Permalink
Fetching contributors…
Cannot retrieve contributors at this time
273 lines (258 sloc) 13.3 KB
                     ______________________

                      NOTES FROM OSCW 2018

                           Hugh Brown
                     ______________________

Table of Contents


1 Overall 2 [2018-09-24 Mon] 3 [2018-09-25 Tue]

1 Overall

  • Many (not all) of the presentations about actual cubesat projects could be best described as "case studies" -- mention (often in passing) of having used open source tools for design, simulation, etc
    • Heavy representation from university projects; next biggest contingents were (I think) community projects (SatNOGS, etc) or space agencies (ESA, NASA)
    • Would be great to see more companies presenting or even just attending (which I've passed on to organizers), for a change of perspective; still learning the field, though, and perhaps there simply aren't that many that are interested
  • Two presentations of visualization tools for space missions (OpenMCT
    • SatNOGS) -- very good to see this, as the focus on UX/UI is often a weak point in Open Source and makes a huge difference for ease of adoption
  • As designs for antenna progress, it will be worth talking to SatNOGS folks about designs/instructions for easy-to-build Phase 4 antenna that can be put into SatNOGS ground stations
    • One interesting possibility might be to market Phase4 radio designs (or better yet, cubesat parts as Bruce has suggested) + SatNOGS ground station as a standard open source ground station for satellite telemetry:
      • visualization via SatNOGS DB
      • telemetry/instrument data via DVB
      • perhaps with built-in amateur radio functionality (whatever the next generation of current AMSAT FM repeaters might look like) in order to get amateur radio community interested as well (though this may need to be balanced against worries of amateur radio folks appearing as scribes/assistants)
      • Note: C&C not part of SatNOGS; they've got some stubs for the functionality, but I'm not sure if this is a priority for them
  • Still a lot of open questions about physical/mechanical/electrical standards (Hyper-cube, pc104); this may complicate ability to provide a board/plan that's anything like "this size fits most"
    • Note: I'm not entirely sure if this has been a priority for the project. My knowledge of mechanical/electrical engineering is incredibly limited, and I have a lot of questions.
  • Some knowledge about Phase 4 Ground project, but not a huge amount
    • This may have been due to the attendees (skewed heavily toward academic institutions, researchers) and/or my choice of who i talked to
    • Might be worth investing in stickers or some such for phase 4 ground/space; I'd be happy to finance this
    • Problems with GNURadio came up a couple of times during the conference; Michelle's offer to help with this may be an excellent way to increase awareness

2 [2018-09-24 Mon]

