oresat-adcs-hardware/ACS Capstone 2018

OreSat ACS Driver Board

Introduction

The OreSat Project

OreSat is an inter university collaboration to design an orbital vehicle. Once in orbit, the nanosatellite will engage in three missions. One of the missions is space testing a relatively new photovoltaic panel. The other missions require the ability to point at and maintain visual contact with a specified location.

Background Motivation

In order to successfully complete these missions, OreSat will need to be able to articulate its orientation as well as its passive “drifting” spin (we’ll discuss what “passive spin” means a little later). Quick and reliable orientation adjustment and control will serve several purposes. It will allow the vehicle to direct its main camera, as well as giving it the ability to execute maneuvers that allow it to maintain visual and radio contact with a specific location. This makes taking photos and videos much easier, but it also allows us to have a segment of time to do large bandwidth data dumps to ground stations. This gives it the potential to do live video streaming.

Active Pointing

The method that was selected for active orientation on OreSat employs the use of internal spinning masses called reaction wheels. The basic principle is simple. If there is a free floating object and a section of it starts to spin, the rest of the body will react by spinning in the opposite direction. The spin rates are related, but not necessarily the same. If you’d like to look more into this, the concept is conservation of angular momentum. So the idea is that we have a pointing vector for Oresat (not poynting, that comes later), and an idea of the pointing vector to the location we would like to image. We simply take the difference, and then rotate the collective mass of all of the reaction wheels in the opposite direction until the satellite camera sees what we want it too.

Passive Spin Control

The term “passive spin” will be used to refer the net angular momentum of the nanosatellite. While the reaction wheels are active, they are able to control the direction and spin of the frame of the satellite, but only by effectively redistributing the momentum from the frame into the reaction wheel masses. The reaction wheels are incapable of altering the net angular momentum. The need for being able to modify the net momentum vector of the satellite helps alleviate the burden of two issues. The issues are thermal regulation and power generation. OreSat only has photovoltaic panels of its elongated faces. To maximize the amount of power, the desired passive spin of the satellite would put the long axis perpendicular to the sun. This maximizes the surface of exposure. There is another part to the total momentum though, and that is its magnitude. The rate of spin effects several things, but most notably it affects the temperature gradient experience by OreSat. If the satellite has no spin, then the side that the sun is hitting will heat continuously until OreSat passes behind the earth. This can cause several issues, most simply that having a hot spot would increase the likelihood of an electrical system to failing. Another issue is that the hot spot will cause different amounts of expansion in different areas of the frame, increasing the likelihood of failure. Another issue that hot spots cause is a drop in efficiency of the solar panels, making rotation partially an energy problem as well. Adding an intentional spin to OreSat will even out the thermal load applied to

The attitude control system (ACS) inside oresat needs to orientate the position of the cubesat in space. This will be acheived in two ways; With brushless DC motors (BLDC) and with Magnetorquers. We are designing a PCB that controls both systems.

Project Overview

Block Diagram

ACS Level 2 Block Diagram Version 3.2

System Block	Hardware Chosen and Reasons Why
Buck Converter	TPS63070 We chose this part because of oresat heritage. This component is being used elsewhere in oresat and we were recommended to use it. It is a buck-boost converter that can take a wide range of input voltage. Overall, it is a robust component.
BLDC Motor	What motor???
BLDC Motor Driver	STSPIN 230 We chose this component for a number of reasons. Namely; - Satisfies thermal and voltage constraints. - Small form-factor. - Half-H-Bridge control over each phase of the BLDC. - Easy interface with STM32F0. - Full Control over Phase Output. - Minimal External Components.
Encoder	AS5047P The AS5047P utilizes an array of Hall effect sensors and a diametric magnet to provide a precise position reading. The sensor provides sufficient precision and provides output data in multiple convenient forms. See the datasheet for details.
MCU F0	STM32F042K6 We chose this MCU because of oresat heritage. It is being used elsewhere in oresat. This MCU has a small form factor and is power efficient.
CAN Transceiver	TCAN330 We were told to use this component because of oresat design heritage. The rest of oresat is using this CAN transceiver.
Brushed DC Motor Driver	STSPIN250 We chose this component for a number of reasons. Namely: - Satisfy thermal and voltage constraints - Small package size - Full H-Bridge control - Basic Magnitude and Phase control - Minimal External Components
Magnetorquer	Still to be determined. The magnetorquer will essentially be an inductor most likely in a flat spiral orientation to save volume inside oresat.

