Large Radius Fillets and Dynamic Stability

[WFWalby]

Was wondering what effect large radius fillets have on dynamic stability? Especially in smaller diameter (38&54mm) minimum diameter rockets. I keep seeing more and more MD rockets with large diameter fillets being used in place of T-to-T reinforcement.
Thanks
William

 

[charrington]

 

[watheyak]

Tip to tip creates its own set of problems and has fallen out of favor.

How large a fillet are we talking? Most I've seen or made on the sizes of rocket you mention are less than a .5" radius. I'm guessing there's a point where the fillet radius creeps up the semi-span so far they reduce the fin's effectiveness.

 

[WFWalby]

Tip to tip creates its own set of problems and has fallen out of favor.

How large a fillet are we talking? Most I've seen or made on the sizes of rocket you mention are less than a .5" radius. I'm guessing there's a point where the fillet radius creeps up the semi-span so far they reduce the fin's effectiveness.

I remember seeing someone use a 0.75" radius fillet on a 54mm MD and that's what got me thinking about this. I'm working on a 38MD to set some PR's for velocity and altitude and want to use large fillets instead of T-to-T. But I want to make sure the filets aren't affecting stability in any significant way.
William

 

[robopup]

"Large" (depending on your definition) fillets aren't a requirement for non tip to tip either. I think good bonding procedure (sanded to a water break, cleaned, and no bubbles in your fillet) is most important.

 

[tfish]

Tip to tip creates its own set of problems and has fallen out of favor.

I'm always the last to find things out ..what's the problems?..with tip to tip

I'm getting ready to do 1/3 - 2/3 on an almost min dia P rocket/flight.

Tony

 

[watheyak]

I'm always the last to find things out ..what's the problems?..with tip to tip

I'm getting ready to do 1/3 - 2/3 on an almost min dia P rocket/flight.

Tony

In my experience, with high mach flights, it's hard to keep the leading edge protected well enough to keep the T2T from peeling back.

Just some fun examples here. First is mine from a few years back, and Robert DeHate's O motor at our last club launch.




And while it's possible to do this with high temp epoxy a la Jim Jarvis, it's quite a bit of work to do neatly. I also started putting the T2T in a recessed pocket machined into the face of the fin. Works great, but again, huge amount of work and also expensive.

A few years ago I started doing fillets with a quality "toughened" epoxy (3M DP-460 NS) sans T2T and had good success. But it didn't have the "cool factor" that the Jarvisonian or recessed T2T did. So I persevered. But...

When all the college teams started showing up at our launch with their spicy (in the college team parlance) MD projects with just fillets, it really proved to me that good fillets are more than adequate and T2T is unnecessary.

That's all the long-wonded rationale for my statement, anyway.

Now on a minimum diameter P, that's a different story and I'm out of my experience range. Robert's was a large O, and to me the damage to that rocket was totally acceptable and even something to be proud of. And you're no stranger to flights like this.

 

[tfish]

In my experience, with high mach flights, it's hard to keep the leading edge protected well enough to keep the T2T from peeling back.

Scott...I agree with ..you and Robert are having problems.

I'm curious what was your and Roberts Mach numbers?

I replied to a text or post about Roberts flight recently..

I agree its hard to find buy epoxy 'good enough' to prevent peeling back like that. There's a real danger zone on leading edges of fins. Even if we found great epoxy we don't have the skills and equipment to make our tip to tip anywhere nears as good as the G10 plate we normally use for the actual fins.
I keep my layup was back from that danger zone. Do we really need to fins thicker or stronger out that far on the fin? It's just dead weight, added thickness and a place that can come loose. When these tip to tip layup peel back they sure eat into to altitude and speed predictions!

I do 1/3 - 2/3 layups. I keep my Raka boat epoxy ..chopped glass fin filets and and hand laid (not vacuum bagged) layups was back of of that danger zone..

I've only been about mach 3 with the above methods..

Tony

 

[charrington]

Check out my threads from blue variable. Pretty sure it’s the highest and fastest a single stage commercial 38mm has flown (mach 3.1, 31,000ish feet). Ive had two attempts at the K record and the fins didn't fall off during flight either time. 1/16th inch High modulus carbon plate fins and hysol 120hp or hysol 9460 for the fillets. My 38mm madcow go devil also does not utilize tip to tip just hysol 120hp for fillets and it hit mach 2.2.

