Three Electronics Bays

Discussion in 'Recovery' started by kbRocket, Feb 26, 2020.

Help Support The Rocketry Forum by donating:

  1. Feb 26, 2020 #1

    kbRocket

    kbRocket

    kbRocket

    Member

    Joined:
    Nov 14, 2017
    Messages:
    22
    Likes Received:
    10
    Location:
    Portland, OR
    I have spent some time evolving thoughts and designs for electronics bays. What I desire in an electronics bay is:
    • RF transparent because I want to put telemetry broadcasting electronics inside the bay. This means no threaded rods the length of the bay that block RF.
    • Maximum space for electronics, uninterrupted by threaded rods
    • Strong in case of an unforeseen event
    • Safe with minimal chance of a connector or wiring failure. A minimal number of connectors are used. Wires are short, secured in place so tugging does not get transferred to terminals, and are all removed in one unit with the sled.
    • Easy to use at flight time: not a lot of fooling around trying to jam everything in and get it connected
    • Easy to arm and disarm at the pad with screw switches accessed through the airframe
    • Redundant dual-deploy
    • Minimal weight
    Here are some principles and construction details of three electronics bays that strive to these requirements. I should note that while these bays work well, they are not the minimum effort to construct.

    Juniper Lander: redundant dual-deploy electronics bay used in a 3” airframe for my minimum-diameter L3 certification. This rocket is a slightly customized version of the Wildman Intimidator 3.
    DSCF3933small.JPG
    Dirt Devil: redundant dual-deploy electronics bay used in a 29 mm airframe. This rocket is a solidly built version of the Madcow Go Devil, with tip-to-tip on the fins and no switch band.
    IMG_7614.JPG
    Effed Up: non-redundant electronics bay for the nosecone of a 29 mm diameter rocket. No pic yet. This rocket is still under construction. It is 100% custom construction.

    I will detail the electronics bay construction details of each of these in separate posts.
     
  2. Feb 26, 2020 #2

    kbRocket

    kbRocket

    kbRocket

    Member

    Joined:
    Nov 14, 2017
    Messages:
    22
    Likes Received:
    10
    Location:
    Portland, OR
    Juniper Lander has a spacious 3” by 9” coupler that makes it easy to mount multiple flight computers.
    whole ebay.jpg

    I mount a sled inside with an Altus Metrum Telemetrum and an Easy Mini. The sled is constructed to be both the backplane for mounting and wiring and also a structural member that is overbuilt to handle tensional forces in the case of a mishap.
    DSCF3946.JPG
    DSCF3945.JPG

    The main body of the sled is a sheet of 1/16” G10 that extends the length of the bay plus forms a tab on either end. For added strength a piece of ¼” Kevlar with loops sewn in each end reinforces the sled for potential tensile loads. The loops are glued inside of the tabs on either end of the sled. They encircle holes D-shackles to connect to tethers. The tabs are thickened with layers of G10 to match the D-shackle opening. 3/16” stainless D-shackles with a 2200 lb load rating are used.

    This sled is strong, lightweight, and does not block RF from the tracking electronics. Also note that in the event of a catastrophe such as a ballistic deployment of the main, the large tensile forces are directly transferred through the Kevlar reinforced spine. In a more traditional electronics bay the end caps become structural members that must transfer force from the threaded rods to the eye bolt.
    IMG_6752.JPG
    At the “sky” end of the sled a G10 centering ring and end cap were glued around the tab. There is a short section of hollow carbon kite stick through which the deployment charge wires pass. This tube is plugged with plumbers putty during use to keep gasses out of the electronics bay. A hole in the tab allows charge wires to be secured with a cable tie.

