Go Devil 38 on a Loki K627, or How I Leaned to Stop Worrying and Love 3D Printing

[Cl(VII)]

I like to fly little rockets fast, but I hate building little rockets because it is frustrating fitting everything in a small package. This is more a limit of my building skills than a true limitation of a 38mm tube, but it is what it is. I’ve turned to 3D printing for these jobs because it provides control over tiny tolerances, is faster to iterate when a part is just not quite right, and takes advantage of my most abundant time resource (sitting at the pool with nothing to do for 90 minutes 3-5 times a week, i.e. CAD time).

The perfect project to test and illustrate all the ways 3D printing can be applied to a build is the Go Devil 38, and to stress test it, the Loki K627.

The Protagonists:
kit_motor by Chris Bender, on Flickr

As you can see the booster is basically going to be all motor when loaded with the Loki 38-1200 case, but I still need to get some things into the booster. Namely these things:
- Internal Motor Retainer
- Shock cord anchor
- Swivel
- Shock Cord
- At least 2” of coupler (preferably more) and all bulkhead/ejection associated stuff
- The electronics bay will be crowded too as an RRC3 powered by a 2S 350 mAh LiPo will need to fit in there.

The payload area will only contain recovery train and the associate 30” TFR thin mill parachute that will function as the main.

The NC will hold an EggFinder mini and 2S 350 mAh LiPo

Note on build technique. I like mechanical attachment of parts whenever possible, so there will be several points where I drill and tap things that could have just been epoxied in, but that is simply personal preference. In my mind that makes it easier to fix things later, but I started this project only planning to fly this rocket once, so I don’t know why that mattered to me.

 

[DavidMcCann]

Friction fit the motor, eyebolt in the forward bulkhead. Boom. two problems solved.

 

[Cl(VII)]

A post about the software, printer and print settings that will be used for all of the 3D printed parts.

The printer is a Prusa i3 Mk2S 3D printer purchased as a kit from the manufacturer (takes about 12 hours to build and calibrate, or you can buy fully assembled for more $): Prusa i3 Mk2S

The filament for all parts is Hatchbox PETG: Hatchbox

All parts are modeled in AutoDesk Fusion 360 which is free to hobbyists: Fusion 360

The models from AutoDesk are sliced for printing using the Prusa version of Slic3r, which is also free: Slic3r

Some basic print settings:
0.4 mm nozzle
Infill: 100%
Bed T: 90 deg C
Nozzle Temp: 245 deg C

If you really want to use this printer and this filament I will send you more details on the settings, but these are the high level settings that would get you started printing similar parts quickly. These settings produce solid parts (no hollow areas) that stick to the untreated bed almost too well assuming you calibrate the Z-axis well.

PETG filament: The stuff is, in my opinion, way superior to PLA for rocket applications for a couple reasons. The biggest reason being that the Tg of PETG tends to run 20 or more deg C higher than PLA from the same manufacturers. That is important when your tracker sled is sitting in a black nose cone on the pad in full sun waiting to fly. When developing the EggFinder mini sled (link to that process) I took the prototypes and let them ride around on my dashboard for what ended up being about a month in the Texas summer with no damage. They were not even softened when I got in the car after work each day. Another advantage is flexibility. Finished PETG parts are slightly flexible where as PLA parts are quite rigid.

PLA does have a big advantage when it comes to ease of printing, and finishing. PLA is super simple to print, and is where I would think everyone starts in 3D printing. You can get nice parts, with almost no strings with a wide variety of settings. With PETG the devil is in the details, and some of the print settings are tough to dial in, and some downsides you will never get rid of completely. The biggest disadvantage of PETG is in finishing. You can’t really solvent smooth it with acetone vapor like you can PLA. You can get parts with no noticeable layer lines in PLA with an easy to make acetone vapor chamber, but PETG is far more solvent resistant…you can’t even really successfully solvent weld the stuff. When I need interlocking parts I work in screw attachments, or bits that will lock and give them some epoxy when I put them together. The other implication of this solvent resistance is strings. Melted PETG is REALLY STICKY which is good for layer adhesion, but when the nozzle moves from one area to another you get super fine strings. These are harmless, but look crappy, and so I don’t use PETG for “pretty” stuff, just functional stuff. If you’ve bought one of my EggFinder sleds you’ve surely seen these, particularly in the full sized sled…that thing looks like a Wookie when it comes off the printer. In theory you can melt the threads with a hot air soldering iron, and get them to fuse to the surface, but the times I have tried this I either wrecked the part or it looked worse afterwards. Like I said, the threads are harmless (non-conductive), so I pull off what I can and live with it because of the desired structural properties of the material…all these parts are inside the rocket after all.