 

[lr64]

Airfoils can be quite thick and still work fine. I would't just assume that fillets degrade stability. Particularly if they are thin at the trailing edge. However, I don't actually know this.

---------
Here's an off the wall idea:
What happens if you stick the glass down with plaster for 1/4 or 1/2 inch near the leading edge? There are high temperature varieties, refractory cement, etc. Or high temp RTV? Maybe a narrow tape wrapped over the leading edge?

 

[watheyak]

Scott...I agree with ..you and Robert are having problems.

Well to be fair, that was many years ago and I learned a lot from that flight. And haven't had T2T problems since.

I think my point was lost in my rambling.

 

[JimJarvis50]

I'm considering a high speed flight with fillets only (no tip-to-tip). My belief is that the value of the tip-to-tip is to help keep the underlying tube round (via the lamination across the tube) and to supplement the thickness of the fins so that the final fin is adequate from a flutter perspective. There may be some additional value in the 1/3, 2/3 approach - I don't really know. So, I think if the fins are sufficient on their own and if the tube will stay round, then a nice fillet should be sufficient, particularly if delamination of the tip-to-tip can be avoided.

Just for fun, I made a test piece consisting of a Wildman airframe tube and a couple of G10 "fins" with modest fillets. I first inserted a coupler tube and a couple of bulkheads into the airframe tube to help keep it round. I then jumped on the piece multiple times, but I was not able to break it. Then, I took out the rings and the coupler tube, and the fins snapped off rather easily. The tube failed on one side and the fillet on the other. Gotta keep the tube round.

Jim

 

[lr64]

I think 1/4, 1/2 probably works a little better than 1/3, 2/3. Loads drop off fast as you move from the root out towards the tip.

 

[MClark]

Popsicle stick is perfect

 

[kramer714]

So, Couple of things to un-pack here from all the posts, in no particular order...

Peeling Skins Tip to Tip.

• When I see pictures of peeled T2T while the G-10 stays in place, there are two or three problems. Low temperature laminating resins are really bad adhesives. I know you are laying up plies but you really need to at worst add a layer of a TOUGHENED epoxy, something with a really high peel strength, at best use a toughened epoxy as the resin.
• Prep, really need to prep the G-10 AND then apply thin layer of adhesive to the surface BEFORE putting anything else on, not cured - but co-cured with the skins. you should never have an adhesive failure it should always be cohesive (adhesive on both sides of the bond after failure, or the substrate failing not the bond).
• Post Cure, almost all of these resins need a post cure if you want thee full properties, really... A hard resin is just that, hard, it isnt necessarily cured, or fully cured. Check the details of the data sheets.

Fillets and Toughness

• Many of the fin failures i have seen are due to folks using the wrong materials for fillets. Again a laminating resin, even if it is filled with fibers isn't necessarily a tough material or for that matter a good adhesive. To make maters worse many of these materials shrink dung cure, causing them to be in tension without any load. A tough material has some allowable elongation and a high peel strength a little bit of give before failure.

Tip to tip vs fillet vs doubler

• I have had this happen more than once where someone will ask me about materials for a fillet. I ask them what are they trying to achieve with the fillet? Same question with Tip to Tip. Surprised by how many people are not really sure why they are doing it. Are you stiffening the joint, stiffening the fin, adding strength or is it an aerodynamic faring?
• Tip to Tip vs a doubler. At the day job (aerospace composites) if we need to relive the stress or stiffen up a T joint, usually it is done with a strip that runs over the joint and just a little bit along the two surfaces. Tip to Tip is adding al ot of material that actually doesn't do anything for you, plus makes applying the reinforcement much harder. There are woven material made just for for this, look up boat tape. A 2 inch wide strip, one inch on the bodytube and one inch on the fin stiffens the joint and adds strength.

Mike (ask why before asking how...) K

 

[Neutronium95]

I'm not a fan of tip to tip because I'm lazy. It's a lot of work to get really good results, and from what I can tell people do it for two reasons, flutter resistance and attachment reinforcement.

Reason one is to improve fin stiffness to prevent flutter. I personally would prefer to just choose a sufficiently thick and stiff fin material to start with. And if I feel like my fin material does need reinforcement, it is simpler to do a flat layup on a plate before cutting and beveling the fins. I also think that a proper press cured composite plate like G10 fiberglass or Dragon Plate will just be of significantly higher quality than I can lay up with the tools and materials I have access to at home.