    IMG_6612.JPG

    At the “dirt” end of the altimeter bay a removable G10 end cap was created that is slotted to pass over the recovery tab. This end also has a short hollow tube for charges wires that faces away in this picture. This removable end cap is held on and in compression with a slotted sealing disk and two #6 screws threaded through a short aluminum closure bar. Plumber’s putty smashed between the slotted seal disk and the end cap creates a good seal around the tab.
    IMG_7659.JPG

    Closing the electronics bay after completing wiring is quick and easy:
    • Insert the sled through the coupler tube
    • Put on the slotted end cap
    • Make a snake of plumbers putty around the tab
    • Put on the circular disk
    • Push the aluminum bar through the square hole in the tab
    • Install one #6 screw that was removed to allow the aluminum bar to come out
    • Tighten the #6 screws, tensioning the electronics bay and smashing the plumber’s putty into a good seal

    This electronics bay is configured with an Altus Metrum Telemetrum and an Easy Mini for redundant dual-deploy and tracking. The flight computers are mounted with #4 stainless screws. These thread into #4 PEM nuts pressed into the G10 backplane and glued in place with G5000 Rocketpoxy. Computers and batteries are mounted by the upper = sky end of the bay to get their mass as high as possible in the rocket. Don’t forget to configure your altimeter for antennae down if it matters.

    The antennae of the Telemetrum is wrapped three times around a fiberglass tube with “pencil” diameter. This inductively center loads the antennae. While it is no longer a ¼ wave whip antennae, this works well and shortens the total length fit better into rockets without going through bulkheads.

    Two sets of charge deployment wires must go from the dirt end to the sky end of the sled. These are separated as far as possible from the Telemetrum Antennae. They pass through holes for good cable management and strain relief.

    Wiring is placed on the back side of the sled. This makes it easier to clearly see how the wires attach to the flight computes and the holes provide strain relief. While the wiring on the back side of the sled looks complex, it is actually fairly simple. Each computer has 4 wire pairs that connect: battery, screw switch, apogee charge, main charge. I provide a rail of 1/16” G10 with a row of 1/8” diameter holes for securing wires. Best practice is to keep the wires short and twist the pairs for RF immunity. This compact design provides stress relief so that wires do not tug on connectors/screw terminals as the rocket is assembled. All switches, batteries, wires, and deployment charge igniters are assembled, secured, and slipped into the rocket as one unit.

    IMG_6750.JPG
    The two screw switches consist of thin brass sheet sandwiching 1/8” G10. On the inside of the switch there is a #4 brass nut soldered to one brass sheet. The fiberglass has a close fit hole for a #4 brass screw. The brass sheet on the screw head side of the switch has a larger hole so that the screw cannot wiggle and make contact until the head is screwed down. One construction note: I start by making the close fit hole in the G10. Then I solder the nut to the brass and drill/tap it. Then I epoxy it to the fiberglass centered around the screw hole and clean it out with a tap. The wires that solder onto the brass sheets pass through holes in the G10 to relieve strain.

    Holes in the side of the rocket and coupler line up with these screws so that the rocket can be armed at the pad with a 1/8” screwdriver. There is no switch band and there are no wires to a coupler or airframe mounted switch. This eliminates long wires that fatigue, get tangled, and must be packed into a traditional bay design.

    I mount a pair of 1s 380 mAH batteries in a fiberglass compartment that I refer to as the “battery coffin”. This is held in place by four #4 screws. I use polarized micro deans connectors to hook up the batteries.

    Note that there are Roman numbers by some holes on the back of the board: I, II, and III are visible. IV is partially covered by screw switch wires. These numbered holes are used to insure the ejection charge wires end up on the proper flight computer terminals. I, and II are apogee charges, II being a larger backup charge. III and IV are main parachute charges.

    When preparing the bay, I ohm out four igniters then number them I to IV on both ends. Apogee igniters I and II pass through the tube at the dirt end of the sled. They are routed through holes to stay in place as far from the antennae as possible. Before trimming the wires to length I verify and mark their number again between the sled and the flight computer terminal. So far I have avoided deploying the main at apogee by this technique. Plumber’s putty is added for sealing the wires in the tubes.

    Some people like bulkheads for attaching igniters. I have tried that and move away from it. With this design it isn’t not difficult to route the igniter wires directly to the flight computers. Fewer connections means fewer chances of failure.

    After assembling the entire bay but before adding any deployment powder, I test the electronics for proper beeps and telemetry information. Only then, with screw switches backed off to disarm electronics, do build the deployment charges and add powder. In this rocket I use surgical tubing and cable ties to hold the powder. I mark the tubes I to IV and make sure each tube has the correct amount of powder during assembly. Then I wrap each with 3 layers of electrical tape.

