75mm Composite Minimum Diameter Build

[robopup]

Hey all,

In the interest of getting a little more active on the forums, and to force me to take a few more pictures than I normally would... I thought I'd do a build thread.

Last year at XPRS, I tried a pretty aggressive (for me anyway) 3" motor. It was 47" of Orange Sunset, which worked out to an ~11k Ns N-1800. Unfortunately, due (I believe) to a couple grains which had poor bonding to the casting tubes, the motor cato'd ~2 seconds into the burn. More unfortunately, I tried this in my L3 rocket and after a couple days and a couple hundred miles driving a grid on the playa, no luck finding the airframe. Here's a pad shot and a still from some video:


The motor group I'm in did some tests with our procedures last winter in an attempt to fix the grain bonding issue - we'd been painting casting tubes for years to prevent this but made a few tweaks which we believe will take care of things.

After licking my wounds for most of the last year, I'd like to ultimately try this flight again. Now that I need a new 3" rocket as well, I'm going to take it as an opportunity to try a few new techniques (configuration layout, nosecone processing and fin bonding) I've been playing around with in smaller diameters. A lot of these techniques have been stolen from folks on this forum, including Tony Alcocer (many projects), watheyak (Mach 3.5 Loki L Altitude Record Attempt Build), AstroAnon (https://www.rocketryforum.com/threads/l3-composite-n5800-flying-case-67kft-msl-m3-5.142693/) and Adrian Adamson (https://www.rocketryforum.com/threads/level-3-high-performance-design-and-build-thread.26383/), along with several other fliers in our northwest group who I'm not sure are active here. Many thanks - I've learned a lot from y'all.

The rocket will be designed to take a shorter length motor as well - my current plan is to:

• fly the smaller motor (which I've had success with in the past) at Blackrock this year - it's 30" of Orange Sunset, a ~7.3k Ns M-1100. If successful, this would still be the highest I've been.
• attempt the larger motor either at the same launch (if I can get my act together with the paperwork in time), or more likely, sometime next year

Design

Below are drawings of both motors in Open Rocket and in Onshape.

• Nosecone will be a homemade 5:1 von Karmen. I'm unhappy with my old mold for several reasons, so I'll be making a new one.
• Motor acts as the coupler for the nosecone
• Tubing will be homemade carbon fiber
• Electronics will be in the front of the nosecone
• Dual deployment from a single break - the main parachute will be immediately behind the ebay on a Y harness, a small drogue parachute will be above the motor. In smaller diameters, I haven't dumped a main with this method but depending on how things feel after I test pack, I may print a piston which I can shear pin to hold the main in.
• This will be a 3 fin design. I've spent the last few weeks noodling on this. I'm fairly convinced at this point that, at least for this design's numbers, 4 smaller fins would be both more efficient and more conservative due to the smaller cantilever. My tower is only set up for 3 fins at the moment though and I doubt I'd get around to building a new base this summer... With both a 3 and 4 fin configuration, I've targeted >1.5 stability margin through all phases of flight, mainly using Rasaero but comparing with Open Rocket. At least per Rasaero, I'm even more conservative than 1.5, but I've had better luck lately giving up a little with larger fins for a straighter boost.
• Fin material will be 0.125" G10, bonded to the tubing and with no tip to tip. If I get to a more optimal 4 fin design, I'll probably take the time to lay up a thinner carbon laminate with some fancy uni pre-preg I have access to through work.

Small motor layout:


Large motor layout:


CAD drawings:


A close up of the ebay, which will be 3d printed. The design is a little wonky because I just changed parameters I had on an existing 38mm/54mm design to upscale to a 75mm, but should work fine.


Altitude/Velocity/Acceleration on the smaller motor:


Stability in Rasaero and OR:



It looks like XPRS is a little over 2 months away, so time to get crackin'...

 

[plugger]

Looks fun. Good luck!

 

[tfish]

cool!
a couple of comments
Be careful how much sweep you put on the trailing edges of your fins. I've seen the tips of these types of fins break off.
I've seen a few to many rockets with 1 caliber fin semi spans not able to recover from mishaps. I go with 1.1 calibers for simi span....
I really feel tip to tip ( which is not really tip to tip..more like 1/8-1/4 and half) will handle much more flight and landing mishaps.
don't be afraid to go with more calibers of stability...2.5 or more...That much fuel has a huge CG change ..and rockets like to switch ends if not ...accounting for that quick CG change.
Here is a video rocket with 40" of swamp gas..1.1 caliber fins and tip to tip.



The rocket lost half of it's exit cone...

Yes, you'll loose some altitude doing the above..but it will have a much better chance at working.

Tony

 

[robopup]

@plugger - thanks!

@tfish - Thanks for the feedback. That chat Bob Yanecek and I had with you at XPRS helped confirm that bonding issue...

Sweep - I'm a little nervous about this, but I've been successful with this motor using a traditional planform already, and with this new planform in 38mm and 54mm to ~mach 2.5 so I'm curious to upscale and see how things go. I did try to not get too crazy though (although maybe I didn't succeed here).

Stability -

• That video is excellent, I'd only seen the still before. Yes, I've also in the camp that has been burned pushing things a little to aggressively with small fins and am convinced now that unless you get really lucky with a boost, fins that are a bit larger are almost always going to perform better because of the straight boost.
• The span on these fins is currently 1.25 calibers, 3.875". Subsonic stability margin is 2 calibers subsonic, and remains above 2 calibers supersonic in Rasaero. Open Rocket drops to 1.5 at burnout, but I'm more suspicious of this number.
• In addition to semi span, do you have a stability margin rule of thumb that you also like? In the past I've simply targeted > 2.0 subsonic. I've only started recently plotting margin against time as suggested in some of the links in my original post. EDIT - sounds like 2.5?

