Drag effects of fin fillets

The Rocketry Forum

Help Support The Rocketry Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

enderw88

New to the game
Joined
Feb 24, 2014
Messages
230
Reaction score
3
I'll post this here since there is not an Aero section. Over in the techniques section we were discussing whether or not it was worthwhile to put fillets on the interior fin-BT joint. The general consensus was that it is worthwhile. There was some disagreement over my claim that external fillets reduce drag overall by reducing interference drag at the fin-BT junction. Rather than moving off topic on that thread the question here is NOT whether external fillets are necessary from a structural point of view, that is undisputed. But whether or not they increase the amount of drag over not having them at all.

Like any hard question there appears to be little published data on this. Using Fluid Dynamic Drag by Hoerner and some high school algebra and physics I explain my claim. The math is difficult to read with the formatting available on this forum (no WebMath or MathML, yuck)

Let me know if you think I erred somewhere or made an assumption you think is cheesy. It's happened before:facepalm:.

View attachment InterferenceDrag.pdf
 
Smooth fillets that are properly sized reduce drag. I didn't think that would be up for debate, I've always believed it to be a well known fact. Less turbulence, less surface area, slightly higher frontal area but offset by the other two. If you leave the joint at a sharp 90 degree angle, there is turbulent flow in the "crease".
 
Your going to get a lot of differing opinions on this. That said I could not pull up your file--likely a problem at my end. Anyway, I'm neutral on this point, from a certain point of view. There is drag produced at a hard fin joint. A small fillet will smooth the airflow over that area. In that case, surface drag is less than the interference drag of the hard joint--so that's a win. BUT, as lite as you make them , fillets add weight and you have to push them up against gravity---soooo now you have a tradeoff. If your subsonic this might be a win. If your flight is supersonic--and you get there fast---read: accelerate fast with a gob of power and push through the high drag envelope---fillets are a waste of time. From a certain point of view! Once you decelerate and coast, the drag goes up--oh yea, that whole altitude and air density thing starts to come into play now. At least in my opinion, you should have small fillets--just enough to clean up the mating surfaces. At that point--if your going for performance--review EVERYTHING in the rocket and shave weight. Everything means everything--seamless joints--minimal paint but smooth, the right cone shape for the flight profile, bt length. You can't just zero in on one thing. Yea, it's a PIA , but it's what we do !!
 
Very interesting question, and your math is good.

But since everything depends on the precise values of the c_Ds involved (e. g. with a higher frontal c_D you want smaller fillets, but will the effect on the fin interference drag still be the same?), I don't think there really is a definitive "yes" or "no".

Oliver
 
Very interesting question, and your math is good.

But since everything depends on the precise values of the c_Ds involved (e. g. with a higher frontal c_D you want smaller fillets, but will the effect on the fin interference drag still be the same?), I don't think there really is a definitive "yes" or "no".

Oliver

Yeah, getting precise C_D is Very Hard and the balance is very sensitive to that. When i have more time I will look a bit more deeply into how it changes over flight.

I was surprised there was much debate over it as well, but I have seen some pretty large fillets that are clearly detracting from flight performance even if they are bettering structural performance.
 
Last edited:
I can see where fin fillets can create tiny tiny bit of drag. But base drag is where it kills performance.
 
Thanks for starting the thread - this is a good topic of discussion.

Your math certainly looks right, with the precise C_D being a key element of the whole thing. Also one must take into account the likelihood that the fillets themselves, having not been created on your average rocket with the most precise of methods, are sub-optimal.

There are two points to note, however:

(1) Hoerner seems to have done most of his analysis with fairly high t/c ratios relative to what we usually see in rocketry, and seems primarily concerned with junctions of foiled struts. If you look at chapter 8, section 4, however, he addresses the intersection of wings, struts, and tail surfaces with a fuselage through the simplified case of a junction with a flat plate. In the accompanying discussion, he states that "The experimental data pictured in figure 23 even suggests that the interference drag will be negative, at thickness rations below t/c = 8% [approx]." I have attempted to attach Figure 23 below.

HoernerDragChap8Fig23.png

(2) Even in the case of figure 29, the optimum fillet radius is quite low relative to the size of the fillets I regularly see on rockets.

Given the low t/c ratios of our rocket fins, I would argue that exterior fillets should be kept minimal for aerodynamic reasons, as well as for weight reasons, unless there is a structural requirement for them such as in the case of surface-mounted fins.

It is possible I have misinterpreted this - it has happened before - so I welcome correction. Our real fluid gurus are out of the office at the moment, so I will consult with them later and update here as best I can.
 
The Late Dr. John Fox once explained to us the benefits of fillets on aircraft wings.
It was explained in a simple manner as most of us in the RC Club didn't attend
college.
Dr. Fox was brilliant. He once did R&D on the Black Bird during it's development.

Fillets reduce drag where 2 flying surfaces meet at 90 degree angles, simply put.

JP
 
Last edited:
The Late Dr. John Fox once explained to us the benefits of fillets on aircraft wings.
It was explained in a simple manner as most of us in the RC Club didn't attend
college.
Dr. Fox was brilliant. He once did R&D on the Black Hawk during it's development.

Fillets reduce drag where 2 flying surfaces meet at 90 degree angles, simply put.

