My Wildly Overcomplicated Level 1 Build

[James Keller]

Hey all,
As a disclaimer this is not my first HPR rocket build. I have built 3 Level 2 rockets with my school that I just graduated from. However I couldn't get certified with the program. I also know this is way over complicated and is way over the top for a level 1 certification rocket, but in my opinion that's what makes it fun. Now that we got that out of the way here are some goals for this rocket.
Goals

• Break personal altitude record and speed record (15k ft and Mach 1.2)
• Have a camera on board for at least the test flights and cert flights
• Minimum Diameter Build
• Certify Level 1 and 2

Basic Plan

• 2 "Forms" of the rocket will be made
• Form one will include 2 CF airframe sections a fiberglass coupler and a fiberglass nosecone. This is the version I will be using to certify with. The coupler will house the camera. While the nosecone houses the electronics and main chute.
  • Motor- This will be flown on a initially on a Aerotech I140 SU motor for L1 Cert, which sims to just over 6K ft and Mach .83. Then eventually on an L2 motor for Cert.
• Form two will get rid of the upper airframe and will be flown without the camera. That way I can squeeze the most altitude and velocity out of it.
  • Motor- This will eventually be flown on a Loki K627 which sims to 25k ft and Mach 2.75 however I will fly it on a bunch of smaller motors before this.
• Structure
  
  • 3 wrap .05 CF
  • 3 Fins .0625 G10 with 1 layer of CF then 1 Layer tip to tip
  • Fiberglass NC 5:1 VK using Robopups technique with a total length of 12in with metal tip
• Electronics
  
  • Eggtimer Quasar in NC for both telemetry and chute ejection
  • 300mah 2s lipo (I know Cris recommends a 500+mah battery but I will be using a switch and turning it on right before the rocket fires)
  • Featherweight Magnetic Switch
  • Runcam Thumb as a camera
• Recovery
  
  • 24in main TF chute, and a 9in chute for the drogue(subject to change)
  • Main will be stored in nosecone and drogue will be stored in the airframe
  • 550lbs Kevlar
• Sim and CAD

Open Rocket Form 1 (I140)

RasAero Form 2 (K627)


Current Progress
Currently I have an airframe curing and am pulling it off the mandrel tomorrow. I also have carbon fiber laid up on g10 tomorrow I will cut them out on my CNC router. After that I think I will move onto the nosecone as I am pretty sure I will need a few tries at that. I also need to make a new fin beveling jig. I might instead try to use my CNC router to bevel the fins. Any input on that would be helpful.

 

[OverTheTop]

This is how I bevel fins:
https://www.rocketryforum.com/threa...ying-and-love-3d-printing.142435/post-1726196

Aldi router table with Dremel attached. Vertical fence is actually not quite vertical.

Remember to not make the edge too thin and fragile.

 

[robopup]

Good luck!

A couple quick thoughts:

• Having done no analysis, those fins look dinky tiny too small to me. In particular the root edge, to get enough area for bonding. Having said that, I would love to see it work; just understand that you're pushing things.
• 9 in drogue sounds very large for a 38mm rocket, especially coming down from the kind of altitudes you're gunning for. I usually fly drogueless or with a very small streamer on my 38's and come in at 50-75 ft/s.

 

[James Keller]

This is how I bevel fins:
https://www.rocketryforum.com/threa...ying-and-love-3d-printing.142435/post-1726196
View attachment 589418
Aldi router table with Dremel attached. Vertical fence is actually not quite vertical.

Remember to not make the edge too thin and fragile.

Hey Overthetop, thanks for the reply. When you say not to make the edge to thin, how thin is to thin? I was thinking 5-10 degree bevel, I think I remember reading that somewhere but it's very possible I'm making that up.

Good luck!

A couple quick thoughts:

• Having done no analysis, those fins look dinky tiny too small to me. In particular the root edge, to get enough area for bonding. Having said that, I would love to see it work; just understand that you're pushing things.
• 9 in drogue sounds very large for a 38mm rocket, especially coming down from the kind of altitudes you're gunning for. I usually fly drogueless or with a very small streamer on my 38's and come in at 50-75 ft/s.

Hey Robopup, I used FinSim to check for flutter and even using just g10 it said it wouldn't flutter. But I think your more talking about landings? It will have 1 layer of t2t if that makes a difference. Currently the root cord is 2in that can be changed easily. How do you do the analysis on your fins? Also thanks for the heads up on the drouge. I think I'll still fly it for my Cert flight just for peace of mind. Also big thanks for your build threads, its been a huge help.

