Low-jitter Reaction Wheel Prototype for Kestrel Eye

The Army is developing the Kestel Eye imaging microsatellite to provide ground imagery directly to the warfighter.    The goal of the program is to provide tactical grade images to forces on the ground at any time and deliver the images fast enough for use in fast moving ground operations. The satellite will provide battlespace awareness for rapidly evolving tactical situations on the ground, for example: the implanting of Improvised Explosive Devices (IEDs); perimeter security of forward operating locations; or movement of hostile motorized forces.

Princeton Satellite Systems is under contract to develop a control system to  meet the exacting standards of stability, satellite location, and pointing accuracy required to meet the needs of the Kestrel Eye satellite.  The objective of our work at PSS is to improve the pointing accuracy or the ground location accuracy of the Kestrel Eye imagery from 60 meters to 10 meters or less.

The features of the proposed control system that are critical to enabling this superb accuracy are:

• Ultra-precise star image centroiding with custom algorithms
• Miniature precision fiber-optic gyro for attitude base and high bandwidth control
• Low-jitter microsatellite reaction wheels utilizing Halbach array motors
• Nonlinear attitude filters incorporating star camera and nontraditional measurements • Composite structure to eliminate thermal distortion
• GPS orbit determination enhanced with two-way ranging

Recently, PSS has completed the design and fabrication of the first prototype reaction wheel.  The wheel is driven by a low-jitter axial flux brushless DC motor, the design of which is currently under patent review.   An important enabling technology is the Halbach array of magnets.  A Halbach array is sequence of permanent magnet segments, each with its magnetic axis rotated from the axis of its neighbor.  The resulting assembly concentrates almost all of the magnetic field on one side, with an almost negligible field on the other side.  This arrangement favors an axial flux motor with a single stationary stator holding coil windings sandwiched between two permanent magnet rotors, each of which has its Halbach field directed toward the stator.  The sketch shows the arrangement.  The stator is green, and the two rotors are red.


We’ve gone through a number design iterations,  settled on a first prototype design, and fabricated it.  We also purchased a simple general-purpose motor driver in order to explore the operation of the motor before moving on to developing custom driver electronics.

We’re very pleased that our first iteration works.  Here’s a video showing the device in action.

We’re already at work on the second-generation wheel incorporating lessons learned in the first prototype.

NEA Scout Toolbox

Near-Earth Asteroid Scout, or NEA Scout is a exciting new NASA mission to map an asteroid and achieve several technological firsts, including being the first CubeSat to reach an asteroid and demonstrate CubeSat technologies in deep space. http://www.nasa.gov/content/nea-scout


NEA Scout will perform a survey of an asteroid using a CubeSat and solar sail propulsion and gather a wide range of scientific data. NEA Scout will be launched on the first Space Launch System (SLS) launch.

NASA asked Princeton Satellite Systems to develop custom MATLAB software based on the Princeton Satellite Systems Spacecraft Control Toolbox and Solar Sail Module to assist with this mission. We just delivered our first software release to NASA!

The NEA Scout module provides MATLAB scripts that simulate the spacecraft. One, TrajectorySimulation, simulates just the trajectory. It includes a solar sail force model and uses the JPL Ephemerides to compute the gravitational forces on the sail. In addition it can use a 150 x 150 Lunar Gravity model during lunar flybys. It also simulates the orbit dynamics of the target asteroid.

AttitudeSimulation expands on this script. It adds attitude, power and thermal dynamics to the model. A full Attitude Control System (ACS) is included. This ACS uses reaction wheels and optionally cold gas thrusters for control. Momentum unloading can be done with the thrusters our using NASA’s Active Mass Translation (AMT) system that moves one part of the CubeSat relative to the other to adjust the center-of-mass so that it aligns with the system center-of-pressure or adds a slight offset to unload momentum. The control system reads command lists that allows the ACS to perform attitude maneuvers, do orbit changes with thrusters and for the user to change parameters during simulations. It adds the rotational dynamics of the asteroid.

The dynamics of the AMT can be modeled either with a lag on the position or a full multi-body model. Dynamics of the reaction wheels, including a friction model, are included in the simulation. The following are a few figures from a typical simulation.

The first figure shows reaction wheel torques during attitude maneuvers. The ACS uses quaternions as its attitude reference. You can mix reaction wheels and thrusters or use either by themselves for attitude control.


This GUI shows the current command and allows you to control the simulation.


The Figure GUI lists all figures generated by the simulation. It makes it easy to find plots when you have many, as you do in the attitude simulation.


The Telemetry GUI gives you telemetry from the ACS system. You can easily add more data to the telemetry GUI which can have multiple pages.


This figure shows solar sail pointing during simulations.


The following figure shows the spacecraft with its solar sail deployed. This is built in the CAD script using the  Spacecraft Control Toolbox CAD functions. The sail is 83 meters square.


The sail is huge but the core spacecraft would sit comfortably on your desk.

If you want more information about our products or our customization services you can email us directly by clicking  Mission Simulation Tools.

