Leading & trailing edges for fins

The optimum trailing edge shape is as tapered as you can stand structurally. I usually recommend a 3:1 ogive cross-section.

The leading edge can be hemispherical for low subsonic but for most rocket applications ogive again is the winner.

Jerry

I think Jerry is right on about the ogives. But a hemispheric leading edge and tapered trailing edge may be easier to execute.

see these construction tips on beveling:

"Fin Beveling [Jig]", by John Coker at

I also came across these interesting articles on the aerodynamics of fin shapes (not bevels)

see the section on "Control Surfaces" in "SMS Guided Missiles, Aerodynamics, And Flight Principles" (Gunners Mate NTRC) at

- iz

Ken Baldw> What's considered the optimal shape of the leading and trailing edges of > rocket fins?

My answer preceeded those tomes being authored and remained my answer after integrating them.

Sometimes the simple answer IS the best answer.

Thanks to both of you, I appreciate the input.

An ogive is going to be something beyond my limited tool shed :-) The tips by Tim Van Milligan should be easy enough to do for model rockets, but a little impractical for larger ones, unless you've got the forearms of Hercules.

I may have to talk to my wife's boss, they run a mill work shop and I suspect they have the tools to fabricate something along those lines for me.

Thanks again,

Ken CAR #S825 replace nospam with telus to reply

It is a taper and the surfaces simply rounded with some sandpaper. It is actually the easiest way. It might not be a "true ogive" but it will be close enough to be better than a pure taper or a pure hemisphere.

I plan on building the jig from John Cokers's webpage myself :)

- iz

Ken Baldw> Thanks to both of you, I appreciate the input.

The best answer is to look at the shape of aircraft that fly at the speeds your rocket will travel. Subsonic bodies and symmetric airfoils are elongated teardrops. Rounded in front, knife edge at the rear. The best natural description I've found is they should look like the top view of a trout. For more detail, I recommend the Soartech #8 publication which has the most extensive testing of low reynolds number airfoils I've ever seen.

At supersonic speeds, double wedge, biconvex, and other knife edge at both the front and back are best.

Bob Kaplow NAR # 18L TRA # "Impeach the TRA BoD" >>> To reply, remove the TRABoD!

This reminds me a LOT of a "hollow ground" knife sharpener. If you examine the result very carefully, you'll find that the bevels this cuts aren't actually flat, but are slightly concave. This is because the curved edge of the sanding disk is what is cutting the finished bevel. Probably nothing really wrong with this on our rockets, but doing similar with the flat of the belt sander wouldn't leave this very minor arc.

Interesting. Still not quite what I want, but pretty close. And definitely not practical for my competition models with 1/64 plywood fins :-)

Bob Kaplow NAR # 18L TRA # "Impeach the TRA BoD" >>> To reply, remove the TRABoD!

thanks, Bob!

from

Abdur-Rasheed writes:>

Bob, All ya need is a 1/128 round over bit and your all set ;-)

RDH8

See NACA 4-digit airfoils:

The last two digits, thickness as percentage of chord, would be low for a rocket to minimize drag.

Back in my DEC days I came across a postscript program that would plot rib patterns for NACA4 series airfoils. I just dug it up, and here it is:

%!PS-Adobe-2.0 EPSF-1.2 %%Title: AIRFOILS.PS %%Creator: G SCHRADER %%CreationDate: 08/20/90 %%Pages: 2 %%BoundingBox: % VER 2.0 % added thickness function % added reverse function % VER 3.0 % modified to use Bezier curves % added leading edge, trailing edge, skin thickness % added mirrorpage function % VER 3.1 % added closed curve smoothing % broken up to support text library command procedure % VER 3.2 % added NACA 4 digit airfoil generator % added rib set support % added extended labeling % added wing jig hole option % added washin/washout support % added thickness tapering support % added elliptical taper blending % added draft mode % added multiple copy support %%EndComments %%BeginProlog statusdict begin /waittimeout 400 def end /AirfoilDict 200 dict def AirfoilDict begin /cl /closepath load def /ex /exch load def /gs /gsave load def /gr /grestore load def /li /lineto load def /mo /moveto load def /rm /rmoveto load def /np /newpath load def /rl /rlineto load def /ro /rotate load def /sh /show load def /sc /scale load def /st /stroke load def /tr /translate load def /tan {dup sin ex cos div} bind def /exd {ex def} bind def /cp /clip load def /vec 300 array def /dst 300 array def /dta 300 array def /currentpoint 0 def /rotatestate 0 def

/thickness { /factor exd [ex /sss 0 def {sss 0 eq {/sss 1 def}{/sss 0 def factor mul}ifelse}forall] } def

/reverse { [ex /sss 1 def {sss 0 eq {/sss 1 def}{/sss 0 def 1 ex sub}ifelse}forall] } def

