Logo
Builders' Plan Gallery  |  Hip Pocket Web Site  |  Contact Forum Admin (Account/Technical Issues)  |  Contact Global Moderator
May 08, 2021, 03:24:46 AM *
Welcome, Guest. Please login or register.
Did you miss your activation email?

Login with email, password and session length
 
Home Help Search Login Register
Pages: 1 [2]   Go Down
Print
Author Topic: General question about washout  (Read 4616 times)
0 Members and 1 Guest are viewing this topic.
RolandD6
Gold Member
*****

Kudos: 24
Offline Offline

Australia Australia

Posts: 862




Ignore
« Reply #25 on: October 20, 2010, 08:35:07 PM »

Thank you everybody for your discussion above. I found it very enlightening and the overall impression I got was that "One has to suck it and see" about the suitability of variable camber and compensating washout or washin for any particular design application.

A while back (ie. pre-broken shoulder) I had done some experimenting with Profili to investigate the likelihood of airfoils based upon log curves being suitable for my particular area of interest, ie roughly peanut sized indoor rubber power scale models. The results were encouraging enough for me to move on to the next phase of this project, namely a device for shaping accurate wing form blocks (wire cut foam or sanded balsa) and finishing sanding the upper profile of a wing structure to the desired contour.

I played around with various ideas until I settled on a variation of devices people build for hot wire cutting of large tapered foam wing cores. The device had to be able to cut/sand wing plan forms varying from parallel LE-TE to a taper ratio of about 3:1. The latter requirement is to accommodate a MiG 3 or LaGG 5 wing. The device needed to accommodate wing chords in the range 33 to 105 mm; variable camber and thickness along the wing span; any desired geometric washin or washout and also swept forward (eg. Le Vier Cosmic Wind) or swept back (eg. Curtiss-Wright CW-21) wings.

The trick is that the hot wire or sanding bar must traverse the root and tip profiles at rates proportional to the relative chord lengths so that there is a smooth transition between the profiles. The difficulty with a hand held sanding bar or hot wire cutter is keeping the motion smooth and correctly proportional on a tapered wing.

At present I am creating CAD drawings of this device (about all that I can do with my current physical restrictions) and I plan to begin a new thread elsewhere on HPA when I have finished them. I plan to use off-the-shelf Dubro or Great Planes 4-40 ball ends and so on for the linkages but some parts will be special purpose. I will be using aluminium extrusions because I have a small machine shop but there should be no reason why the device could not be built from timber.

At present I have nothing much to show other than some chicken scratchings on paper and conceptual CAD drawings but at the risk of boring you with more polar diagrams, I will post a few below as an indication of the results I got with Profili. It looks like log curve based profiles will be suitable for my build list of biplane prototypes but the device described above will not be limited to them. There should be no restriction on creating a form block that accommodates a 'semi-symmetrical' profile at the wing root and an 'undercambered' profile at the tip end or vice-versa.

Paul
Logged
RolandD6
Gold Member
*****

Kudos: 24
Offline Offline

Australia Australia

Posts: 862




Ignore
« Reply #26 on: October 20, 2010, 08:55:16 PM »

The first airfoil based upon log curves.

This profile probably represents the upper limit (or nearly so) of a range of possibilities that suit my particular area of interest. The section is rather thick and some instability in the airflow is becoming evident.

I have labeled the profile 1.8%_over_7%_with_3.4%_UC_0.9%_LE which means the following. All percentages are relative to the chord length:

* 1.8% thick laminated or steam bent upper rib bow, over

* 7% thick logarithmic curve convex template profile;

* 3.4% logarithmic curve under camber, and

* 0.9% leading edge radius.

The name ignores the TE thickness which is about 1.3%, ie. about 0.025" for a 51.5mm wing chord (as per the 1/25.4 scale Banhidi Gerle project)

The resulting height of the upper surface of the profile is about 8.8% above the aerodynamic chord line.

