This page contains some important letters/answers I did on some interner forums dedicated to C/L stunt.

... Feel fre to tell me how stupid I am (as some others did) if you think I am not right :-)

... Or just feel free to add any other important inputs to this thoughts. - I will be lucky to keep it here.

"Stunt model dimensions" topic on RCOnline  ~February 99

There are some lift and moment curves showing lift and airfoil moment jumps of flapped airfoils used on wings and tails of C/L stunt models. Those jumps can cause shape errors of some figures.

Wind tunnel measured lift curve of FX71 airfoil. It is 15% airfoil with 20% flap.

The lift curve shows there is serious problem around 0 deg angle of attack if flap is deflected over 15 deg. It can happen in corner. The flap is deflected and wing angle of attack is from 0 to 5 deg depending on elevator to tail ratio, because of circular airflow. It is evident that more flap deflection makes larger jump.

Theoretical (calculated) lift curve of 20% airfoil like used on C/L stunt models with 20% flap at 20 deg.

Second graph shows both lift (positive values) and moment curve (negative values around -0.2). Both curves explain what is going on. The airflow is separated on negative pressure side at hingeline if angle of attack is too high. The jump height is around 10 % of lift. The flap is around 20% of chord length. If airflow is separated at 80% of chord on one side, the 10% lost of Cy is expected. The same is on moment curve. The value -0.2 (the negative pitching moment of flaps) drops to half at the same angle of attack because of one side separation.

July 99

Here is an example of airfoils for wing and elevator without bumps on polar.

The airfoil is optimized for 30 degrees flap deflection and maximal angle of attack 17 degrees. The drag was minimized at the same conditions rather that at no flaps and 0 angle of attack. It will help to keep constant speed.

The tail airfoil is optimized for maximal lift at 0 degrees flap deflection necessary for exact fly off from the corner and good sensitivity to the corner at 30 degrees flap deflection. The complication is low thickness. Thick airfoil is worse because of less critical RE number (stab of the elevator is on its level). Any way I am still not sure whether it enough, so more main chord of the stab, or wire turbulator with 0.45mm diameter on LE can help (I recommend to try it).

Here is polar table for root wing airfoil at RE=500 000 and 30 degrees flap deflection:

The root airfoil has maximal lift coefficient 2.44 at 15 degrees AoA. The drag is almost constant below that AoA, but I note that the table shows data for constant 30 degree AoA, so it can not be applied for level flight.

The separation appears at 18 deg AoA (stall point), but it is smooth without jumps (visible on SU column) making bumps on polar. Anyway, the airfoil has enough lift at stall condition up to 20 degrees AoA over 2.

Also TU column shows that no turbulator on LE can help.

"Engine / Prop Performance Program"  topick on RCOnline  ~February 99

Some peoples asked me to send prop HP curves. So they are here:

If anybody need some of those graphs in better resolution, let me know.

"Undercambered versus flat"  topick on Compuserve  ~October 98.

On pictures are two airfoils looking like those used on our props. It only looks, so take it only qualitatively, not quantitatively. First airfoil has flat bottom GO 795, second is undercambered GO 342. Both are measured at RE around 80 000 which is appropriate for tip of thin 12" prop at 10 000 rpm.

??? what about it ???

 igor

Hi All,

It looks that elevator/flap ratio is somehow obscure. But basically everything is clear. Here is how I do it:

1/ The wing has to produce maximum possible lift possible in corner, while it should have as small as possible area in level flight. It is because for level flight we not need so much lift and too much wing area makes model sensitive for turbulence and for side wind (as we have wept back wing – wind from outside rolls wing toward the circle). So it looks we need as large flap as possible. The limitation is feedback to handle. The force necessary to deflect flaps must not exceed available line tension overhead.