/mirrorpage { clippath pathbbox pop ex pop add 0 tr -1 1 scale } def

/fade {dup 3 -1 roll mul ex 1 ex sub 3 -1 roll mul add} bind def

/makevectors { 0 2 dsize 1 sub { /i exd /ifp i dsize add 2 add dsize mod def /ibp i dsize add 2 sub dsize mod def % distances /dxb dta i get dta ibp get sub def /dyb dta i 1 add get dta ibp 1 add get sub def /dabsb dxb dup mul dyb dup mul add sqrt def /dxf dta ifp get dta i get sub def /dyf dta ifp 1 add get dta i 1 add get sub def /dabsf dxf dup mul dyf dup mul add sqrt def % vectors /dx dxb dabsb mul dxf dabsf mul add dabsb dabsf add div def /dy dyb dabsb mul dyf dabsf mul add dabsb dabsf add div def /dabs dx dup mul dy dup mul add sqrt def vec i dx dabs div put vec i 1 add dy dabs div put % back distance /dstb dta i get dta ibp get sub dup mul dta i 1 add get dta ibp 1 add get sub dup mul add sqrt def % front distance /dstf dta ifp get dta i get sub dup mul dta ifp 1 add get dta i 1 add get sub dup mul add sqrt def

dst i dstb dabs dstb dstf add div dup 2 div 2 add div mul 2.0 mul put dst i 1 add dstf dabs dstb dstf add div dup 2 div 2 add div mul 2.0 mul put } for } bind def

/smoothcurve { np dta 0 get dta 1 get mo 2 2 dsize 1 sub { /idx exd /ifp idx dsize add 2 add dsize mod def /ibp idx dsize add 2 sub dsize mod def %control point 1 dta ibp get vec ibp get dst ibp 1 add get mul add dta ibp 1 add get vec ibp 1 add get dst ibp 1 add get mul add %control point 2 dta idx get vec idx get dst idx get mul sub dta idx 1 add get vec idx 1 add get dst idx get mul sub %endpoint dta idx get dta idx 1 add get curveto } for } bind def

/draftcurve { np dta 0 get dta 1 get mo 2 2 dsize 1 sub {/idx exd dta idx get dta idx 1 add get li} for } bind def

/curveoffset { 0 2 dsize 1 sub { /i exd dta i dta i get vec i 1 add get cofs sca div mul sub put dta i 1 add dta i 1 add get vec i get cofs sca div mul add put

} for } bind def

/DrawCoords { gs /lbl ex dup type /stringtype eq{[ ex ]}if def % convert to array if necessary /sca exd /yofs exd /xofs exd /curv exd /ymax -100 def /ymin 100 def /ymaxx 0.3 def /dsize curv length def

/lbl [lbl aload pop ribnumber 0 ne {( R#) ribnumber 5 string cvs}if ( Ch) sca 100 mul round 100 div 15 string cvs AOA 0 ne {(" AOA ) AOA 100 mul round 100 div 10 string cvs}if ] def

curv dta copy pop /stx dta 0 get def 2 2 dsize 1 sub { /i exd i 1 add curv ex get dup ymax gt { dup /ymax exd /ymaxx curv i get def } if dup ymin lt { dup /ymin exd } if pop } for

xofs yofs tr rotatestate 0 eq {-90 ro}if offsetmode 0 ne {0 ymin sca mul neg tr}if

stx 0.5 lt {sca 0 tr -1 1 sc}if AOApoint ymax ymin add 2 div sca mul tr AOA ro

holesize 0 ne { frontholepoint AOApoint sub holeelevation sca mul 0.01 mul tr np 0 0 holesize 2 div 0 360 arc holesize 0.75 mul dup 0 mo dup neg 0 li dup 0 ex mo 0 ex neg li st rearholeoffset 0 tr np 0 0 holesize 2 div 0 360 arc holesize 0.75 mul dup 0 mo dup neg 0 li dup 0 ex mo 0 ex neg li st rearholeoffset neg 0 tr frontholepoint AOApoint sub neg holeelevation neg sca mul 0.01 mul tr }if

AOA neg ro AOApoint neg ymax ymin add 2 div sca mul neg tr stx 0.5 lt {-1 1 sc sca neg 0 tr}if

sca dup scale

/chrhgt 0.03125 def /Helvetica findfont chrhgt scalefont setfont

makevectors draftmode {draftcurve}{smoothcurve}ifelse st draftmode {draftcurve}{smoothcurve}ifelse cl cp

/ymax ymax 0.5 add def /ymin ymin 0.5 sub def np stx 0.5 gt {leading}{trailing}ifelse sca div dup ymax mo ymin li 1 stx 0.5 lt {leading}{trailing}ifelse sca div sub dup ymin li ymax li cl st np stx 0.5 gt {leading}{trailing}ifelse sca div dup ymax mo ymin li 1 stx 0.5 lt {leading}{trailing}ifelse sca div sub dup ymin li ymax li cl cp

/cofs skin stx 0.5 lt {neg}if def curveoffset draftmode {draftcurve}{smoothcurve}ifelse st

offsetmode 0 ne {0 ymin sca mul tr}if rotatestate 0 eq {90 ro} if 1 sca div dup scale xofs neg yofs neg tr gr /height ymax ymin sub sca mul def