Further extensions to the file names of the attached polar diagrams should be self evident.
Attached files Thumbnail(s):
Re: General question about washout
Re: General question about washout
Re: General question about washout
Re: General question about washout
Logged
RolandD6
Gold Member
*****

Kudos: 24
Offline Offline

Australia Australia

Posts: 862




Ignore
« Reply #27 on: October 20, 2010, 08:58:19 PM »

Next is the same profile but with a 1.2% undercamber.
Attached files Thumbnail(s):
Re: General question about washout
Re: General question about washout
Re: General question about washout
Re: General question about washout
Logged
RolandD6
Gold Member
*****

Kudos: 24
Offline Offline

Australia Australia

Posts: 862




Ignore
« Reply #28 on: October 20, 2010, 09:01:50 PM »

Next is a similar profile but with a 1.25% rib bow.

1.25%_over_7%_with_3.4%_UC_0.625%_LE
Attached files Thumbnail(s):
Re: General question about washout
Re: General question about washout
Re: General question about washout
Re: General question about washout
Logged
RolandD6
Gold Member
*****

Kudos: 24
Offline Offline

Australia Australia

Posts: 862




Ignore
« Reply #29 on: October 20, 2010, 09:04:11 PM »

And again with a 1.2% under camber
Attached files Thumbnail(s):
Re: General question about washout
Re: General question about washout
Re: General question about washout
Re: General question about washout
Logged
RolandD6
Gold Member
*****

Kudos: 24
Offline Offline

Australia Australia

Posts: 862




Ignore
« Reply #30 on: October 20, 2010, 09:13:56 PM »

And last is a different profile again that is indicative of the similar results possible with this type of airfoil.

1.4%_over_6.8%_with_1.2%_UC_0.7%_LE

The interesting thing with these profiles is that Xfoil is predicting much the same zero lift Alpha of about -1.5° for them all for expected Reynolds Numbers so there may be little need for geometric wash-out to compensate for aerodynamic wash-in. Possibly 0.5° is all that is needed to allow for slight errors in construction, perhaps a bit more for a low wing airplane. The Re values or 6000, 8000 & 20000 are plotted to show possible behavior outside the anticipated operating range.

For what it may be worth Undecided Grin

Paul
Attached files Thumbnail(s):
Re: General question about washout
Re: General question about washout
Re: General question about washout
Re: General question about washout
Logged
OZPAF
Palladium Member
********

Kudos: 189
Offline Offline

Australia Australia

Posts: 6,228



Ignore
« Reply #31 on: October 23, 2010, 06:13:03 AM »

A bit late again but thanks TMAT. I should do some checking on the current approach with F1B wing design - it all sounds very interesting as is this thread.

Roland - thanks for posting your plots - interesting and I agree with your basic conclusion - re the lack of any great difference in max Cl with camber and a much CL as well and the difficulty of making the higher camber/washout approach work at your low Re's.

I would suggest having a look at moving the thickness high point forward on your lowest camber thin foil for use on the tip. Its likely to give you a slightly better CL/CD polar near the stall, with better handling in turbulence particularly and less drag overall.

Also as the wing will require a lower CL near the tip anyway - you could use this to your advantage by lowering the camber even further to reduce tip drag and thus wing drag as much as possible.

If you haven't already had a look - check out XFLR5 which can give a spanwise lift graph to help with your tip design. Alternatively a simpler approach is to use "LIFT/ROLL"
John
Logged
RolandD6
Gold Member
*****

Kudos: 24
Offline Offline

Australia Australia

Posts: 862




Ignore
« Reply #32 on: October 24, 2010, 03:31:54 AM »

John

Thanks for tip in regard to XFLR5. I had not heard of it.

I have downloaded both v5.00 and v6.01 beta and documentation from sourceforge.net to have a bit of a play.