JP

That is their intention. But how big? At what speeds (Reynold's numbers) and configurations (t/c ratio, local boundary layer effects, etc)? I have some experience which has shown situations where this is not the case, or at least where the benefits of the fillets are outweighed by their costs. I do not think the words "simply put" apply to the problem.
 
That is their intention. But how big? At what speeds (Reynold's numbers) and configurations (t/c ratio, local boundary layer effects, etc)? I have some experience which has shown situations where this is not the case, or at least where the benefits of the fillets are outweighed by their costs. I do not think the words "simply put" apply to the problem.

I agree. There was enough hand waving in my simplified look that if I were absolutely going for max performance I would want to look deeply at the the specific design. If I am designing a sport rocket I would go for a 5% radius fillet using lightened epoxy and make it pretty.
 
IIRC, GH Stine addressed this issue in HoMR, and even had a ROT to go with it. I am not close to a copy, so someone will have to look it up to confirm.

Greg
 
And the Blackbird didn't have fillets where the engines met the wings.

You're right, there's a speed factor to consider. Sorry I can't do the math,
I'll bow out gracefully and just observe.

Usually I run one pass of epoxy then use a filler to finish the shape with a pass
of thin CA on top. I go for effects more so than strength for the external fillets.

JP
 
With turbulence around the body of the rocket and the very light weight of fillets, I doubt it would be possible to measure any flight performance effect due to fillets.

Math is always highly dependant on initial assumptions.

To meaningfully answer this question I believe wind tunnel testing would be required.
 
Thanks. I never noticed that the inner nacelle-wing junction didn't have a fillet. I wonder why? Kelly Johnson never missed a trick, there must be a good reason.

There are a lot of factors that go into design. While they wanted good aerodynamics, there was also considerations related to thermodynamics, mass management, maintenance, manufacturing, and radar cross section.

A-12-model.jpg

Greg
 
It seems like the drag effects of differing sizes of fillets would be the kind of thing that you would worry about after you had exhausted every other possible performance factor. It reminds me of the backpackers who trim the margins off of their maps to save weight. I suppose if you add together enough negligible effects, you can get a non-negligible effect. But it seems like in this case, the effect is not only negligible but also ambiguous.
 
interesting discussion! Dr Gerald Gregorek discusses this in the Estes tech report TR-11. He doesn't apply any hard and fast numbers to the drag reduction...it would be interesting to break it down like he did the other parts of total drag. as mentioned above, base drag is your biggie, boat-tailing helps that.
 
"Radius of fillet should be 4-8 % of the length of the root"


I have supplied this quote in half a dozen discussions on this topic .

Screen Shot 2014-11-20 at 12.54.40 PM.png
 
Last edited:
The radius of my fillets matches the radius created by the end of a popcicle stick held at a convenient angle.
 
Build lite, Keep the paint lite. Paint the legs before you attach the descent stage. Keep it lite, add an extra rubber band to each leg, build it lite, keep a tight fit on the leg fairings, make sure it's lite.....

It is ridiculous to think that a small range of numbers will fit 100% of cases. That number may be true in sub sonic flights but will probably not be true in super sonic flight.

The issue of fillet size is a very complicated advanced fluid mechanics problem, and most likely can't be solved analytically.
 
I just finished reading Skunk Works by Ben Rich and according to that book, the chines and fillets on the Blackbird were to reduce radar cross-section, not drag.

I had always taken for granted that fillets reduced interference drag, but this thread is causing me to think a little more critically about it.
 
I wish he had provided some sort of explanation for that number and how he arrived at it.


Most likely this:
According to (HOENER, 1965),
“Interference drag [also] originates at points where wings, struts or tail surfaces join or
penetrate the fuselage, or where different parts of the tail assembly join each other.” He
carries on saying, “interference drag can appreciably be reduced by fairings properly installed
in the corners”. In fact, the interference drag can be reduced by up to 10% (or more) over an
un-faired junction, depending on the fairing radius, which has been found to be at an optimum
when it is small, at between 4-8% of the chord length.
 
Most likely this:
According to (HOENER, 1965),
“Interference drag [also] originates at points where wings, struts or tail surfaces join or
penetrate the fuselage, or where different parts of the tail assembly join each other.” He
carries on saying, “interference drag can appreciably be reduced by fairings properly installed
in the corners”. In fact, the interference drag can be reduced by up to 10% (or more) over an
un-faired junction, depending on the fairing radius, which has been found to be at an optimum
when it is small, at between 4-8% of the chord length.

Missed that.

As I've stated above, I think that might be a misreading of Hoerner, given his additional treatment of Wing-fuselage interference with low t/c sections in part 4 Of Chapter 8 of The drag book.

I would love to have somebody address that graph, which seems to me more applicable to typical rocket sections than most of Hoerner's work on the subject, which generally is concerned with sections that have a t/c in the neighborhood of 0.3. Surely it is not outside the realm of possibility that the aerodynamic need for fillets is somewhat reduced when the thickness of the sections is vanishingly small...

Sent from my iPhone using Rocketry Forum
 
Last edited:
Part of our problem is that we would need someone to run some aero trials in a wind tunnel with different rockets, with different kinds of fillets, and at different velocities to create a matrix of Cd's so that at least we had some objective data that could be referenced. Until then, we have the aforementioned ROT and it's handmaiden TLAR.

Unless someone invents OpenFillet ... :wink:

Greg
 
Back
Top