 

[OverTheTop]

Hey Overthetop, thanks for the reply. When you say not to make the edge to thin, how thin is to thin?

You are in the right ballpark with the angles, but I was referring to the final edge width. If you make the fin edges really thin they are too easily damaged. I leave them no thinner than about 0.5-1mm, to make them a little more robust. Considering your fins are only about 1.6mm thick total I would guess you would need to go a little finer than I did. Try to find the balance between fineness and strength that you are happy with. It is your rocket so you get to choose .

The good news with your design is that they are a little forward of the aft end and less likely to be damaged.

 

[robopup]

Hey Robopup, I used FinSim to check for flutter and even using just g10 it said it wouldn't flutter. But I think your more talking about landings? It will have 1 layer of t2t if that makes a difference. Currently the root cord is 2in that can be changed easily. How do you do the analysis on your fins? Also thanks for the heads up on the drouge. I think I'll still fly it for my Cert flight just for peace of mind. Also big thanks for your build threads, its been a huge help.

I don't have a good way to nail absolute numbers for structural design on fillets beyond build and break. I think it's actually a pretty tough problem to do this on for various reasons:

• loading - I think you can get in the ball park for static loads from supersonic airfoil and wing calcs. For any dynamic loads like flutter, personally I don't have a great feel for getting ballpark. I have generally been disappointed with the various flutter calcs floating around the forum, I find they're very conservative compared to what I've actually flown.
• materials - In general, datasheets for composite materials lie like rugs. You also have to factor in that the materials on the datasheets are not what you're actually building. Your material(s) have some other properties which are unique to your processing, likely with the datasheets as upper limits on what the actual values are. I'd love to run some as build fillets through a test machine (which I recently got some bit of access to!), but I haven't yet.
• construction - at least for me, actually building fincans which approximate the design is one of the harder things on the build to do accurately.

Instead, I do a lot of comparison on my designs. I usually start out with some requirement for where I want my cp/stability margin. Then I cook up a few proposal fin designs which meet my stability requirement while varying whatever parameters I'm interested in (3 fins, 4 fins, root length, tip length, span, sweep, thickness etc). Finally I do simple, quick static and normal modes analysis on these. I don't know whether the absolute numbers are correct but I can usually say things like "these designs have the same stability, similar drag, but this one has a first natural frequency which is x% higher than the other, and displacement at the tip which is y% lower than the other, I'll pick this one." It also helps to have a large bank of fin data which I know worked, because I flew them.

Accordingly, I think you should give it a try! I think one of the tricks to this forum is to filter through the tribal knowledge and use it where it is necessary or good, but to always prefer cold hard data (like flying this rocket) over guys like me saying "those fins don't pass the eye test."

I should add - all of above assumes that your fins are sized appropriately for your stability requirements. How are you determining your cg? The first image looks reasonable but the K627 cg looks very far forward compared to how my rockets usually come out.

 

[James Keller]

You are in the right ballpark with the angles, but I was referring to the final edge width. If you make the fin edges really thin they are too easily damaged. I leave them no thinner than about 0.5-1mm, to make them a little more robust. Considering your fins are only about 1.6mm thick total I would guess you would need to go a little finer than I did. Try to find the balance between fineness and strength that you are happy with. It is your rocket so you get to choose .

The good news with your design is that they are a little forward of the aft end and less likely to be damaged.

Awesome, I'll do some testing for sure.

I don't have a good way to nail absolute numbers for structural design on fillets beyond build and break. I think it's actually a pretty tough problem to do this on for various reasons:

• loading - I think you can get in the ball park for static loads from supersonic airfoil and wing calcs. For any dynamic loads like flutter, personally I don't have a great feel for getting ballpark. I have generally been disappointed with the various flutter calcs floating around the forum, I find they're very conservative compared to what I've actually flown.
• materials - In general, datasheets for composite materials lie like rugs. You also have to factor in that the materials on the datasheets are not what you're actually building. Your material(s) have some other properties which are unique to your processing, likely with the datasheets as upper limits on what the actual values are. I'd love to run some as build fillets through a test machine (which I recently got some bit of access to!), but I haven't yet.
• construction - at least for me, actually building fincans which approximate the design is one of the harder things on the build to do accurately.