Regarding moving the maximum thickness further forward:
1. Forming a relatively thick upper rib bow over a logarithmic curve template does this a small amount;
2. I found that moving the maximum thickness further forward by using a different profile did not provide much advantage (at these very low Reynolds Numbers) and there was increased risk of separation of the laminar flow over the profile (at least according to Xfoil). I am trying to avoid sudden transition from laminar to turbulent flow and therefore the likelihood of unsteady flight.
3. I have not bothered to examine the possible performance of these profiles at normal stalling angles (ie 13°- 15°) since my intention is to trim for flight with the AoA around 3° to 5° which should (I hope) produce much more 'scale looking' flight patterns. My models are for indoor flying and ultimate duration is not an ambition.

It is all a challenging exercise for the old brain which has been struggling a bit lately.

Regarding changing camber along the wing or not:

All options are open. My current project is to develop a mechanised way for producing wing form blocks quickly and accurately so that different ideas regarding airfoils and their distribution along a wing can be experimented with without indulging in many (challenging for me) hours of carving and sanding etc.

Paul
Logged
OZPAF
Palladium Member
********

Kudos: 189
Offline Offline

Australia Australia

Posts: 6,228



Ignore
« Reply #33 on: October 24, 2010, 06:50:16 AM »

Paul,

Just for interest I had a quick look at a aerofoil similar to your best - the 1.25%over 7% with 3.4%UC 0.625% LE. This is a 5.43% Thick at 35% and 5.16%Camber @ 37%. As you are looking at best performance at around CL of 0.4 or approx 4deg AOA then my impression was you may be better of with far less camber.

Bearing this in mind I processed a source foil to develop a section similar to foil and then did another couple of quick mods for the same camber airfoil with the same thickness(for strength) but with the high point moved forward.

The third effort has the camber reduced to 1.94% and the same thickness again but moved considerably forward. it is by far the best performer. It should give an easier to trim wing at these Re's- due to the shape of the curve at min drag(the low drag bucket width). The min drag is also much lower and should thus require less power.

This second mod section is only 1.91%camber appearing to suit the working CL much better.

Keep working on it and good luck - its a good use for XFOIL/Profili.
Hope this is not taking the thread to far off the subject.

John
Attached files Thumbnail(s):
Re: General question about washout
Re: General question about washout
Re: General question about washout
« Last Edit: October 24, 2010, 07:10:06 AM by OZPAF » Logged
Duco Guru
Guest

« Reply #34 on: October 24, 2010, 07:01:27 AM »

Paul, Hope this is not taking the thread to far off the subject.
John

 Roll Eyes Roll Eyes Roll Eyes Roll Eyes Roll Eyes Roll Eyes Roll Eyes Roll Eyes Roll Eyes Roll Eyes Roll Eyes Roll Eyes Roll Eyes Roll Eyes Roll Eyes Roll Eyes Roll Eyes

Guru
Logged
Yak 52
Titanium Member
*******

Kudos: 77
Offline Offline

United Kingdom United Kingdom

Posts: 2,734


Jon Whitmore



Ignore
« Reply #35 on: October 24, 2010, 08:38:29 AM »

Paul,

Camber is really not your friend at these Re! I would rule out anything above 4% and as suggested you will still have ample performance at your desired CL.

At Re below 20,000 pretty much everything will be laminar flow - so worrying about transition for steady flight is probably less important than looking at hysteresis problems and laminar separation bubbles.

Here's an interesting paper:

http://icas-proceedings.net/ICAS2004/PAPERS/192.PDF

Regarding thickness distribution there may be gains available by reducing the pressure gradient in the recovery part of the upper surface - ie flattening out (or even hollowing a little) the last 40% chord of the upper surface. This is sometimes called a 'bubble ramp' as it should help prevent separation bubbles. It has an effect on stall behaviour but as you said that's not too much of a concern for you.

Bear in mind that computed models have serious limitations at low Re - XFLR5 is one of the best but none of them are great at predicting laminar separation. Interestingly when flow is fully laminar, Re 10,000 or less, XFLR5 is said to be pretty accurate!

