuff and the Comanche had this all the time through its triple-redundant digital fly-by-wire. So if you are flying American helos with the good cockpits, the only one that should really have CC OFF if you want to be realistic is the Kiowa, and if you're flying coaxials, you'll want FM 0 so you can yaw in them. With tweaks, EECH really gets right now the feel of an AFCS helo, so, IMO, in for a penny, in for a pound.

While FM 0 & 1 also have weak ground effect and translational lift, especially the latter, FM 2 is much worse in this regard, even when you set translational lift to OFF with FM 2 to get it to work with its current bug. FM 0 & 1 have at least 2X the yaw maximum they should that requires you to divide your available yaw axis input, but FM 2 has a fraction of the yaw maximum it should on coaxials with no mitigation currently possible other than not using coaxials or switching to FM 0 or 1. FM 1 has literally no realistic aerodynamic streamlining in the yaw, though, which leaves FM 0 with the tweaks I have previously outlined in the ini, requiring some means of dividing your available yaw axis, and finally now also requiring crosscoupling ON as the best current universal FM setting for all helos... and this way now it rivals FSX and X-Plane AFCS 'digital FBW' helos, I may add, as a potential successor to Longbow 2. That's the good news, though there's a dance involved :-)

With no tweaks or mitigation methods, the FMs are 'functionally' pretty terrible currently, but the code for the dynamics modeling in EECH appears to be generally deeper than I'd thought, and those superficial vibrations are helping things, too. It does leave some conundrums, though, in light of the crosscoupling OFF only working properly on FM 2 right now. Firstly, it means we must keep at least FM 0 and 2 for the foreseeable future while still trying to find ways of improving both, such as limiting yaw input in FM 0 and increasing available yaw in FM 2, among other things. Preferably for FM 0, we might be able to find a way to use speed (or rather the lack of) inside the sim to divide the yaw axis gradually, since having a lot of yaw authority during high speed is desirable but not desirable in a hover. At RBS speed you'd have full yaw authority, while at 0 speed you'd have a fraction of that... with some curve applied, probably log. It might also mean that instead of an ini setting there should be an automatic switching between FM 0 and 2 on-the-fly based on if you're flying a coaxial helo (or planning to) or whether you have crosscoupling OFF, meaning you wouldn't be able to do both and you don't need to anyway, and it could switch automatically.

So again, here's the weird criteria...

Crosscoupling off AND you won't be flying coaxials = use FM 2 with ini tweaks (including dmrl=1.1), translational lift showing OFF, and don't expect enough yaw on coaxials if you do jump into one

Crosscoupling on OR you will be flying coaxials = use FM 0 with ini tweaks (including dmrl=0.9) plus some means of (preferably manually & dynamically, like the divider on a throttle thumb rotary) dividing available yaw by about 1/2 on non-coaxials and down to 1/3 (66% reduction) for coaxials

I know that's a lot to remember, but that's how to get the best out of EECH 1.16. Please make your selection in the poll to help prioritizing which FMs to improve.

***

Edit:

Upon further tests, I have discovered that FM2 is not simply reversing the translational lift value in the Options - Dynamics in-sim menu, but is also changing lift while in a hover and it's not just some wind effect. So FM2's translational lift appears to be broken quite a bit worse than I realized and in order to even get hover lift to work right you must currently have translational lift OFF, not just to get its anemic translational lift working. That prior thought about automatically changing FM based on whether you are flying a coaxial (use FM0) or have crosscoupling OFF (use FM2) looks like it would actually work. If you absolutely must have manual tail mixing with your feet all the time, then FM 2 with translational lift set to OFF (to get it on) is your only option right now in EECH 1.16. The crosscoupling OFF lift calculations are simply malfunctioning too much when pitched forward to do manual tail mixing yourself in FM0.

In case you haven't read the "FLIGHT MODEL or where is it ?" thread, these are the ini dynamics values:

flight_model=0 # FM 0 generally and especially for coaxials. FM 2 if you insist having crosscoupling off plus remember to turn off translational lift setting. FM 1 is poor quality
drbs=1.0 # Retreating blade stall
drv=1.0 # Rudder value, both its drag, fake yaw damping, and coaxial differential, thus directly affecting coaxial and inversely affecting non-coaxial yaw rates
dra=1.0 # Rudder acceleration, pedal ramp up
drd=0.1 # Main rotor drag, resistance to Y & X rotation of disc, but also loss of speed with pitching
dmrl=0.9 # Main rotor lift. 0.9 for FM 0, but 1.1 for FM 2. Results in about 90% torque 2500ft 7730kg Apache OGE hover, just over 70% torque rolling takeoff (IGE+ETL) at 50% collective, and also ability to still barely fly after failure of one of two engines
dtrd=10.0 # Tail rotor drag, aerodynamic streamlining
dyal=2.5 # Yaw altitude loss. You may prefer 5.0 if you have crosscoupling off
dyad=3.0 # Global air density modifier. 3.0 with above settings to get just over 20000ft ceiling at 100kts in avg Apache load

