Design and Build Thread: ARO-B Two-Stage

The Rocketry Forum

Help Support The Rocketry Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Kane

Well-Known Member
TRF Supporter
Joined
Nov 29, 2020
Messages
174
Reaction score
214
I've been working on the design of my first two-stage project over the past couple of months and am about to start fabrication. It's only fitting that I take the opportunity to post my first build-thread on TRF despite having more than a few scratch-builds in my fleet. I am absolutely welcoming input from those with far more experience than me - especially the staging gurus who have supplied ample late-night reading for me over the years. That said, I've already purchased most of the material and am beginning fabrication. This first post is somewhat posthumous with regard to the design process.

The name of this project is a not-so-subtle riff on one of my all-time favorite rockets: the Aerobee High. I actually purchased and built a Cosmodrome Aerobee kit years ago that promptly found its way into a soy-bean field where it resided over the winter and spring; returning to me somewhat worse for wear. Hopefully, this build will not share the same fate.

The entire vehicle has been modeled both in Fusion 360 (for fabrication as well as exports for 3D printed parts) and RockSim for flight simulations. The images below are taken from the Fusion 360 model.

Project Goals:
1. Design and build to fly on a wide range of motor combinations (including sustainer only) that stay under local field waiver (East Coast field).
2. Utilize construction methods developed on non-staged projects that combine fiberglass and 3D printed components.
3. Minimum diameter on the booster so 54mm booster motors can be flown (this is driven more by the Aerobee High configuration that had a smaller diameter booster than sustainer).

View attachment ARO-B - ARO-B.jpg
View attachment ARO-B - DIMENSIONS.jpg


BOOSTER
The booster airframe is 54mm fiberglass minimum diameter with 3 swept trapezoidal fins. I like this plan form due to its long root cord that provides longitudinal strength. It also pushes the bulk of the surface area aft benefitting CP, but not so far as to risk breaking in rough landings (such as a swept delta). Booster motors will be retained via a minimum diameter retainer and 5/16" threaded rod to a threaded/plugged closure. An adapter can be used for 38mm motors. Following is a cut-away view of the booster with the motor retainage (the airframe has been hidden in this view).

View attachment ARO-B - MOTOR RETAINER.jpg


INTERSTAGE COUPLER

The interstage coupler is constructed from a 54mm coupler that houses the booster avionics and GPS tracker. The booster nose cone is 3d printed PETG with 50% infill. The nose cone will be epoxied into the 54mm avionics coupler. The interstage base of the sustainer coupling is also 3d printed PETG with 50% infill and will be epoxied into a 65mm fiberglass coupler. These two pieces are then tied together with 3 high strength, thick-walled aluminum struts. An Eggtimer quantum will control the booster deployment event. For higher altitude flights, the booster parachute (24" semi-circumference ultralight hemispherical) will be deployed via Jolly Logic Chute Release. The ISC avionics bay also includes an Eggfinder Mini GPS transmitter. Both the Quantum and the Eggfinder mini are mounted to 3d printed PETG trays. In order to mitigate potential impacts on the radio and GPS antennae, the aft bulkhead of the booster avionics bay will be held in place with knurled nuts and nylon threaded rods. Given the relatively low mass of the booster, I’m comfortable that the nylon rods (and threads) have sufficient tensile strength to hold everything together. The ISC sustainer coupling extends into the aft end of the sustainer 1.3 calipers. The 3d printed sustainer base is printed with a well for the separation charge.

E-matches for the separation charge and air start will run down internal conduits made from 3/16" diameter Garolite.

Following are views of the ISC.

View attachment ARO-B - INTERSTAGE COUPLER.jpg
View attachment ARO-B - INTERSTAGE COUPLER SECTION.jpg


SUSTAINER

The sustainer has a 65mm airframe. It incorporates 3 swept trapezoidal fins similar to the booster, though with a longer root cord and shorter semi-span. These fins will be through-wall and epoxied to the 38mm motor mount and centering rings.