I have found that the real gains in model performance at these sizes is by addressing wing loading. This has the obvious benefits on sink rate but also at low Re the CLmax tends to be much lower - often only 0.6-0.7.

At low Re, operating near CLmax comes with separation, hysteresis, bubbles etc so I go for something like CL0.2-.0.3 at very low wing loadings. This means low camber and low drag.

It's also born out in the peanut competition circuit - low wingloading is the holy grail!
Logged
RolandD6
Gold Member
*****

Kudos: 24
Offline Offline

Australia Australia

Posts: 862




Ignore
« Reply #36 on: October 24, 2010, 06:05:26 PM »

I think Guru is right. We are beginning to diverge from the intent of the original post, ie washout. Embarrassed

Earlier posts opened a discussion about variable camber and its aerodynamic washin or washout effect. I posted a few polar diagrams to show that some airfoil profiles under certain circumstances may not need washout to compensate for aerodynamic washin. Now we may be heading down a completely different path unrelated to the original intent of Roman's question.

So, I recommend that one of you open a new thread with an appropriate title. I feel certain that there are others who will be interested in what you have to say as well as myself.

Meanwhile I will followup on your most recent advice.

Regards

Paul

PS. I have downloaded the paper 192.pdf and copied your comments to my computer so that I can study the information therein more comfortably.
« Last Edit: October 24, 2010, 06:21:15 PM by RolandD6 » Logged
TimH
Nickel Member
*

Kudos: 0
Offline Offline

United States United States

Posts: 1



Ignore
« Reply #37 on: May 02, 2021, 05:55:26 PM »

For washout as well as many other design factors, it is important to realize which parts of the flight envelope will be affected, either aided or degraded, by changes to the design factor in question.   This is also why details like washout may be discussed and debated endlessly - ultimately they become part of a blend of many parameters going into a single design, often with little or no hope of isolating any performance change that may  result directly or indirectly from any single design feature.  Naturally it seems a shame to build-in wing twist if it adds drag especially at higher speeds.  On the other hand, a single ugly tip stall on downwind for landing can be fatal to the model, so we have to make our conceptual design decisions based on experience as well as on the heritage with whatever design we are developing.  When starting out completely from scratch, 'better safe than sorry' may be a good motto.  Later iterations on the same overall vehicle platform provide opportunity for 'tuning' the configuration depending on the wing loading, power loading, scale and other factors.   Framed up wings that use shrink covering for torsional rigidity may also provide adjustability for washout after the build, by getting out the iron or heat gun and 'setting' a different twist into the wing for fine tuning.

One approach to the tip stall issue at lower speeds or higher angles of attach is the Flap-A-Lator configuration.  This concept uses a relatively large wing flap for pitch control, changing the relative angle of incidence between the wing and horizontal stabilizer, or 'decollage' angle.  The stab is set at a fixed angle depending mainly on its own airfoil, camber & surface area, as well as on the CG location of the overall model.  An example is shown at:
https://www.youtube.com/watch?v=7clsTfbXpC4
from June 2020 flight test sessions of a 1.5 m span HL glider that was scratch-built in 1998.  This model turned out to be very responsive to pitch control input, especially at lower loading and further aft CG location (significantly behind the flap hinge).  The limit of this aft CG has not yet been established.  In the flight test sessions shown in the linked video, aft CG was flown for limited duration due to the smaller battery required to achieve that balance state. There was also significant turbulence, making higher wing loading advisable due to those on-site conditions (June in the Outer Banks barrier islands of the Carolinas, East Coast USA).

