reported Artificial horizon

[jef32]

It's a question that I have, since I've bought P-51, years ago:

 

 

The succion is OK and despite that, the horizon loose its good position very often. Just when I'm taxying in exemple, even if I've uncaged it before moving, the instrument loose the horizon with the first turn.

 

 

 

And in flight it's the same: Very often I must stabilize my flight Vz=0 and with my sight i must cage/uncage again to have an approximate horizon ( more difficult in a mountain zone with the false appreciation of the horizon by the human sight )

[saburo_cz]

Do not taxi with unceged horizon!

Uncege it before take-off after you line up with RWY.

 

During flight, just fly with ball in the middle.

 

These will help to keep horizon works properly.

 

This instrument is old technology without correction of errors caused by turns. (means rotation around the vertical axis - Yaw)

[jef32]

Ok for uncaging just before take-off. But do you think that uncaging before moving can damage the artificial horizon ? And that's why after, it loose its good behaviour during flight ?

[saburo_cz]

But do you think that uncaging before moving can damage the artificial horizon ?

No, it will not damage horizon.

It will cause horizon misalignment due to plane movement in third axis, Yaw.

Horizon measures and shows only Pitch and Roll movement, but Yaw movement have effect on horizon gyroscope too.

 

And that's why after, it loose its good behaviour during flight ?

Align plane, just try to fly straight without any roll, pitch and yaw movement, and then align horizon, there is no another advice or help..

[Andy1966]

Ok for uncaging just before take-off. But do you think that uncaging before moving can damage the artificial horizon ? And that's why after, it loose its good behaviour during flight ?

 

yes, never taxi uncaged. every time you change direction, it throws it further out of alignment, I usually leave mine caged until after take off then get well trimmed straight and level then uncage it. during violent maneuvers, it can get out of alignment, and just get straight and level and cage and uncage it will correct it.

[-0303-]

On 3/2/2019 at 4:42 PM, saburo_cz said:

This instrument is old technology without correction of errors caused by turns. (means rotation around the vertical axis - Yaw)

To unbury the lede the DCS AH does not work correctly in any old plane up to and including the F-86. This because a key modelling of Artificial Horizon (AH) self correction is completely missing rendering the DCS AH useless for any real world simulation. It drifts and it shouldn't.

~

It is not true that WW2 Artificial Horizon (AH) technology lacks correction. The correction mechanism is the pendulous vanes that continuously self correct toward the gravity vector.

Bench example of actual P-51 AH self correction. Un-mute.

It's true that small momentary errors are induced by a 180 degree turn (for example). But adding an additional 180, ie doing a full 360, the opposite error will cancel the first error out. So forever turning will not skew the AH erroneously. The actual error is caused by the centrifugal force acting on the inertia of the free-hanging pendulous vanes.

If one performs a 180 turn and then flies straight and level, the small AH error will quickly self-correct by the pendulous vanes. Actually it's wrong to think of it as having to fly straight and level  to self correct. The pendulous vanes works continuously at all times though they can be momentarily overcome by violent maneuvers. It's obviously insane to have to fly perfectly straight and  level for AH self-correction.

An additional important limitation applies to WW2 era AH (and in to the 50's). For mechanical protection the AH has built in mechanical stops at +-60 pitch and 110 roll. In practical terms the AH will go bananas, coco-for cocopuffs at every roll or looping. DCS AH doesn't do this. It should.

This real life Spitfire video shows this. Roll at 17:01, 27:15, 31:40, 32:40, 37:50, 38:47, 39:08, 39:40. Also note self-correction (full 90 degrees) over 9 minutes from 18:05 - 27:00.

 

This video training video explains much of this. How it works, momentary errors caused by turns  (180, 270, 360), by acceleration or deceleration. Also mentions the  60 and 110 degree mechanical stops. First 11:25 minutes covers vacuum driven (WW2) AH.

 

This bug is more than 10 years old (in that thread read kablamoman).

Not just P-51 as mentioned, every old old plane up to and including F-86. Testing this is easy in a plane  with a cageable AH:

Cage AH, roll 30, 60, 90  whatever, uncage, roll level. Fly straight and level for some minutes (10 min should be enough). If it doesn't self correct it's wrong. Tested P-51, Spitfire, Bf109, F-86 myself. All fail.

Edited February 26, 2024 by -0303-

[kablamoman]

@NineLine As -303- mentioned, it's been over a decade. Any chance we can actually get a fix for this long-standing problem?