2/ Flap deflection gives more lift AND drag. As the drag is not so important with current power train, we can go to deflection where the lift coefficient Cl does not grow anymore (almost). It is at around 30 deg. If we have to take care to drag (underpowered model), it is better to end at 20 deg. Under 20 deg. the Cl grows more than Cd (drag coefficient). (and now we got it - compromise)

3/ Now we know what is geometry of wing. Thus we also know the Cl/alfa polar of wing (lift as function of AoA). The Cl coefficient of an airfoil grows 0.11 per a degree of AoA until separation begins (until a maximum lift – it is little before stall) it is somewhere between 14 and 16 deg. AoA with our 18% flap airfoils. The task of the tail is to keep the wing at begin of corner little under that value. If it gives more, it could stall and not to enter circular path. If it gives less you do not use whole potential of the wing. As the model enters the circular path, the AoA of wing goes down as AoA OF AIR ENTERING LE OF TAIL has higher than 0 value due to circular flow an distance of tail to wing. So AoA of wing in corner is at only around stall safe 5 or 6 deg. This disproportion could be moderated by length of tail little bit. Longer tail needs more deflection of elevator, but goes better out of corner. It is another compromise.

4/ Now we come to ratio itself. The tail has to produce lift (yes, negative :-)  ) overcoming CG front of AC of wing and pitching moment of wing and minus pitching moment of tail while incoming airstream comes at some AoA from circular flow. It happens at some angle of deflection with correlation to size of elevator in relation to size of stab. So it is clear that if really want to speak about ratio, we MUST also tell ratio of stab to elevator as it has same function.

What I wrote is very simplified, but I hope that it makes picture. It is possible to figure up the ratio, but there is a lot of inputs not known before building, so my recommendation is – build exactly a tested, “debugged” model (from begin to end – including the weight), or use adjustable horns for playing.

But anyway, do not think you will get significant improvement if you change ratio from 1:1 to 1:1.1. If you are able to feel that difference you also know how it should be.

Regards

igor

Hi All,

It looks I missed something here past week as I have only limited access to RCONLINE. I now I am too late, but anyway some thought:

First I would like to say that there was written almost everything, but not all of you speak about the same thing. While Al uses terms like center of lift and dynamic stability, Ted uses aerodynamic center and static stability. It leads to many confusions. So let’s tune to same language (I will try to write more words – this can solve problems with my English :-)  … sorry if not …  ).

Many of following was already written by Alfredo, but I will collect it together:

Ac – aerodynamic center is place at 25% of main chord of flying surface – wing. It comes from definition and is chosen for common communication and calculation around flying surfaces. It means all aerodynamic forces are measured in Ac. It means if we have lift (cy), or drag (cx) or moment coefficient (cm), all of them are measured in this point. Cy reflects vertical forces (lift) in Ac, cx reflects longitudinal forces (drag) in Ac and cm reflects pitching moment in Ac - to the leading edge (LE) – yes, it needs another point making the arm. The value 25% was chosen because symmetric airfoils has moment coefficient close to 0 on that place, and undercambered airfoils has almost constant moment coefficient with changing angle of attack (AoA) on this place.

Cl – center of lift is place where “it looks” is lift concentrated. By other words where we have to fix flying surface to produce vertical lift only without pitch to any side, by the other words where to put center of gravity (CG) to get flying wing without tendency to pitch down or up. If we know vertical force - lift of wing and pitching moment of wing, it is very easy to figure up how far we have to put CG aft or front of Ac to balance pitching by the CG with weight equivalent to the lift on arm to Ac making exactly same moment as pitchin moment of airfoil  … and thus where is center of lift.  … HUH - I think it is necessary to read it again … or may be twice

… The cm is almost constant with changing AoA so the pitching moment is also constant, but lift grows with growing AoA so Cl of wing is really moving with changing AoA as was stated by Alfredo I think (if not - sorry). This is why Cl is not used for theoretical calculation and I can only agree with Al that you can only guess where is your Cl in level flight.

Np – neutral point. It is place where fixed body in stream does not show any moment to fixing point in any direction. It means that if we put CG at this point the body will not turn to some “predefined” front. I guess everybody understand that if we move CG to any direction from Np it will become “front” in stream.

If we speak about airplanes and if we use airfoils without pitching moment and with 0 coincidence it is place at 25% between MAC of wing and MAC of elevator.