Instead, I do a lot of comparison on my designs. I usually start out with some requirement for where I want my cp/stability margin. Then I cook up a few proposal fin designs which meet my stability requirement while varying whatever parameters I'm interested in (3 fins, 4 fins, root length, tip length, span, sweep, thickness etc). Finally I do simple, quick static and normal modes analysis on these. I don't know whether the absolute numbers are correct but I can usually say things like "these designs have the same stability, similar drag, but this one has a first natural frequency which is x% higher than the other, and displacement at the tip which is y% lower than the other, I'll pick this one." It also helps to have a large bank of fin data which I know worked, because I flew them.

Accordingly, I think you should give it a try! I think one of the tricks to this forum is to filter through the tribal knowledge and use it where it is necessary or good, but to always prefer cold hard data (like flying this rocket) over guys like me saying "those fins don't pass the eye test."

I should add - all of above assumes that your fins are sized appropriately for your stability requirements. How are you determining your cg? The first image looks reasonable but the K627 cg looks very far forward compared to how my rockets usually come out.

I think I'll play around with them a little more, I am going to try to stick to a smaller fin. On my last rocket which hit Mach 1.2 the fins were tiny and they survived. Thats not to say they can't be better optimized however. The cg was taken from an open rocket measurement. I will do more detailed CAD once I get some more parts specifically the electronics built. But I always put my fins on last. Finding out your cg isn't where its supposed to be after you put you're fins on sucks. Also, I've read and it seems logical to me that using multiple different types of materials like G10 then a carbon laminate makes the frequency at which the fins flutter at smaller because there are multiple materials. I noticed however that you typically only use G10, why is that?

 

[jqavins]

Something doesn't add up.

• Minimum Diameter Build
• Motor- This will be flown on a initially on a Aerotech I140 SU...
• Fiberglass NC 5:1 VK using Robopups technique with a total length of 12in with metal tip

View attachment 589407
View attachment 589408

The I140 is a 38 mm motor, and the diameter is 1.6" in the OR picture, so that jives.
The nose cone is to have a 5:1 aspect; 5×1.6" is 8", not 12". The RAS Aero picture does, indeed, show it at 8" (as near as I can tell).

 

[James Keller]

Something doesn't add up.

The I140 is a 38 mm motor, and the diameter is 1.6" in the OR picture, so that jives.
The nose cone is to have a 5:1 aspect; 5×1.6" is 8", not 12". The RAS Aero picture does, indeed, show it at 8" (as near as I can tell).

Yeah sorry in hindsight that wasn't explained well. What I meant was it has an 8in "nosecone section" and a 4 inch body tube section to hold the parachute. So the overall length is 12 inches. That's what the second airframe is in RasAero.

 

[robopup]

Awesome, I'll do some testing for sure.

I think I'll play around with them a little more, I am going to try to stick to a smaller fin. On my last rocket which hit Mach 1.2 the fins were tiny and they survived. Thats not to say they can't be better optimized however. The cg was taken from an open rocket measurement. I will do more detailed CAD once I get some more parts specifically the electronics built. But I always put my fins on last. Finding out your cg isn't where its supposed to be after you put you're fins on sucks. Also, I've read and it seems logical to me that using multiple different types of materials like G10 then a carbon laminate makes the frequency at which the fins flutter at smaller because there are multiple materials. I noticed however that you typically only use G10, why is that?

W/r/t multiple materials and natural frequencies, this is false. Here's a quick model I threw together comparing two fin layups:

1. Carbon epoxy only. Representative properties for a 0.1 in thick plate made of plain weave carbon epoxy
2. [Carbon epoxy/Fiberglass/Carbon epoxy] where there are two layers (0.01 in each layer) on each side of a fiberglass core (0.06 in). Representative properties for plain weave fiberglass and using the same carbon material above.

Here's what the model looks like. I fixed the root edge (certainly too stiff, but for our purposes of comparison only, it'll do)


I ran two subcases:

1. Linear static, where I have a constant pressure load on the fin, normal to the picture above
2. Normal modes to find first natural frequency

Results are:

• Linear static - laminate with carbon epoxy only is ~10% stiffer
• Normal modes - first natural frequency mode is ~10% higher for laminate with carbon epoxy only.

On some projects I use G10 only. In general, it's my favorite material because a) I feel more confident in my G10 leading edges than I do pulling vacuum on my own flat plate, b) it's pretty cheap and c) it still has reasonable stiffness and strength compared to other materials like carbon. If I'm going for stiffness only, then I do have access to some old prepreg T700 and a shop to process it but it's time consuming to do it right.