I won’t even get into ABS because printing it is a whole other level of Jedi training. That stuff warps if it cools too quickly, particularly on long prints, and has more shrinkage on cooling that PETG. My PETG parts typically show less than 0.3% shrinkage in any direction upon cooling…many parts show 0.1 mm or less variance in all axes relative to the source model…which is basically the theoretical resolution of the printer.

That got wonky, but the strings drive me crazy, and I felt the need to justify them.

 

[Cl(VII)]

Friction fit the motor, eyebolt in the forward bulkhead. Boom. two problems solved.

Hate friction fit...I'm a worrier. Don't like tape either...ugly. Don't fret, we'll get there.

 

[Cl(VII)]

The fin guides that came with the rocket were completely jacked up. I’d like to give a better explanation than that, but they were off in multiple ways…I notified MadCow, and they said they would check their stock. I also told them not to worry about new for me as I would just print some, and I did.

These are big for my bed, particularly because I buggered the surface coating in a couple places and need to replace it (not terrible, but probably 4 hours work) to regain the entire printable area. I want to install quieter linear bearings on the bed too, so I will replace the surface when the bearings get replaced. Because of the available print space I printed these one at a time, and am pretty happy with them.

The fins were beveled by hand, and notches cut into the root edge with a razor saw. The body tube was roughed up with 90 grit in the fin areas, and all bonding surfaces were wiped down with ethanol. The fins were attached one at a time using Aeropoxy 1 cal from the aft end. The motor was used to balance and weight the assembly, so that the guides would sit flat on the bench and force alignment of the fins.



Fillets were then done on each side, one at a time, using Aeropoxy thickened with milled fiberglass. A scoopula was used to pull and shape the fillets. These are rather small for a Mach 2+ rocket as I will be doing t2t also.

 

[Cl(VII)]

In all things epoxy prep is the most important step. All surfaces where CF was to go were rough sanded, and wiped down with ethanol.

The CF (2k 199 GSM) was laid out and taped off for the full size pieces of CF (top layer) so that there was at least 1 in overhang on the fin edges. The easiest way I know to cut these is with a roller cutter on a cutting mat. Smaller pieces without a tape edge were then carefully cut. These will be the first layer, and do not go all the way to the edges. You can see the two templates in the pic. The yellow is the actual fin to fin size, and the pink is the first layer.



I was going to bag these, and then got lazy. Instead they were done one side at a time using tape to protect the undersides of each fin, and the body tube where I didn't want epoxy. I used the same materials I would have for bagging, but instead of pulling a vacuum on the bag I weighted it with a sack of tiny lead shot. Worked pretty well, but bagging definitely gives better results. OK for here as this thing ain't going more than about Mach 2.5 regardless of what I put in it, so a little imperfection is fine. As I said before, prep is everything when laminating, so everything was set up in advance:



Epoxy: US Composites 635 Medium Hardener
Layers: CF (2), perforated release film, bleeder layer, vacuum bag material, weight.



After the epoxy starts to setup, and is almost to the leathery stage, all of the layers (except the CF of course) are removed. I then used an xacto knife to cut clean lines fore and aft in the CF to make it flush with the fin ends (the protective tape on the back side of the fins and body tube was removed at this time too, and any dribble cleanup was done with ethanol soaked paper towels at this point. The epoxy then setup for at least 8 hours, and the remainder of the excess off the fin edges was cut away in this leathery stage.

The laminated fins were given a thin coat of laminating epoxy to be a sacrificial sanding layer, and after curing 2 days in the back of my car were sanded up to 1500 grit wet. The forward and aft edges were also sanded to a smooth transition at this point.



No photo of the final product, sorry...but it looked sweet.

 

[Cl(VII)]

Like I said above, I don't like friction fit or taping the motor in. If that works for you, great, but it ain't my thing. An Aeropack internal retainer would work, but is expensive, and relatively large. So I made a simple version that is much thinner, and cheaper. I simply printed an 8mm thick disk that had 3 holes equally spaced around it's circumference to accept 4-40 set screws. The weakness of 3D printed materials tends to be in the Z-axis, so to minimize any risk of a tearout on ejection or other such shock I sandwiched this disk with washers that were almost the diameter of the tube (had some laying around, so good there). These were then held on with a 1/4-20 eyebolt, and a lock nut. The shank of the eyebolt was cut so as to threadinto the forward closure of the motor, and there you have a cheap Aeropack. The 3D printed disk is basically just to give the set screws something to grab.