Static stability – is fact that body described in Np sentence turns to its “front” and shows moment keeping it in this direction positively dependent on its angle.  …  This is how arrow flies.  :-)

Reserve of static stability – how far is CG front of Np. This is important for flying planes with more than one wing. It is number which tells us how much is CG front of Np in % of MAC of !!!airplan!!!. Mac of flying plane is distance between Ac of wing (or wings) and Ac of elevator. Example was done by Alfredo, but unfortunately I did not understand very well … Alfredo, I think the picture is really sometimes better … Our models with huge tails has values from 20% to 25% this is far more than we can afraid of. Any movement few % aft, even aft of wing Ac will not hurt static stability. The table of calculation of my Next Model is on my home page at:

www.Netax.sk/hexoft/stunt/thenext.htm

Al wrote that any wing needs a tail balancing its pitching moment. This is simplified, and commonly not true, as we know flying wings.  Properly written the pitching moment could be balanced by CG. But as we know, the pitching moment of our stunt flapped wings is negative, so  CG should be aft of Ac and this is not possible as we will lost static stability. It is clear that for regular undercambered airfoil we definitely need tail – front or aft, but truly another surface. Exception is airfoil with zero or positive pitching moment – it looks like black magic, but true is that we can take any undercambered airfoil, and make him little up deflection on TE and pitching moment will go down.

Dynamic stability – this is why free model flies. This is ability to keep wanted AoA of model, even disturbed from outside. Whole think is based on hope that if model is disturbed, “something” appears which moves model back to its AoA. This “something” could be input to elevator by pilot or stabilizing automate (or anything else what Al do not like to see), or aerodynamically (this is what is important for us). Have a look to lift curve (or cl/alfa polar) of any airfoil. If you do not have, go to my homepage at “notes” section at:

www.Netax.sk/hexoft/stunt/notes.htm

There are 2 such polars. It shows how lift grows with growing AoA. It is always a hill. Almost any airfoil around 0 AoA shows 0.11 cy per 1 degree. Does not matter what is the shape. As we go to higher and higher AoA some airfoils earlier, some later, but everyone once start losing its lift because of separation and shows less and less cy per a degree until reaching top of the hill – this is place of maximal lift coefficient. If we will go to higher AoA we will come to decreasing of the lift to its half because of total separation on upper surface = stall. The aerodynamic stability is based on fact that lift curve looks like a hill. Assume that elevator is at -1 AoA and wing has +1 degree of AoA, model flies exactly level with maximal lift coefficient of the wing (on top of the hill). Forget drag for simplicity. This is situation expected by Al. Elevator makes negative lift, while wing caries weight. Now assume that model got an impulse which pitched him +1 degree. The wing can not produce more lift as it is already at top of the hill, and +1 deg AoA gives same lift as before. But tail was 2 degrees under the top, so +1 will give him more lift -> it will pitch the model DOWN, back to the wanted AoA. This is true not only for top of the hill, it works also on whole its left side, where is enough disproportion of cy/alfa between wing and tail. It enough if lift on tail grows more. Fortunately it works fine also for changes in speed – if we fly at higher speed, wing gets more lift thanks linear dependency to speed variation. It turns it up and thus the speed does not grow anymore and the rest is equivalent to variation in pitch. I note that there are also different methods then just AoA. It enough to use properly chosen airfoils this includes flying wings with autostabile airfoils. Or move CG far UNDER the wing like we have on rogallo wings.

It is clear that top of the hill is most stable place, and free models flies at maximal lift coefficient to keep itself as much possible TIME not a distance. The gliding ratio is not important, Al is right that it flies close to stall (top of the hill), but the reason is not in stability. I hope that it is clear that such controlling system works independently on static stability. Better word is that it has nothing to do with static stability. !!!!!!! Do not smile !!!!! It is true!!!!! If you move CG aft of neutral point, model will turn itself backwards, it will fly backwards, but will have some dynamic stability -  with little luck you will get stabile canard.

Everything depends on how strong is negative feedback (which makes dynamic stability). The CG counteracts to any feedback or controlling because of static stability (CG front of Np). It means if we move CG aft, the feedback becomes stronger. If the CG is too aft, the feedback is stronger than necessary and model will oscillate while converging back. If CG is even more aft, oscillations will have increasing amplitude and model will never come back an will fall down like a piece of paper. The same could happen if CG is too front. In this case the feedback is not strong enough to move nose back, and model goes down while accelerating till the drag balances it. If CG is even more front, model will also pitch down while accelerating – in this case the drag has no chance and model will brake with nose in the ground. From this point of view the most stabile conception is clearly canard. The canard “tail” could be heavily unloaded (aft CG with small tail) and can fly past its maximal lift point or even at stall point. The wing could carry all the weight and so if tail exceeds stall point and falls down it will never allow stall condition on the wing. This is what we can not risk at “aft-tailed” models.