Schematic

Rev1.1:

• The “flatsat” version that we manufactured and tested.

Rev2:

• Minor changes. Rev2 was never manufactured. Skip and go reference Rev3.

Rev3:

• Rev3 was manufactured and tested. Note: R206 must be changed from 100kΩ to 150kΩ.

Board Layout

• The design of the board layout was a significant hardware challenge due to the physical space constraints in the cubesat. Our entire system needed to fit within a volume of 10x10x4 cm. Our system needs to drive four BLDC motors and three magnetorquers. Additionally, each motor needs an encoder positioned right next to it in line with the motor’s axle.
• Our proposed solution was to design a board that contains all the hardware needed to drive a single BLDC motor, a single magnetorquer, run the encoder, and also run the CAN communication. We’d use four of these boards to run all four BLDC motors and also the three magnetorquers. Having a single board for each BLDC motor allows us to mount the encoder directly on the board and then mount the board directly in line with the motor’s axle.
• The volume constraints drove us to create our board layouts as compact as possible.

REV1:
The goal of REV1 was to make a very compact board as a prototype so that we could get a good idea of how much area our board will take up. REV1 ended up with an area of 42x43mm. We passed this information onto the structure team so they could start prototyping how to fit everything into the cubesat frame.

Rev1: Top Rev1: Bottom

Rev1.1:
The goal of REV1.1 was to test all the hardware components and prove that they would work when all integrated together. REV1 was manufactured and assembled. All the hardware worked as expected with only minor design errors discovered.

Rev1.1: Top Rev1.1: Bottom

Rev2:
The goal of REV2 was to get a head start on developing better packing geometries that utilized both top and bottom layers for component placement. Upon reviewing the design for developing a mechanical mount, it was determined that with how effectivly the space was used, size should be deprioritized in favor of how it was to be mounted. This layout was then abandoned due to the amount of modification necisary for the mounting system that was developed shortly after.

Rev2: Top Rev2: Bottom

Rev3:
We received input from the structural team which informed us that we need to redesign the shape of the board and change the locations where thermal contacts occurs. The total area of REV3 came out to be a very small 38x35mm.

Rev3: Top Rev3: Bottom

CAD photos for satellite mounting:

Board Stuffing SoP

Both Rev1.1 and Rev3 can have the top side fully stuffed and baked in the reflow oven.

1. Prepare all tool and supplies: Tweezers, solder-paste, napkins, surface mount components, schematic and board layout references.
2. Prepare solder-paste stencil: Place the bare PCB on the table and place the L-shaped supports around it. The L-shaped supports should come with the solder stencil. Tape the supports to the table so they do not slide around. Now align the solder stencil on top of the board and tap the stencil down to the L-shaped supports so that the stencil does not slide around. Apply solder-paste and squeegee/scrape off the excess.
3. Organize surface mount components in whatever is most convenient.
4. Place components one by one on the board with tweezers by referencing the board layout. This is the super long tedious part. Be patient. It’s better to work slow than to mess up. You don’t want to clean a pasty mess.
5. After all the components are placed, carefully remove the board and inspect everything visually under a microscope. Make small alignment fixes where necessary.
6. Carefully place the board in the reflow over and follow reflow procedure.
7. Remove board (CAUTION - HOT). Inspect visually with the microscope. Look for solder bridging. Reflow and remove solder where necessary. Pro tip 1: look at tricky components at an angle so you can see underneath them better. Pro tip 2: use the Metcal solder station. It dumps heat very quickly and regulates its temperature well.
8. Proceed to stuff the bottom of the board by hand with the solder iron. Pro tip 3: don’t use solder irons with tiny tiny tips. The tips rarely get hot enough to melt solder properly. Pro tip 4: For soldering surface mount resistors and capacitors; tin one pad with solder and leave the other one clean. Bring the component close with tweezers in one hand and with the other hand use the iron to reflow the solder while aligning the component into its place. Solder the other pad once the component is being held in its proper place. A similar method should be used for an mcu. Only tin one pad. Aling and place the mcu. That one pad will hold it aligned in place. Then solder the rest of the pins. Don’t worry if you bridge. The Metcal is good at heating up braid for removing solder bridging.
9. Congrats! You are done. Now you can move on to testing and debugging.