In the avionics bay, an Eggtimer proton will control the primary sustainer deployment events as well as the separation charge and air start. The proton has a barometric sensor and a 200g accelerometer. For air-starts, the proton incorporates a barometric/acceleration deviation qualification for lock-out. The barometric readings are used to compute the vertical distance from the ground while the accelerometer data is integrated to arrive at the total distance traveled. If the difference between these values deviates beyond a preset threshold, the air-start is locked out. The proton will be connected to a telemetry module that will send real time flight data to a ground station. While the proton's arming features are Wi-Fi enabled, a pull-pin switch will be used for additional safety at the pad and to satisfy any RSO's that insist on a mechanical switch or shunt on the air-start e-match. An Eggtimer quantum will control redundant recovery deployments.

The drogue will be a 6.5" parabolic. The main will be a 42" (semi-circumference) hemispherical. All tethers will be 3/16" tubular Kevlar.

The nose cone is a 5:1 Von Karmen with aluminum tip. The sustainer GPS transmitter is housed in a bay in the nose cone and mounted to a 3d printed PETG sled.

Following are views of the avionics bay and the nose cone GPS bay.

View attachment ARO-B - SUSTAINER AVIONICS.jpg
View attachment ARO-B - GPS BAY.jpg



Next post will include static and dynamic analysis of the design.
 
Last edited:
Question/comment regarding the interstage and the aluminum struts with the upper and lower adapters.

With (3) radially splayed struts into matching splayed blind sockets into the upper and lower adapters. How do you anticipate to install more than (1) of the aluminum struts?

Geometry will not allow more than (1) strut to be installed. There is no means to get the other (2) struts into the upper and lower sockets so long as the struts and sockets are splayed.

Nice models and drawings BTW. What did you do the work in?
 
Last edited:
Question/comment regarding the interstage and the aluminum struts with the upper and lower adapters.

With (3) radially splayed struts into matching splayed blind sockets into the upper and lower adapters. How do you anticipate to install more than (1) of the aluminum struts?

Geometry will not allow more than (1) strut to be installed. There is no means to get the other (2) struts into the upper and lower sockets so long as the struts and sockets are splayed.

Nice models and drawings BTW. What did you do the work in?
Geometry be damned.
Just a matter of oversizing the socket holes just enough to allow some flex.
 

Attachments

  • 16963008320448764814614222474932.jpg
    16963008320448764814614222474932.jpg
    1.5 MB · Views: 4
Big caveat: I am not an aeronautical engineer (although I did spend my first year in college in aerospace, I think that counts for almost nothing). I have a healthy distrust of "rules-of-thumb" so I find employing some modicum of nuanced analysis to be cathartic.

First slate of analysis data is based on the lowest thrust/impulse motor combination that I would plan to fly in the ARO-B: an Aerotech I357T to an H123W. I will also post an identical analysis using the highest thrust/impulse combination.

Most of the graphs below ghost in the velocity curve for this motor selection for reference. I've simulated a 2 second air-start delay (for initial simulations only - the actual delay will be honed with final build weights and after shake-down flights). Simulated launch conditions are set relative to East Coast flight conditions at roughly 75' AMSL in wind conditions varying between 8 and 14 mph.

With a loaded weight of 7.5 pounds, the I357T is just over a 10:1 TTW ratio. Simulated speed off the rail would be roughly 72 ft/s.
The projected altitude with this motor combination is 3,700' with a maximum velocity of 450 ft/s.

From at static analysis standpoint, the stability margin is between 4.6 and 5.2 calibers from the beginning of boost to burnout. Sustainer flight maintains a stability margin around 2 calibers until shortly before burnout.

02 Thrust to Weight.png

01 Static Stability.png

Looking at the dynamic data, the vehicle oscillates between 2 and -2 degrees (averaging out the initial spike) corresponding to an underdamped vehicle. The oscillations do continue somewhat into the sustainer burn. The lowest corrective and damping coefficients corresponding roughly to the end of the oscillations and the point of sustainer motor ignition.