Tip to tip - I believe it does buy you margin if done correctly, but I've been flying this way for a couple years and have yet to have a fin bonding root cause (sorry) failure. I work at a company which writes structural simulation software. I'm in a group which is responsible for all composites related capability. A big area right now is modeling of progressive damage in a laminate.

A couple years ago I put together some detailed models of a fin root + fillet and pulled on it like a cantilever beam. I modeled two typical fin laminates:

• traditional tip to tip - this had I believe 8 plies in the fin stock, quasi-iso laminate with an additional tip to tip ply with fibers perpendicular to the fillet on each side.
• no tip to tip laminate - this had the same 8 ply fin stock, but the additional 90deg plies were also placed in this laminate. The fillet had the same dimensions.

I arbitrarily chose a uni T700 pre preg carbon fiber (I don't have test data on my wet lay process and I'm sure there is significant variance if I build different structures). The result was that the traditional tip to tip performed about 20% better if I remember correctly. The rub is that the analysis assumes excellent processing, which I have no doubt I cannot achieve with my wet lay tip to tip process - I suspect more than a few min diameter rockets fail because of poor tip to tip processing.

If I was really going to push things, I would tip to tip, but only the fillet area and not the full span as you say. I would also only do this now with prepreg and a vacuum bag setup. So long story short, with good processing I do believe it helps some, but I suspect I'm already so overbuilt with the 1/8" fins and fat fillets that I don't need the extra performance for all that extra work.

 

[robopup]

Nosecone mold design:

I've gotten started on the tools for the mold. I'm going to use a similar process to what watheyak used in his 54mm build here. I'll use a male plug and a parting board to lay up the first half, then lay up the second half directly onto the first. The one major difference is that I'm going to use thin 3d printed inserts to apply pressure to the fabric and the seam, rather than a bladder. Here's my 54mm version as an example:


This method is time consuming but I'm a big fan because the completed tool is bulletproof. I have 29mm, 38mm and 54mm versions and they give a tight seam, don't warp over time, release easy and are easy to clean up and reprep. The layup is:

• gelcoat 1
• gelcoat 2
• 2 layers 2oz mat glass fabric (to prevent print through to gelcoat)
• 2 layers 4oz mat glass fabric
• 2 layers 6oz plain weave carbon
• 1.5 in square steel tube on either side of the plug - I'll put bolts through these to hold things together during a layup
• epoxy/sand to fill volume between tubing
• 2 layers 6oz plain weave carbon (I'm paranoid enough about symmetry to include these layers, but not quite enough to additionally do another two 4oz and 2oz glass layers)

Here's a quick sketch of the mold plug, parting board and steel tubes:


Here are what the major components of a completed nosecone will look like:

Exterior - the molded nose will ultimately be bonded to a short length of body tube:


Printed inserts (I may additionally glass on the outside of the aft insert to get some fibers from the molded nose to the body tube:


Forward insert removed to show ebay sled:


I've finished the prints for the plug. I have a Prusa mk2. The filament I used was a PETG cut with carbon - I believe by Atomic because it's what I had around. Now that Protopasta carries this material I have a couple rolls on order and will use it going forward (I've been really happy with all their other filament and like shopping local).

It took a few test prints to get something I felt okay about not warping. I've been using a honeycomb infill with a pretty low infill percentage of 10%.

My printer:


Completed prints:


Test fit:


Now I need to bond these together and get going on the finishing work.

 

[robopup]

Body tube:

I use Tony Alcocer's process almost exactly. The layup starts with a mandrel:



For release, I use 2mil mylar. I believe I bought this stuff from either Tap Plastics or Aircraft Spruce. Usually I use two wraps around the tube, but have found with the larger diameters, it's a little easier to slide the finished tube off with 3 wraps:


Five wraps of 6oz plain weave carbon fiber, purchased from Soller Composites. I prefer plain weave because it holds form quite a bit better than the twill weave - I'm willing to take the minor hit on stiffness vs twill, spread tow, or uni. Thickness for each wrap is ~0.01 in.


I've been really happy with Adtech 820 long cure as a laminating resin. Very long pot life, low viscosity, wets fabric out well, decently high Tg and not super expensive. Also purchased from Soller Composites. A trick I learned a while ago for stirring resin is to use two cups to make sure everything is thoroughly mixed. I start with the first cup and popsicle stick for about a minute, then pour everything into the second cup and continue for another minute.


Tube beginning to wrap on. I wrap in the same direction I put the mylar on and try to keep things pretty tight as I go.



I finish with a 2 layers of peel ply to help squeeze out excess resin. I got this stuff from Aircraft Spruce.


Into the hotbox. I do an initial ~12hr cure at 35 or 40 C, then do a 2hr post cure at 80 C.




Completed, trimmed tubing. One down, one to go.


I forgot to measure the final fabric weight, and the final tube weight before trimming to check how I did on my volume fraction. I'll try to remember for the next one.

 

[robopup]

It's been a busy month or so for me. Since I started this project I went on my honeymoon and spent another week on a backpacking trip. Life is hard.

Still, work on this project goes on. Here's the fincan build:

Fins cut from 0.125" G10 using a tile saw. Final dimensions ended up being similar to the original design, I'll attach final as built drawings when the project is finished.


I've been fiddling with this fin beveling jig for the last few years. It gets marginally better each time I use it, to the point I'll probably build a cleaner v2. Fins are held in with a piece of G10 bolted to the jig. I pass the jig through the saw using some hand holds on the other side. The extra pieces of G10 you see here are just to keep the bolted G10 reasonably flat.

It works well for the smaller fins I typically do. The leading edges on these were kind of tricky though because they're so long compared to what I typically do - my wood isn't quite flat enough anymore, so I had to ease into the depth I was cutting a bit more carefully than I usually do.