And here is an example of collective 'nonlinearity' settings (it's not really nonlinear, just changing the min and max):

nonlinear-collective-zone1=0.4 # Use a value like this to offset zone1 bottom percentage
nonlinear-collective-zone2=0.9 # Depends on the throttle you're using to get 120% at just exactly full rearward
nonlinear-collective-percentage-at-zone1=54.0 # 6% minimum torque. 50.0 gives more realistic reduced deceleration but shows 10% instead of the real-life 6% minimum.


Last edited by Reticuli; 03/26/21 08:22 PM.

The term "necroposting" was invented by a person with no social memory beyond a year. People with a similar hangup are those o.k. with the internet being transient vapor.

http://www.openfuelstandard.org/2011/12/methanol-wins-open-wager.html

Saitek X65 and X52, Glide, Winx3D, and GlovePIE Profiles http://library.avsim.net/search.php?SearchTerm=reticuli&CatID=miscmisc

http://library.avsim.net/register.php

X52 + Silicone Grease = JOY stick
#4554469 - 02/01/21 06:36 PM Re: Crosscoupling poll [Re: Reticuli]  
Joined: Jun 2005
Posts: 2,132
Reticuli Offline
Member
Reticuli  Offline
Member

Joined: Jun 2005
Posts: 2,132
Dayton, OH, USA
This obviously doesn't address automatic attitude hold (auto trim, dynamic trim update, etc) or having a yaw divider axis, but both of those can be addressed with GlovePIE in the meantime. It also does not address the coordinated turn yawing during a bank below ETL speed (which should just be side-slipping), but that will have to be addressed in the source code. Potential automation we might be able to hard code into the sim that would not be user-accessible:

Crosscoupling ON or flying a coaxial:

flight_model=0
drv=1.0
dra=1.0
drd=0.1
dmrl=0.9
dtrd=10.0
dyal=2.5
dyad=3.0

Plus Translational Lift Options-Dynamics setting to ON.

***

NOT flying a coaxial and you have crosscoupling OFF:

flight_model=2
drv=1.0
dra=1.0
drd=0.1
dmrl=1.1
dtrd=10.0
dyal=5.0
dyad=3.0

Plus Translational Lift Options-Dynamics setting to OFF.

Last edited by Reticuli; 04/26/21 02:53 AM.

The term "necroposting" was invented by a person with no social memory beyond a year. People with a similar hangup are those o.k. with the internet being transient vapor.

http://www.openfuelstandard.org/2011/12/methanol-wins-open-wager.html

Saitek X65 and X52, Glide, Winx3D, and GlovePIE Profiles http://library.avsim.net/search.php?SearchTerm=reticuli&CatID=miscmisc

http://library.avsim.net/register.php

X52 + Silicone Grease = JOY stick
#4555064 - 02/06/21 01:22 AM Re: Crosscoupling poll [Re: Reticuli]  
Joined: Sep 2016
Posts: 585
Phoenix54C Offline
I am just a cowboy
Phoenix54C  Offline
I am just a cowboy
Member

Joined: Sep 2016
Posts: 585
Emerald Isle
Ok so no1 else came back 2 u but i like your enthusiasm...

Just tried on and off and cant see any diff .

However i really have no idea of what crosscoupling is?

dra=1.0 is a must with flight model 0 for me as the dam choppers just go crazy if u leave it at default 8.

Way more used 2 Jets but what beats a good heli sim for a change...

You guys rock and this is the Ultimate Combat helo sim to date,

Keep up the good work and stay safe...

#4555121 - 02/06/21 07:45 PM Re: Crosscoupling poll [Re: Phoenix54C]  
Joined: Jun 2005
Posts: 2,132
Reticuli Offline
Member
Reticuli  Offline
Member

Joined: Jun 2005
Posts: 2,132
Dayton, OH, USA
"Just tried on and off and cant see any diff ."

You tried what on and off?

"However i really have no idea of what crosscoupling is?"

