Glider Vertical Stab Volume

[Sail 'n Soar]

Anyone got any data on model aircraft and glider design practices and statistics relative to vertical stabilizer volume, Av*Lv/(Sw*b)?

Current full scale sailplane practice is generally for .045 < Vv/S*b < .08, with a few exceptions, such as the PW-5 ~.03 - .04. Using the "that looks about right" approach, I usually design to a vertical stab volume of ~.035. But checking out a couple of published designs I calculated vertical stab volumes significantly less, as low as .02.

[Sparky Paul]

You gotta stop reading those book! :)
Try it and fly it!

[Sail 'n Soar]

Originally posted by Sparky Paul
You gotta stop reading those book! :)
Try it and fly it!

I used your suggested approach in highschool, but have gotten more conservative over time - and stubborn! ;) I really do like to understand what's going on. But vertical stab area seems still to be more art than science. So OK, Sparky, what's your approach?

[Sparky Paul]

First, look at how critical the cofficient might be.. a range from .045 to .08 is 2:1. Av or Lv or b or Sw can change by a factor of 2, or any combination of them, and still be in the recommended envelope.
Lots of room inside, and more than likely room outside to wiggle around in.
Then, it's not only the tail volume but other things like dihedral which will influence lateral stability and control.
Combining all the things that influence the situation, the very broad area of acceptable size
is an indication of the problems involved in sizing.
That successful planes fly way outside the envelope is another indication of the witchcraft involved in vertical sizing.
.
I believe there's a treatise on lateral stability at Mark's Charles River site. That's probably the most reliable source for model airplane aerodynamics.

[BMatthews]

I did a cut and try on a hack aileron 2 meter model way back when. I didn't record the sizes but there's no doubt in my mind that models want to use a much smaller coefficient than full sized craft. Also that many models that used the TLAR approach using full sized as a guide are too spirally unstable with the large verticals. The old Windsongs and other Dodgson models were amongst these. Or perhaps many designers just used the old poly fin sizes because they "looked right" to them.

Aim slightly smaller than you think and use a sacrificial cut and try set. Done to the right size you'll loose a lot of that need for "top aileron" in the turns. Go too small and you'll notice a tendency for adverse yaw and lots of sideslipping in the turns.

PS: Sparky, there's a wide range from design to design in a table but I found that for the one model the too large to perfect to too small was about 2 or 3 sq inches on a 20 sq inch vertical. That's sensitive enough in my book.

[Sparky Paul]

The biggest problem I've noticed with cutting vertical area is at low speeds, like launches.
Made a couple of MonoKoted stick-bags with too-small verticals... :D

[BMatthews]

Oops, I can just imagine..... :eek:

The hack ship got a bit touchy and hunting when I was a tad too small too. In my case a bit of glue to replace the last peice and some Monokote "trim" for the rudder tip and it was all better.

I was truly amazed at how much the vertical tail size affected the flight.

By the way Sail n' Soar, I don't see anything about the CG location in that formula. Trust me, it makes a BIG difference. The CG location needs to be in tune with the fin area. The further back the balance the smaller the vertical area.

[Ollie]

I think it would be useful to consider some of the conflicting objectives that should be balanced in the design of the vertical tail. See:
http://www.charlesriverrc.org/articles/design/donstackhouse_rollcontrolstability.htm
This rather length discussion explains why vertical tail area is contingent on many other things.

For most of us, cut and try will be the easiest and best way to get the results desired.

[nuevo]

If you're using the rudder for roll control, even the weight of the wingtips affects how large the vertical stab/rudder should be. That never shows up in any of the equations. ;)

[Dick Huang]

Sail 'n Soar,
See National FreeFlight Society Sympo 1996 where 44 F1A Nordic.36 F1B Wakefield and 28 F1C power Models were analysized for their static stability. There are data on center of gravity boundary for longitudinal stability(horizontal tail volume vs.
CG location) and Powr lateral-directional stability boundaries (Vertical tail volume Vs. Rolling moment due to sideslip). This figure defines "Unsafe,spiral dive" , "Safe" and "unsafe Dutch roll" regions for the above 108 models. You can surf the NFFS website and find a copy.
Dick Huang:)

[BMatthews]

Originally posted by JonStone
If you're using the rudder for roll control, even the weight of the wingtips affects how large the vertical stab/rudder should be. That never shows up in any of the equations. ;)

Sorry JonStone. Actually the weight of the wingtips would have no effect on the requirements for vertical area of the fin and rudder. Stability is the issue here, not roll response. Make the vertical tail too large and you'll have spiral instabillity, too small and you have dutch roll and adverse yawing. Heavy wingtips only cause the model to respond more slowly to the effects that you would have anyway. This is why it doesn't show up in any of the equations.