To install this the motor was setup with the retainer, and the spot of the rocket marked. A really tightly fitting drill guide (holes spaced 120 deg) was threaded on, and this line extended around the rocket. The fin locations were marked, the guide was then slid down to the marks, and the holes drilled in line with the fins. The holes were then tapped 4-40, and the motor inserted with the retainer. Here I'm using button head screws, but when my 4-40 set screws arrived I subbed them out.



A length of 1/8" kevlar was tied onto the bolt before it was put in place as that would have proven difficult later. Once attached the motor easily screwed on and off, so retainer and hardpoint done.

 

[Cl(VII)]

The nose cone will house the tracking for this guy, which is an EggFinder-mini, and its associated 350 mAh LiPo. Not surprising I'm using the Nose Cone sled system that I developed (along with sound advice from the beta testers) and sell for just this job. Most important feature: NO ALL THREAD. This thing is going high, so I need all the range I can get.

The kit:

Plenty still available incidentally :wink:

The EF mini mounts to the sled along with the battery. The base of the sled has an 11mm nut capture which is perfect for a 1/4-20 lock nut, or ok for an M6 lock nut. Everything smaller needs shimmed to some degree or potted in epoxy. Between the printed bulkhead and the sled it is recommended to put either a bulkhead or a large washer. This is about keeping the plastic in compression along its Z-axis.



You start by drilling the NC shoulder to match the bulkhead hole locations. There is an a drill guide, that must be used with the thin side down! It is also best to drill the holes 1 at a time, and then insert a screw into the hole so the guide doesn't slide, and your holes end up not at 120 degrees. Once these are drilled you can tap them M4 as I do or go brute force and thread the screw and let it thead itself.



Now the bulkhead slides in, and is retained by three M4 set screws.




I said I like mechanical attachment, so the NC is mounted to the shoulder with three 4-40 set screws. I used the bt drill guide again to mark evenly spaced holes and drill them through the NC and shoulder. These were done all at once since the fit of everything was so snug I wasn't worried about losing reference. The holes were then tapped 4-40 and set screws added (they had arrived in the mail by this point).

 

[Cl(VII)]

I was in a serious rush by this point, so you are only getting a few pics, and some post flight.

A sled was made to carry the Magnetic Switch, RRC3 and LiPo to power it. Same battery as for EF-mini, I'm trying to standardize around it as I bought a bunch last time they were in stock. It was difficult to make these fit in the 38 mm bay, but it worked out. You can't really see it in the pic but me TRA and NAR numbers are printed into the sled for whatthat is worth...probably nothing.



To give the most room possible in the drogue section only 2" of coupler is aft of the vent band. The band epoxied on with Aeropoxy, and is the only other use of adhesive outside the fins. The body tube drill guide was used to space the 3 vent holes, and they were drilled to a diameter I don;t remember right now.

The bulkheads end up needing a lot of stuff. the drogue needs 2 charge wells, recovery attachment point, sled all thread cap nuts x2, and terminal blocks. Yes, I know there are ways to get around all of these like through the bulkhead ematches, and glove finger charges, but I like everything very securely attached, so no free floating charges. To make spacing easy a drill guide was prepared for the bulkhead that snapped onto the center hole, and outer edge. Shown is an earl version, but a second version included 4 holes for the terminal block connections too. The Charge wells were 3D printed too. A double well for the drogue compartment (Drogue and Aux on the RRC3), and a single well for the main. To hold these on the have the same three holes as the enter and sled all thread, and get sandwiched under all three. The guides were used to drill the holes, and the terminal blocks were solder to wires and potted in epoxy into the bulkhead.



The charge wells contained FFFFg. Drogue: 0.25 g, Aux: 0.5 g (NOT coming in ballistic, maybe in pieces, but not ballistic), Main: 0.7 g. The wells have held up after ground testing and flight with little damage...they look gnarly, but after you scrape off the crud they are fine. I would guess these would be good for about 10 flights, and are easy to replace as they are only mechanically attached.

 

[djs]

This is very awesome. What sort of speed/altitude are you looking at on the K627?

