NACA 6-series cusp

Discussion in 'Aircraft Design / Aerodynamics / New Technology' started by PiperCruisin, May 2, 2019.

Help Support HomeBuiltAirplanes Forum by donating:

  1. May 2, 2019 #1

    PiperCruisin

    PiperCruisin

    PiperCruisin

    Well-Known Member

    Joined:
    Jan 18, 2017
    Messages:
    91
    Likes Received:
    28
    Location:
    Idaho
    I imagine this subject has been somewhat hashed out on the forum from what I've seen, but here it goes.

    I've been looking into some airfoils for a design, but starting to confuse myself (not an aero guy). I have Riblett's book and was comparing his airfoil to the 6-series using Javafoil.

    I found the NACA 6 series to give some fairly good results (mainly cd vs cl; oldies but goodies?) while the Riblett results were a bit disappointing. Maybe it is how the software runs. It also showed much degraded results when you take away the cusp from the NACA foils which I understand is desirable for aileron forces.

    So, my questions:
    1. Is a typical observation?
    2. Is there a blended version between cusp and no cusp?
    3. If I was going to spend real money on an airfoil analysis program, what would you recommend (VisualFoil, Profilli 2, DesignFoil)?
     
  2. May 2, 2019 #2

    pictsidhe

    pictsidhe

    pictsidhe

    Well-Known Member

    Joined:
    Jul 15, 2014
    Messages:
    5,915
    Likes Received:
    1,506
    Location:
    North Carolina
    I am somewhat skeptical of Riblett's method for adding thickness and camber line. So far, the results using his (easy) method look ok for my XFLR foil experimentation. Though I want to get around to trying them the NACA way when I get a round tuit. I firmly believe his good results in his GA30 airfoils are entirely from using the a=0.5 6 series camber line with 4 digit thicknesses rather than his method of combining them. I have not tried his 6 series in XFLR5. XFLR5 was not impressed by his modified 4 digit thickness.
    While XFLR5 has some accuracy issues, I think that they are fairly consistent so it is useful for comparing similar airfoils.
    I believe that profili uses the same airfoil code as XFLR5

    The cusp-less 6 series have a couple of benefits.
    1. the cusp airfoils are very thin at the TE, this can involve them being fragile. I think this was the major reason for the cusp-less airfoils.
    2. Cusped airfoils also have higher stick forces. If you are building an LSA, this may even be a good thing. I am a long way away from scale ailerons on my 103, partly so that they need more than a sneeze to fully actuate...
     
    PiperCruisin likes this.
  3. May 2, 2019 #3

    fly2kads

    fly2kads

    fly2kads

    Well-Known Member HBA Supporter

    Joined:
    Jan 2, 2010
    Messages:
    1,411
    Likes Received:
    496
    Location:
    Justin, TX
    I am not any pro, but I am an airfoil nerd, and I have slept in a Holiday Inn sometime in the past. That said, here is my take:

    1. When I have looked at the Riblett airfoils, I have had to do some smoothing first before getting a good analysis. The point distribution was too coarse, and led to some discontinuities in the panel angles that threw off the results. Once that was fixed, they performed well.
    2. I haven't seen any blended versions, but some airfoil tools give you the ability to roll your own, setting geometry somewhere between two starter airfoil files.
    3. I haven't seen the need to spend real money on an analysis program yet. I use XFOIL most of the time, XFLR5 for some things, and have experimented with Xoptfoil with good results. (The latter two are based on X-Foil.)

    An additional observation on cusps in general...the cases I'm aware of where a cusp led to undesirable aileron forces were on aircraft that had both high wing loadings and high cruise speeds. Aircraft that are less highly loaded (like just about every sailplane made since the '60s) have worked just fine with cusped airfoils. So you will need to assess your particular application, see where you fall on that spectrum, and determine how much hinge moment is too much.
     
    PiperCruisin likes this.
  4. May 3, 2019 #4

    DeepStall

    DeepStall

    DeepStall

    Well-Known Member

    Joined:
    Dec 26, 2010
    Messages:
    136
    Likes Received:
    44
    Location:
    Colorado
    What Reynolds number? Are you running free boundary layer transition, or tripping it to turbulent somewhere? Are you able to anchor your calculations against NACA data in Theory of Wing Sections/Report 824 (or Report 903 for the 6-series / no cusp "A" modification)? Reports available at ntrs.nasa.gov

    1. The NACA data should tell you what trends to expect from cusp vs. no-cusp
    2. Haven't heard of anyone reducing the size of the cusp
    3. Ditto on not paying for airfoil analysis. Unless you're doing something very exotic, there are usually only a few % improvements to be had for designing your own foil vs. running with a "handbook" foil that's reasonably matched to your application and well-characterized (ideally wind tunnel tested, thanks NACA!). Risk starts popping up when you rely solely on computations for a foil and can't validate calculations.
     
