Servo-Ball Valve

For the 2022-2023 rocketry season, we decided to transition away from the pyro valve and instead use a servo-actuated ball valve.

CSI has a poor history with servo-actuated ball valves. For the 2018-2019 rocket, they experienced a failure to launch due to 3D-printed gears seizing. This led to a transition to a pyro valve for the 2021-2022 rocket, which had a successful static fire, however, the pyro valve took significantly longer to set up compared to a ball valve system.

In addition, the servo-ball valve system provided improved flow characteristics. The pyro valve functioned by using a piston to obstruct flow. Once it was time to launch, they used gunpowder to fire the piston forward, obstructing the flow with a smaller area and thus enabling flow from the inlet to the outlet. However, the pyro valve had a head loss of 100 PSI and thus reduced overall thrust. A ball valve, on the other hand, provides a linear channel for fluid flow and is the industry standard for these applications.

As such, this year leadership decided to move forward with a servo-actuated ball valve, however, with new measures to ensure smooth actuation at cryogenic temperatures.

The first step was buying a Swagelok 3/4″ ball valve. Then, I machined a motor horn that attached to the ball valve axel using a large bolt.

The motor horn is offset from normal by 45 degrees to prevent binding during opening/closing. However, we then ran into a minor setback. Using a 3D-printed housing designed by another team member that attaches the servo to the ball valve, the torque from the servo was transferred into the case rather than the linkage.

Solving the housing problem required going in an entirely different direction. Because of the 90-degree internal corners, machining the existing housing design was impossible with our tools. However, using a waterjet it was entirely possible to create thin aluminum plates with the required geometry. Then, using bolts and spacers to connect the plates creates a rigid housing.

Waterjetted housing for ball valve.

After initial tests with the actuation assembly (which I sadly cannot find footage of), we decided to move with this design, except with a larger motor.

The next step was upgrading the torque on our ball valve. I chose a 125kg/cm servo from Amazon that provided ample torque for the ball valve. Furthermore, after testing the v1.0 design we became concerned with airframe tolerance. Since we were constrained to an OD of 6″, the horn on the servo would collide with the airframe at full extension. To offset this risk, I revamped the design to have the servo sit lower. While this required remanufacturing several components, I think it demonstrates how rocketry design requires knowledge of several subsystems to fulfill the team’s goals.

In tandem with the redesign, we successfully validated that the servo linkage design was viable using the existing housing. Not only did we achieve successful actuation several times, but we did it extremely quickly.

The next step is assembling the linkage and motor housing using my brand-new design, then testing it in a static fire!

To iterate on the housing design, I made some design improvements. The current housing, when mounted on the rocket, would collide with the airframe during actuation. Improving this required redesigning the motor horn and housing to sit lower down.

Valve horn v2 fresh from the mill
Housing 2.0 assembled

Housing 2.0 required a through hole for the axle on the ball valve horn, so I added extra bracing on the side for security. Additionally, while the servo isn’t touching the ball valve the gap was reduced to accommodate airframe tolerance. The housing also more easily permits access to the Swagelok interface for mounting the completed assembly on the rocket.

Completed fluids stack with housing 2.0

The next crucible was testing the new system at -40C. As I mentioned in the beginning, the reason for transitioning to the new linkage system was a past failure in actuation. Pictured below, the legacy system using 3D-printed resin gears. When the ball valve reached -40C, the gears froze together and it failed to actuate, forcing the 2019 team to scrub the launch.

Pictured: Rocket co-lead Ryan Wu comparing the legacy system (designed for a 4.5″ ox tank) to the 2022-23 valve.
Fluids system at -40C! Almost the entire tank was covered in frost. Although the servo got quite cold, it never failed to actuate during our testing regime.

The next step was validating the system under launch conditions! Our valve cam captured great footage of the servo operating during our April 26th static fire. All footage is courtesy of Jorge Casas, Rockets Co-Lead. The audio is extremely loud so please be mindful.