The other reason is to improve the fin attachment to the tube. I've found that decently sized fillets with a proper structural resin provides adequate fin attachment strength. I've never had a fin come off in flight.

 

[lr64]

I'm not a fan of tip to tip because I'm lazy. It's a lot of work to get really good results, and from what I can tell people do it for two reasons, flutter resistance and attachment reinforcement.

Reason one is to improve fin stiffness to prevent flutter. I personally would prefer to just choose a sufficiently thick and stiff fin material to start with. And if I feel like my fin material does need reinforcement, it is simpler to do a flat layup on a plate before cutting and beveling the fins. I also think that a proper press cured composite plate like G10 fiberglass or Dragon Plate will just be of significantly higher quality than I can lay up with the tools and materials I have access to at home.

The other reason is to improve the fin attachment to the tube. I've found that decently sized fillets with a proper structural resin provides adequate fin attachment strength. I've never had a fin come off in flight.

With tip to tip, only the material that's fairly close to the base of the fin helps against flutter. Material out at the tip adds mass in the worst possible place and doesn't make the tip any stiffer in bending. For a subsonic rocket, you can add the thickness with an airfoil shape, while possibly even reducing the drag compared to a flat plate. Even a foil that's 10 percent thick won't add to the drag.if you're going to glass the surface, you can make the fin out of foam and end up with something lighter, stiffer, and stronger than thin G10. Another thing to improve stiffness is to have some of the fibers at +/-45° to the span, for torsional stiffness. Carbon is probably a better fabric on wood fins, because glass won't get anywhere near its full strength before the wood starts to fail.

Most of the increased strength of something like G10 is because the proportion of fiber to resin is higher. If you had the same number of layers of glass as G10, it might be heavier, but it would be about as strong.In bending, it would probably be stiffer and stronger because of the increased thickness.

If your fillet has glass tape over it, it ought to be much stronger. If you put that on within a couple of days, laminating resin ought to stick fine. Or so I've read.

 

[nickt.mrc]

I'm with Neutronium95 on this, although my reason is not because I'm lazy, but because, on the one occasion I tried, the result was a mess!
My solution is to firmly attach the fins to the motor tube and centering rings and to over specify the rings. This moves the forces away from the body tube, onto the strongest component of your rocket. Normal fillets will then suffice.
If you use 3k carbon flat plate for the fins, you already have fibres laid out at 3 angles.

 

[drewnickel]

I'm with Neutronium95 on this, although my reason is not because I'm lazy, but because, on the one occasion I tried, the result was a mess!
My solution is to firmly attach the fins to the motor tube and centering rings and to over specify the rings. This moves the forces away from the body tube, onto the strongest component of your rocket. Normal fillets will then suffice.
If you use 3k carbon flat plate for the fins, you already have fibres laid out at 3 angles.

I think the discussion of a T2T layup is generally geared towards surface mount fins. With TTW fins, you have two points of contact on the rocket which can distribute the load, vs a single point on a surface mounted fin.

 

[kramer714]

I'm with Neutronium95 on this, although my reason is not because I'm lazy, but because, on the one occasion I tried, the result was a mess!
My solution is to firmly attach the fins to the motor tube and centering rings and to over specify the rings. This moves the forces away from the body tube, onto the strongest component of your rocket. Normal fillets will then suffice.
If you use 3k carbon flat plate for the fins, you already have fibres laid out at 3 angles.

Dont assume it is in three directions. Typical plate is only 0/90 (G-10 is 0/90). If you want three directions you need to get 'quasi-isotropic' (sometimes called Q-I). Quasi isotropic has plies in 0/90 and in +/-45.

Having said that, couple of things,

• for a fin do you want 4 angles? Is the +/- 45 helpful? A 0/90 will be stiffer in the zero direction than a QI plate.
• 0/90 plates are a lot cheaper since there is very little material waste for the manufacturer.

Balanced Plate.

• Composite plates need to be balanced when they are made to prevent warping during manufacturing and also with temperature changes. . so for an 8 ply QI laminate it would be.
  • ply 1 +45
  • ply 2 -45
  • ply 3 0
  • ply 4 90
  • ply 5 90
  • ply 6 0
  • ply 7 -45
  • ply 8 +45
• Or it could be
  • ply 1 90
  • ply 2 0
  • ply 3 +45
  • ply 4 -45
  • ply 5 -45
  • ply 6 +45
  • ply 7 0
  • ply 8 90
• The point is in a thin laminate the outer fibers add more to the stiffness than the inner ones. Ideally the outer fibers are running spanwise. If you have a QI panel that has the +/-45 as the outer layer, to really optimize it you want to cut it on a bias.