 

[Cl(VII)]

Lastly the electronics bay had to be attached to the payload tube as the NC will eject for main. Also, a single 2-56 shear pin was used for both break points, and the heads cut off for flight...drag is evil. Lastly, a static vent needed to be drilled in each recovery bay. The payload tube was held on with a 4-40 set screw.

These were all done on the same line (OCD) with those taking screws being tapped along the way. I don't have a 2-56 tap, so I just ran a SS screw in and out a bunch of times until threads had been created.



The drogue shock cord was 15' long with a swivel tied in the middle. Everything was knotted, so no quicklinks were used to save space. Barely got that in the drogue bay.

The main was a 15' cord with a swivel tied in about 1' from the electronics bay. A TFR 30" thin mill was used for the main.

The rocket was given a coat of Duplicolor Clear, and polished with Maguire's Ultimate Compound a few days later.

 

[Cl(VII)]

Got the mighty Loki K627 LR and #25 nozzle from Chris at Airfest. Built the motor in the hotel room, and had it ready to go on Sunday.

Pad weight 4.9 lbs.

If you were at Airfest on Sunday, this is the little rocket that came out of the tower on the 40s bank probably around or just before noon.

The flight was dead straight, and absolutely hauling ass. When the ignitor lit the motor smoked for a second or so, and then left that tower like a scalded cat.

The EFmini of course stopped sending data on boost, but picked back up just before apogee. Maintained data all the way to landing, and picked it back up when we got within 1/4 mile of it on ground. Walked right up to it about .8 miles from the pad. Rocket basically undamaged except the leading edge of each fin has pitting running it's length...not surprising given the speed. I'm actually glad I didn't see any delamination. The decals were melted/peeled almost completely off. I know they produce drag, but I wanted to rip them off in flight...kind of a Mach trophy.



The data (yes barometric, so not fully accurate, but certainly in the neighborhood).

Top speed: 2571 ft/s (Mach 2.28)
Broke Mach 1: ~600 ft based on hiccup in data
Alt: 16871

Lower than expected, but again baro only so good on high performance flights, and fun regardless. Also, I angled the tower away from the crowd probably 4-5 degrees. If something went wrong I didn't want it coming down anywhere near people.



So, that's it. This was my fastest build ever at about 2 weeks. I know there were ways to do things differently, but built "my way" would have been impossible in that time without the guides and parts I was able to print. No parts were damaged in flight, and once it is cleaned up, and the leading edges repaired it will be ready to fly again, but I don;t know what to put in it now. Maybe a higher acceleration Loki.

 

[Cl(VII)]

This is very awesome. What sort of speed/altitude are you looking at on the K627?

Thanks, and I cheated. It already flew.

Sim was 21000 ft at Mach 2.0 in OR, but with an unflown rocket I had no idea how good that sim was. I set everything to polished, but this didn't reach true "polished" I just didn't have enough time.

I'd like to take some time, and clean up the surface flaws, and fly it again with a Raven (and no decals obviously) to get a more accurate picture of performance.

 

[AeroAggie]

Congrats! I've been sitting on a 95% complete GD38 for many months, trying to decide how I want to set up avionics. I was hoping to make Airfest with it but I decided to wait and see what others have done. Glad I did, and now I have a go forward plan.

 

[captbk]

Super flight Chris, loved the melted lettering too. After enjoying a successful flight with my own GD 38, I had to start building my GD 54 when I got back home. Cant wait till next Airfest.

 

[BDB]

Great build and flight! I really like the 3D printed charge wells. They are almost enough inspiration to make me force my kids to take more swimming lessons. (I assume that's how you found the pool time.)

 

[Cl(VII)]

Great build and flight! I really like the 3D printed charge wells. They are almost enough inspiration to make me force my kids to take more swimming lessons. (I assume that's how you found the pool time.)

I was skeptical about the charge wells until I ground tested them. They hold up as well as micro centrifuge tubes, or cryogenic cell storage vials.

Yep, oldest has been on a swim team for 2.5 years now, and the young one just made the team too. They swim 48 weeks a year...didn't know that when I said, sure you can try a swim team. My car needs a "Swim Taxi" sticker on it.

 

[Cl(VII)]

Super flight Chris, loved the melted lettering too. After enjoying a successful flight with my own GD 38, I had to start building my GD 54 when I got back home. Cant wait till next Airfest.

MD rockets are addictive. I melted/peeled the decals off my 54 Mongoose at Airfest too. It went Mach 2.1 to 18,450 on a Loki L1400. Warning, those Loki motors are REALLY addictive!