[grafspee]

This issue always got me feeling that it is something very wrong in modeling, i had no proof for this. But when i flew in clouds even if i stayed very gentle artificial horizon could not provide me enough stability to go through cloud layer. Fact that this issue is decade old is jut straight forward outrages. And i put all my money that ED team knew this.

Edited February 26, 2024 by grafspee

[=475FG= Dawger]

I know that it being buggy sucks but you dont really need it to fly through the clouds for short periods. 
 

Other instruments provide all the information you need to fly blind. 

[-0303-]

22 hours ago, grafspee said:

This issue always got me feeling that it is something very wrong in modeling, i had no proof for this. But when i flew in clouds even if i stayed very gentle artificial horizon could not provide me enough stability to go through cloud layer. Fact that this issue is decade old is jut straight forward outrages. And i put all my money that ED team knew this.

Though all documentation supports DCS AH being erroneously modeled, all the proof one needs is simple common sense. The DCS AH lacks self correction, it drifts and therefore it's useless.

Doesn't matter if it's cageable or not (manually correctable in clear skies). In zero visibility neither is correctable.

Unless doing acrobatics one never  need to correct the AH because it self corrects. Even after acrobatics it'll correct itself, it might just take a few minutes.

I think the vast majority of DCS users fly in clear skies and ignore the AH because they never need it.

The AH behavior annoyed me from the very beginning but for the longest time I thought I was missing something, that I just didn't "get it". I couldn't, didn't want to believe DCS had modeled it wrong. Until I thought about it and realized this behavior is stupid on it's face and can't possibly be correct.

ED team can not not know this, the AH function must be 101 elementary for any professional or serious pilot because the AH is one of the six basic flight instruments.

I'd prioritize this bug over hundreds of minor graphic flaws.

[PawlaczGMD]

I found this thread before rereporting this, so let's bump.

 

@BIGNEWY @NineLine

As outlined in this thread, all older (WWII) artificial horizons in DCS are lacking the self-correcting pendulous vanes mechanism, which makes them not very useful.

The drift due to maneuvers is modelled, but no self-righting is modelled, so the AH just build up drift. In reality, they should correct if you're flying the aircraft straight and level. So errors caused by maneuvers should be there, but they should decrease gradually when you stop maneuvering.  They don't do that.

I don't have specific data on the Ka-50 but it is rather obvious that it must also have some self-righting property - this one is the most useless of them all (backup one). 

[Bowie]

Came here to post on this very problem RE: the P-51D-25-NA.

@NineLine

Please Advise.

 

Bowie

 

[NineLine]

Thanks all, while I do not think it is game-breaking on the P-51D as you can reset when you need to, I have requested a review of all our Artificial Horizons and see if we can improve the modelling. 

[ssn]

Artifical horizon drifts on shutdown and newer corrects itself neither stationary on the ground with engine running nor in a level flight.

 

 

spitfire-gyro-broken-by-shutdown.trk

[Art-J]

Gyro AH not self - correcting in any phase of flight has been a DCS problem for many years and affects all warbirds. Brit ones are just affected the most since they lack cage/uncage AH feature that US and German ones have. We've been recently told it might be reviewed in the future. More info here:

 

[ssn]

Thanks.

So a six year old bug that affects multiple planes doesn't get fixed? Depressing.

[-0303-]

The bug is 11 years old at least and affects not just Warbirds but all old planes up to and including 1959 F-5 (reportedly, myself confirmed the bug in F-86.

Attention NineLine. Besides the Artificial Horizons (AH) not self-correcting to the gravity vector it also doesn't tumble on roll exceeding 110 degree or pitch exceeding 60. It spins around smoothly 360 degrees and it shouldn't.

Compare the AH in a real Spitfire below. Recently discovered this video showing this very clearly again and again and again...

Watch the "roll angle" indicator (DCS manual name) along the AH lower instrument periferi. Every time it exceeds ~110 degrees the AH tumbles. This Spitfire just keeps rolling and looping never giving the AH the time it needs (10 minutes?) to self correct. While being and staying "wrong" after the first roll, one can still observe how it tumbles any time the indication reaches 110 degrees to the left or right.

The indicator looks like it "bounces back" while the horizontal line rapidly moves up or down. This happens a ~dozen times.