Any relation to our stunters? :-(

But to thoughts in thread:

This is probably thread with most exactly (even not always properly) written thoughts. I thing after all it is clear that there are statically and also dynamically models with CG far aft the Ac of the wing and also models with two flying surfaces producing positive lift.  (Sorry AL). The example could be also R/C stunter without flaps. I am not sure, but they uses CG positions somewhere at 30 or 40% MAC on symmetric airfoils, am I right?

But there are still some open questions like:

>>>Discussion of canards, tandem wings, etc. are not irrelevant to the subject of stability of CL stunt airplanes, they are crucial to answering that question you posed right at the beginning - if the CG is behind the aerodynamic center of the wing, what is holding the tail up? Answer: the tail.<<<

This is not true, we have also negative pitching moment of the wing which can keep tail up. --- Ted and Iskandar, Do not kill me right now --- Read following before:

Al asked how the model could make positive lift if CG is at Ac (I mean AC of the wing for next time). Alfredo wrote him that it is easy – it enough to have some positive AoA. But my question is how we can keep positive AoA if we have CG at AC and it means also front of Np and it means we CAN NOT have any positive AoA. For this situation we definitelly needs deflection on elevator = deflection on flaps = negative pitching moment ballanced by negative lift of tail as the CG is at AC. This is what Al wrote. To come to positive lift on tail we need to move CG even aft (ant this is aesily possible as we have 25% reserve of static stability) – to Cl but Cl if moving with AoA -> it needs adjusting deflection on elevator-> makes deflection on flaps-> changes of pitching moment-> even more moving Cl ... I would say we have to stay at 25% MAC and count with negative lift on tail. ... And I would like to tell to all upcoming jokers that I really know we fly also invetred loops with definitelly positive lift on tail, but now we speak about steady level flight.

This is also answer to Doug. If the model stays on main (front) gears by almost whole weight, and aft wheel under tail is loaded only little and is mounted upper than main gears (isn‘t it also smaller on Extra?). The model stays on ground with relively high AoA, so you must count that if it moves on the groung, the tail and stab at such AoA produces lot of lift. But tail is loaded only little so it goes up earlier than wing could bring up the rest, that is all. But again, we speak about level flight, nobody wrote that also in inverted loops ....

Alfredo, I definitelly did not understand your last message. (I do not want to say that conclusion is wrong!)

What is dCmg/dCl and how it change sigh from + to - ? You wrote that this must be true: dCmg/dCl < 0 But it can change only if one of them changes sign ... I am out of synchronization ...

Regards

igor

Do not kill me ... See (read?)  you next week.

Or OK, kill me, it will bring more light wanted by Alfredo.

Hi All,

Speaking about LE only is too simplified. It is necessary to take whole stab and elevator. It is still some kind of black magic for me as it surprises me on almost every new model, anyway here is how I see it:

I did many tests passed 5 years. Unfortunately I was too lazy to make removable stab as David did, so I did not get so much data as he got on ONE model with sveral stabs. But if I will add also his publiced experiments I believe it is as follows:

1/ I will little disagree with Ted. I do not think that the case of wire on LE is well defined stagnation point. The stagnation point is enough defined by pressures on upper and lower side of LE made by incoming air. So if we have an exact and streight stab we have also exact and streight stagnation point. It could look that if the LE is blunt, the stagnation point goes down with growing AoA of stab. It is true of course, but as Ted well explained us in his Doctor article, in this case the airfoil makes more lift as it works like underchambered. For more theoretically based guys I would say that air path from such a point toward the TE is longer on upper side and shorter on bottom side so also speed on top is higher than on bottom and from bernouli equation it makes more lift – and also moves little the stagnation point up. I know it is too theoretical, but from practical point of view we can count that almost all airfoils make lift 0.11 per 1 degree until ... :-)  . It hink this until is much more important here.

Have a look to following video comming from Martin Hepperle home page which shows how the stagnation point can be dependent on flap (or elevator) deflection, but without change of AoA. It also answers why the wire on LE can act as an turbulator. If stab is at 0 AoA and elevator is deflected down, the stagnation point is UNDER the wire, so airstream running from stagnation point to low pressure top of the airfoil must reach the wire - it turbulates him and make him more stable.