With polyhedral models you'll seldom see the spiral instabillity show up as a tendency to fall off the straight and narrow and into a spiral dive but it often shows up as a tendency for the model to be neutral (no rudder input needed) when turning at a certain bank angle.

I think you may have confused the stability factors with the issue of where to put the rudder HINGE LINE so as to achieve a good handling model. But the proper cure in any case is to build light in the wingtips and tail.

[Sail 'n Soar]

Originally posted by Dick Huang
Sail 'n Soar,
See National FreeFlight Society Sympo 1996 where 44 F1A Nordic.36 F1B Wakefield and 28 F1C power Models were analysized for their static stability. There are data on center of gravity boundary for longitudinal stability(horizontal tail volume vs.
CG location) and Powr lateral-directional stability boundaries (Vertical tail volume Vs. Rolling moment due to sideslip). This figure defines "Unsafe,spiral dive" , "Safe" and "unsafe Dutch roll" regions for the above 108 models. You can surf the NFFS website and find a copy.
Dick Huang:)

Found the web site. Thanks. Wish I could just order the specific paper vs. the entire '96 Symposium report. Oh, well.

[Dick Huang]

Sail 'n Soar,
When you get your copy of sympo96 and want a to run some numbers I woud be happy to send You the *.exe file. You can e-mail me for the request.
Dick Huang;)

[markdrela]

Originally posted by Dick Huang
See National FreeFlight Society Sympo 1996 where 44 F1A Nordic.36 F1B Wakefield and 28 F1C power Models were analysized for their static stability. There are data on center of gravity boundary for longitudinal stability(horizontal tail volume vs.
CG location) and Powr lateral-directional stability boundaries (Vertical tail volume Vs. Rolling moment due to sideslip). This figure defines "Unsafe,spiral dive" , "Safe" and "unsafe Dutch roll" regions for the above 108 models. You can surf the NFFS website and find a copy.
Dick Huang:)

There is something very suspicious about this analysis. I don't have the 1980 NFFS Sympo, so I can't examine Bogart's method. According to the graph in the 1996 report, increasing the tail arm length will make the airplane more spirally unstable. This is exactly opposite to what really happens.

By far the most reliable predictor of spiral stability is Blaine Rawdon's criterion. You first compute the "B" parameter:

B = EDA x (tail_length/span) / CL

EDA is the Equivalent Dihedral Angle (in degrees), discussed by Blaine in his Modal Aviation article series "Dihedral" a while back.

Anyway, the airplane is spirally stable if B>5, and spirally unstable if B<5. A longer tail increases B, and hence increases spiral stability. The vertical tail area doesn't even appear. Good rudder/elevator RC gliders have enormous vertical tails by FF standards, and are spirally stable regardless, just as the criterion above predicts.

Here are some relevant RCSE posts on the topic:

http://www.mail-archive.com/soaring@airage.com/msg37346.html

http://www.mail-archive.com/soaring@airage.com/msg37462.html

http://www.mail-archive.com/soaring@airage.com/msg37652.html

Some of Blaine's articles are at www.rc-soar.com

[BMatthews]

With all due respect Mark if the equation doesn't take into account the vertical tail area then I don't see how it can determine the spiral stability factor.

I DO know that the tail area makes a difference. For decades free flighters have been increasing the fin area to eliminate Dutch Roll or decreasing it to stop unwanted spiral dives. Also that equation doesn't take into account the balance point and I know from my own cut and try 2 meter glider project that it DOES make a drastic difference to how the model reacts in this regard.

The Rawdon Equation may be a tool but I don't see how it pertains as does the tail volume coefficient.

And not all the poly ships are totally spirally stable. I had a Top Flite Metric that was trimmed to as rearward a CG as it and I could tolerate. It would try to tighten a turn at any bank over about 10 degrees. When I moved the CG forward a smidge to reduce the pitch wandering and get the trim function to be less than 2 click critical the spiral wind in was delayed to about 60 degree banks.

I have to read those links yet but I do know from my own experiments where I changed only one item at a time that the area and the CG at least are linked. And since these fit into the tail volume coefficient equation I tend to lean towards that being a better analysis of what's happening.

I'm looking forward to your expanation of how these all link together as I know you've got more knowledge about aerodynamics in your left pinky than I've got in my head but in this case I just can't see the light.

[Sail 'n Soar]

markdrela

Thanks for the references. Straying from the subject only slightly, what are the useful CL ranges for the AG 35, 36, and 37 airfoils at a reduced RE of ~90.000? The Mike Garton and David Orman site compares low RE tunnel data on a number of foils, but none for yours, that I can tell.