    PiperCruisin likes this.
  5. May 3, 2019 #5

    PiperCruisin

    PiperCruisin

    PiperCruisin

    Well-Known Member

    Joined:
    Jan 18, 2017
    Messages:
    91
    Likes Received:
    28
    Location:
    Idaho
    Hi DeepStall...
    Re 1e6 to 3e6
    1. I've looked at the NACA docs with the cusp, but have not had a chance to dig up the no-cusp data...I'm assuming it exists.
    2. The Cherokee has a 6 series airfoil, but I don't believe it has the cusp. I have not heard of blending them. I tried that with a combination of Excel and Matlab, but it produced some wierd geometry at the T.E. which caused the results to go berserk.
    3. I would not feel bad about $100 if it got me some better results and some good tools to work with though I do like Javafoil. I see a lot of emphasis on Riblett or other typical glider foils (you know where I am going with this...), but the NACA 6-series seems to work well for my application though I would like to see if I could make some gains.
     
  6. May 3, 2019 #6

    pictsidhe

    pictsidhe

    pictsidhe

    Well-Known Member

    Joined:
    Jul 15, 2014
    Messages:
    5,915
    Likes Received:
    1,506
    Location:
    North Carolina
    The 6 series works pretty well. NACA went on to play a little with asymmetric airfoils in the 7 series. You could do something simular but without the nasty maths by using the top of one foil on the bottom of another. Laminar flow is easier to achieve on the bottom...
     
  7. May 4, 2019 #7

    Retiree

    Retiree

    Retiree

    Well-Known Member

    Joined:
    Nov 29, 2010
    Messages:
    78
    Likes Received:
    12
    Location:
    USA
    Riblett did publish cusp as well as non-cusp airfoils (GA Airfoils book).
    The cusp is the effect of aft loading of the airfoil. Aft loading gives higher lift coefficients, but with higher negative pitching moments.
    The NACA airfoils (Theory of Wing Sections book) provide a great set of airfoils with different characteristics to work with.
    Doug
     
    BBerson likes this.
  8. May 4, 2019 #8

    pictsidhe

    pictsidhe

    pictsidhe

    Well-Known Member

    Joined:
    Jul 15, 2014
    Messages:
    5,915
    Likes Received:
    1,506
    Location:
    North Carolina
    The cusp is to do with the pressure recovery and hence the thickness distribution, not the mean line. The mean line determines the loading.
     
  9. May 4, 2019 #9

    wsimpso1

    wsimpso1

    wsimpso1

    Well-Known Member

    Joined:
    Oct 19, 2003
    Messages:
    5,627
    Likes Received:
    2,866
    Location:
    Saline Michigan
    Something is wrong someplace. Harry's 35, 37, and 40 series airfoils are the NACA 6 foils with the leading edge reworked from very slightly reflexes into slightly drooped. They should be very similar for nominal flight except at high AOA. I suspect faceting influences...
     
  10. May 4, 2019 #10

    wsimpso1

    wsimpso1

    wsimpso1

    Well-Known Member

    Joined:
    Oct 19, 2003
    Messages:
    5,627
    Likes Received:
    2,866
    Location:
    Saline Michigan
    I do not understand this comment. The combination methods from the NACA and Riblett are the same. Set a camber curve to get the base Cl, and then add a thickness distribution to it. The difference comes in at how the thickness distribution happens at the leading edge. NACA fooled with the shape a bit for some reason that is unclear to me. Harry's assertion is that NACA ended up with a slightly reflexed leading edge and that this makes it stall sensitive. Harry then went on to make his thickness distribution more straightforward, then slightly droops the leading edge of the camber curves, then just adds them together.

    We have it on several good authorities, that Harry's foils stall soft and fly nice, and work as laminar flow foils. Whether that is because of Harry having the right theory or program or if he just got lucky is anyone's guess.

    Billski
     
  11. May 4, 2019 #11

    Retiree

    Retiree

    Retiree

    Well-Known Member

    Joined:
    Nov 29, 2010
    Messages:
    78
    Likes Received:
    12
    Location:
    USA
    Actually not quite the same. NACA added the camber to the symmetric airfoil perpendicular to the camber line. Riblett just added the camber to the y coordinate of the symmetric section. So Riblett's airfoils have somewhat more camber than the equivalent NACA sections.
    Doug
     
  12. May 5, 2019 #12

    pictsidhe

    pictsidhe

    pictsidhe

    Well-Known Member

    Joined:
    Jul 15, 2014
    Messages:
    5,915
    Likes Received:
    1,506
    Location:
    North Carolina
    Riblett simply adds the thickness and camber together. NACA takes the camber line and then add the thickness to that curve. Riblett's book has a good explanation of the two methods.
     