2/ Turbulators.

I did tests with turbulators on LE. I tried glued wire on LE, Wire glued only on few points (not sealed an ceratinly shaking while flight) and wire glued on points, but each second point was little moved to upper side an all others to lower side – the wire crossing the LE. It means definitely turbulating, but definitely not „pointing“. I found the second was most signifficant. It very well fits theory of turbulators. The fact is, that our stab flies at very low RE number. It means it could dramatically change flying properties in some cases. The turbulator helps. The proper construction of turbulator is: height (wire diameter) of 1/2 of boudary layer what is about 0.4mm placed where the air flow is laminar yet – it is in our case only LE as it often separate there (will be explained next). The question is where it could help. I found two different things. First is level flight with flat stab. The LE of flat stab is very blunt so it is adept to be able to make lot of lift. But „unproperly“ shaped airfoil where arc changes to flat surface can start laminar separation at very low AoA due to low RE number. The result is that little deflected elevator causes higher pressure gradient at LE (as tail has to produce lift – I note that around half of lift is created by first 10% of chord). In this case a „buble“ can apear at place where the round surface goes to flat and this can downgrade the lift. It means the pilot feel it is necessary to deflect elevator even more. This is positive feedback which can lead to oscilation in controlled system = hunting around level. The second effect is sometimes evident in flying of figures. Mostly corners. As explained by Alfredo, the stab flies at relatively high AoA in circular incoming flow. It is more AoA than sharp or thin airfoil could last without separation. We have to count with separation, but low RE number is ticket to troubles also here. The turbulator can help to more predictable switch between separated and not separated (will be spoken in next again). The result is, that if you move handle 100 times same way, you will get 100 same corners. Without turbulator it often happens that 99 are same and one ends few degrees out, or sometimes worse.

So I would collect this point: turbulator does not help everytime; I did not get worse result with turbulator than without.

2/ LE radius <-> RE number.

For a while I believed on triangular cross section for stab. I knew troubles of low RE number and I also knew the sharp LE moves critical RE number to lower RE (safe for us) numbers as sharp LE disturbs laminar flow like turbulator – yes, because the stagnation point is not the front point of airfoil – the flow must sometimes overrun the sharp LE from lower side around LE to upper side. (can you imagine it?) Also growing thickness (in direction of chord) keep safely high pressure gradient on both sides until hingeline – acts against laminar separation in level flight. May be somebody remember - this speach was already some years ago on Copuserve. After that I saw David‘s article (eyeopening for me) about stab, I was very surprised that he got good results with thin flat stab. It was very provocating for me, so I looked to polars of flat thin airfoil. And I was surprised even more. The flat (up to 3%) airfoil has NO critical RE number. Better said it does not change properties around critical RE number (whar is one of things causes hunting as I believe), even better said it so bad airfoil that it can not be worse under critical RE. But lift generation is not our problem, as we do not need so much lift (we need around .3 to .4 while it can make up to .5) on our relatively large tail.

After this knowledge I decided 3 things: I will make stabs with long chord to be at as high as possible RE number; if not (for example small models like my Tiny) I will use flat 3% airfoils; I will use as sharp as possible LE what moves critical RE numbe lower.

3/ Separation at LE.

It was already said that LE of the stab flies at variating AoA. If the elevator is deflected up, to the corner with positive G, and model entered circular path, the LE of stab flies at POSITIVE AoA up to 10 degrees. It can be too much for sharp LE, so it will separate at UPPER side of LE. But tail make negative lift to the corner, so it will make a „buble“ between LE and hingeline where air pressure grows again due to deflected elevator. This buble causes MORE negative lift on stab. The rule .11 per degree is not usefull as we have changed quality of airflow. If we will compare it to blunt LE of flat but thick stab, the 10 degrees can not make probles and such stab is much more resistent till some 15 degrees. The result is, that sharp stab has better (quicker) response to the corner than blunt. It allows to make model with longer or larger tail (or CG more to the front). Then we have comparable response of both stabs, but the shapr one is more stable at low deflections – level or stright fly or loops. This I used on my Next model which has longer tail than usual and even smaller elevator than usual. Anyway I have been told by some flyers who buit it that it is too sensitive and they moved CG even front to 17% MAC. I think it is question of preferences. Ted wrote he do not have Arnold‘s muscles, so he definitely like sharp stab. But there are some guys who likes to feel or overcome resistence to the corner.  