Also known as automatic tail-mixing based on collective position. On older or hybrid-designed helos with tail rotors that have this it's either a gear ratio adjustment that the collective also affects (in addition to the main rotor uniform pitch) or it's a heading hold mode the SAS adds inputs for. In an AFCS 'digital' FBW helo like Comanche, it's purely computer-controlled. If you turn it off inside the in-sim options-dynamics menu, you have to keep the pedals pushed most of the time to maintain heading at low speed for the helicopters with a tail rotor. As I said, turning crosscoupling off with FM 0 causes a malfunction of the tail and lift calculations when in forward flight. For instance, you can pitch the nose down 45 degrees, add pedal, and maintain altitude. So if you have CC off and aren't flying a coaxial helo, use FM 2. The instructions in post 3 shouldn't lead you astray.

Last edited by Reticuli; 02/06/21 08:08 PM.

The term "necroposting" was invented by a person with no social memory beyond a year. People with a similar hangup are those o.k. with the internet being transient vapor.

http://www.openfuelstandard.org/2011/12/methanol-wins-open-wager.html

Saitek X65 and X52, Glide, Winx3D, and GlovePIE Profiles http://library.avsim.net/search.php?SearchTerm=reticuli&CatID=miscmisc

http://library.avsim.net/register.php

X52 + Silicone Grease = JOY stick
#4555243 - 02/07/21 07:07 PM Re: Crosscoupling poll [Re: Reticuli]  
Joined: May 2001
Posts: 5,462
Polak Offline
Hotshot
Polak  Offline
Hotshot

Joined: May 2001
Posts: 5,462
New York, NY
I am pretty much new to this stuff so bear with me. I like crosscoupling setting off but that induces slow rotation when hovering, which I can only counter with my Z axis on Saitek x-52 stick. I am trying to find and program tail rotor trim to get rid of this rotation. Any hint what key commands I am missing to cure that? Thanks

#4555244 - 02/07/21 07:13 PM Re: Crosscoupling poll [Re: Polak]  
Joined: Jun 2005
Posts: 2,132
Reticuli Offline
Member
Reticuli  Offline
Member

Joined: Jun 2005
Posts: 2,132
Dayton, OH, USA
Why would you want crosscoupling off and have some program automatically compensate for you? That's literally what crosscoupling is. When you deactivate the AP, you'll see it get confused briefly. There is actually a crosscoupling ratio in the dyn files that probably is how much you want it to do it, so if you put it at like 0.8, I would guess it would result in similar to Longbow 2 where with very hard collective changes you get a little yaw. You, of course, have the ini setting for altitude changes based on yaw, which is the reverse effect.

Last edited by Reticuli; 02/07/21 07:13 PM.

The term "necroposting" was invented by a person with no social memory beyond a year. People with a similar hangup are those o.k. with the internet being transient vapor.

http://www.openfuelstandard.org/2011/12/methanol-wins-open-wager.html

Saitek X65 and X52, Glide, Winx3D, and GlovePIE Profiles http://library.avsim.net/search.php?SearchTerm=reticuli&CatID=miscmisc

http://library.avsim.net/register.php

X52 + Silicone Grease = JOY stick
#4555247 - 02/07/21 07:52 PM Re: Crosscoupling poll [Re: Reticuli]  
Joined: May 2001
Posts: 5,462
Polak Offline
Hotshot
Polak  Offline
Hotshot

Joined: May 2001
Posts: 5,462
New York, NY
Ok this is all little too confusing to me. I need to study your posts some more.

#4555257 - 02/07/21 09:38 PM Re: Crosscoupling poll [Re: Polak]  
Joined: Jun 2005
Posts: 2,132
Reticuli Offline
Member
Reticuli  Offline
Member

Joined: Jun 2005
Posts: 2,132
Dayton, OH, USA
OK.

I've been working with one of the mods to try and decipher some of what the source code does, but it's still a lot of guesswork so far. This thread increasingly enforces my previous hunch that most of the users of EECH over the years have probably rarely even adjusted the in-sim options settings, let alone adjusted the EECH.ini settings much. Understandable, because it's kind of a hassle. Hence my idea that the next step probably should be automating some of this stuff based on the helicopter people jump into to fly in the sim. That way people don't have to fart around with the ini settings every time they want to change the CC setting in the sim or use a different helicopter. It's shocking just how deep the EECH flight model code is given how bad it flies by default. Just deviating post #3's settings advice a little and suddenly your Apache can cruise at 50,000ft or requires 110% torque to hover or starts spinning wildly on the ground and your rotors rip off. The ini REALLY has to be tweaked just right to get it not to fly like garbage, and people currently need to also pay attention to what their in-sim options settings are, like crosscoupling (CC) and translational lift (ETL)... the latter which is actually wrongly called transitional lift right now in the sim options menu.