  13. May 6, 2019 #13

    Starflight

    Starflight

    Starflight

    Well-Known Member

    Joined:
    Nov 7, 2010
    Messages:
    51
    Likes Received:
    2
    Location:
    Saskatchewan..."GO RIDERS"(pride)where's my banjo?
    I used TOWS to design an airfoil section for a very light U/L. Wanted to see what the shape would be, using the NACA a=0.4 mean line and combining it with a laminar 63-515 thickness form.
    Drew it out on paper with ordinates perpendicular to the camber line, and a rather beautiful airfoil developed. Theoretical results looked very promising when I used Javafoil to analyse it :)
    Written out, the designation would be: NASA 63(2)-515 a=0.4 modified
     
  14. May 6, 2019 #14

    Retiree

    Retiree

    Retiree

    Well-Known Member

    Joined:
    Nov 29, 2010
    Messages:
    78
    Likes Received:
    12
    Location:
    USA
    Show us a picture please. If you want to, please tell us the U/L performance characteristics you used for the design of the airfoil.
    Doug
     
    BJC likes this.
  15. May 6, 2019 #15

    PiperCruisin

    PiperCruisin

    PiperCruisin

    Well-Known Member

    Joined:
    Jan 18, 2017
    Messages:
    91
    Likes Received:
    28
    Location:
    Idaho
    Here is what I found so far. Below is a comparison between cusp and no-cusp with the NACA report and Javafoil. Javafoil did well except for the drag with the no-cusp. Might have something to do with the image below when it is generated (entry by hand produced the same result, a bump on the top surface near the TE). Manual smoothing helps a little (cd ~0.008) which is about what I get for the Riblett. Would like to find no-cusp data for something like a A415 to A618. Anyone?
    upload_2019-5-6_8-27-18.png
    upload_2019-5-6_8-34-53.png
     
  16. May 7, 2019 #16

    pictsidhe

    pictsidhe

    pictsidhe

    Well-Known Member

    Joined:
    Jul 15, 2014
    Messages:
    5,915
    Likes Received:
    1,506
    Location:
    North Carolina
    That bump at the back is wrong.
     
  17. May 7, 2019 #17

    PiperCruisin

    PiperCruisin

    PiperCruisin

    Well-Known Member

    Joined:
    Jan 18, 2017
    Messages:
    91
    Likes Received:
    28
    Location:
    Idaho
    Yeah, I know. But when you enter the data from the NACA report or generate it, that is what I get. I manually smoothed it which improved things a bit, but still not along the lines of what the test data says.

    So, does anyone have a NACA 6A series data that performs like the test data?
     
  18. May 7, 2019 #18

    pictsidhe

    pictsidhe

    pictsidhe

    Well-Known Member

    Joined:
    Jul 15, 2014
    Messages:
    5,915
    Likes Received:
    1,506
    Location:
    North Carolina
    TOWS has both airfoils. Javafoil generates 6 series that look correct. It's not my favourite analysis program, but it's not awful. It looks like a single bad datapoint in your foil above.

    Last 3 upper surface points for the 642A215:

    90.076 2.046
    95.039 1.039
    100 0.032

    By inspection, that's a pretty straight line. if you still get a kink, something is wrong with your drawing
     
    Last edited: May 8, 2019
  19. May 7, 2019 #19

    Starflight

    Starflight

    Starflight

    Well-Known Member

    Joined:
    Nov 7, 2010
    Messages:
    51
    Likes Received:
    2
    Location:
    Saskatchewan..."GO RIDERS"(pride)where's my banjo?
    Sorry about the lack of an illustration, Retiree (Doug) but I am dismal at handling software graphic operations directed at online forums. :-( It is easy to generate with Javafoil, and it is a valid program when comparing other candidate airfoils against one another. Tabulated data results could have significant error when applied to aircraft design. Hello, "pictsidhe". Can you help Retiree with that custom a=0.4 NASA mean line airfoil he would like see; please :)
     
  20. May 8, 2019 #20

    pictsidhe

    pictsidhe

    pictsidhe

    Well-Known Member

    Joined:
    Jul 15, 2014
    Messages:
    5,915
    Likes Received:
    1,506
    Location:
    North Carolina

Share This Page

arrow_white