    One comment on strength: on a test flight, this electronics bay survived a premature separation of the rocket at about Mach 1.4 with no damage and proper electrical function.

    This bay also flew really well on my L3 certification. An M1350W took the rocket up to 23,783’ at Balls 2019. All four charges went off at proper height and in the proper order safely bringing the rocket down.
     
    Steve Shannon likes this.
  3. Feb 26, 2020 #3

    kbRocket

    kbRocket

    kbRocket

    Member

    Joined:
    Nov 14, 2017
    Messages:
    22
    Likes Received:
    10
    Location:
    Portland, OR
    The stock 6” long 29 mm coupler in my Dirt Devil has a much more constricted space for the same redundant dual-deploy electronics. I admit that the electronics bay is full, but it is really 5 pounds of stuff in a 5 pound bag. It’s not really more difficult to use than the spacious bay of Juniper Lander.

    I am going to provide fewer words because of similar construction techniques.
    IMG_7642.JPG
    This design eliminates the aluminum closure bar and instead uses two #4 screws to hold on the endcap.
    IMG_7616.JPG IMG_7625.JPG

    There is plumbers’s putty smashed between the removable end cap and an inner centering ring to provide a good seal. PEM nuts were pressed into the bottom side of the inner coupler and glued in place for the closure screws. You can see residual white plumber’s putty in places.
    IMG_7626.JPG
    IMG_7635.JPG
    The fixed endcap has two short pieces of carbon fiber kite tubing. One for deployment charge wires, the other is to provide a sealed channel for the Telemetrum Antennae to extend slightly beyond the endcap. While carbon fiber is conductive and should never be used to enclose an entire antennae, I reasoned that using a little at the tip of the antennae would not block signal. Fiberglass tubing might be a preferable, but I haven’t had any problems with signal strength.

    For strength, this sled also has a sewed and glued piece of ¼” Kevlar encircling the tether attachment points. There is a short piece of hollow tubing inside of the Kevlar at each end serving as a mandrel while the glue hardened. It also protects the Kevlar from scuffing.
    IMG_7631.JPG
    There are two battery compartments with G10 fiberglass half disks that keep batteries in place. There is no room for a battery coffin, but also no place for the batteries to flop around. I put a piece of blue tape on each battery to hold it in place while it is inserted into the coupler.

    Finding a good place for screw switches was a challenge. They ended up being built on a 1/16” G10 glued to the sled baseplate and upper end centering ring. Same construction technique as Juniper Lander: #4 brass nuts soldered to brass plate.
    IMG_7633.JPG IMG_7634.JPG
    There is a trick to help get the holes drilled in the airframe rocket and coupler to line up with the screw switches. The sled, coupler, and airframe aligned and secured in their flight configuration before adding the screw switch holes.
    IMG_7630.JPG
    IMG_7635.JPG
    Next predrill the screw switch holes with a small diameter drill bit through the airframe, coupler, and stop after making marks on the G10 where the screws in the switch need to be. This helps get perfect alignment.
     

    Attached Files:

    mpitfield likes this.
  4. Feb 26, 2020 #4

    kbRocket

    kbRocket

    kbRocket

    Member

    Joined:
    Nov 14, 2017
    Messages:
    22
    Likes Received:
    10
    Location:
    Portland, OR
    Here is more of Dirt Devil with the final 2 images of the electronics bay.

    First the electronics bay is built without the brass parts of the screw switches. The nearly complete electronics bay/coupler is secured to the forward airframe. In this rocket a #4 flat head screw secures the coupler to the airframe. It screws into a PEM nut inside the coupler.
    IMG_7644.JPG
    IMG_7647.JPG
    I like to use a non-conductive foam cushion from the hobby shop between the electronics and the sled. If you do this make sure that you provide cutouts so that the foam does not block the ports on the barometric pressure sensor.

    I admit that the separation between the antennae and other RF opaque components is not great. There are batteries on the back side of the sled. There are battery connectors and a pair of charge wires on the opposite edge of the sled. However, in a traditional electronics bay construction, threaded rods would also be present to further attenuate RF. There also wouldn’t be enough space to fit all of these components and batteries into the coupler.