Gerry

[markdrela]

Originally posted by BMatthews
I DO know that the tail area makes a difference. For decades free flighters have been increasing the fin area to eliminate Dutch Roll or decreasing it to stop unwanted spiral dives. Also that equation doesn't take into account the balance point and I know from my own cut and try 2 meter glider project that it DOES make a drastic difference to how the model reacts in this regard.


Yes, I've flown FF quite a bit, and I do know that a large rudder will make a spiral dive more likely. However, I've come to the conclusion that this is more a failure to roll out properly than due to true spiral instability. Excessive rudder area will prevent the rapid yawing motion required for a quick "flick" into the glide at the top of a climb. Instead, the model will go over the top like on rails, and will more likely end up in some bad initial glide orientation, e.g. banked and nose down. Because simple FF models must be trimmed with very little pitch stability (very aft CG) to prevent looping on launch, such a bad initial orientation can easily result in an overspeed, resulting in a pitch tuck-in while doing a loose turn. This looks and smells like a spiral dive, but technically it isn't. If you take that same "spiral"-prone model and start it in a proper glide, it can quite possibly glide indefinitely long with no entry into a spiral.

Perhaps the difference is a bit academic, since whether it's a true spiral dive or a turning pitch tuck-in, the result is the same -- the model augers in. But the difference may be important on FF planes which have auto-elevator in the climb, which naturally allows more a more forward CG, and less tendency for pitch tuck-in. I suspect such planes, with a suitable CG setup, can tolerate a lot more vertical tail volume.

BTW, have a look at the Photon DLG.

http://www.netmeister.net/~jerry/

click on "Photon" link.

This is much like an F1A, but the vertical tail is perhaps 5-6 times bigger. I assure you it exhibits no sign of spiral instability.

[markdrela]

Originally posted by Sail 'n Soar
Thanks for the references. Straying from the subject only slightly, what are the useful CL ranges for the AG 35, 36, and 37 airfoils at a reduced RE of ~90.000? The Mike Garton and David Orman site compares low RE tunnel data on a number of foils, but none for yours, that I can tell.


The bottom corner of the low-drag bucket is at CL of 0.08 to 0.10, which would occur in a fast penetration glide at about 3x the minimum glide speed, which is about 35 mph with a 5 oz/ft^2 loading.

Max usable CL is about 0.9 in straight flight, but you normally don't want to fly there. The min-sink CL for most gliders will be in the range 0.7 to 0.8, depending on the aspect ratio and the turn radius.

[Sail 'n Soar]

Originally posted by markdrela
The bottom corner of the low-drag bucket is at CL of 0.08 to 0.10, which would occur in a fast penetration glide at about 3x the minimum glide speed, which is about 35 mph with a 5 oz/ft^2 loading.

Max usable CL is about 0.9 in straight flight, but you normally don't want to fly there. The min-sink CL for most gliders will be in the range 0.7 to 0.8, depending on the aspect ratio and the turn radius.

My anticipated wing loading is closer to 10 oz/sq, AR ~10, with a larger cross-section, scale-like fuselage, pushing the min-sink CL closer to 0.9 -1.0+.v Since your last reply I've down-loaded XFoil4 and have been examining the predicted polars for 110% and 120% thickness ag35 variants. Both look quite good at my anticipated RE sqrt(CL) ~93000. Slightler higher Cd's, but higher predicted L/D and stall, with reduced min-sink compared to the ag35. Your thoughts?

[markdrela]

Originally posted by Sail 'n Soar
Since your last reply I've down-loaded XFoil4 and have been examining the predicted polars for 110% and 120% thickness ag35 variants. Both look quite good at my anticipated RE sqrt(CL) ~93000. Slightler higher Cd's, but higher predicted L/D and stall, with reduced min-sink compared to the ag35. Your thoughts?

When the AG35 is scaled up by 1.06, it becomes very nearly the same as the AG34, which is the thickest section in the AG3x series. I wouldn't go much thicker than the AG34 if you want to maintain the good speed range.

This assumes you have a RES ship. You can go thicker if you will have TE camber control.

[Sail 'n Soar]

When the AG35 is scaled up by 1.06, it becomes very nearly the same as the AG34, which is the thickest section in the AG3x series. I wouldn't go much thicker than the AG34 if you want to maintain the good speed range.

I down-loaded the AG35 - 38 from either the Charles River or UUIC site. Neither include the AG34 coordinates. Is there a source for that coordinate file? Or based on actual construction tolerances, is scaling the AG35 close enough?

This assumes you have a RES ship. You can go thicker if you will have TE camber control. RET only powered glider, no spoilers.

[Ollie]

Go to:
http://groups.yahoo.com/group/Allegro-Lite/?yguid=108420033
Join the group if you are not already a member. This will give you access to the files section. In the files section you will find a folder on the Bubble Dancer. In that folder you will find the coordinates of the AG34.

[Sail 'n Soar]

Thanks.