With bevels cut I moved to tacking on fins and adding fillets. I usually do my fillets with JB Weld but I bought some EA-60HP, which I'd never used before. I'm a big fan of its properties and the gun, although disaster nearly struck...

Bodytube sanded with 180grit (and then cleaned after this picture) before tacking fins on:


Fins tacked on. I have an upper and lower diameter ring for each body tube size I build, and then 6 printed guides for each fin thickness I've used. The rings set the spacing of the guides. The guides hold the angle perpendicular to the body tube. I'm finally starting to get some of the originally hoped for "modularity" out of the design...



Fillets were also done with EA-60HP. Everything was prepped similar to flynfrog's excellent sticky on this forum:https://www.rocketryforum.com/threa...our-government-doesnt-want-you-to-know.58389/

Sanded with 180grit:


Cleaned with acetone:


Taped off:


I went with a 1.5" diameter on the fillets. I'm a big fan of spheres to pull fillets because you don't have to hold the tool at a constant angle. A quick print for the tool:


After pulling tool through and removing tape. I was really happy with five out of six of them. But on one, I didn't get a full cure - it was firm, but not hard like the others. I scraped the fillet out out, cleaned everything well and repeated the process. I believe what happened was I ran out of glue in the gun halfway through the problamatic fillet, and probably squeezed too long when it was running low before reloading a new cartridge. Continuing with the theme that life's hard, life is hardest when yer dumb...

 

[robopup]

Nosecone mold prep:

I'm ready to make the first half of the nosecone mold tomorrow. I have a finished plug and parting board, 1.5" steel tube, board end dams and fabric cut.

Last time I left off with the plug fresh off the printer:


I bonded these pieces together and then finished with a high build epoxy boat primer and topside urethane paint. It's Interlux Prekote 404 primer and Perfection paint. I hate spraying two part paints, fortunately both these products can be brushed. The results aren't *chef's kiss* but they're surprisingly good.

The primer goes on terribly, especially the first coat. I'm still playing with my technique, but it's so easy to toss coats on and the primer sands so well if I take care of it within a day or two after painting that I'm fine with this. I used 3 coats on this plug, sanding inbetween each:



I finish with two coats of the topside paint, sanding in between. It was probably a little too warm here the days (and hours) I painted. Usually I get one or two drips but I got a ton on both coats this time around. I buffed the drips out, sanded to 1500 grit and then polished. Here's the fresh second coat:


And polished. At this point I've also polished in about 8 coats of Partall Paste #2 wax to help with release:


I used melamine for the parting board. It finishes... okay, but I had some on hand. Cut out parting board and end dams on a table saw and then band saw for the curvature. I'm not too careful when I cut out the curved portion on the band saw - I actually make the fit pretty sloppy because I'll go back and fill with epoxy later. Cut out and screwed together:



To get a sharp seam line, I fill the gap between the plug and the parting board with an epoxy+West Systems 410 low density filler, mixed to heavy peanut butter consistency. To get a decent bond to the parting board I knock the corners off with a dremel and then sand:


Then I paint the plug with 2 coats of PVA for release and squeeze the thickened epoxy in using a bag with a cut off corner:



After curing, I tape the parting board off so I don't go through it and sand the epoxy flush. After I do this, I usually end up with several pinholes so I paint some unthickened epoxy on and sand flush again:


All the parts for the plug + parting board. The black pieces are supports to hold the plug at the correct depth against the parting board:


Assembled. This will get 2 coats of PVA tonight and I'll begin making the first mold half tomorrow.

 

[robopup]

Ran out of room on my previous post, so here is the last image I wanted to post:

Parting line close up. I'm not going for a class A painted in the mold finish, so the plug and parting line aren't perfect, but they'll be good enough for my purposes. You can see several of those pinholes in this image which have been filled with epoxy.


Since I have room in this post now, I also got grains cast on Tuesday (with much help!):

 

[G_T]

Some suggestions for the mold:

1) Put an epoxy relief channel perhaps a half inch back from the edge of the nosecone on the mold. You can use layers of striping tape for this. Extend out to the edges of the mold. This gives a gas/epoxy evacuation path with a little lower resistance.

2) Put alignment points into the mold. Marbles, large ball bearings, something like that set at the four corners guarantee only one possible alignment.

3) Put a ramped slot into the mold somewhere, so you can insert a flat bladed screwdriver for starting to pry the mold halves apart.

4) Do not trust that Partall by itself is going to make for a nice release. It would be good to use some PVA (diluted about a third with water) for additional insurance.

5) After you've put down your tooling coat, it is worth putting carbon tow down along the edge of the mold next to the nosecone. This helps reduce the odds of breakout of the sharp edge by providing some structural reinforcement.

6) Use a fairly hard curing epoxy for your mold. Not West Systems... It won't take heat and it dents too easily. I used surplus System 2000 epoxy, since I needed a couple gallons for the mold in the pics. Anyway a number of epoxy systems will work, and there are those intended for mold making. Note, these completed molds are HEAVY!

7) Don't let your molds warm up from epoxy exotherm during the curing process. More heat -> more warpage. I used an IR thermometer to monitor.

Layup this way was about a 36 hour process per half mold.

8) The tooling surface of your part from which you are making a mold is unlikely to be perfect. You can use 2000 grit or finer wet/dry sandpaper wet, with a thin balsa sheet backing, to gently smooth/polish your mold when done. If you get a pinhole bubble, you can fill it carefully with thick CA, and sand it back this way. Use very fine sandpaper and plenty of water. Don't allow any swarf to build up. And make sure the mold halves are really cured first - say for a week.