04 Angle of Attack.png
05 Corrective Moment Coefficient.png
06 Damping Moment Coefficient.png

The natural frequency over velocity (taken during the coast phase of the sustainer) seems to be on the low end as best as I can tell from the literature I've looked at. Almost all of my high power birds maintain a NF/V in this range and fly well. The fins are admittedly on the smaller side. Increasing the fin span up to and beyond the body diameter (per one of those dreaded rules-of-thumb) had minimal effect on this which is likely due to the relatively high velocity experienced by the sustainer.

08 Natural Frequency.png

The damping ratio is above the 0.05 threshold but well below 0.3 which implies a design optimized for altitude (though that was not necessarily a design goal).

09 Damping Ratio.png
 
It's been a minute since my last post. Despite the slow pace, I have made some progress on the ARO-B.

Sustainer motor mount assembled with 3D printed epoxy dams. The holes through the centering rings are for the air-start and separation charge conduits.
PXL_20231017_005409641.jpg

3D printed interstage sustainer cap coated with high temp epoxy.
PXL_20231022_160113915.jpg

Completed interstage coupler (and construction diagram).
PXL_20231119_202817709.jpg

Completed booster avionics tray (sans altimeter and GPS tx).
PXL_20231024_012505099.jpg

Sustainer avionics bay in progress with fully assembled Eggtimer Proton.
PXL_20231119_202916828.jpg
 
While working on fabrication, I've been diving a little deeper into the static and dynamic flight analyses.
While evaluating multiple motor combinations, I discovered that the damping ratio was dipping dangerously close to the minimum recommended value.
I also started reevaluating the static stability margin as a percentage range of the delta CP-CG to airframe length. Using the range of 8% to 18% (8% being a value reserved for a fineness ratio less than 1:10 and 18% being an upper limit reserved for ratios greater than 1:40). The full stack has a fineness ratio of roughly 32 where the sustainer alone has a fineness ratio of roughly 21. Through extrapolation, I'm shooting for a SSM of the full stack around 16% and around 12% for the sustainer alone. At the same time, I need to keep in mind that the sustainer will fly as a single-stage at times and not benefit from any boost velocity with regard to its CP.
Here are two charts illustrating the SSM percentage (left vertical axis and shaded regions) and the damping ratio (right vertical axis and single lines) with two different fin plan forms: one at 2.25" span and one at 2" span. Each tracks the values with three different motor combinations from the least total impulse to greatest. Since I'm pretty comfortable with the full stack and booster fin plan form, I kept the same sustainer motor in each example. It's definitely interesting to see the relationship between static stability and damping.

FIN PLAN FORM WITH 2.25" SPAN
FIN PLAN FORM 2-25.JPG


FIN PLAN FORM WITH 2.00" SPAN
FIN PLAN FORM 2-00.JPG
I prefer the fin plan form with the 2.25" span since it provides a more comfortable SSM for the sustainer - especially when it will fly as a single-stage. However, the damping ratio basically hits the minimum recommended value at about 1.8 to 2 seconds into the flight (just after booster burnout which occurs at 1.6 seconds). Maybe this short dip in the damping ratio is nothing to work about .... ?
 
Progress has been slow with life getting in the way. I was able to nearly finish the sustainer avionics bay.
Primary recovery deployments (drogue and main), separation charge, and air start are all controlled by the Eggtimer Proton. Redundancy for recovery deployments are handled by an Eggtimer Quantum.
The Proton (and Quantum) do not supply power to any of the output channels without "switching/arming" via the WiFi interface. Regardless, I added a pull-pin switch attached to the positive deployment power jumper pads. In this configuration, the Proton cannot even "see" that there is anything connected to the output channels so there is virtually no chance of inadvertently igniting the sustainer motor during pad setup.


ARO-B Avionics Progress 01.jpg
Eggtimer Proton mounted on the 3D printed PETG sled. The diamond shaped structures to the rear will retain the two 7.4V 400 mAH LiPo batteries and provide a "perch" for the Eggtimer Telemetry Module.
There are four ring terminals that will get soldered to conductors for separation and air-start charges. My plan is to use modified JST connectors to connect the lead from these ring terminals to the e-match conductor.


