L3 Build - Harakiri

The Rocketry Forum

Help Support The Rocketry Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
One comment, you don't need to make your motor tube as long as the longest motor you want to use, just the fin can (plus the length your ebay sticks into it)

rick

I guess I should cut it off then.... ;)

Think - Flying M1 Abrams (The A-Team)

[video=youtube;EK2eV6KEgko]https://www.youtube.com/watch?v=EK2eV6KEgko[/video]
 
Last edited:
This is the most unnecessarily awesome build thread I have seen yet. Awesome work! A man after my own overengineering heart.
 
This folks, is how a machinist builds a rocket. Tougher than a brick $:!?-house. Not only great rocket, but awesome thread. I'll keep coming back to this one.
 
Thanks guys :) I wouldn't go as far as to call myself a machinist though. Apprentice maybe...

As I'm prepping to mate the two sections, there's an item on my build checklist that needs to be done lest I forget. There are four holes in the lower motor tube that I used to secure it to the mandrel. As I'm going to inject epoxy for the final interior fillets, the holes need to be plugged. I put some blue tape on the motor tube interior and sealed all four with some 50/50 mix. I'll let it cure overnight.

gl_fillets08.jpg
 
This guy is a machinist, retired from aerospace IIRC. Unfortunately, his works of art are no longer in production.
 
Awesome! I've always wondered and wanted to see someone use a rotary table for aligning fins. I also always wanted to be a machinist when I was young, but I'm not. Still, it looks magically spectacular!
-Ken
 
Those are cute!

Yes they are. I wish they were still in production. There's a lengthy thread buried on cnczone somewhere for his prototype.


I also always wanted to be a machinist when I was young, but I'm not.

No time like the present. I took several semesters of gunsmithing over at Lassen College in Susanville, CA. That's where I learned my way around the machines. I had some metalshop in highschool but that was many moons ago. The very first day we were handed a chunk of round cold rolled steel and a file. The task was to make a cube. The instructor gave us a .002 tolerance. Most of us used the whole chunk. There was one student who was a machinist - he cut the chunk in quarters and made a tiny cube. He was done in a fraction of the time.

Grizzly has some reasonably priced equipment and with some TLC, can be made into useful equipment.

cube01.jpg
 
Keep up the great work, I too love the shape of those fins, I'll keep checking in on this build. I wish your shop was my shop!
 
Thanks! My shop is let's say, cozy. It's particularly challenging getting around when my two dogs, 95 and 105 lbs, come to visit. I'll be working over the weekend so no progress until next week.
 
Back at it.

I decided to epoxy, then drill. The railbutton mounts are in place. I also wrapped a small fillet around each of them to keep them in place should the epoxy come loose.

gammaL_railbutton_mount07.jpg

gammaL_railbutton_mount08.jpg
 
I've mated the fincan and lower airframe. I like to sign it and get a picture of the fincan just before mating. Hopefully, I'll never see the insides of it again.

The rear of the airframe is clamped with a hose clamp. That forces the slotted sections against the lower centering ring for a good seal. The upper parts were clamped with some C clamps and angle stock.

The forward coupler is in place. Once the epoxy cures, I'll attach the upper bulkhead.

gammal_fincan_mating01.jpg

gammal_fincan_mating02.jpg
 
I'm getting some of the preliminary work done on the avbays. There are two avbays, one in the upper airframe and another in the nose cone coupler. Both will have machined aluminum bulkheads.

The avbay in the upper airframe will be enclosed with a 12" section of coupler. The bulkheads will not have lips around the edges. Instead, there will be an o-ring seal to keep the deployment gasses out of the avbay. The interior assembly of this avbay will slide out of the coupler section fully assembled. It will be secured to the coupler with flathead screws. This avbay will contain two altimeters, a GoPro, and a few other bits of electronics.

The nosecone avbay will have lips around the edges of the bulkheads and will require partial disassembly to remove the contents of the avbay from the nosecone coupler. It is housed in a 7.5" of coupler tube as supplied by Performance Rocketry. This avbay will contain a Big Red Bee 70cm GPS module.

Attached are some early Rhino renderings of one of the avbay assemblies.

gammaL_avbay_assembly1.jpg

gammaL_avbay_assembly2.jpg

avbay_main_bulkplate_assembly5.0_redundant1.jpg

avbay_main_bulkplate_assembly5.0_redundant2.jpg

avbay_main_bulkplate_assembly5.0_redundant_ghosted.jpg
 
Now, how to get from plan to reality.

Back in an earlier post I posted an image of a widget. This widget is actually a zeroing fixture for my minimill.

A note about my minimill. One of my other hobbies is CNC machining. The minimill is a heavily modified Harbor Freight 44991 minimill and is a work in progress. Some of the modifications include an extended Y axis, precision cut ways, brass gibs, belt drive, computer controlled spindle and a stepper/ball screw driven CNC conversion.

