This is the story of when I built an advanced rocket engine static test system / development environment.
Upon having decided to do it, the studying began around late April 2017.
Study followed by system design, consulting friends and friends of friends for their rocket expertise and safety inputs…
System design followed by detailed engineering, testing & fabrication…
Followed by buying materials…
And tooling…
Having acquired all the instrumentation, I began the software development process.
I love doing software development first, because the iteration cycles are hyper-fast, giving you confidence and the right mindset to move on to hardware.
A key decision was to go with a graphical user interface (GUI) for the test system. The idea is that a GUI speeds up the actual rocket engine development, because data can be visualized and analyzed immediately.
After completing the GUI, I hooked up a router to a Raspberry Pi 3, to 3 Arduinos, analogues for the system which would ultimately do data acquisition and control.
Working day…
And night…
Once the software was squared off, I proceeded to disassemble the test stand I built for electric aircraft engine testing, and paint it white (mainly for rust-proofing but also style)!
A key component in the test system is a manually controlled override controller. Best not to trust Windows to safety critical systems… Best to built those from scratch and have ultimate control.
This unit, dubbed HACS (Human-controlled Analog Control Station) has (left to right) a:
- Controller Voltmeter
- LCD screen for status
- Guarded ARM switch
- STOP button
- STATUS LED
- Fire suppression toggle switch
- (Later added: RSSI meter)
The primary frame was 3D printed out of PLA, and then Aluminum was used extensively to beef up the structure.
By the end of the 1-day build process (spent all night printing and soldering)…
It was done! Pretty isn’t she? I eventually decided to scrub the LCD screen for time.
Initially tested with a Raspberry Pi…
Eventually a LoRa controller established the wireless connection between the main test control board (ADCS) and the HACS.
Awwwww yeaaa! Gorgeous!
Further electronics engineering ensued – mainly to drive the 4 relays required for the system (more on those later), as well as instrumentation level shifting, acquisition, and logging.
This alien-like structure was the optimal shape for fitting the Pi3 and Feather LoRa into the ADCS box (seen in a previous image)
Picked up this massive case at AllElectronics for housing electronics and protecting them nearby the test stand.
Of course, had to paint it white!
Installed the instrumentation interface, as well as the high power connector version 1. A more water-safe version of this was later designed.
Here the test stand is in it’s “version 1” test configuration. The HVAC tubing acts well as a flame duct.
This is the fire suppression hose. A solenoid controlled by a 12V relay, which in turn is switched by a high power transistor, in turn switched by the primary control microcontroller. Two hose nozzles pointing in the key directions provide water suppression. The only worry, is that in a power outage, all fire suppression is lost (a problem fixed later on)
Initial tests
To shakedown the system and help out a friend, Joe Barnard of BPS.space, I ran 4 Estes F15-0 solid motors, mainly looking for thrust profiles.
A series of videos [1, 2, 3, 4] on my YouTube channel capture the system in use. Data from these tests can be found for free in the description box of each video.
A robust series of procedures and checklists were developed to operate the system, and I particularly trained all emergency procedures prior to each test. Safety trumped literally each and every decision. With that said, iteration between tests took approximately 1-2 days (a problem fixed later).
These shots show the GUI and HACS working together, as well as the work station / mission control after one of the solid motor tests.
I photographed and inspected each fuel grain prior to each test.
Ignitor in place (ignitor holder designed by Joe Barnard)
I love seeing the charred ignitor holders… They’ve been through hell and back!
Immediately after these tests I joined my family on a vacation in Yellowstone. Some pretty pictures! This turned out to be a good time to consider what to revamp in the test development system.
Rocket Development System Version 2.0
After about 3 weeks developing GFOLD Python, I got back to the improvements in the rocket development system.
Chiefly:
- Redesign the test GUI for ease of use and simplicity. (video)
- Clean up the wiring and make all critical system wires low strain and high reliability
- Add a secondary electrical bus (in addition to the main 12V supply)
- Implement current shunts to measure the current load on the 12V electrical buses
- Redesign the power interface, and generally all interfaces in the outdoors to be waterproof or at least water resistant
- Build a backup, battery powered fire suppression system (dubbed “F2”), in addition to the primary one. This system MUST be 100% seperate from the primary system
- Separate laptop for actuation, on the more reliable Ubuntu Linux
- Separate, isolated network for control signals
- Separate power supply (battery)
- Build a nitrous oxide tank instrumentation unit
- Pressure instrumentation
- Temperature instrumentation
- Wireless to reduce number of extraneous connections and allow movement of nitrous tank.
Redesigned power interface, with a rubber seal and properly oriented power terminals.
The redesigned instrumentation interface. Green chips are load cell amplifiers, blue chip is the venerable Arduino Mega.
Used a 100% waterproof circular pin connector for the updated cable of wires going to the test stand.
The IU (Instrumentation Unit), name borrowed from the Saturn V, houses all on-test-stand electronics in a water-resistant housing
This is the nitrous oxide instrumentation unit. Inside is a thermocouple amplifier, plus two level shifters (one up, one down) for converting everything into usable voltages for the 3.3V Adafruit WICED board. It sits in a custom 3D printed housing with a LiPo battery underneath.
Some physical upgrades were also made, such as using U-braces for the cross-brace element.
Also, dug the flame duct into the ground for easier test stand securing and transport.
In one shot, this is the test stand hardware during a night test of a hybrid rocket motor.