ARO-B Avionics Progress 02.jpg
My wiring "code": Red/Black = Power :: Green/Black = main deployment :: Yellow/Black = drogue deployment :: Red/Red = Power Switching :: Yellow/Yellow = Separation Charge :: Blue/Blue = Air Start.
The four aluminum screws/nuts at the bulkhead are for separation charge and air start. These are attached to ring terminals on the outside of the bulkhead which get wired to the e-matches.


ARO-B Avionics Progress 03.jpg
Eggtimer Quantum for redundant recovery deployments.
The pull-pin switch pin fits through two of the static pressure ports in the switch band of the avionics coupler.
Visible on this bulkhead are the four Lab Rat through-bulkhead terminals for drogue and main deployment wiring.
 
Question from someone else working on my first 2-stage. I have a similar layout as you (but not nearly as well designed or documented!).

If the sustainer ignition wire runs from the bulkhead of the sustainer bay through the conduit to the sustainer motor, what causes that wire to break when the drogue is released? Do you use a pin block on the outside of the bay bulkhead, or some other connector that stays together until the drogue charge causes the wire to be yanked and then it comes apart?

(fantastic thread BTW. thank you!)
 
I've seen a lot of different solutions for this on TRF: everything from twisting the wire from the bulkhead and the initiator together and tapimg to utilizing more exotic pull-apart connectors and even conductive contacts on the coupler and airframe. I have been testing a modified JST connector. If you cut off some of the "sheathing" on the male side connector (this sounds like a circumcision), it reduces the friction required to pull the two pieces apart. They stay together well enough to maintain continuity, but pull apart easily enough with an appropriate drogue separation charge.
 
Really like your circumcised jst connector - easy and cheap.

I’ve been using magnetic connectors. Not as cheap, but work well - https://www.amazon.com/Charging-Connector-Spring-Loaded-Magnetic-5Pin/dp/B0BRF72XFD

I would really be worried about the petg nosecone and coupler. Those aluminum struts don’t seem to be transferring the load to the fiberglass airframe. Looks like the would be driven through the petg like a nail instead of transferring the load. Have you thought about using a stepped fiberglass or aluminum ring?

I could be missing something but I’ve seen a few strut based interstage designs that were all aluminum and still collapsed from the stress, like Jerry O Sullivans Iris

 
Really like your circumcised jst connector - easy and cheap.

I’ve been using magnetic connectors. Not as cheap, but work well - https://www.amazon.com/Charging-Connector-Spring-Loaded-Magnetic-5Pin/dp/B0BRF72XFD

I would really be worried about the petg nosecone and coupler. Those aluminum struts don’t seem to be transferring the load to the fiberglass airframe. Looks like the would be driven through the petg like a nail instead of transferring the load. Have you thought about using a stepped fiberglass or aluminum ring?

I could be missing something but I’ve seen a few strut based interstage designs that were all aluminum and still collapsed from the stress, like Jerry O Sullivans Iris


Cool video!
The magnetic connectors are sick. I'd love to see how you've used them.

I had similar structural concerns (though the Aerobee line of sounding rockets did it successfully). While my grasp on aero forces is not so strong, I'm not too bad with structural forces (my day job is architecture). Luckily, this is mostly a statics problem (barring any eccentric forces resulting from sub-nominal flights).

First, the rocket in that video is WAY bigger than my AERO-B. The AERO-B is 2.6" in diameter. It will be 10 lbs fully loaded (both stages, both motors). That Iris in the video is probably 8" in diameter. I have to imagine it's well over 100 lbs - if not 200. I don't think I'd want to venture the success of any 3D printed airframe parts on a bird that big and powerful.
Second, mine will fly with WAY less thrust than what it appeared Jerry was loading in his Iris. The biggest booster I will fly will be a J800T. I'm not sure what was loaded into the booster of the Iris, but it looked like 8 or 9 29mm motors around a central 75 or 98mm? Complex N? (I don't think you can fly complex O's at MDRA). Regardless, that's gotta be orders of magnitude greater thrust than what my ARO-B will see.
Third, the struts on that Iris seemed REALLY thin compared to the overall size of the vehicle. This is admittedly a visual judgement since I don't know the specs. They may be true bars and not tubes, but that won't help much if they are too long and skinny.