 

[OverTheTop]

Nice simulations. What does it do to the damping of resonances if you have the mixed CF/G10 stackup?

 

[jqavins]

W/r/t multiple materials and natural frequencies, this is false. Here's a quick model I threw together comparing two fin layups:

1. Carbon epoxy only. Representative properties for a 0.1 in thick plate made of plain weave carbon epoxy
2. [Carbon epoxy/Fiberglass/Carbon epoxy] where there are two layers (0.01 in each layer) on each side of a fiberglass core (0.06 in). Representative properties for plain weave fiberglass and using the same carbon material above.

Here's what the model looks like. I fixed the root edge (certainly too stiff, but for our purposes of comparison only, it'll do)
View attachment 589603

I ran two subcases:

1. Linear static, where I have a constant pressure load on the fin, normal to the picture above
2. Normal modes to find first natural frequency

Results are:

• Linear static - laminate with carbon epoxy only is ~10% stiffer
• Normal modes - first natural frequency mode is ~10% higher for laminate with carbon epoxy only.

On some projects I use G10 only. In general, it's my favorite material because a) I feel more confident in my G10 leading edges than I do pulling vacuum on my own flat plate, b) it's pretty cheap and c) it still has reasonable stiffness and strength compared to other materials like carbon. If I'm going for stiffness only, then I do have access to some old prepreg T700 and a shop to process it but it's time consuming to do it right.

I don't mean to criticize your results, as I don't know more than very little about this, but I do have two potentially important questions. First, I note that the thickness for your all CF case is 0.1 inch, while the total thickness for the laminated case is 0.08 inch. Since, all else being equal, stiffness goes with the third power of thickness, that may make a difference in the ballpark of a factor of two.

Second, the same as OverTheTop's question: damping. One of those things I do know is that lamination, as a general rule, causes a lot of damping due to the change of the speed of sound in the materials, the "sonic index of refraction", if you will. And an increase in damping would lead to a reduction in the amplitude of oscillation at the resonant frequency; but the frequency is the only result you reported. (I dropped a handful of change on my night table one evening, and there was a strangely clear, sustained bell-like ringing. I looked through the change and found, as I had suspected, and old silver quarter, which rings much better than a clad quarter.)

 

[robopup]

I don't mean to criticize your results, as I don't know more than very little about this, but I do have two potentially important questions. First, I note that the thickness for your all CF case is 0.1 inch, while the total thickness for the laminated case is 0.08 inch. Since, all else being equal, stiffness goes with the third power of thickness, that may make a difference in the ballpark of a factor of two.

RE: Thickness - the thickness in my models is constant. For the carbon/FG/carbon stack, there are two carbon layers on EACH face => 0.1 in thickness (same as the carbon only) per my description above. The math which is used to model the stiffness of these materials in linear cases like my is well understood, has been used for decades and correlates well with test.

Nice simulations. What does it do to the damping of resonances if you have the mixed CF/G10 stackup?

Second, the same as OverTheTop's question: damping. One of those things I do know is that lamination, as a general rule, causes a lot of damping due to the change of the speed of sound in the materials, the "sonic index of refraction", if you will. And an increase in damping would lead to a reduction in the amplitude of oscillation at the resonant frequency; but the frequency is the only result you reported. (I dropped a handful of change on my night table one evening, and there was a strangely clear, sustained bell-like ringing. I looked through the change and found, as I had suspected, and old silver quarter, which rings much better than a clad quarter.)

RE: Damping - I do very little dynamic analysis so this quickly gets outside my realm of expertise. But I do want to emphasize that carbon panels and G10 panels are already multiple materials (some sort of fiber and some sort of resin). If you start with a G10 panel, you have glass and some sort of epoxy. If you add carbon layers to this, you have a third material - the carbon fiber - and likely a forth material, whatever resin was used for the carbon layup. Finally, the homogenized fiberglass-epoxy properties and carbon-epoxy properties are similar enough I'd be surprised if you had any constrained layer dampening hapenning. Here's a quick calc of speed of sound in these materials along with @jqavins quarter example (I didn't read enough about quarters to know what the ratio of nickel/copper/nickel is):


I don't mean to turn @James Keller 's build thread into a fin material discussion - I suggest we start another thread after this post if we're going to continue. I don't have anything else to say, unless I get around to @OverTheTop 's damping request but that will probably be a while if I do it.