Just to make matters worse, woven material has different properties in the warp direction (long direction on the roll) and weft direction. On G-10 it is pretty negligible, but with woven carbon panels not so much.

Attached is a quick ply layout I did for materiel optimization. Gives you an idea about the waste, can be a little better with multiple plies nested together but still alot more expensive than just slitting a roll of prepreg to the exact width.




Mike (composites guy) K

 

[drewnickel]

Attached is a quick ply layout I did for materiel optimization. Gives you an idea about the waste, can be a little better with multiple plies nested together but still alot more expensive than just slitting a roll of prepreg to the exact width.



View attachment 681319
Mike (composites guy) K

One thing to note about the waste from a ply layout like this is, if you swapped to something like a unidirectional prepreg, you could substantially reduce the waste. Cutting the ply parallel to the fiber orientation doesn't really weaken the final product, as you are only cutting the un-cured matrix, so you can piece together off-cuts of unidirectional prepreg laying the plies parallel to each other for each layer. The stipulation with using off-cuts is the fibers still need to maintain a continuous un-broken span for the length/width of the panel/part you are making, so shorter off-cuts could be used for corners on the +/- 45 deg layers.

Back to OP's original question, which relates to the impact of fillets on stability/etc. rather than the runaway T2T and structure discussion we're having, fillets that are sized for aerodynamics rather than purely structural characteristics can reduce interference drag/losses between the fin root and airframe. With properly sized fillets you can at least in theory get the same lift performance for a fin with a shorter span as more of the fin is generating useful lift and you don't have the additive effects of the body and fin boundary layers combining in the corner between the fin and airframe. Exactly how much of an impact it will have will depend largely on your particular rocket (fin profile, rocket diameter and length, flight regime, etc.), but I've seen references for the optimum being somewhere between 4-6% of the fin root chord length for subsonic flight or 2-5% of the airframe diameter for supersonic flight. On subsonic flights a more gradual transition is beneficial as it delays flow separation, while supersonic flights also have to compete with added wave drag from large radius fillets.

Edit: adding this diagram for a simple explanation of combining the pre-preg sheets for less waste.

 

[WFWalby]

I'm considering a high speed flight with fillets only (no tip-to-tip). My belief is that the value of the tip-to-tip is to help keep the underlying tube round (via the lamination across the tube) and to supplement the thickness of the fins so that the final fin is adequate from a flutter perspective. There may be some additional value in the 1/3, 2/3 approach - I don't really know. So, I think if the fins are sufficient on their own and if the tube will stay round, then a nice fillet should be sufficient, particularly if delamination of the tip-to-tip can be avoided.

Just for fun, I made a test piece consisting of a Wildman airframe tube and a couple of G10 "fins" with modest fillets. I first inserted a coupler tube and a couple of bulkheads into the airframe tube to help keep it round. I then jumped on the piece multiple times, but I was not able to break it. Then, I took out the rings and the coupler tube, and the fins snapped off rather easily. The tube failed on one side and the fillet on the other. Gotta keep the tube round.

Jim

View attachment 681089

Since I'm working with minimum diameter airframes, I would think that the motor case would provide enough additional support to keep the airframe round. So in your example as long as the fins are stiff enough, then the fillets would be sufficient on their own. Given the proper adhesive and size. Yes/no/maybe??

 

[kramer714]

Since I'm working with minimum diameter airframes, I would think that the motor case would provide enough additional support to keep the airframe round. So in your example as long as the fins are stiff enough, then the fillets would be sufficient on their own. Given the proper adhesive and size. Yes/no/maybe??

yes.

 

[kjhambrick]

Maybe Off Topic, but Mike T opened the layout geometry can-o-worms so here goes a Q that's prevented me from cutting my purty plates into fins ...

I bought some relatively cheap 1.5mm, 2.0mm and 2.5mm plain weave carbon fiber plate:

Amazon > Kalolary Carbon Fiber Plate 300 X 300 X 1.5 MM, Carbon Fiber Board 3K Full Carbon Fiber Sheets Plate Plain Weave and Glossy Finish - Available in 0.5mm 1mm 1.5mm 2mm 2.5mm 3mm 4mm

When I cut out wooden fins, I learned as a kid to align the leading edge with the grain.