} bind def

/copies 1 def /draftmode false def /ribcount 3 def /rootAOA 0 def /tipAOA 0 def /AOA 0 def /rootAOApoint 0 def /tipAOApoint 0 def /AOApoint 0 def /rootchord 10 def /tipchord 10 def /chord 10 def /rootfrontholepoint 0 def /tipfrontholepoint 0 def /frontholepoint 0 def /rootholeelevatiion 0 def /tipholeelevation 0 def /holeelevation 0 def /rearholeoffset 0 def /holesize 0.25 def /rootthickness 1 def /tipthickness 1 def /elliptical 0 def /offsetmode 0 def

end

%%EndProlog

%%EndSetup

AirfoilDict begin

/cambercalc4A { dup maxc sub ex 0.001 add maxc ge {maxcam} {maxc} ifelse div dup mul neg 1 add camber mul } bind def

/thickcalc4A {

dup 0.3 le {0.3 sub 0.3 div dup mul neg 1 add 0.329 exp 1.25 mul -0.25 add} {0.3 sub 0.7 div 1.67 exp neg 1 add 0.98 mul 0.02 add} ifelse 0.5 mul th mul } bind def

/intercept {

/inty2 exd /intx2 exd /intm2 exd /inty1 exd /intx1 exd /intm1 exd

inty2 inty1 sub intm1 intx1 mul add intm2 intx2 mul sub intm1 intm2 sub div dup intm1 mul inty1 add intm1 intx1 mul sub

} bind def

/NACA4 {

/NACAnum exd /th NACAnum cvi 100 mod 100 div def /maxc NACAnum 100 div cvi 10 mod 10 div def /camber NACAnum 1000 div cvi 10 mod 100 div def /maxcam 1 maxc sub def

/NACAstr [(NACA) NACAnum 10000 add 8 string cvs 1 4 getinterval] def

%tip /rad 1.1 th dup mul mul def /rang 0 def /mrang rang tan def /EF 1.5 def %upper le /x2 0.1 def /y2 x2 thickcalc4A EF mul def /m2 -1 0.11 thickcalc4A EF mul y2 sub 0.01 div div def

/ah 175 def /al 95 def 0 1 10 { pop /am ah al add 2 div def /x1 rad dup am cos mul add def /y1 rad am sin mul EF mul def /m1 am tan EF div def m1 x1 y1 m2 x2 y2 intercept /y0 exd /x0 exd x1 x0 sub dup mul y1 y0 sub dup mul add dup /r1 exd x2 x0 sub dup mul y2 y0 sub dup mul add dup /r2 exd le {/al am def} {/ah am def} ifelse } for /a1 y1 y0 sub x1 x0 sub atan def /a2 y2 y0 sub x2 x0 sub atan def /r0 x2 x0 sub dup mul y2 y0 sub dup mul add sqrt def /arad y1 x1 rad sub atan def

%lower le /ao2 135 def /a3 225 def

%build array [1.00 0.00 0.95 -0.05 0.11 {dup dup cambercalc4A ex thickcalc4A add } for

/f exd /fa a2 a1 f fade def

fa cos r0 mul x0 add dup cambercalc4A fa sin r0 mul y0 add EF div add

} for

arad 180 arad sub 4 div 360 arad sub 1 sub { dup cos rad mul rad add dup cambercalc4A 3 -1 roll sin rad mul add }for

/f exd /fa a2 a1 f fade def

fa cos r0 mul x0 add dup cambercalc4A fa sin r0 mul y0 add EF div sub

} for

end

AirfoilDict begin %%%% begin first page %%%% save

/AF 0009 NACA4 def /copies 1 def /labelfoil (NACA-0009) def /skin 0 def /leading 0 def /trailing 0 def /draftmode 0 0 eq def /ribcount 1 def /rootthickness 1.0 def /tipthickness 1.0 def /rootAOA 0 def /tipAOA 0 def /rootAOApoint 0 def /tipAOApoint 0 def /rootchord 10.0 def /tipchord 10.0 def /rootfrontholepoint 0 def /tipfrontholepoint 0 def /rootholeelevation 0 def /tipholeelevation 0 def /rearholeoffset 0 def /holesize 0 def /elliptical 0 def

restore end

%%Trailer

Bob Kaplow NAR # 18L TRA # "Impeach the TRA BoD" >>> To reply, remove the TRABoD!

Shape 'em like this and it too may go 10975'

~1½" bevel on all external surfaces and sanded to a (pretty) hard point.

But that's for this bad girl which needed a spank> What's considered the optimal shape of the leading and trailing edges of > rocket fins?

Very nice Gene, is the ~1 1/2" bevel the length of the exposed surface or from the edge of the fin back (or forward)? From your photo's it looks like it's the exposed surface.

Did you use Eric Vandergriff's technique for fillets? I just tried the old popsicle stick method on one of my model rockets, so far so good. I did use his Elmers Fill & Finish method on a kit I got for Christmas, seems to be a nice way to fill the little imperfections. Much easier to sand out too.

Can't wait to get to the point I can deal out those spankin's ;-)

Ken Baldwin CAR S825 replace nospam with telus to reply