 

[jqavins]

Good point about the damping in the composites.

W/r/t multiple materials and natural frequencies, this is false. Here's a quick model I threw together comparing two fin layups:

1. Carbon epoxy only. Representative properties for a 0.1 in thick plate made of plain weave carbon epoxy
2. [Carbon epoxy/Fiberglass/Carbon epoxy] where there are two layers (0.01 in each layer) on each side of a fiberglass core (0.06 in). Representative properties for plain weave fiberglass and using the same carbon material above.

RE: Thickness - the thickness in my models is constant. For the carbon/FG/carbon stack, there are two carbon layers on EACH face => 0.1 in thickness (same as the carbon only)...

Perhaps 0.06 was a typo. I agree that we've threadjacked enough.

 

[James Keller]

Hey guys, haven't been able to get to much rocketry lately but I finally got some time to get some done. Firstly airframe 1 got done. Still need to do final cut but its polished and pretty much ready for fins. I ended up bumping the root cord to 2.5in and got those cut out. I ended up screwing up on the programing for one of the fins so I glued that fin to a spare piece of airframe to see how strong the epoxy bond would be. With a 3/8" R filled on the fin I am very pleased with the bond strength (ignore the messiness of the epoxy I got rushed). That being said I wasn't as happy with the stiffness of the fin. I am now very confident that the 2nd layer of CF for the t2t is something I want. It's very likely I could get away without it but there's nothing wrong with making it stiffer. I also am preping the start the NC mold. I have the plug made. It's not quite as smooth as I should probably make it. But for a first attempt I think its acceptable. I plan to do a 1-2 layers of gelcoat then a 3-4 layers of 4oz Fiberglass. Then 1-2 layers of CF. I'm unsure whether or not I should go though the extra effort of making it "steel reinforced" similar to how robopup or watheyak do it since I'm not going to be inflating a bladder so stiffness isn't as important. Any input on that would be very helpful.

 

[robopup]

Looking good!

If this is the first mold you've made, I probably wouldn't worry about backing it with steel and sand. It makes the whole thing a bigger, messier and more expensive project. I only did mine that way because I wanted to leave the option of bladder molding open, and also to have something rock solid that will last me years because I'd previously made many other straight fiberglass molds to dial in what I wanted. I wouldn't have done it otherwise (and may not again if I ever need to remake one of the diameters I have one of these molds in).

Having said that, if you're not going to do a bunch of backing you're going to want to pay more attention to:

• mold thickness - @watheyak and I are able to get away with very few layers because of the backing we used. It's been a long time since I made an unbacked mold, but I would probably shoot for more like 1/8in to 1/4in thickness. @tfish is still doing these and probably has some good numbers for you.
• mold stacking sequence - if you're not doing a backing which is significantly thicker than the fiber layers, you want to pay much more attention to laminate balance and symmetry to reduce warping. This also may include fabric shearing around curvature.
  • Balance - The angles of your stack should add up to zero. And obviously -0/0 and -90/90 are the same thing. For example, [-45/-45/45/45] and [90/0/0/90] are balanced and [45/45/45/-45] is not balanced.
  • Symmetry - The stack should be equivalent in either direction of the midplane of the laminate. For example, [90/0/0/90] is symmetric and [0/90/0/90] is not symmetric
  • A typical balanced and symmetric laminate could look something like [-45/0/45/90/90/45/0/-45].
  • Note also that material type counts for above. So if you're using fiberglass and carbon fiber, [-45FG/0FG/45FG/90FG/90CF/45CF/0CF/-45CF] is not symmetric or balanced.
  • Even gel coat technically counts for balance and symmetry but I probably wouldn't worry about it for this attempt.
  • EDIT - I should also add, if you're using a typical woven product, each layer counts as a +/-θ for our purposes (for example, a 0/90 or a -45/45)

 

[James Keller]

Looking good!

If this is the first mold you've made, I probably wouldn't worry about backing it with steel and sand. It makes the whole thing a bigger, messier and more expensive project. I only did mine that way because I wanted to leave the option of bladder molding open, and also to have something rock solid that will last me years because I'd previously made many other straight fiberglass molds to dial in what I wanted. I wouldn't have done it otherwise (and may not again if I ever need to remake one of the diameters I have one of these molds in).