 

[Cl(VII)]

Congrats! I've been sitting on a 95% complete GD38 for many months, trying to decide how I want to set up avionics. I was hoping to make Airfest with it but I decided to wait and see what others have done. Glad I did, and now I have a go forward plan.

Thanks. I'd definitely recommend a accelerator based alt if you have experience with one. I went with the RRC3 because I am really comfortable with it, and trust it completely. I was most apprehensive about the range of the EF-mini, but I couldn't have asked for better performance. After about 20 seconds with no data I looked at my Dad and said "We ain't getting that one back", then the LCD box started beeping...oh me of little faith.

 

[OverTheTop]

Great job Chris! Very nice outcome with the flight too.

I have one of these I got in the November sales last year. I hope to get it built over our Summer. I might have to steal your motor mount method and give it a go .

3D printing is very versatile. I started using it for structural parts a couple of years back, and I got looks from people that told me that they thought I should have been in (or perhaps come from!) an insane asylum. It is really a process/material like any other and you work around the constraints that come with it. The number of people averse to using it in HPR are reducing, thanks to nice builds (like this one:wink that show what is possible.

The use of 3D printing for drill and alignment guides is not to be underestimated. It is something that is sometimes forgotten if you are used to using other techniques, and it can give a superior result for accuracy if you use printed guides.

 

[AeroAggie]

Thanks. I'd definitely recommend a accelerator based alt if you have experience with one. I went with the RRC3 because I am really comfortable with it, and trust it completely. I was most apprehensive about the range of the EF-mini, but I couldn't have asked for better performance. After about 20 seconds with no data I looked at my Dad and said "We ain't getting that one back", then the LCD box started beeping...oh me of little faith.

I'm planning to use a Raven (or two). It's pressure based but has accelerometer based logic gates as safety checks and backup channels.

 

[Cl(VII)]

I'm planning to use a Raven (or two). It's pressure based but has accelerometer based logic gates as safety checks and backup channels.

It will give you an accelerometer based flight trace though won't it?

 

[AeroAggie]

It will give you an accelerometer based flight trace though won't it?

Yep, along with pressure and all the logic gates as well.

 

[OverTheTop]

It will give you an accelerometer based flight trace though won't it?

It will chart either accel and/or baro altitude IIRC. Lots of charting options. I don't normally look at accel based altitude, so I could be wrong.

 

[DavidMcCann]

Hate friction fit...I'm a worrier. Don't like tape either...ugly. Don't fret, we'll get there.

looks like you nailed it


after I jammed a G78 into a rocket and fought to get it out with just having the motor label jamming things up, I haven't worried about friction fitting, esp on long motors. It's a pain in the butt, but if your not flying a pocket full of motors, its no big deal to jam one in to me.

nice flight

 

[ColumbiaNX01]

Great flight. Really no need to do tip 2 tip. I used to be a believe in that. Of coarse its great insurance but for most flights not necessary. I flew a 3 inch min diameter on a m 1850. exceeded mach 2. No tip 2 tip. I did use some expensive epoxy. Hysol (loctite) HP-120. From my experience that is all that is required. Now if I was doing a 98mm min diameter and pushing the envelope I would do tip 2 tip. But anyway nice rocket.

I want to do a 54mm min diameter with the loki M with head end deployment. I would probably build that without tip 2 tip but use the expensive epoxy.

 

[GrouchoDuke]

This is awesome, Chris. Congrats on a great build & flight!

 

[Steve Shannon]

Nice!


Steve Shannon

 

[Eric]

Just beautiful..... I'm excited to do my GD29 I just picked up. I will get the Go Devil 38 next. Any chance you want to sell some 3D printed stuff for it?

 

[TBob]

Great flight. Really no need to do tip 2 tip. I used to be a believe in that. Of coarse its great insurance but for most flights not necessary. I flew a 3 inch min diameter on a m 1850. exceeded mach 2. No tip 2 tip. I did use some expensive epoxy. Hysol (loctite) HP-120. From my experience that is all that is required. Now if I was doing a 98mm min diameter and pushing the envelope I would do tip 2 tip. But anyway nice rocket.

I want to do a 54mm min diameter with the loki M with head end deployment. I would probably build that without tip 2 tip but use the expensive epoxy.

The Go Devil comes with 1/16" fiberglass fins. They are very flexible. Tip to tip is highly recommended with fins this thin and flexible on flights over mach. Expensive epoxy or not fin flutter is a real thing.