BTW, I made my own tooling coat from the epoxy plus atomized aluminum plus a little graphite. This produced a very slowly flowing almost paste. Make sure it gets in the edges without air bubbles then pour up layers. I used two layers of tooling coat. It's the grey color in the pics. Inclusion of graphite in the tooling coat has plusses and minuses. Include more than a little and it gets weaker. I did use graphite in the sand/epoxy mix. It really helps to see that everything is well mixed. Some experience is useful to see how much sand to add into the epoxy. The surface should get glistening but not puddle, when given a few minutes to settle. Then you've got it right. I used disposable buckets for mixing, and stout sticks.

I don't know whether you plan to just lay up the nosecones, or pressure mold them. If the latter, the mold needs to be very solid.

In either case, after the tooling coat is down and partially cured, the next layers should be either a very fine weave fabric or scrim fabric (unwoven) so you have minimal print-through of a fabric weave pattern. Use plenty of layers before moving up to medium fabric, and then heavy. Do everything you can to avoid air bubbles.

Each layer needs to go on when the previous layer is at the green stage. Otherwise the mold is weak.

I like using washed and thoroughly dried (kiln) sand with epoxy for the bulk fill. This goes in layers perhaps a quarter inch thick each at most. Each should reach the green stage before the next goes on. It is also probably a good idea to run some steel bars in the mold - two near the bottom and two near the top. Epoxy, and all polymer cure systems, shrink during cure. The more rapid the cure, the higher the shrinkage. For epoxy the shrinkage occurs predominantly in the last 10% of the cure.

Odds are your plug is not absolutely perfectly rotationally symmetric. Put a witness mark on it so when you mold the second half, the plug is in exactly the same place. Or don't remove the plug at all.

The second picture shows where I'm preparing to make the second mold half. These molds were used for a short bit of production work. I probably pulled about 30 parts from them. Their expected lifetime was 100-150 parts before print-through was expected to be an issue. They were used under pressure... I settled in the end for 43psi after a bunch of testing. That's why the molds are stout. If they fail, they'll launch chunks right through a wall. I'm not showing where I have a row of clamps made from 1.5x1.5 oak that go all along the molds before much pressure is applied via bladder and regulated compressor.

Gerald

 

[robopup]

Thanks for the info Gerald. You glider guys make some seriously nice tools and parts, most of my mold making process has been stolen from threads over at rcgroups.

Most of the points you bring up I handle in the same, or a similar way. I'll have more detailed info when I write a post after this first half is layed up, but a couple quick comments:

1. I'm already committed on the first half without channels. Perhaps I'll try one as you suggest on the other half.
2. Yep - I carefully dremel small keys somewhere a bit less than 1/4 the length of the tool from each end. Then I polish, layup the second half, knock a little of the top off the second half side and repolish.
3. I confess I just mark one side so I'm consistent, way at the corner out of the way and give it a wack with a flat blade screwdriver. Haven't had any problems (yet) doing this...
4. Yes, I use PVA as well. Curious whether you only use 1 coat or multiple? I've never stuck a part, or come close, using one coat, but my releases aren't the cleanest either so I've moved to 2.
5. Yep, I do this after the first tooling coat has kicked and before I paint the second coat.
6. I'm using Adtech 820 for the mold. Medium cure on the tooling coats and long cure for the fabric and backing.
7. I've been nervous about exotherm in the past when I first started using this process, in particular for the sand backing, but have never had a problem with the Adtech. This is the biggest mold I've made though. I think I'm going to still do the sand backing in one shot, but monitor for an hour or so after.
8. Not going for a clean painted in the mold finish, but I'll keep this in mind for the future.
9. Tooling coat - I do something similar, I use wyowindworks (Adam?) recipe: 9% graphite and 28% (I'm actually doing more like 22% nowadays) West Systems 404.
10. Your description of the sand backing is about what I thicken to. I need to check my notes for the weight ratio I fill to. For this mold I'm using playground sand, because I have extra hanging around.
11. Correct - plug is not perfectly round (as much as I wish it was and tried to get it to be). It has marks for replacement for the second half. I'm going to try to keep it in the first half when I pull the parting board, we'll see how it goes...

 

[G_T]

Epoxy is exothermic cure. It's also not too bad an insulator. The cure rate for epoxy doubles about every 10C warmer it gets. The result is a thin layup seems to not get warm and takes a while to cure. A thick layup can get hot enough to smoke, and melt plastic. Easily.

Mix up a batch of what you intend to use, and put it in a plastic SOLO cup or equivalent. Put it outside, with aluminum foil tray under it. See what it does. Most times it will melt the cup and rather visibly smoke. You don't want to breathe the fumes.

I only use one coat of lightly diluted PVA, brushed on and allowed to flow out. HOWEVER I polish the mold several times with Partall and make sure the mold is aged at least a week before the first use. If you are going to have serious adhesion problems, it is the first couple of uses of the new mold. Once a mold is seasoned I generally only wax every two or so times, if that often. PVA of course is every time.

You can use a semi-permanent mold release instead of Partall. But you need to make that decision before first applying Partall.

Make your first layup of just some light fiberglass or Kevlar. Not thick; not expensive. I consider it a breakin step. Plus if you cut the fabric to your templates (you are making templates, right?) it gives you a chance to validate your templates and all your processing steps before laying up a part you might care about. Less stress that way!

I made up a fabric board with outlines of each shape/type of fabric, with written instructions for how many and what fabric. I had templates for each shape of fabric, and used a rolling cutter for everything that wasn't Kevlar. Having a layout board like that makes it visibly obvious that you have all the fabric pieces you need ready to go.

Heck, my tooling coat was I think at least 50% aluminum, and a few percent graphite. I forget the numbers. It's fairly hard, and polishes up decently. Any polishing is before putting on Partall, of course. PVA does smooth things a touch, but starting out very smooth and polished is best.