    So far Dirt Devil has had six successful flights. There were no problems at all with telemetry. Tracking signals were robust both in the air and on the ground. Five of the flights were above 12k feet including 19,177’ on a J510W. This flight achieved 53 Gs, Mach 1.8. Without proper strain relief or securing of components high Gee forces can cause faults. I recommend attaching your fins well if you try a J510. I had to rip off the apogee streamer to fit the long motor in the airframe.
     
    mpitfield likes this.
  5. Feb 26, 2020 #5

    kbRocket

    kbRocket

    kbRocket

    Member

    Joined:
    Nov 14, 2017
    Messages:
    22
    Likes Received:
    10
    Location:
    Portland, OR
    Effed Up is a rocket I am building for an elegant flight using a long-burn Apogee F10 motor. This 29mm rocket will have just one flight computer mounted in the nose of the rocket, a Telemetrum for redundant apogee deployment and tracking.

    The computer sled mounts in the nose cone with two #4 screws.

    IMG_7554.JPG
    IMG_7556.JPG
    IMG_7557.JPG

    To save space, holes are cut through the sled so that the terminal block, power connector, beeper, and GPS receiver poke through the sled. I also eliminated the PEM nuts for gripping screws. Instead the sled is drilled and tapped.
    IMG_7555.JPG

    This rocket tries to concentrate as much of the mass as possible at the front end of the rocket. I thought of using a smaller battery, but decided to keep my usual 380 mAH battery since it is close to the front end. This rocket requires a little nose weight.
    IMG_7558.JPG
    IMG_7559.JPG
    A single hole in the base of the nose cone serves two purposes: accessing the screw switch and providing a vent port for the altimeter pressure sensor. While I do not recommend a vent port at the base of the nose cone, it is a necessary evil in this case. During flight this area can be subject to turbulence that may confuse the flight computer and cause premature deployment. I have experienced this on a different rocket with a nose cone port at supersonic speeds. Even though this rocket is not supposed to break Mach, present simulation are > 0.9 Mach so there could be air turbulence problems. I intend to use Mach lockout until the speed is predicted to be back below something like 500 fps.
     
    mpitfield likes this.
  6. Feb 26, 2020 #6

    kbRocket

    kbRocket

    kbRocket

    Member

    Joined:
    Nov 14, 2017
    Messages:
    22
    Likes Received:
    10
    Location:
    Portland, OR
    One final remark on construction. These electronic bays contain a lot of little pieces causing multiple steps of cut, glue, repeat. I significantly reduce the cycle time by curing the glued parts in a warming oven at about 140 degrees. This sets the epoxy in about 20 minutes. I mix up a small batch of G5000 Aeropoxy and put it in the fridge to extend the working life to something like 4 to 6 hours. This allows me to keep adding parts in quick cycles with minimal waste.
     
  7. Feb 26, 2020 #7

    TimothyG

    TimothyG

    TimothyG

    Well-Known Member

    Joined:
    Nov 1, 2018
    Messages:
    294
    Likes Received:
    110
    Gender:
    Male
    Some great inspiration. I never thought of the fiberglass eyelets like that. Great way to reduce mass and get rid of metal.
     
    grouch likes this.
  8. Feb 26, 2020 #8

    kbRocket

    kbRocket

    kbRocket

    Member

    Joined:
    Nov 14, 2017
    Messages:
    22
    Likes Received:
    10
    Location:
    Portland, OR
    The fiberglass eyelets are probably strong enough in most rockets without Kevlar reinforcement. I added the Kevlar as an extra safety factor for the L3 cert. Even though Dirt Devil is significantly lighter and less energy, I added 10 inches of Kevlar because it doesn't weigh much. G10 and fiberglass, especially s glass, is really strong in tension. Probably much stronger than your tether.

    I had early thoughts of using fiberglass rod through the ebay instead of metal rods. It still has potential, and could even be threaded. A loop of Keval just seemed superior.
     
    Steve Shannon likes this.

Share This Page

Group Builder