I'll admit, post #3 is a lot of dance to go though, but it should get you on the right track, even if you're using a professional pole cyclic with a mechanical trimmer, a joystick with its spring removed, a FFB stick, or just plain don't want to bother with that PPJoy + GlovePIE stuff in the flight model thread. If you have a normal joystick with a spring (used to be me) or you have a pressure stick (me now), I HIGHLY recommend giving GlovePIE a chance to get auto trim and yaw limiting. Again, it gives you a crude Longbow 2-like rate-command / attitude-hold that was the Comanche's default flight law, and you get to divide your yaw axis on-the-fly if you have something like a spare thumb rotary to put the divider. Otherwise, you can just divide the available yaw by 2 or 3 when using FM 0. All you gotta do is get your axis assignments in PPJoy to match up to the script. The explanation is in the script comments.

That said, we did find each axis in seemingly-unused 'pressure' versions (pressure cyclic X, pressure cyclic Y, etc) in the source code, though, so it's possible those might have originally been programmed for a similar purpose of manipulating in-sim virtual controls where you 'adjust' the intermediary (the manual trimmer button and trim hat assignment does that already), rather than directly control them... or they might unfortunately just be something related to FFB. It will be interesting to find out what these pressure versions of each axis actually do. Considering how a lot of stuff seems to be in the sim code, if they aren't FFB-related, then I would guess they're not coded quite right and will need repairs to get them to work as well as the GlovePIE script.

Last edited by Reticuli; 02/08/21 02:47 AM.

The term "necroposting" was invented by a person with no social memory beyond a year. People with a similar hangup are those o.k. with the internet being transient vapor.

http://www.openfuelstandard.org/2011/12/methanol-wins-open-wager.html

Saitek X65 and X52, Glide, Winx3D, and GlovePIE Profiles http://library.avsim.net/search.php?SearchTerm=reticuli&CatID=miscmisc

http://library.avsim.net/register.php

X52 + Silicone Grease = JOY stick
#4555261 - 02/07/21 10:29 PM Re: Crosscoupling poll [Re: Reticuli]  
Joined: May 2001
Posts: 5,462
Polak Offline
Hotshot
Polak  Offline
Hotshot

Joined: May 2001
Posts: 5,462
New York, NY
I will give some further time to check all in EECH. It is one of the most sophisticated helicopter simulation bar none however its strength is to my very ignorant and very initial feeling is in FM, avionics and gameplay. Other areas of this sim are also good but there are some which are outright annoying, but I will not elaborate on them.

When you are talking about the depth of certain items in the source code it is quite obvious and commonly found confirmation that the creators of those simulations some 20 years ago were truly pioneering and ingenious folks who deeply cared about the depth and sophistication of their projects. Having now more time I am truly amazed to discover how deep and complex some of them were and how ignorant and wasteful our reception as the end users/consumers has been, Truly sad if not outright tragic situation and outcome.

So my applause and admiration go to all of us who are willing to tinker with all that is so good yet nearly abandoned and forgotten in those simulation gems. No wonder why EECH source code had become available and many many thanks for that! I hope to stick some longer around and continue this discussion.

Last edited by Polak; 02/07/21 10:30 PM.
#4555269 - 02/08/21 02:44 AM Re: Crosscoupling poll [Re: Reticuli]  
Joined: Jun 2005
Posts: 2,132
Reticuli Offline
Member
Reticuli  Offline
Member

Joined: Jun 2005
Posts: 2,132
Dayton, OH, USA
"Mustang 1-2 seeking firing position."
"Mustang 1-2 seeking firing position."
"Mustang 1-2 seeking firing position."
"Mustang 1-2 seeking firing position."
"Mustang 1-2 seeking firing position."

That can get annoying, as can jumping in a landed helo with the gear retracted that starts on fire, not being able to share target data in the flight, etc, etc. Lots of room for improvement.