For actual parts layup my first coat in the mold is generally laminating resin thickened with cabosil enough that it isn't very runny. I paint that layer in. It fills fabric weave a bit and also helps stick the fabric in place on the mold so it doesn't swim around as much. Both features are useful.

Then I mix up the next small batch of laminating resin, generally enough for one side of the mold. Then another small batch for the other side. Making smaller batches helps prevent cook-off, and gives you lots of working time.

Generally put the fabric in the mold dry, on top of damp or slightly wet. Then use a pretty dry brush to form the fabric in place. The fabric wets out from the bottom, thereby avoiding trapping any air bubbles.

Once the last layer is in, I put a layer of paper towels in and dry the layup with a dry brush. If the paper towels come out wet, I repeat. Note I do pressure molding for the bigger molds. You would be surprised how much epoxy gets pressed out after I've done the drydown!

Hopefully all this helps a little. Good luck with your project!

Gerald

 

[kalsow]

BTW, I made my own tooling coat from the epoxy plus atomized aluminum plus a little graphite. This produced a very slowly flowing almost paste. Make sure it gets in the edges without air bubbles then pour up layers. I used two layers of tooling coat. It's the grey color in the pics. Inclusion of graphite in the tooling coat has plusses and minuses. Include more than a little and it gets weaker. I did use graphite in the sand/epoxy mix. It really helps to see that everything is well mixed. Some experience is useful to see how much sand to add into the epoxy. The surface should get glistening but not puddle, when given a few minutes to settle. Then you've got it right. I used disposable buckets for mixing, and stout sticks.

Heck, my tooling coat was I think at least 50% aluminum, and a few percent graphite. I forget the numbers. It's fairly hard, and polishes up decently. Any polishing is before putting on Partall, of course. PVA does smooth things a touch, but starting out very smooth and polished is best.

Ok, I'm curious! Why the aluminum? Long-term durability? Easier to polish? Thermal properties? Something else? Amateur mold-makers are wondering....

 

[G_T]

Epoxy compresses easier than aluminum, is not as durable, and the polish is easier to mar. Aluminum also improves the thermal conductivity but that isn't of much help just in a tooling coat. IMHO an aluminum saturated tooling coat is more mechanically stable than epoxy or gel coat with lower solids saturation. BTW the highly aluminumized tooling coat also polishes out nicely with wet sanding with wet thin balsa backing (so it bends and conforms nicely). Use only finer grades of good quality wet/dry paper, like you'd use for car finishing. I forget what grit size I was using for mold touch-up and polishing. One can always start too fine and go coarser. Starting too coarse though and you're stuck with a long annoying job of removing the scratches! Perhaps start around 4000 grit or finer. And, if there is no need, don't do it at all!

To make the tooling coat, you basically keep adding aluminum and mixing well until it gets pretty sluggish flowing. Ideally then you'd vacuum or ultrasonic degas to remove any fine air bubbles since they won't flow out quickly. HOWEVER since it is applied in coats that are not very thick, you just pop any bubbles with a needle. There's plenty of time to do so using a good laminating resin. Ignore West's destructions on how to use the aluminum, and don't use West's epoxy for mold making.

West sells 420 Aluminum Powder. https://www.westsystem.com/420-aluminum-powder/ I may not use their epoxy much, but I definitely use their fillers a lot!

If you don't have aluminum at all, cabosil can be used and does provide benefit. It will also improve the mechanical properties of the tooling coat.

If you aren't going to pressure mold, you can make clamshell molds. Tooling coat, 2nd coat tooling coat, unwoven or super light fiberglass, many layers, medium fiberglass many layers, then heavy fiberglass many layers. Get it built up to at least a quarter inch thickness, and do it in stages where one layup batch gets to green cure (rubbery, dentable with fingernail) before adding the next layer. It makes a mold much more quickly than sand layup and it's still a fine mold. You just can't use it for pressure molding since it's too flexible. That's what the sand is for in a sand based mold - such molds are very stable once cured. But, they are also often overkill, heavy, brittle, more expensive, more time-consuming to make... Clamshells are much cheaper and much faster to make.

I should mention, that if you are going to post-cure your parts in the mold (so they don't warp) then (1) BEWARE WAX! (2) Your mold needs to have a Tg higher than your post-cure temperature for your parts. That means you need to use a higher temperature epoxy. That's where polished aluminum molds usually come in.

Gerald

 

[G_T]

I destroyed - totally destroyed - a mold one time making a first Kevlar part where I post-cured in the mold. The mold could take it, but I mistakenly used Partall or some other wax. It melted and diffused into the part. The part epoxied itself permanently onto the mold. I ripped it off in chunks and strings with vice grips (being Kevlar) but could only get some of it off.

When I screwed up another mold half one time I put the mold on the Bridgeport I had at the time, and manually milled out the part down to about a paper's thickness of the surface. That was nerve-wracking 3D manual milling! The I was able to get the remainder out without doing much to the tooling coat. I did have to repolish it.

If you've invested enough in a mold, you'll go to extremes to save it. That includes milling out a section, and putting the plug back in to remold the damaged area. Hence my suggestion for the carbon fiber tow along the edges, on top the tooling coat, a few posts back. Then I wouldn't have had THAT problem!

I'll shut up now. This isn't my thread!

Gerald

 

[robopup]

I'll shut up now. This isn't my thread!

Gerald

No problem at all, thanks for all your insight!