The above notwithstanding, I did do the math.
Each strut, fully loaded at maximum thrust is going to see about 274 N of axial force. Each strut is 0.25" diameter, .049" wall 6061 aluminum tube. There is a slight eccentricity to the loading due to the angles of the struts that will introduce some shear and bending stresses - but it's basically negligible and the bending and shear stresses don't exceed those that 6061 aluminum can perform with. The struts are not long enough to be limited by buckling (slenderness ratio is less than 40 - values equal to or greater than 40 require evaluation using Euler's column formula). Since buckling is not a failure mode, crushing will be. 6061 aluminum's bearing yield strength is 386 Mpa which works out to an acceptable axial load of 5,776 N given the cross sectional area and a health safety margin.
PETG's compressive bearing yield strength is 70 to 100 MPa. Using the same cross sectional area for the strut as the concentrated area for loading (and a health safety margin), the PETG can bear 1047 N.
My next concern was one that I think you may have alluded to: transferring the load from the struts to the PETG and then into the fiberglass airframe - the latter transfer taking place through the epoxy used to adhere them together. Since I really don't have a way to even ballpark the adhesive strength between the PETG and fiberglass, I added some belt and suspenders by passing permanent set screws through the fiberglass, aluminum strut, and PETG on the sustainer end of the ISC. The booster nosecone has a "lip" where the thrust loads travelling through the booster fiberglass can transfer to the PETG.

Thanks for the comments! Keep them coming.
 
The booster nose cone is 3d printed PETG with 50% infill. The nose cone will be epoxied into the 54mm avionics coupler. The interstage base of the sustainer coupling is also 3d printed PETG with 50% infill and will be epoxied into a 65mm fiberglass coupler. These two pieces are then tied together with 3 high strength, thick-walled aluminum struts

I'm probably missing something obvious because I've seen this a few times, but I have to ask a dumb question. Why bother making a nosecone for the booster that then has the "interstage base of the sustainer coupling" permanently fixed right on top of that nosecone? Am I right in thinking that after the separation charge goes off, the booster will have the "separation charge cavity" as the top most exposed part (and not the booster nosecone part). I hope my question makes sense, and again I suspect I'm missing something obvious.
 
I'm probably missing something obvious because I've seen this a few times, but I have to ask a dumb question. Why bother making a nosecone for the booster that then has the "interstage base of the sustainer coupling" permanently fixed right on top of that nosecone? Am I right in thinking that after the separation charge goes off, the booster will have the "separation charge cavity" as the top most exposed part (and not the booster nosecone part). I hope my question makes sense, and again I suspect I'm missing something obvious.
There is no functional reason. The "look" is inspired by the Aerobee sounding rocket booster. TBH, I'm not even sure why the Aerobee boosters had this open interstage.
 
Finally got back to doing some work on the ARO-B.
After a lot of analysis paralysis, I finalized the plan form for both sets of fins and cut them out of 3/32" G10 using a carbide tipped, fine tooth blade on a table saw. I then beveled each exposed edge by hand. I don't have a stationary belt sander so I have to be as mechanical as possible with my sanding regime. I 3D printed all my fin guides.



All the fins cut and beveled and dry fitted to the airframe.

PXL_20240207_023734240.jpg



Detail of the sustainer fins.

PXL_20240207_024502722.jpg



3D printed fin guides.

PXL_20231203_205452072.jpg
PXL_20231203_205514838.jpg
 

Attachments

  • PXL_20231203_205514838.jpg
    PXL_20231203_205514838.jpg
    3.1 MB · Views: 0
Continuing with the fin assembly, I started the glue-up process. I'm apparently not very good at gluing and photographing at the same time, so the initial glue-up is not documented, only the fillets.
I sanded with 220 grit all the contact surfaces of the fins and airframe and wiped with acetone. My epoxy of choice is Rocketpoxy. I honestly have never tried any other epoxies for rocketry. I have not had any reason to. Why mess with perfection?

