PDA

View Full Version : working on spin



dmaher
September 14th, 2010, 17:48
I’m working on spin for an RV7.
I’ve had pretty limited success :)
It will spin, but it doesn’t settle into a stable spin.
It will sometimes even go horizontal into a series of snap roll.
In a spin the forces should balance out eventually reach a very stable condition (as far as I understand)
A good description here...http://www.ozaeros.net/spin/ostivxx.html


I’m following some guidelines in AAM, I’m looking at tables: 451, 456, 460, 464

To start I’ve…
Slightly inverted roll dampening @ stall (tbl. 456)
Reduced roll moment due to slip @ stall (tbl.451)

These are some of the flight characteristics I’d like to get to.

Controls…


rudder hard over
elevator full back


features…


steady developed stall
steady vertical path
fixed airspeed
fixed pitch angle
fixed slip angle
fixed roll rate
fixed yaw rate


I guess I am trying to match airfile parameters to each features as best I can.
Does anyone have any good insights into tuning a stable spin?

Thanks,
Danny

Brett_Henderson
September 15th, 2010, 02:53
A few years back; I set out to get MSFS to replicate a spin. The flight model just isn't sophisticated enough, without setting control surfaces/flaps/, dynamic-CoG, and invisible "engines" that only come into play (via invisible gauge XML coding), during the spin. I.E.. an extreme flap setting with bizzare, lift/drag/pitch.. and that setting not available by flap switch, but only by the XML coding triggered by spin conditions.

And even if you get that to work, spin recovery is equally complex, and outside the ability of the MSFS flight-model.

fliger747
September 15th, 2010, 17:06
Waiting to get on a flight in Seattle, so not enough time to delve into this now, but it is possible to model spinning, though it takes some iteration. The F7F that I am working on for Milton will fufill the above criteria. It can be very hard to recover, though not always impossible given a good hand and enough altitude.


FS does approach these things in sort of a backhanded manner, but it is possible......

Cheers: T

dmaher
September 15th, 2010, 21:16
Thanks, guys…any thoughts are helpful.
For the most part it is spinning vigorously….it’s just not stable.

Some random thought …not sure if there’s a connection.


The AOA/pitch oscillates…needs to be dampened somehow (?)
Speed fluctuates...due to pitch oscillation (?)
The stall is intermittent… due to pitch oscillations/elevator authority (?)
The path is an extra long phugoid spiral…not sure (?)
Full rudder aggravates problems... rudder authority and rudder propwash too strong (?)
Occasional horizontal departures…lift from exaggerated rudder (?)

fliger747
September 16th, 2010, 09:21
Very very small adjustments make a big difference. Perhaps an adjustment to the weathervane stability might help. The planes do tend to spin a little flat. To observe, fly near an airport and use the tower view to evaluate the spin in playback.

T

Ivan
September 16th, 2010, 18:19
I have built a few Combat Flight Simulator aircraft that can spin pretty well. As an example, the F4U-1A Corsair I built many years ago will sustain a spin all the way to the ground. I am not entirely certain what the proper relationships are to accomplish this though. I don't believe this was one of the things I ever thought about enough to quantify.

The most reliable way to enter a spin is also a bit strange: Go into a vertical climb and pull up hard as the speed drops. A slight touch of rudder and the plane starts to spin. If the controls remain centered, the plane will spin from over 10,000 feet all the way into the ground. The recovery isn't difficult though.

If you are curious about the flight model, the plane is downloadable here and I believe that version only differs from my current version in the visual model.

- Ivan.

sparks
September 19th, 2010, 08:56
I'm not sure how this post is going to work out, but here goes...

First, the reason an aircraft spins is that either one wing stalls and the other doesn't, or that both wings stall and one wing stalls more than the other. What you end up with is both a rolling motion due to wing to wing differences in lift, and a yawing motion due to wing to wing differences in drag. The severity of the spin depends on how much of difference there is in stall from one side to the other. Also, because the wing is stalled, the ailerons usually aren't very effective in stopping a spin.

Now you could say that there is only one wing in MSFS, so how could you possibly model a spin? Well 'one wing' is correct with respect to some things like lift and drag. But these are linear forces, and what we need for a spin are rotational forces. Rotational forces in MSFS are determined by other stability coefficients, not lift and drag coefficients.

In order to model a spin in MSFS, you need to know what stability coefficients determine the roll and yaw forces and what happens to them when one wing stalls and the other doesn't.

One thing that can happen to the roll damping stability coefficient (Cl_p in 1101 and section 456 Cl_p vs AoA) is it can rapidly change signs when the wing stalls. The effect is that, instead of providing a resistance to rolling, it provides autorotation. Here's an example of what section 456 might look like in this case:

19235

The aileron control section (1535) would also need to reflect the fact that the ailerons are ineffective after the wing stalls.

The values I've show in section 456 would also tend to produce a tendency to snap roll.