 

Real Spitfire video notes:

• 0 - 28 seconds. Intro, Chris Hadfield sings (ugh) rolls twice and the AH tumbles because it encounters the mechanical stops at 110 degree roll.
• 6:35 - 6:50. Engine start. AH at rest is about 45 degree roll misaligned but spins up and self corrects in 15 seconds (much like the P-51 AH here, 12 seconds).
• 7:41-7:46. Taxiing, Left turn. AH misaligns slightly, 1-2 degrees. This AH error is caused by the centripetal effect on the inertia of the pendulous vanes and the slightly bottom heavy instrument housing (see educational video, timestamp 9:10 - 11:22).
• 8:25 - 8:35. Taxiing, Left turn. Same thing, more pronounced.
• 10:28. First roll. Hard to see but at 10:31 when it reaches ~110 degrees it tumbles because of the mechanical stop. When we see it again at 10:34 it's wrong but stable until ... 10:36 when AH again reaches 110 degrees and it tumbles again.
• Hadley keeps rolling and looping, one can see how the tumble coincide with 110 degree indication again and again. Some of the tumbles caused by 60 degree pitch as well I'm sure.

Ed/add: A nice touch would be the AH randomly misaligned on spawn and then rapidly self correct on engine startup (like above and just about every Warbird video engine startup I ever seen...)

~

Adding boilerplate" for Artificial Horizon (AH) bug threads. Old "classic" AH needs minimum two functions to work correctly. DCS AH lacks the second. Hence DCS AH doesn't work at all on any old plane up to and including the F-5.

1. A spinning gyro that provides "rigidity in space".
2. Pendulous vanes that constantly corrects the gyro towards the gravity vector.
There's also a third attribute which DCS does not model at all. AH tumbling whenever roll exceeds 110 degrees or pitch exceeds 60 degrees. This because of built-in mechanical stops that protects the mechanism. Source educational video timestamp 02:52.

A simple test if any AH works correctly:
    - Airstart
    - Bank steeply
    - Cage, uncage, roll level.
    - Fly some minutes straight and level. If AH doesn't self correct it doesn't work correctly.
If uncageable, like the Spitfire, just do above after noticing misalignment, the AH **always** self-corrects it's part of its basic function.

End "boilerplate".

Edited November 11, 2024 by -0303-

[Lau]

Awesome report thanks a lot !

 this new DCS pilot, bought the spitfire just to find out about the artificial horizon being broken with no support from ED for over a decade, not the way to do business.

Edited November 11, 2024 by Lau

[Rudel_chw]

28 minutes ago, Lau said:

with no support from ED for over a decade

 

the Spitfire appeared on 2016

[Lau]

Funny 

Chistoso 

Gracioso 

I believe you speak spanish if not pls disregard 

Edited November 11, 2024 by Lau

[Slippa]

Welcome to DCS SSN ,

and welcome to the vintage prop preservation society. We live in hope, we have to as we don’t really get the care and attention afforded to more modern aircraft.

The Spit’s really well done but does have a few things that need sorting out. You’ve pointed out the artificial horizon and 303’s covered this above really well but we just never know when or if we’ll get things like this worked on. A lot of things aren’t even acknowledged for a long time and on the odd occasion that something does get done, we never know if they’ll improve things or make them a lot worse.

We’re told the ‘WWII team is very small’. This could mean there are only 25 people from a workforce of 150 on the team but it more than likely means there’s one fella named Mr.Team who is very short and can only code one finger at a time. ED seem to have more balls to juggle than they can manage. Every new module adds more to an ever-growing queue of things vying for attention or fixes. There’s more enthusiasm for fast jets or helis, more sales, so more virtual pilots’ voices to answer to. The ‘small team’ has to choose which to work on next.

It’s frustrating having paid for something that doesn’t work as it should. It’s more than frustrating having to argue the case for a repair, sometimes backed up with plenty of evidence and still get nothing done about it. Too many things are left broken for years. Years. That’s just wrong.

Every now and then we do get a surprise. A while back somebody noticed that the boost gauge was marked incorrectly. Unless I’m mistaken it was fixed pretty quickly no trouble at all so things can get done. Another surprise was a whole rework of the cockpit interior. Some good things, some not so good.

We’ll see how this one goes I s’pose. Enjoy flying it in the meantime. 

 

 

[Lau]

Thank you Slippa for cheering up my day, we are blessed with this community.

[BIGNEWY]

Threads merged we will ask the team to review the issue. 

thank you