The sustainer fins are set into "pockets" with epoxy dams. I fill these with enough epoxy so that when the fins go in, the epoxy fills the entire void in the pocket.
Gluing the sustainer fins is a bit of an orchestration: apply a ring of epoxy on the inside of the body tube for the forward centering ring of the motor mount, filling the epoxy pockets (all three) 2/3 with epoxy, epoxying the aft centering ring, quickly sliding the whole assembly into the body tube, and finally installing the fins.
Because of the robust through-wall mounting of the sustainer fins, I like to keep a minimal fillet: it's only about 1/8" in radius.
The booster fins are a different story. Since the booster if minimum diameter, the fins are surface mounted. I prepped all the surfaces well, tacked the fins on, and then treated them to generous fillets (5/16" radius).



Taping out for the booster fillets.

PXL_20240211_004510571.jpg



Like I said, why mess with perfection?

PXL_20240211_024650923.jpg




Sustainer fin fillets

PXL_20240211_192713469.jpg
 
As the glue is drying, I started to sketch out my thoughts on testing, shake-down flights, and launch prep.

Ground Testing:
  1. Recovery event tests to appropriately size charges for sustainer drogue and main.
  2. Recovery event test to appropriately size charge for booster main.
  3. Separation event test to size separation charge.
Shakedown Flights:
  1. Fly the sustainer alone for a couple of flights first to get a feel for how it handles. First flight will be a higher initial thrust motor followed by increasingly longer burn, lower thrust motors.
  2. Fly the sustainer + booster stack without air start (i.e., motor in the booster only, dummy weight in the sustainer motor mount). This would be flowing with a higher thrust motor.
  3. Fly the full stack to the lowest simulated altitude - probably with an I357 booster and H123 sustainer.
Prep Procedures:

I always use a thorough checklist for all my high-power flights. Most are two to three pages of procedures organized by pre-prep inspections (in the workshop), prep table (at the field), and launch pad. Each of those sections' procedures are broken into sub-sections such as airframe, recovery, avionics, tracking, motor, etc. I also include all critical specifications such as airframe length, dry weight, loaded weight, motor specs, thrust-to-weight calculations for initial and average thrust, CG and CP locations, etc.

This excerpt is taken from another rocket of mine.

checklist.JPG
checklist specs.JPG
I expect the checklist for the ARO-B to be longer and it will evolve as I work out best practices.

I think the most complex is working out how to safely prep and install all the event charges and initiators and arming electronics.

Here's an initial diagram and low detail step-by-step plan from the prep table to the pad:
View attachment ARO-B - AVIONICS AND EVENT CHARGE PREP.jpg
Installing the sustainer air-start igniter while the vehicle is vertical seems to be one of those tasks that require at least three hands. I will fabricate a "hold apart" from some all-thread and some 3D printed seats to keep the booster and sustainer apart while I fiddle with the air-start initiator.
 
Last edited:
Waiting for the weather to turn so I can start painting.
In the meantime, thought it would be nice to have an onboard camera so I designed and printed a custom shroud that will house a RunCam Split, 400mAh LiPo and an Eggtimer WiFi Switch. The outer shell was printed with supports so will need a couple layers of tabletop float epoxy.

1000006925.jpg

1000006924.jpg
 
First shake-down flight of the ARO-B sustainer stage flown at LDRS 42 on Saturday, June 8. Flown with an Aerotech H178DM. Winds weren't great: between 12 and 18 mph. Nonetheless, the flight went great. Little bit of wiggle off the rail, but probably just due to winds. Predicted altitude was around 1900' and the actual was around 1450'. After the next flight, I'll compare these again and make adjustments to the Cd as necessary (this will be an iterative process). I did not fly this with the onboard camera since I wanted to see how it flew without. Next flight will get the camera.

Here's video of the flight.


The Eggtimer Proton has an accelerometer so I was first interested to see how it compares to the filtered velocity. The only odd thing is that the accel derived velocity continues to increase to -150 ft/s during the decent - even after the main deployed. The filtered velocity shows what happened more accurately.

FVeloc v VAccel.png

Chart showing acceleration and velocity:
VAccel v Accel.png

An finally the altitude and velocity:
Altitude v Velocity.png
 
Last edited:
Back
Top