I got the first half of the mold layed up yesterday. Things mostly went off as expected. I try to do the mold halves in 1 day, allowing the epoxy to get tacky between each step, so that I get the best bond between layers. High level, the list of material that goes on is:

1. Brush PVA layer 1 on. Allow to dry. This is for release.
2. Brush PVA layer 2 on. Allow to dry.
3. Gelcoat 1 - Adtech 820 resin with West Systems 404 22% by weight and graphite powder 9% by weight
4. Carbon tows wrapped around edges at parting line.
5. Gelcoat 2. Same as above. In addition, after painting on the gel coat I add some more 404 to peanut butter consistency and wipe a fillet around the edge of the plug so that it's easier to lay fabric.
6. Fabric - 3 layers 2oz fiberglass mat, 2 layers 6 oz fiberglass mat, 3 layers 6oz plain weave carbon
7. Steel tubing, attached with Adtech 820 thickened to peanut butter consistency with the 404.
8. Sand/epoxy mixture. I did 3.25 batches of 236g resin to 750g sand.

Release:
I've used semi-permanent releases (mainly frekote) on other projects and I'm a big fan. But for a hobby mold that will get irregular use, I didn't want to deal with the cost or their other pecadillos so I stick with wax + PVA.

On new plugs I use at least 8 coats of Partall #2 (which I had previously buffed on), followed by PVA. The night before the layup I painted on 2 coats of PVA, letting the first dry completely before going on with the second. I just use a chip brush to do this. When I first paint it on it's kind of icky, but the bubbles pop and it self levels if I get it on and then don't mess with it. It doesn't like the melamine nearly as much as the plug surface, so there are drips and it's not great here, but that's okay. I tip the tool at an angle so that it doesn't pool as much in the parting line.



Gelcoats:
After first coat has gone on. I make my own gelcoat using a recipe I found on the glider forums. It's laminating resin, 22% West Systems 404 (original calls for 28% by I find I don't have to really pop any bubbles at 22%) and 9% graphite powder by weight. I try to paint this first layer on as thin as possible, and then let the epoxy kick before adding layer 2.


Before the second coat, I tuck 2 carbon tows around the edge of the plug and parting board to improve the durability of the edge. Then I paint on the second gel coat, thicken the mixture even more, and add a fillet around the plug/parting board edge:



Fabric:
Then fabric - 3 layers of 2oz mat fiberglass to help reduce print through, 2 layers of 6 oz mat fiberglass and 3 layers of 6 oz plain weave carbon fiber: I forgot to take a picture after just this step, but I let the epoxy kick before adding the steel bar.

Steel reinforcement:
I use the steel to help keep things rigid, to take up volume which would otherwise have to be filled, and to provide a face to bolt against when I close the mold. It also acts as a dam for the sand/epoxy step.

This is where I ran into some trouble on this mold. It somehow hadn't occurred to me that 1.5" tubing on a 3.095" plug will result in a plug (well the fabric on top of it) which protrudes above the steel and therefore will not have any sand backing. Unfortunately this laminate is kinda thin too, because I was expecting the sand backing. Measure once... layup twice?

I ended up removing the steel, and building up 6 layers of thing 6 oz mat fiberglass strips on each side to at least get it flush. I considered adding 3 more layers of carbon after the sand to build thickness and at least maintain symmetry of the layup w/r/t the carbon. A quick simulation of the current laminate showed this wouldn't do much good with extension/bending coupling due to (non) symmetry, and I didn't want fiberglass mat on the outside so I decided to leave it as is. I don't have any immediate plans to do any bladder molding and the mold is so stiff on all surrounding sides that I shouldn't see much warping. I also feel like not messing with things now yet gives me the most options if I do decide to build more thickness later. I'll see how things feel when I demold.



Sand/epoxy backing:
After the tubing step kicked, I filled the remaining volume with sand and epoxy. The idea here is to build thickness in a cheaper (well, less expensive) and fast way. I used playground sand because I have a lot of leftover I've been trying to use up from a separate gardening project. had been baking the sand in my curing oven at 80C for the previous 24hrs or so to help reduce moisture. I mixed at a ratio of 236g epoxy to 750g sand. It took 3 of these batches and then one 1/4 batch to fill things up.

As G_T mentioned above,I was slightly worried about heat caused by the exothermic reaction while the epoxy cured because of the thickness. I've made several molds of this style using the Adtech 820 in the past and never had an issue, but I did babysit for an hour or so after the step with a fan on it to make sure nothing happened. It was anticlimactic - I could just feel a small amount of warmth if I left the fan off for 5 minutes or so. YMMV though - I have one other boating epoxy I got on sale from Fiberlay a long time ago which is VERY sensitive, particularly with additives. It was exciting the one time I mixed up a ~100g batch of that stuff with filler and it started smoking on my work bench...


Now this half is curing in the heat of my garage. I'll let it sit the rest of today, then pop it open tomorrow and see how things went.

 

[robopup]

First half came out clean. I popped things open Saturday evening.

Release of the parting boards was easy. I had to mess with the plug for a while. I printed it at pretty low density filler in an attempt to reduce warping. This worked well on the smaller molds I've done, but in my attempts to pull this one out using the 5/16" hole I have in the half end of it, it began to mush the hole before it released so I stopped. After a couple hours trying to drip water into any cracks I could find (there weren't really any), I finally gave up and committed the cardinal sin by giving the end a wack with a flat blade screwdriver. Release was easy after this, but I do have a ~1/4" divet at the back of the mold. This shouldn't be an issue as I build in about an inch of slop at the back of the molds so I can cut the nosecones as needed.

Here it is trimmed, and with alignment keys dremeled in:


With the plug reinserted after trimming. I couldn't think of a good way to trim without removing the plug:


I did the second half layup Sunday. Now the whole mess is sitting in my hot box. I'm slowly (5degC per hour) ramping up to 75C which I'll hold for 2 hours, before slowing ramping back down.

I also spent some time at a buddy's place place last weekend to turn a few nose tips. His shop is awesome - his CNC'd lathe (and expertise) made pretty short work of these - one is for this 75mm project, another is a 54mm I'll also be flying at XPRS.