You'll also need to consider what to do with section 0460 Cn_beta (weathervane effect) vs AoA and 0464 Cn_r (yaw damping) vs AoA for yawing forces in a spin.

Also, since the roll and yaw motions are factors of AoA, pitch stability affects AoA and therefore plays a large part in ease or difficulty of stall initiation and recovery.

All other things being equal MOI affects how fast spins occur and how long a recovery takes. This is also true of station loads and fuel tank locations (and how full the tanks are).

Given the same flight model, spins work better and are more convincing in some versions of MSFS than others. FS9 can be really bad with some flight models when severly stalled. Any stalled aircraft should eventually fall to earth, but I've seen FS9 models continue to gain altitude as long as the model was stalled. (I seem to remember CFS2 and 3 being prone to similar behavior) In general, moderate seems to work better than extreme.

fliger747
September 19th, 2010, 09:54
Good explanation Jerry! Yes, I do remember some FS9 planes climbing in a spin, and also some that would yaw through 180 degrees, but not produce a full rotation. Sanderson's Hurricane did seemed to be a good spinner, even in FS9.

As to ailerons, mishandling of them at high AOA can actually initiate a spin, stalling the wing rather than picking it up. Why did it originally drop? Probably through a yaw angle introduced through improper rudder use. Rudder is the proper tool to pick it back up with.

Some of the WWII naval aircraft have interested me because of their low speed handling (which was not always good), P-factor, torque, sometimes nasty wing drops and bad spin manners. Approaching at slow speed, with stick and rudder in various, perhaps opposite corners of the cockpit....

Cheers: T

sparks
September 19th, 2010, 16:24
Tom is right. Although difficult to quantify, P-factor and torque do influence spins in MSFS, usually in contributing to the 'spinabilty'. I think it's much harder to get a jet flight model to do a convincing spin.

You mentioned WWII naval aircraft? Here's a dusty old 2004 fraps video of a Corsair I was working on for FS9 doing a spin and then recovering:

http://www.mudpond.org/f4u1_spin.wmv (http://www.mudpond.org/f4u1_spin.wmv)


I did the flying for this R/C style using a tower view, and I had no idea how close the ground was. :)

Brett_Henderson
September 19th, 2010, 16:28
I'll just add another, experience-bourne (both in sim-modeling and real airplanes) 2-cents..

A spin doesn't initiate until both wings have stalled. One stalling before the other is one way for initiating the yaw needed , that would take you into a spin AFTER the stall becomes total.

Now.. in order to have enough difference between the two wing's airspeeds for one to stall before the other, you have to be either turning very steeply near the accelerated stall speed, or intentionally yawing at normal stall speed... either case would likely be when you'd be trying to induce a spin.. and MSFS doesn' allow for this phenom, unless you mod the flight dynamics to the point where normal flight gets messy.

And the kicker is.. the only reason you'd be trying to induce a spin; is to practice recovery techniques .. and this is where MSFS really falls apart... especially with the mods in place that allowed the pseudo-spin to happen in the first place.

It's an admirable goal, and has been sought after for what.. a decade or more ?

If any of you come close, I'd be honored to test it.. but I don't think it's possible sans dynamic, XML input.

fliger747
September 19th, 2010, 19:08
A really nice Corsair model I did for a certain vendor (they didn't like it) had about all the standard Corsair nasties, including fliping over the top from an accelerated stall into a spin.... The current underway project, Milton's Tigercat, will do it too. But of course the most fun intentional spins are snap rolls.

But Brett is certainly correct, in the civilian sim we seldom explore those corners of the flight envelope.

Cheers: Tom

sparks
September 19th, 2010, 21:06
Attempting to explain complicated subjects in one short post never seems to satisfy everyone, so for those who would prefer an in-depth discussion of stalls and spins by a true expert I would recommend:

http://www.av8n.com/how/htm/spins.html

fliger747
September 20th, 2010, 07:41
Interesting article! A salient point that I hadn't fully appreciated was the different points on the lift/drag curve of each wing, with the more stalled wing not only producing less lift, but also more drag, which contributes to the turning moment. To be able to recover, sufficent rudder power must be available to counteract this drag differential which maintains, or even tightens the spin.

Cheers: T

dmaher
September 20th, 2010, 13:16
Guys, you have painted a very clear picture of spin dynamics.
And I see Av8n.com is an excellent online reference – must be one of the best there is.
I’m just getting back on task now…I accidentally got caught up in those propeller tables – it’s ok now :)
So, sorry I was absent for a few days.
I appreciate the time everyone has taken to post. Thanks. Danny

bstolle
September 21st, 2010, 13:28
If any of you come close, I'd be honored to test it.. but I don't think it's possible sans dynamic, XML input.