I can not state definitive numbers here, but at 4“ stab I foud that 3mm RADIUS – again, DADIUS is over and makes resistent model. 1mm radius makes separation bubble in corner – it makes sensitive but stabile model. I recommend to make a test on blunt LE – glue a wood rod with diameter of 2mm (radius 1mm) by TAPE matching it to airfoil ove all span. You will feel much easier turn to corner than without rod.

I never tested less diameter like 0.5mm. But I do not think it can bring something even better. But I hope Ted will tell us. The wire at front of blunt LE with less diameter than 0.5 does not change the thing this way – it is too small to cause separation and acts only like a turbulator.

Unfortunatelly I have to write I also found one negative point here. There is definitelly a small range of LE radius, where model does not make corner properly. It is somewher between, when some corner is done little tighter than others. In this case some corner is done with the buble while another without. It means the corner is difficult to reproduce. I also find that no CG position or flap/elevator ratio does not help. It this case the change of LE radius is necessary or – it enough and I did it on one of my models, add the turbulator to LE – it is much easier.

... little longer ...

... I hope helps ...

Regards

igor

Two notes here.

 Some month or two ago we spoke about thermal expansion of wood. We found that necessary AoA of stab is much higher than necessary to overcome gyroscopic moment. I calculated it and it looked the hot pipe could be the reason. Afterwards I did some measurements on model. It looks the AoA is no more than 0.15 degrees more on hot model than on cold. I have measuring tool showing scale 0.1 deg as 2 mm, so that measurement is not so precise, but enough - all values was between 0.1 and 0.2. It means together and together:

1/ Gyro moment needs around 0.1 AoA on tail. BUT – it is almost balanced by drag of landing gears and thrust line over the AC of wing (forger tail, its drag is almost 0). Balanced but only in level. The drag in corner makes negative pitching moment. Anyway it is not a reason to add any AoA to stab.

2/ The thermal expansion needs about 0.15 deg. If I will take 0.5 degrees AoA as an average value for optimum, there is still some 0.2 degree on not known down pitch.

3/ As we know that positive AoA of stab helps inside as well as inverted, the only reason I see is downwash from top of the wing airfoil (the elevator over the wing). Until now it is still open thing for me. Any idea here? Did anybody find optimal AoA on model with inline stab?

4/ The same goes on with nose. The engine nose down offset 0.xx degrees will also help.

Regarding aligning wing and stab AC: I agree, the stab Ac should be aligned to wing AC not to make roll. But do we know how aligned? In direction of flight (who knows what is the yaw in flight and how is changing in figures?)? Or in direction of fuselage?  ???

I would say that in line of flight. If we count little yaw out, and if we try to keep AC of wing somewhere close to center, the geometrically centered stab with little right AC will almost match flight position of wing AC.  Or not?

igor

Something about airfoil thicness: if you want compare a 18% airfoil and 16% airfoil, you will found that drag is around 10% less, but also maximal lift is around 10% less. It means you need 10% wing area and you are back with drag and maximal lift. Unfortunately the lift of any airfoil around zero lift (which we use in level or straight flight) is linear with angle of attack an does not depend on airfoil shape or thickness (OK, almost). It means in human language that larger wing with equivalent model weight is more sensitive for disturbed air - in wind and turbulence. So for me - I would say it is better to use better airfoil, does not matter what drag it produces and use only necessary wing area. AND we can still have stronger engine if necessary :-)

igor

It is a common question - what does a model in a corner, and how tight cornes we can fly, I hope following picture makes it little more clear. It is construction of several frames taken by my digital camcorder (PAL-50fps) composed to one picture. The quality is not best, but I thing it is eye opening picture anyway. 

First picture shows angle of attack in middle of corner. The estimated angle of attack of the "Next" model data sheet fits it well. 

Second shows diameter of the corner. Shown lenths are measured in picture metrics. The fuselage length corresponds to 1100mm, so the radius is little under 4m.

[ Home ]

Send mail to igor-hexoft@netax.sk with questions or comments about this web site.
Copyright © 1998 Hexoft
Last modified: októbra 21, 2001