What I am after with my new carbon fiber fins is mostly resistance to flutter ...

Which way does the 'grain' run in plain weave carbon fiber ?


i.e. which way should I align the leading edge of my fins ?

Left-Right ? Top-Bottom ? Diagonal ?

Or am I asking the wrong Q because maybe I should have shelled out more bucks for Dragon Plate Dragon Plage ?

Thanks !

-- kjh

 

[Rschub]

The stiffest direction on your image is 45 degrees to the squares, if that helps.

 

[manixFan]

Was wondering what effect large radius fillets have on dynamic stability? Especially in smaller diameter (38&54mm) minimum diameter rockets. I keep seeing more and more MD rockets with large diameter fillets being used in place of T-to-T reinforcement.
Thanks
William

My son is an aerospace engineer and I asked him about fillets several years back. I should ping him again but he's pretty busy right now so I'll have to go from memory. I recall he said that unless the fillets started having a dramatic effect on the overall diameter of the rocket or reduced the surface area of the fins considerably (by basically turning the root into an extension of the body tube) they would have minimal effect. I think he also said that fillets were not heavily modeled in 'real' rockets because they were not really used as the fins were bolted, welded, or otherwise attached in a fashion that did not require a fillet, so there wasn't a lot of worry about them.

And referencing @drewnickel 's post above, I've read the same stats about fillet radius, but I can't find any original source material for the numbers. The first time I saw those numbers they were on a PowerPoint presentation, and I've seen them quoted dozens of times. It would be great to find the original source for those numbers and why they were found to be optimal.

EDIT: I went back to the PowerPoint and found the original references – they are later in this thread – I've posted links:
https://www.rocketryforum.com/threads/large-radius-fillets-and-dynamic-stability.189750/post-2671567

Tony

Here's a good example of a modern suborbital rocket that shows there are no fillets of any kind along the fin root:


Source:
https://www.esa.int/ESA_Multimedia/Images/2024/10/SpaceForest_Perun_rocket

 

[sharkbait]

My son is an aerospace engineer and I asked him about fillets several years back. I should ping him again but he's pretty busy right now so I'll have to go from memory. I recall he said that unless the fillets started having a dramatic effect on the overall diameter of the rocket or reduced the surface area of the fins considerably (by basically turning the root into an extension of the body tube) they would have minimal effect. I think he also said that fillets were not heavily modeled in 'real' rockets because they were not really used as the fins were bolted, welded, or otherwise attached in a fashion that did not require a fillet, so there wasn't a lot of worry about them.

And referencing @drewnickel 's post above, I've read the same stats about fillet radius, but I can't find any original source material for the numbers. The first time I saw those numbers they were on a PowerPoint presentation, and I've seen them quoted dozens of times. It would be great to find the original source for those numbers and why they were found to be optimal.


Tony

Here's a good example of a modern suborbital rocket that shows there are no fillets of any kind along the fin root:
View attachment 681339

Source:
https://www.esa.int/ESA_Multimedia/Images/2024/10/SpaceForest_Perun_rocket

Tony,
Could your son score me one of those launchers

 

[tfish]

Tony,
Could your son score me one of those launchers

Launcher or rocket?

https://www.northstateauctions.com/...fa7CONEDT3lrcXV-tg_aem_lvnmYs_G9N1R3xGdfHAUsg

The other Tony

 

[manixFan]

Tony,
Could your son score me one of those launchers

He works for SpaceX, so the best he could do would be a Mechazilla, but it doubles as a landing zone!


Tony

 

[kramer714]

um

The stiffest direction on your image is 45 degrees to the squares, if that helps.

Um not sure you are right about that..

In the fiber direction, (left and right in your picture) assuming you are a plate with all the plies in the same direction, and assuming 60% fiber by volume,

Exx direction 9.9 MSI tensile modulus
Eyy direction 9.9 MSI tensile modulus
Exy direction 2.5 MSI tensile modulus
Eyx Direction 2.5 MSI Tensile modulus

In pure bending / thin skins (like a fin) the tensile modulus in the skin drives the bending stiffness,



Bigger E smaller deflection

You want the X (or Y) direction spanwise.