The second image is of a tooling plate, also for the minimill. It took a month of sundays to drill/tap 105 10-32 holes on a 1x1 grid. The tooling plate bolts to the stock table on the minimill and provides a convenient platform upon which a variety of fixtures can be mounted. The third image shows the zero fixture mounted on the tooling plate. The indicator is a coaxial indicator and is used to align the center of the spindle with the hole in the zeroing fixture. Once that's done, the machine XY zeros can be accurately set.

The fourth image shows the milling fixture which will be used to machine the avbay bulkheads and internal mounting components. The fixture serves several functions. First it secures the workpiece to the table. Secondly, it protects the tooling plate from inadvertent damage should I make a programming error or if something bad happens. In this respect it is sacrificial.

Next are a stack of .75x6" blanks from which the bulkheads will be machined. They were predrilled to allow mounting to the milling fixture.

Each bulkhead will receive two machining sequences. The first is the pocketing. The last image is of the pocketing toolpath as generated by the madcam Rhino plugin. The second sequence will cut the partially machined bulkhead out of the blank and is written in simple g-code.

Given the size of the minimill, one cannot make very aggressive cuts. Each cut was .050" using a 1/4" carbide endmill at a fairly slow feed rate. Needless to say, it takes some time to machine a bulkhead.

tooling_plateA01.jpg

gammaL_blukplate_machining01.jpg

gammaL_blukplate_machining02.jpg

gammaL_blukplate_machining03.jpg

pr_cbp5.0_.75_base_toolpath.jpg
 

Attachments

  • zerofixture02.jpg
    zerofixture02.jpg
    112.8 KB · Views: 14
Here are a couple of shots of the machining in process.

As the machine is a work in progress, I've not yet completed the way covers. The shop towels are used to keep the swarf off of the ways and out of the ballscrews and limit switches. They are attached to the table with a few 10-32 SHCS.

The machine does the work but requires some babysitting with a vacuum to clean up the piles of swarf as it's generated. A periodic squirt of wd40 helps things along as well.

The second image is a completed pocketing pass.

gammaL_blukplate_machining04.jpg

gammaL_blukplate_machining05.jpg
 
Next, the bulkhead is cut out from the blank using the second machining program. Note that the pocket interior is now bolted to the fixture. This is the reason for the predrilling of the blanks. Once the bulkhead is cut out, it must be secured to the table else bad things could happen.

The second image shows the completed cutout. There are some washers between the blank and the fixture. As such, the fixture was untouched by the endmill during the cutting process.

As this was the first bulkhead to be cut out, some fitting was required. I used the published figures for the tube dimensions. I ended up taking a few thou off of both the bulkhead OD and the OD of the lip.

Last but not least, the bulkhead. Some additional manual machining will be required to drill/tap holes. I'm also considering jeweling the exposed surfaces.

One down, three to go...

gammaL_blukplate_machining06.jpg

gammaL_blukplate_machining07.jpg

gammaL_blukplate_machining08.jpg

gammaL_blukplate_machining09.jpg
 
On to body work and prep for the external fillets. First the slots at the end of the airframe tube needed to be filled in. I used a bit of glass/epoxy mix to do it. I'll let it cure overnight, clean them up and start on the fillets tomorrow.

gl_fillets09.jpg

gl_fillets10.jpg
 
awesome! Will b watching this thread!

Also, if I may make a suggestion (unless you already have the GoPro)...The Contour Roam 2 shoots a higher frame rate at 1080p than the GoPro Heros 1,2 and 3. (3 being the white edition, of course. The black edition is awesome, that's a bit too much for a rocket camera to me...) :smile:




Braden

Not to mention ease of mounting. I've been looking at the Contour. Its cylindrical shape allows for a very simple mounting scheme. The Hero would need to be mounted on its side, causing the video to be sideways. One could edit the video and rotate it 90 degrees but then there's a significant loss of resolution.

I've been toying with some mockups of the avbay internals. Pictured is what it might look like. There's a crude model of a Contour, a Telemetrum and a Raven 3 along with some basic brackets.

BTW - The specs show 1080p at 30fps, not 60. Where did you see that it has a higher frame rate than the Hero?

gammaL_avbay_assembly3.jpg

gammaL_avbay_assembly4.jpg
 
Last edited:
WOW - Suddenly I feel inadequate some how.

Nah. I'm just obsessive compulsive and do things the hard way :)

The rate I'm going I'll be lucky if this thing gets off the ground with an N. I'll probably have to go with an 80" Iris chute.
 
Last edited:
Next, external filleting. First I roughed up the area with some 60 grit paper. I did the full surface area of the fins as well as the fiberglass tube. This area will eventually receive tip to tip CF. Then I washed it down with some denatured alcohol.

I masked the area with blue tape. I used some 3/8" lathe tool blanks as a guide. I chose 3/8" as the tool I used to smooth the fillet is 3/4" in diameter. On smaller rockets I've used 1/4" lathe tools with 1/2" fillet tools. The filleting tool is a length of delrin rod, rounded at the end. It can be seen at the far left center of the last image just inside the stand. I dip the tool in a bit of acetone to help lubricate the tool and smooth the fillet.