Milviz Cessna 310. No XML etc...just plain careful designed aerodynamics. There are interestingly different ways to achieve a spin in FSX. You can alter the rotation rate, pitch attitude etc...
Unfortunately what works for a GA single may not work at all for a GA twin.
LOTS of hours of trail and error

Best regards

Bernt

fliger747
September 21st, 2010, 16:40
Bernt:

The C-310 goes into a nice auto rotation, but the geartly increasing airspeed seems to suggest this is a really a spiral rather than a true spin.

Having this particular issue with the DH80 at the moment.

Cheers: Tom

bstolle
September 21st, 2010, 21:44
Depends on the fuel load, (due to the heavy outboard main tanks) pax load and spin entry.
She does everything between a flat spin and a spiral dive.
It would never occur to me to call a spin that becomes a spiral dive a 'spin'....
With its new FDEs even the Flight Replicas Super Cub is now capable of a stable spin.
Furthermore you can design the recovery effectiveness rather easy by adjusting the rudder effectivity vs. AoA

fliger747
September 22nd, 2010, 01:14
For the 310 what airspeed range do you get in the spin? In the F7F-1 I get a cyclic variation of 57-72 knots. In the current version of the Puss Moth I have the an acceleration which, as in the 310 has the autorotation, but an increasing airspeed well above a normal stall. Didn't yet monitor the G loading.

Off on another trip..........

Cheers: Tom

bstolle
September 22nd, 2010, 02:47
For the entry I use a full stall. Have to check when I'm back home for the rotation speeds.
Greetings from KIAD in the meantime.
BTW according to our MET guys the NAT crossing should have been awful with sev turb in the vicinity of Igor.
The special briefing took an extra 20min....Actual we had ISA +22 but no turbulence at all....

Cheers

Bernt

fliger747
September 22nd, 2010, 17:34
Hi from Charleston SC.... Off to fly the Dreamlifter tomorrow for an airshow at McConnel AFB, Wichita Kansas. Dang, we leave before the airshow, the airplane remaining on static display.... Fly by would have been a hoot!

I just picked up a new Saitek Cyborg X for travelling. Handle unscrews, base folds, very compact, has dual throttles!

Interesting to investigate the AOA one can maintain in an attempted spin. If Ican get the 310 up to about 20 deg AOA and keep the stick back, it will maintain a high AOA, though with an increasing airspeed, and will loose the high AOA and eventually fly out of the autorotation. Hard to get into the deep stall, requires some accelerated "snap". Without the snap it stayed at about 12 deg AOA and accelerated a lot.

Will look at som other planes... Cheers: T

fliger747
September 23rd, 2010, 05:13
Checked out the F7F-1, which if one does not do a careful recovery, you can spin all the way into the ground. In a spin it goes into a very deep stall, 50-70 deg AOA. The critical item to maintain this spin and not want to fly out into a dive (without maintaining up elevator) is to have a neutral or pitch up moment (post stall) which keeps the nose from dropping.

Cheers: T

bstolle
September 24th, 2010, 01:19
Well, flying the Dreamlifter in front of the crowd must be quite something! Maybe there's another chance for you as the plane is rather new, I'd say.

Concerning the spin. e.g. with the default loadout (which is even a bit nose heavy) and empty main tanks) she spins at around 50kts with an AoA between 50-60deg with full up elevator and pro spin rudder.
Elevator trim full up of course as FSX unfortunately doesn't simulate a normal elevator with a fixed stabilizer.
To have a realistic pitch authority one needs to take the awful 'FSX trim simulation' into account.

That's why I prefer planes which have an adjustable stabilizer IRL like the Super Cub.
BTW, the 'new' Super Cub spins at around 30kts and remains in the spin with all controls in neutral.

Cheers

Bernt

fliger747
September 24th, 2010, 05:18
Actually the Dreamlifters are rather ancient 747-400's that have given their lives to a reincarnation.

The condition I was experimenting with in the 310 was full fuel and 4 pax, two in front and one in each row, giving a quite aft CG of something like 38%. Of course one of the items we have to contend with as far as CG goes, is that FS pretty much ignores the stability issues that an aft CG engenders and still relies on the seperatly derived tables and scalars.

I think the first FS9 plane I saw that was a good spinner was Sanderson's Huricane, lots of people said it couldn't be done. Indeed possible, but not an easy formula to approach! I suppose Jerry could program it into "Airwrench".....

Cheers: Tom

sparks
September 25th, 2010, 05:34
I suppose Jerry could program it into "Airwrench".....

Cheers: Tom

Should be there already. On the Tuning tab, move Spin Stability slider towards 'Unstable' and move Stall Roll off slider towards 'Positive'. Results will vary depending on other settings.

fliger747
September 25th, 2010, 08:24
Jerry:

I am probably about 5 versions out of currencey with Airwrench. A good program and one I reccomend a lot.

Looking at the varying AOA of the aircraft during a spin, it seems to oscilate over a range, the oscilation being coincidental with the rotation, one oscilation cycle per rotation. I presume that this is derived from the way FS "computes/resolves" the different aerodynamic conditions of each wing?

Cheers: Tom