NIAC Orientation

I had a great time at the NIAC orientation in Washington DC last week, where I got “mugged” with program manager Jason Derleth:

Stephanie Thomas and Jason Derleth posing with a NIAC mug

Stephanie receiving her NIAC mug from Jason

The meeting was at the Museum of the American Indian, which was a great venue with so much beautiful art to see, and a cafe featuring unusual native foods from across America (elderberry sauce on the salmon). I had the opportunity to meet the other NIAC Fellows, and put names and faces to the other creative projects selected, as well as meet the illustrious NIAC external council. These experienced folks provide advice and encouragement throughout the NIAC process from their experience as physicists, engineers, biologists, science hackers, and even science fiction authors.

I have to say, my poster on the fusion rocket engine was popular, and everyone wanted to know how it works, why it hasn’t been funded already, and how soon the engine can be ready. Of course, we have yet to actually demonstrate fusion using Dr. Cohen’s heating method, but that is why we need the NIAC study – to flesh out the science and engineering of the rocket application to help bring in funding for building the next generation machine. And yes, let’s get to Pluto in only 4 years the next time! I’m really looking forward to working on the project in the next few months and presenting it at the NIAC symposium in August!


PSS NASA project on SciShow

We just discovered that our NASA NIAC project on the DFD mission to Pluto was covered in a SciShow episode from June 14, 2016.

Hank Green does a great job talking about our project, and I love that he called it a “Pluto Explorer”, which rolls of the tongue better than “Pluto Orbiter and Lander”. However, he did get our fuel wrong: we are using deuterium and Helium-3, a reaction which produces no damaging neutrons. Hank cited “two types of heavy hydrogen”, which would imply deuterium-tritium fusion; this produces most of its every in very damaging neutrons, and is a reaction we go to great lengths to avoid in our machine. There will always be some tritium produced from the side reactions of deuterium with itself, but our machine is designed to exhaust it before it can fuse.

The comments from the viewers were interesting, including several along the lines of, “wait, did I miss fusion becoming a working technology?” Of course the fusion rocket is still theoretical, but it’s based on a real plasma heating experiment going on now at Princeton Plasma Physics Lab! And its true that many people don’t realize that fusion itself has been achieved in many machines, just not break-even fusion. Our machine is very different from the large tokamaks most people are familiar with.

Lunar Orbit Insertion Maneuver

New functions in the Lunar Cube module in 2016.1 allow you to easily plan lunar insertion and orbit change maneuvers. In the following pictures you can see a lunar orbit insertion from a hyperbolic orbit. In all figures the lunar terrain is exaggerated by a factor of 10.


The same maneuver looking down on the orbit plane. The green arrows are the force vectors.


The following figure shows a two maneuver sequence. The first puts the spacecraft into an elliptical orbit. The second circularizes the orbit.


Lunar Cube Module for 2016.1

We are adding the Lunar Cube Module in 2016.1 to our CubeSat Toolbox for MATLAB! It allows users to analyze and simulateCubeSats in lunar transfer and lunar orbit. It includes a new dynamical model for CubeSats that includes:

  • Earth, Moon and Sun gravity based on the JPL ephemerides
  • Spherical harmonic lunar gravity model
  • Reaction wheels
  • Thrusters
  • Power generation from solar panels
  • Battery energy storage
  • Variable mass due to fuel consumption
  • Solar pressure disturbances
  • Lunar topographic model
  • New graphics functions for lunar orbit operations
  • Lunar targeting function
  • Lunar mission control function for attitude control and orbit control

The module includes a script with a simulation of a 6U Cubesat leaving Earth orbit and reaching the moon. The following figure shows the Earth to Moon trajectory.


This figure shows the transfer orbit near the moon. The lunar topography is exaggerated by a factor of 10 to make it visible. It is based on Clementine measurements.


Here are results from the new LunarTargeting function. It finds optimal transfers to lunar orbits. The first shows the transfer path to the Moon’s sphere of influence.


The next shows the lunar hyperbolic orbit. In this case the transfer is into a high inclination lunar orbit.


Contact us for more information!

Reaction Wheel CAD Project


I’m a sophomore at MIT who joined PSS as an extern over Independent Activities Period (IAP). Free to choose how to spend the month of January, students can take an extended vacation,  attend short, intensive classes, do research in MIT’s various labs, etc. Many like myself choose to participate in short internships with MIT alumni – the correct lingo for this type of job experience is “externship”.

I was assigned the task of 3D modeling a reaction wheel for a 25 kg satellite. Essentially, the wheel controls the orientation of the satellite in space. Comprised of a small axial flux motor and a flywheel for added inertia, the wheel sits at 40 mm tall and 80 mm wide. It must spin in both directions, and meet tight dimensional constraints. I believed I really had my work cut out for me.

Continue reading

Pluto Spacecraft

Here is a picture of our DFD transfer vehicle. You can see the lander on the front and two Deep Space Optical Communication System (DSOC) assemblies mounted on trusses. There are 2 DFD engines.


A picture of the Pluto Lander. The solar panels are illuminated by a laser from the orbiter. The lander has a dry mass of 150 kg.


Both were designed in the Spacecraft Control Toolbox v2015.2.

You can get more information about the Pluto orbital mission on Slideshare.

Lunar Topography

If you are sending a spacecraft to the moon, you will be interested in lunar topography. A new function in the Spacecraft Control Toolbox lets you superimpose a height map onto any sphere.

MoonTopoThe function RSHMoon.m gives you the Clementine spacecraft topographic data using a spherical harmonic expansion of the rangefinder data.


A new function, PlanetWithTerrain.m, lets you superimpose this data onto a sphere.


Continue reading