 

[robopup]

I got busy with both work and finishing this project, so I'm a couple weeks behind with posting. But the rocket is fully prepped, motor is built and it's on its way to Nevada with friends and my Pops who are helping out with ARLISS (actually, it might have arrived on the playa by now). On with a few more pictures of the rest of the build...:

Nosecone:

The second half of the mold was layed up exactly like this first half. No issues, separated fine and got the plug out much more easily this time. After the second half layup, I let it cure in the heat of my garage for a couple days. Then I post cured to 75C in my hotbox. I did very slow 5C ramp rates both up and down from temp, and held at 75C for 2 hours. Here are both halves, waxed with 6 more coats of #2 Partall and with a layer of PVA, ready for a nosecone:


Since I started this project, I've continued to fiddle with my nosecone layups in smaller diameters. I've never been totally happy with the seams I get without inflating a bladder, both dimensionally and w/r/t stiffness and strength so I've continued to experiment with alternatives. The layup I settled on for this nosecone was:

• Thin IML liners manufactured which I lay up on. For the cone section, this is 3D printed at 0.035" (with additional inner features which hold the ebay in place). For the tube section, this is 1.1 wraps of 6oz carbon fiber, rolled exactly like I make my usual body tubes above.
• A combination of fiberglass sleeves on the nosecone:
  • 2 are the length of the cone section
  • 3 butt splice with rolled carbon fiber which covers the tube section, and about half of the cone section to where the ebay begins (my tracking is in the ebay so I only want fiberglass at this location).
• 3 layers of rolled carbon fiber on the tube section, and nose section up to where the ebay begins.
• A fiberglass sleeve which covers the full cone on the outside.

I was able to get the stacking sequence symmetric at all zones in the nose section. The tube section ends up being 4 layers of carbon fiber with the single layer of fiberglass on the outside - I didn't want to sacrifice a layer of CF for fiberglass on the ID to keep symmetry.

Liners before being superglued together:



Prepped for the layup:


Before the final layer of fiberglass sleeve:


All fabric layed up:


This whole mess was placed in one of the mold halves, and backed out 0.25 in or so. I closed the mold, bolted it loosely, pushed the cone as far as I could, and then tightened the bolts. This cured tip down in my hot box at 38C for 12 hours, and then was post cured at 65C for 2 hours.

Luckily, I think because of quite a bit of practice on smaller diameters, the first part is very usable.

The exciting part - splitting the halves open:


Released, cleaned and trimmed:


Last step was to cut the nosetip off to integrate the aluminum tip. To get the right location, I printed a tool with the diameter of the base of the aluminum tip and an OD the same as the tube section of the cone to hold things in the right place while I whacked the tip off on my chopsaw:

 

[robopup]

Electronics:

I've been exclusively flying my own altimeters for the last few years. They have a couple accelerometers (higher sensitivity 16g and lower sensitivity 200g which switch based on acceleration conditions), a barometer, 3axis gyros which I eventually will do tilt sensing with and 3 output channels. I use Atmel SAM21D chips which I put the Arduino bootloader on. Code is written in Visual Studio Code and I use Platform IO to compile and upload to the microcontroller.

I was short a board, so I made one (and an extra for good measure) for this event.

I get my boards manufactured at oshpark.com. Before starting:


I use a stencil to smear solder paste on:




All components placed on with tweezers:


I blast these on a (dedicated) frying pan to solder, and then solder on the supercap, connectors and buzzer by hand after:

 

[robopup]

Ebay:

The ebay is placed in the forward half of the nosecone - many thanks to watheyak (Mach 3.5 Loki L Altitude Record Attempt Build) for the inspiration for this design.

In part I wanted the 3d printed liner above because I have some features which lock the ebay in place both rotationally and along the axis of the rocket. A 10-24 bolt passes out the front of the ebay which the nosetip threads onto. 1/8 in kevlar is also tied onto this bolt and passes through a slot in the ebay out the aft end.

CAD of the design from above:



Printed - I used some cheaper PLA I had lying around for this one, and printed at 100% because I don't mind having a little extra weight up front.


Ematches connect directly to the altimeter outputs and pass out a hole in the aft bulkhead. I use rope caulk to help seal the holes the kevlar and ematches pass through.

 

[ksaves2]

Good gosh! Now that is REALLY starting from scratch!!! Best of luck on the flight outcome when the time comes.
I wonder, did you make duplicate parts for future use and extra electronics? Seems to me it would easier to do while you're in the current "building mode" of this project in case you wish to repeat it in the future. Kurt

 

[Hobie1dog]

Wow, making your own altimeters? printing custom bays, making your own motors. You are way advanced in this hobby. We are not worthy

 

[robopup]

@ksaves2 - No complete duplicate parts on this one so far, although I did build this in parallel with a 54mm so I doubled up on things like electronics and propellant casting. The majority of the effort on this project went into dialing in the various tooling though, and not the parts themselves. Once I get tooling set up for a given part, usually it's not a ton of effort after that. The trouble is my designs keep changing! I'm hoping I continue to be happy with this general configuration for my builds so I can continue to use these tools...

@hobie1dog - Thanks. I think my favorite thing about this hobby is that there are about a dozen disciplines you can go as deep into as you wish. Most of what is presented here is stolen from all the awesome content on this forum and others like it. Perhaps someone finds something useful; that's my hope anyway.

Recovery:

In an attempt to pack smaller, I've been making my parachutes for the last year or so using Adrian's excellent process here (although it looks like he's tried something new now?): https://www.rocketryforum.com/threads/simple-efficient-diy-chute.27000/ These are approximately hemispherical, feel pretty draggy and pack small. My wife inherited an old National sewing machine which we fixed up over COVID. I'm not very good on it, and the fabric I'm using is kind of hard to work with but the results aren't bad.