The term "necroposting" was invented by a person with no social memory beyond a year. People with a similar hangup are those o.k. with the internet being transient vapor.

http://www.openfuelstandard.org/2011/12/methanol-wins-open-wager.html

Saitek X65 and X52, Glide, Winx3D, and GlovePIE Profiles http://library.avsim.net/search.php?SearchTerm=reticuli&CatID=miscmisc

http://library.avsim.net/register.php

X52 + Silicone Grease = JOY stick
#4555477 - 02/09/21 07:23 PM Re: Crosscoupling poll [Re: Reticuli]  
Joined: Jun 2005
Posts: 2,132
Reticuli Offline
Member
Reticuli  Offline
Member

Joined: Jun 2005
Posts: 2,132
Dayton, OH, USA
Issue: Yaw authority is too high with FM0 at very low or hover speeds. Not everyone right now wants to use PPJoy+GlovePIE to divide the yaw axis, and it should be possible to code this to occur automatically.

When using FM0 and without Hokum we want 0 indicated airspeed (IAS) to cause a yaw input divider of 2, but with IAS=Vne we want divider to be 1. With FM0 and actually with Hokum or Hind, we want the max divider (when IAS is 0) to be 3. With FM2 or Blackshark is flown, we want the original yaw input to get through all the time, and therefore we want the yaw input divider always just at 1.

So here is a hypothetical EECH code for this using a combination of arbitrary and guessed names for the values:

If FM=2 or Blackshark=yes, then YawPart = 0, else If Hokum=no and FM=/=2, then YawPart = 1
If (Hokum=yes or Hind=yes) and FM=/=2, then YawPart = 2
YawDiv = YawPart + 1 – IAS / Vne * YawPart
NewYawInput = OldYawInput / YawDiv

The use of an exponential or log calculation may be necessary to eventually get the most out of this. I suspect the relationship between IAS and the yaw input divider should probably be nonlinear with nearly full available yaw input at substantially less than Vne, but this simple linear relationship is a good first start.

For more beyond what is in this above thread, see:

https://SimHQ.com/forum/ubbthreads.php/topics/3781630/5/flight-model-or-where-is-it

I believe by implementing the hardcoded conditional ini section earlier in this thread and something like this script, that much of the flight model problems will be resolved from the user perspective.

The next step after that I think ought to be investigating those pressure versions of the input axises and also to figure out a way to get coordinated turns (i.e. yaw with bank) not to begin until after ETL threshold speed rather than at any speed as it does right now. If that's not possible, maybe we find a way to make coordinated turns one axis and pure yaw-only its own axis. Then trying to improve the vibration characteristics, such as when passing through the ETL threshold or doing extreme input changes would be good. Then perhaps do something about the not-entirely-correct VRS recover stuff would be in order. Promising.

Edit: BTW, I was slightly wrong about the yaw authority issues in FM 0 & 1. You also need to divide the Hind yaw by 3 like you have to do on the Hokum, but the Blackshark requires a divider of 1 (no reduction). So it's not as simple as just all coaxials needing only a third of the yaw to not get screwy. All the other helos are 'controllable' at a divider of 1.5-ish, but they reach US Army ADS-33 standards of minimum authority needed at a divider of 2 (half authority) when in a hover. So while they don't spin apart divided by 1.5, it's still more yaw than necessary. As such, I've corrected the above content.

Last edited by Reticuli; 02/14/21 02:44 AM.

The term "necroposting" was invented by a person with no social memory beyond a year. People with a similar hangup are those o.k. with the internet being transient vapor.

http://www.openfuelstandard.org/2011/12/methanol-wins-open-wager.html

Saitek X65 and X52, Glide, Winx3D, and GlovePIE Profiles http://library.avsim.net/search.php?SearchTerm=reticuli&CatID=miscmisc

http://library.avsim.net/register.php

X52 + Silicone Grease = JOY stick

Moderated by  RacerGT 

Quick Search
Recent Articles
Support SimHQ

If you shop on Amazon use this Amazon link to support SimHQ
.
Social


Recent Topics
Carnival Cruise Ship Fire....... Again
by F4UDash4. 03/26/24 05:58 PM
Baltimore Bridge Collapse
by F4UDash4. 03/26/24 05:51 PM
The Oldest WWII Veterans
by F4UDash4. 03/24/24 09:21 PM
They got fired after this.
by Wigean. 03/20/24 08:19 PM
Grown ups joke time
by NoFlyBoy. 03/18/24 10:34 PM
Anyone Heard from Nimits?
by F4UDash4. 03/18/24 10:01 PM
RIP Gemini/Apollo astronaut Tom Stafford
by semmern. 03/18/24 02:14 PM
10 years after 3/8/2014
by NoFlyBoy. 03/17/24 10:25 AM
Copyright 1997-2016, SimHQ Inc. All Rights Reserved.

Powered by UBB.threads™ PHP Forum Software 7.6.0