Having said that, if you're not going to do a bunch of backing you're going to want to pay more attention to:

• mold thickness - @watheyak and I are able to get away with very few layers because of the backing we used. It's been a long time since I made an unbacked mold, but I would probably shoot for more like 1/8in to 1/4in thickness. @tfish is still doing these and probably has some good numbers for you.
• mold stacking sequence - if you're not doing a backing which is significantly thicker than the fiber layers, you want to pay much more attention to laminate balance and symmetry to reduce warping. This also may include fabric shearing around curvature.
  • Balance - The angles of your stack should add up to zero. And obviously -0/0 and -90/90 are the same thing. For example, [-45/-45/45/45] and [90/0/0/90] are balanced and [45/45/45/-45] is not balanced.
  • Symmetry - The stack should be equivalent in either direction of the midplane of the laminate. For example, [90/0/0/90] is symmetric and [0/90/0/90] is not symmetric
  • A typical balanced and symmetric laminate could look something like [-45/0/45/90/90/45/0/-45].
  • Note also that material type counts for above. So if you're using fiberglass and carbon fiber, [-45FG/0FG/45FG/90FG/90CF/45CF/0CF/-45CF] is not symmetric or balanced.
  • Even gel coat technically counts for balance and symmetry but I probably wouldn't worry about it for this attempt.
  • EDIT - I should also add, if you're using a typical woven product, each layer counts as a +/-θ for our purposes (for example, a 0/90 or a -45/45)

Okay that's very good to know. What if I did around 12 layers of FG and got rid of the CF? I could also to a few more layers as well. Also what do you recommend as a gel coat? If I can get away without doing a gel coat and instead paint and polish it it might help with the overall finish of the mold.

 

[James Keller]

Something I forgot to mention in my earlier post and I would like to document is the airframe weights. Fresh out of the peel ply I was at 129 grams, first layer of epoxy got me to 132 grams. After the second layer of epoxy and after finishing the final weight is 133 grams, with a length of 32.75". I got around a 49:51 ratio of epoxy to which I am very pleased with.

 

[robopup]

Okay that's very good to know. What if I did around 12 layers of FG and got rid of the CF? I could also to a few more layers as well. Also what do you recommend as a gel coat? If I can get away without doing a gel coat and instead paint and polish it it might help with the overall finish of the mold.

What weight and thickness is your fiberglass? That sounds kinda thin. You'll be in the ball park if you get to >1/8 inch.

I still like the carbon for at least a few layers, I find it builds stiffness very quickly. And it can be had pretty cheap at places like Soller composites. Certainly not necessary though. And I like lighter weight fiberglass immediately behind the gel coat to help avoid fabric print through. Whatever materials you choose, just make sure your layup is balanced and symmetric.

I just thicken laminating resin for my "gel" coat. For my purposes, it works great. I can't remember my recipe off the top of my head but I believe I posted it in my thread here: https://www.rocketryforum.com/threads/75mm-composite-minimum-diameter-build.173626/

 

[jqavins]

I still like the carbon for at least a few layers, I find it builds stiffness very quickly... And I like lighter weight fiberglass...

This suggests another question. Is there a thread somewhere (let's not hijack this one) about the pros and cons of various composites, especially CF vs. FG? Tensile strength, compressive strength (when it matters, which is probably not often), stiffness, density, abrasion resistance/hardness, heat/fire resistance, etc. Of course, some of these depend as least as much on the resin (various epoxies, various acrylics) as on the fabric.

Every time I read that someone is mixing layers of FG and CF, I wonder why. What is it that each contributes that the other doesn't such that mixing them is a good idea? Again, a reference to another thread (or outside source for that matter) would be better than further threadjacking.

 

[James Keller]

It's been far to long but some progress has been made. I'm going to try my hardest to launch October 16th for my cert. The NC has been a royal pain in my ass. I went though 2 iterations on the mold and 3 iterations on the NC itself. While the final product is far from perfect it's good enough to fly. The mold was unfortunately a little out of round, so I used a 3d printed spacer about 1/32" thick to correct it. When I go to make a new 38mm NC I'll make a new mold but for now it will do. I used PVA for the plug that way I could soak it in water and pull it out. Currently the NC has a coat of thickened laminting resin on it, which will get sanded to 320 grit. Then another coat light coat of laminating resin, then to wet sanding. I also made a change to the fins, I ordered some Uni directional CF from sollercomposites and will make a plate from that. Then do a t2t with 2x2 twill because I like the look of it better. Got my work cut out for me but the hard part is out of the way I think.