I hadn't thought too much about it before, but I really went down the rabbit hole on what a parachute diameter even means when you start trying to compare different chute designs and what an appropriate drag coefficient and area to use is when you're calculating drag. Mostly because this was the max size fabric I found easily, this design is 60" in diameter (circle cut out) and has a 37" diameter when inflated as a hemisphere. The basic process is:

• Cut out a circle of fabric - a soldering iron on an old circuit board works well for this. The result also doesn't need to be hemmed which is a big time and volume saver.
• Mark and pin the fabric for sewing. This takes an appropriate amount of fabric out of the cut circle to approximate a hemisphere. These parachute end up with 8 equivalent panels.
• Sew the pinned sections above
• Remove the additional material which was sewn, again using a soldering iron
• Sew shroud lines.

I whipped up a little spreadsheet to calculate how much of the diameter of the circle I needed to remove to convert to a hemisphere, and which implements other parachute rules of thumb like shroud line length.

I'm using 0.75oz ripstop nylon I got on sale. I start by tracing out a circle for my chute size.


Cut out using a soldering iron, pressing against an unused circuit board:


Next I fold into eighths so I can get 8 equal panels after sewing:


I cut out a template from some extra peel ply which I use to mark each location I'm going to sew.




After pinning it, I inflate as a test. I got things a little too tight on top in this photo, so I loosened up my pins and markings after:


Then, per above, I sew the eight marked sections, trim off the excess material, and then sew on shroud lines. I'm using ~25 lb (?) kevlar I found on Amazon. I do this part without power, spinning the machine by hand, with about 1.5 in of line sewn onto the chute for each line, so that I make sure each stitch passes through the shroud line.

Finished and inflated. Apologies for the (even more than usual) terrible photo of the finished product. My wife was out of town last week for work, so I was solo dolo in the house - was hard to get it inflated and snap a usuable picture at the same time:


What a loose packing job for this parachute looks like:


And the full recovery gear laid out. I'm going for a single break dual deploy using a "Y" harness, which has worked well in smaller diameters. The apogee charges are packed immediately below the main chute and blow everything but the main out at apogee. The main chute charge is packed immediately below the ebay bulkhead and blows only the main out. Loops tied on ends get double bowlines (hold onto nosecone, eyebolt on motor and main paracute), loops in the middle of the cord get alpine butterflies (drogue and chute protectors).

 

[robopup]

Finishing:

Like the nosecone plug, I used Interlux Primekote 404 primer on both the fincan and nosecone, 2 coats this time. In the interest of time, I finished with a single coat of the topside Interlux Perfection. I brushed both on. After getting a ton of runs on the topside coat like with the plug, my guess is my paint is just too old - same process I've always used, day wasn't abnormally hot or humid, so I'm kinda scratching my head. I sanded down the resulting drips and polished the whole thing. It doesn't look great but it's fairly smooth.



I used a slightly oversized dummy 3d printed nosetip while finishing to keep the aluminum one clean.


Along with a 54mm L I'm also going to give a shot:


Final numbers:

• Total pad mass: 7.75 kg
• Motor total mass: 6.35 kg
• Propellant mass: 3.72 kg
• Total length: 51.5 in

Stability still looks similar to the original design. I'm still dialing in my new sims, but if (god willing) I hold together I think I'll be over 40k. I went with a modified OS formulation a friend put together and has been flying for a few years which is a bit higher total impulse, so the motor is ~8200 Ns in Burnsim.

 

[Flare]

Great design and workmanship, thanks for this thread. Learning a lot! Excited to see the results of this thing flying.

 

[waltr]

Nice build and photos.
I see you did taped bundles on the cord to the fin can. You should also have these on the main chute crod as this really helps reduce recovery system stress when deploying.

 

[robopup]

@Flare - glad it helped
@waltr - yes, the main chute has a couple of these as well. I have to make sure I don't get carried away though because the main cord must remain longer than the nosecone attachment so that the main doesn't get pulled out.

 

[ksaves2]

You lost me after you cut the circle out. Am not a sewing person but I was trying to figure out if you were sewing stays into the chute and/or sewing all the deployment lines in, in one piece My mother who's 93 years old and is still an expert on a sewing machine, she couldn't see it either. (Might be an age related thing but sheesh she's 93 so I can't hold it against her for crying out loud!) The chute looks really nice and I applaud your effort. Your results look great even though "stupidhead" here couldn't figure out how you did it by your post. Not your fault. Just maybe my 3d "brain" spatiality isn't so good visualizing sewing parachutes which I've never done before. Best regards, Kurt Savegnago

 

[robopup]

@ksaves2 - Good question and no problem. I had to noodle on how this all worked the first time I read about it. I do also suggest reading Adrian's thread I linked above, it has more (and better) images than mine does.

No stays as far as I understand the term (TBH I had to look up what these are). In short, we're trying to convert a flat circle into something that is 3D and approximates a sphere. Hopefully these two images along with the math I used to calculate how much fabric to use help. Apologies for the bad paint art.

Circle diameter: C_d = 60 in
Circle circumference: C_c = C_d*PI = 188.5 in
Sphere circumference: S_c = C_d*2 = 120 in
Total fabric to remove to get from C_c to S_c: L_t = C_c - S_c = 68.5 in
Fabric to remove per shroud line/panel: L_sl = L_t/8 = 8.56 in
Length of marking in image below (because we're folding chute in half): L = L_sl/2 = 4.28 in

I follow the process below 4 times so I end up with 8 "panels". I sew first per the marking below and then cut off the extra fabric.


After sewing, it will look like below. Shroud lines are sewn about 1.5 in on each of the 8 locations I sewed from above:

 

[ksaves2]

Thanks loads! Kurt