The fillet was comprised of around 8g of the glass/epoxy mix.

gl_fillets11.jpg

gl_fillets12.jpg

gl_fillets13.jpg
 
The filleting is done. I have a bit more body work to do and then I'll inject the internal fillets.

Time to plan out the T2T. Here's another excerpt from the design doc regarding fin flutter and why the fins are designed the way they are. Thanks to DixonTJ for his guidance with FinSim.

Fin Flutter Analysis

The fins are constructed of a .1875" G10 core with a four-layer lamination of 5.7oz
3K 2x2 Twill Carbon Fiber cloth using T2T techniques. Material analysis was performed
via "The Laminator" which performs a Rules of Mixture analysis, returning equivalent
values for the Modulus of elasticity, Poissons ratio and Yield strength of the composite
material. Additionally, the density of the composite was calculated by hand. These
values along with the fin geometry were entered into FinSim as a custom material to
determine critical velocities.

Modulus: 1.006e+007
Density: .065
Poissons Ratio: .272
Yield Strength: 8.275e+004

This rocket is being designed with a Vne of Mach 2.0. While M motors will be used for
certification, simulations were performed using more aggressive N motors (excluding the
N5800). Simulation with an N3301 motor predict a Max Q occurring at 1500AGL at
1975fps. The flights of this rocket will occur at Black Rock, NV. The elevation at launch is
3900MSL. As such, FinSim was programmed to perform the flutter analysis at 5000MSL.
The Classic 2-D lift-slope analysis predicts a divergent velocity of 2231fps and a flutter
velocity of 3035fps. The more optimistic Barrowman 3-D lift-slope analysis predicts a
divergent velocity of 6077fps and a flutter velocity of 8267fps. The author of the software
has indicated that the actual value is somewhere between the 2-D and 3-D models.

Given my construction techniques, the velocity predicted by Rocksim with an N3301 is
most likely quite optimistic and will likely be well below 1975fps. Given this, I believe I
have an adequate safety margin even with an N3301. Refinements to the Rocksim
model will be made as construction proceeds.

As it turns out, the analysis is pessimistic. The analysis was performed assuming the G10 was .1875" thick. It's actually .203" thick.

I've created a pattern using DraftSight, exported it to a pdf and converted the single pdf sheet into a full sized pattern consisting of four 8.5x11 sheets with pdfposter. I'll transfer the pattern to some poster board and cut out the patterns, one for each CF layer.

fin_pattern3.jpg

View attachment fin_pattern3_multisheet.pdf
 
You mention Black Rock. Is this going to be LDRS, or one of the Aeropac launches?

Also, with the G10 you actually have, will it withstand the N5800? You know you have to try (once you cert, of course).
 
You mention Black Rock. Is this going to be LDRS, or one of the Aeropac launches?

Also, with the G10 you actually have, will it withstand the N5800? You know you have to try (once you cert, of course).

One of the Aeropac launches. Definitely not Mudrock. Aeronaut maybe but most likely XPRS even if I have to take a week off to finish it and it comes naked :) I have to make sure my TAPs can make it. With the fin can nearing completion, most of the hard stuff is behind me. I have one more bulkhead to machine and then that's done. Lots of holes are still left to drill though. And the T2T...

The G10 alone wouldn't survive. As designed, the N5800 would push it to M2.28 which is outside of the design limits. It might survive, and probably would. I'm not sure I want to risk it and join the victims list with this one though. A baby O is certainly a possibility. I'll have to crunch the numbers with the thicker G10 but I don't think it will change things too much.
 
Wow, this is an amazing build thread. Great job Don and good luck with the L3! I have no doubt you'll achieve your goals.
 
Thanks for the kind words soxfan!

I've finished the machining of the bulkheads. The two on the left have SMA bulkhead feedthrough fittings. They will be used on the top of the two avbays for 434MHz antennas. The aluminum bulkhead will provide a reasonably good groundplane for the quarter wave whip. The nosecone avbay will have a BRB transmitter, the main avbay a Telemetrum. The second image shows an antenna mounted on top of the nosecone avbay.

These avbays are huge compared to any other I've used to date. There's tons of space inside.

The last two images are of an internal mount for the avbays. The various boards will mount to them. I'm thinking about fabricating a backplane to hold both the Raven and Telemetrum. The backplane would have quick disconnect cabling and would mount to the internal mounts. This would allow me to easily swap the altimeter stack from one rocket to another.

gammaL_bulkplate_machining10.jpg

gammaL_bulkplate_machining11.jpg

pr_cbp5.0_mount4x_01.jpg

pr_cbp5.0_mount4x_02.jpg
 
Man those bulkplates look like they should be installed on a Dragon capsule, not a hobby rocket--great job with the build!
 
Back
Top