 

[James Keller]

Updates, 2 steps forward 1 step back. After seeing @Adrian A's fillets on Stratospear I knew I wanted to try something similar. I don't trust my hand sanding enough to do it precisely enough so instead I machined out 2 little fin extensions, had I thought it out beforehand I would have included it in the fin but it's not structural so I'm not too worried about it.

The fins are made of 6 layers of t700 uni carbon fiber, and 1 layer of 2x2twill 6oz carbon fiber. This comes out to approximately .055" thick fins. The layer direction is 0,45,-45,(0/90),-45,45,0 with the 2x2 twill in the center. The fins came out fairly stiff. A fair amount stiffer than some .0625in G10 I had lying around. Also, I did a 7deg bevel on the fins. They came out fairly similar. The thickest leading edge is .017" and the thinnest is .01". I'll sand and then add a layer of laminating resin to the leading and trailing edge. I did a 1/4" radius fillet on the fins with thickened laminating resin (slightly modified Jim Jarvis recipe). They look a little dirtier in the picture then in person. Soon I'll do a tip to tip. I have a feeling that the fillet may be a little on the small side. After the tip to tip though I think I will be okay.



After the resin is cured I will sand the fillet with 220 until the leading and trailing edges are a similar thickness (maybe a little thicker) than the fins.

Onto the step back. The nosecone I had previously made was destroyed in the tip-casting process. I put a 1/4-20 threaded insert for woodworking in the tip of the NC when casting and suspended it using some all thread. Long story short the threaded rod isn't ever coming out. For one I added too much epoxy to the tip and 2nd I didn't use PVA and wax. Lesson learned, this however gives me a reason to make a new NC mold. This time I'm going to do it right. I'm going to make it with the intention of using a pressure bladder. I'm referencing both @watheyak post (here) and @robopup post (here). The only real difference is I'm backfilling with 4lb 2-part foam from US composites. I think it will be easier and cheaper than backfilling with sand. This NC will likely not fly on the maiden flight in October. I really don't want to rush it, so I will likely just 3d print a NC and either coat it in resin or throw a layer of fiberglass sleeve over it. But the fiberglass NC will definitely be ready in the spring when the big waiver becomes available again and send it with a J1026. Has anyone hit Mach 1.5 with a 3d printed NC? That's the only thing I'm mildly worried about for the Cert flight.

 

[NTP2]

Has anyone hit Mach 1.5 with a 3d printed NC? That's the only thing I'm mildly worried about for the Cert flight.

For the 3d printing I think that if you have it done by someone who knows what they’re doing you will be fine.

 

[robopup]

Has anyone hit Mach 1.5 with a 3d printed NC? That's the only thing I'm mildly worried about for the Cert flight.

I have - max for me with 3d printed NC was M2.1 on 54mm K250. 5:1 von Karmen, thickness was 0.08 in. I use this material for all my printed airframe components: https://proto-pasta.com/collections/all/products/carbon-fiber-petg This was done on my Prusa mk2. Nothing fancy, just make sure your print settings are dialed so layer bonding is good. Seen here with crispy BBQ paint and ablated nosetip (it started off quite a bit sharper).



And good attempt with the mold! Keep going, you'll get it. The one I made which you reference was, I believe, my 10th... I wish there was an easier, faster, cheaper way but so far I haven't found it - the results I've gotten with these split molds are superior to anything else I've tried.

EDIT - also, just curious, what's the reasoning for the single layer of woven material at the core of your fin stock?

 

[James Keller]

I have - max for me with 3d printed NC was M2.1 on 54mm K250. 5:1 von Karmen, thickness was 0.08 in. I use this material for all my printed airframe components: https://proto-pasta.com/collections/all/products/carbon-fiber-petg This was done on my Prusa mk2.

That makes me feel much more confident. What nozzle do you use when printing that filament? Do you notice a big difference in strength between the carbon fiber petg and regular petg? Did you coat it in anything or just fill in the layers and paint?

EDIT - also, just curious, what's the reasoning for the single layer of woven material at the core of your fin stock?