Many advantages of the Flap-a-Lator include:
* No elevator linkage routing to the tail, eliminating that critical few grams weight on what may be a very long tail boom,
* lower yaw inertia due to less required tail hardware, and/or a longer tail boom because of less required mass, allowing a smaller fin & rudder for lower drag, etc. - these are all inter-related,
* variable camber of center wing panel allows a low-% thickness airfoil that will still have effective camber when 'up elevator' (downward flap deflection) is applied, but has very low drag at low angle of attack,
* very good speed range since the thin airfoil and small tail both contribute to very low drag at higher speeds (at lower positive angles of attack), and
* outstanding low speed directional control because adding back stick control deflection (equivalent to 'up elevator') adds effective washout, since the outboard panels are at fixed incidence relative to the fuselage and tail.  But the washout is variable, so it goes away at low alpha (angle of attack) when you don't want it there adding drag, and only comes into play when you need it while riding back-stick or high alpha.

In addition, the pitch control response is also very fast because changing the flap angle is an effective means to change the wing's overall angle of attack immediately (i.e. a primary result of the control deflection).  This is in comparison to a conventional hinged elevator or even a fully flying stab which has to change the angle of attack of the horizontal tail which then accelerates the airframe in pitch to change the wing's relative angle of attack (i.e. a secondary result of the control deflection).  There is also a secondary effect of the flap changing the effective airflow direction (incidence angle) over the stab, but the primary effect is obvious as soon as one touches the stick in pitch.  Response is immediate.

There is a further benefit to this configuration: at lower speeds when angle of attack and induced drag are higher, changing the flap angle remains highly effective for pitch control (rapid and authoritative response).  This is not typical for conventional elevator control as airspeed decreases.  Apparently the Flap-A-Lator directly leverages the increased magnitude of the bound vortex (essentially the induced drag component of lift from the wing) and remains highly effective even when the flow rate across the wing surface is reduced.   These types of results are exactly why it can be very interesting to experiment with variations when designing and building.  Of the 4 attempts at various Flap-a-Lator configurations, two have been successful, and two were duds, so that is the potential down-side to the experimental process.  Both straight-wing attempts with ailerons were poorly controllable, while both 3-panel polyhedral attempts showed good pitch and steering authority using flap and rudder respectively.  Interestingly the first attempt was 3-panel polyhedral with a relatively large T-tail stab on a relatively short tail boom, and the T-tail worked well. 

Naturally the torsional stiffness of the wing's center panel spar is critically important for this structural config.  With the framed-up wing structure shown, the torsional stiffness and torsional strength may be varied by choice of covering material.  The example model uses thick covering for the fixed forward portion of the center panel, with a thinner covering on the flap and outboard wing panels.  The relatively small rudder at extended tail moment arm of the model in the video provides excessive steering authority even at low speed on this prototype, requiring softening exponential in the programmable transmitter. 

Reducing drag is possible by many different design & construction methods.  The same goes for improving control effectiveness.  In the end, many of these methods and ideas can be 'over-thought' or over-emphasized, with time and energy spend where it doesn't need to be.  Don't be afraid to experiment with different configurations, because you will often find a result that is more to your liking than the model that you never get to fly because it is still sitting on the building board.  Stability doesn't always have to be the enemy of controllability, so if you are trying RC just realize what elements of each quality you are looking for, and then try to design in options that will give the best blend of both.  Ultimately that preferred 'blend' is only realized by trial and error.  Getting into the air is the only way to know for sure, and even then you need calm air and steady conditions for test flights to know what the model is doing on its own vs. what the wind or terrain are doing to the model.
Logged
OZPAF
Palladium Member
********

Kudos: 189
Offline Offline

Australia Australia

Posts: 6,228



Ignore
« Reply #38 on: May 03, 2021, 09:55:41 PM »

Interesting input and model Tim. Thanks.

John
Logged
dosco
Platinum Member
******

Kudos: 10
Offline Offline

United States United States

Posts: 1,106



Ignore
« Reply #39 on: May 04, 2021, 08:18:56 AM »

Tim:
Neato!

-Dave
Logged
Pages: 1 [2]   Go Up
Print
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.21 | SMF © 2015, Simple Machines Valid XHTML 1.0! Valid CSS!