2 reasons. 1 to build up material. The woven was .015" thick whereas the uni was .006" thick. Second and more importantly in my mind was reading Adrian's record attempt and decided that creating stiffness parallel to the root chord was not nearly as important as creating stiffness perpendicular to the root chord and in the axial direction perpendicular to the root chord. So I used 1 piece of woven in the center where it matters the least that way I would have some stiffness parallel to the root chord, while also maximizing stiffness perpendicular to the root chord. In hindsight I suppose I could have done all uni in the 0,45,-45,90,0,90,-45,45,0 orientation and had a similar thickness. Hope that made some semblance of sense. Do you think that a 1/4in fillet is sufficient? This is my first time working on a HPR this small so I just scaled what worked well for me in 3in rockets, not sure if that carries though. Thanks for all of the knowledge you continue to share with all of us!
-James

 

[robopup]

I use the hardened nozzles from Protopasta. I've been meaning to stick this filament in a test machine I have access to, but (I believe) no difference in strength, but obvious and significant improvement in stiffness. No coating, just sanded then primer and paint.

I see what you're going for with the weave now - I would probably just drop it and replace with a single uni layer in the direction parallel to the root - it's much easier to get very flat plates without the woven material. Also, I prefer to rotate my uni's in 45 deg increments where possible, it makes for a more robust laminate. When I make fin stock for small stuff it's usually [-45/90/45/0]s (or just do a single 0 in the middle as you propose), where the reference is the rocket axis and the outer 45 fibers are approximately parallel to the fin leading edge.

Fin fillets - having done zero analysis, that seems a little thin but don't listen to me - I'd test it. For my fillets, I usually find I'm towards the larger end of the recommendations here: http://s605282183.onlinehome.us/wp-content/uploads/2016/01/Nose-Cone-Fin-Optimization.pdf

 

[NTP2]

This is my first time working on a HPR this small so I just scaled what worked well for me in 3in rockets, not sure if that carries though. Thanks for all of the knowledge you continue to share with all of us!
-James

It does, but it might make it a bit overkill but it’s a cert flight so a bit overkill is what you want.

 

[James Keller]

- it's much easier to get very flat plates

On my first attempt making stock I did all uni and it came out warped. I sandwiched it between 2 tiles and then put it in vacuum and let it cure for roughly 5hrs and then put it in the oven at 170(it's medium 820 from soller).It came out very warped. Any idea why this would happen? The plate that I actually used came out a little warped as well. Maybe 1/8" over 8".

http://s605282183.onlinehome.us/wp-content/uploads/2016/01/Nose-Cone-Fin-Optimization.pdf

This is a great resource thanks.

It does, but it might make it a bit overkill but it’s a cert flight so a bit overkill is what you want.

Hopefully fingers crossed this will fly on a Loki j1026 and maybe k627. That's why it's so crazy over built

 

[NTP2]

Hopefully fingers crossed this will fly on a Loki j1026 and maybe k627. That's why it's so crazy over built.

Ok that is a good reason.

 

[robopup]

On my first attempt making stock I did all uni and it came out warped. I sandwiched it between 2 tiles and then put it in vacuum and let it cure for roughly 5hrs and then put it in the oven at 170(it's medium 820 from soller).It came out very warped. Any idea why this would happen? The plate that I actually used came out a little warped as well. Maybe 1/8" over 8".

It's because the material (at least at ply and laminate scales) you're working with is orthotropic. The interactions between plies in the laminate can produce non intuitive (at least to me) behavior of the plate under load, just based on the stacking sequence you choose (and ultimately, what you actually manufactured).

Consider a couple simple examples:

1) Laminate which is unbalanced, symmetric [-45/-45/-45/-45]. When you pull on it in tension, the laminate will deflect in the direction you're pulling, but also in the direction (in this case) 90deg from the load. This is because the fibers are trying to straighten themselves out. This behavior is called extension shear coupling:


2) Now a laminate which is balanced but unsymmetric [0/0/90/90]. When you pull on it in tension, the laminate will deflect in the direction you're pulling, but also out of plane. This is called extension bending coupling:


These examples were with a mechanical load, but this load can also be thermal. When you cure at 170deg then pull it out to room temp, you have a deltaT of about 100deg. The carbon fiber and the epoxy have thermal coefficients which are mismatched, usually the epoxy's is at least an order of magnitude larger. The effect of this is that if your laminate is not balanced and symmetric (not that there aren't usecases where this is desired), when you pull it out of the oven and it comes back to room temp, it will warp just like in our examples above. All this behavior can be predicted with "classical lamination theory."

It's really hard for me to be precise in my garage which is why I usually don't make flat panels. When I do though, I only use uni (and it's usually prepreg) because it's much easier to get close on orientations. And I've altogether stopped using woven products for flat panels because it's so easy to shear the fibers when you're laying them up.