  • Intro from Artur and Manthos
    • "We need lots of people involved in open source and space -- and not only space engineers."
      • \o/ there's hope for me yet
    • Two open access journals, Software X and Hardware X (both from Elsevier :O) offering to publish papers from the workshops free of charge
    • Maybe 20% at this workshop were here last year
    • Mention of sdrmaker.space ([https://sdrmaker.space/2018/08/sdrmakerspace-initial/]), partnership between LibreSpace and ESA to fund work on using SDRs for satellite communications
  • Bruce on open source
    • On manufacturing cubesats for cheap: "Why use a fake satellite in your classroom? A university can afford a real one."
  • Formation flying talk
    • Calibration of telescopes very important -- they're sensitive enough that systematic effects are the biggest source of uncertainty.
    • There are natural sources that have well-known intensities, but that's not true of polarization -- and that's what they want to mention.
      • That's where the cubesat comes in
    • Used open source tools for design
      • OCDT: client/server; Concorde for concurrent design; ESA standard for workstation communication? ESA models for different submodules?
        • [https://ocdt.esa.int]: "OCDT is a client / server software package developed under a European Space Agency (ESA) contract to enable efficient multi-disciplinary concurrent engineering of space systems in the early life cycle phases." Avail only to people at institutions in ESA member states.
  • MM wave polarizatoin (Mansur Tisaev)
    • Another open source case study (OS tools used for design, antenna simulation)
  • Break
    • Artur: continuing conversation with ESA (ESAC?) director about open source. No, he won't require it, but can still prompt.
  • Cansat
    • Model rocket payload -- excellent introduction to working with cubesats
    • cansat.eu: ESA website
    • open-cosmos.org
  • Ilias from SatNOGS: pq9ish
    • Wasn't satisfied with RS485 bus in pq9, so designed with CANBus: gets to treat data like packets; mult-imaster message , message notification comes free; would be really hard to implement this on RS485 & would have to be done from scratch
    • used STM32 ARM chip; heritage from UPSat
    • Q: launch options to LEO? A: Piggyback on cubesat, or use cubesat as deployer
  • SatNOGS update
    • They are doing a lot of work to "close the loop" for telemetry from satellites: not only to collect raw radio signals, but to decode them to data frames and from there to telemetry
    • This includes beginning work on graphical display of satellite telemetry (dashboard showing temperature, battery levels, etc over time); still work in promise, but looks amazing
      • Based on Grafana for display, kaitai ([https://kaitai.io]) parsing
      • This depends on knowing the data format for a given cubesat -- this can be reverse engineered in a pinch, but immensely easier (and much more likely to happen) if the project actually publishes details
      • Q: what do you need from satellite operators to be able to decode the data frames? A: Publish encoding and data spec -- otherwise we have to reverse engineer
  • TUDSat
    • We've had a lot of problems with GNURadio - lots of installaton probs, bugs, changes between versions, etc
    • Q to audience: what has your experience w/GNU radio been? Response (two people): Stick with GNURadio; some friction but once that's done everything is in one codebase.
  • MAGIC:
    • open source connection mainly sharing scientific data from instrument?
    • Talked to Dr. Chiara Palla, who made this presentation; she created the MAGIC instrument and helped integrate it into another team's cubesat
      • That integration took a lot of work: there were some things (software and hardware, IIUC) that were not well-specified, and a fair amount of iteration to get things sorted
      • I'm unsure if this is case for standardization, or for better comms between teams
  • ESA Summer of Code in Space: Red (moderator) is one of the organizers of this
    • didn't happen this year (timing? funding?)
    • goal is to have this not only in ESA but other places too
  • Python workshop
    • Focused on pros and cons of using Python, particularly for astronomy/space data
    • Included discussion of ease of use, which is a concern that extends beyond Python
      • Docker is a good solution, but you can't count on other people having it.
        • This was quite surprising to me; in current job/industry, if you don't know Docker it's expected you'll learn at least basic use of it quickly. Another (welcome!) reminder that not everyone has the same experience/imperatives you do.

3 [2018-09-25 Tue]

  • Conversation with Milenko Starik from TUDSat; asked for guidance re: troubleshooting unexpected performance problems with LimeSDR. Pointed him to Michelle, who was able to offer guidance
  • Conversation with Argentina Radio Club about their project to build up cubesat interest in their country, and about Phase 4 Ground
  • Team Anant
    • open source cubesat project from India
    • Using Zboard for POC, PetaLinux (prov. by vendor) for distro
    • Zboard too big for flight; wll use same SoC on final board
  • Poquet Cube - Nepal
    • Another vote for cansat for training
  • OpenMCT
    • NASA mission control dashboard; used by lightsail 2 among others
    • flexible enough to be used by lots of things
  • Ruediger Gad: use of non-space components
    • used message-oriented middleware (openwire, stomp) to translate between ie java programs used by esa and vr libraries written in latest hotness
    • looks like a very interesting way to do this kind of translation; good to keep in back pocket
  • Break
    • talked to Xabier Crespo Álvarez about idea of analyzing telemetry from SatNOGS DB to find signs of impending failure; should be good to use as first training set; volunteered to help with setting up pipeline, etc
  • Orekit
    • most cubesat projects don't need a lot of flight dynamics analysis; "as long as it does not fall...", and as long as you've got screen to display beautiful keplerian trajectory for VIPs
    • "clearly TLEs are not an option for high accuracy positioning"; i had no idea
  • Marcin Stolarski, hyper-sat.com
    • Had chance to talk with him afterward, so notes integrated here
    • New standard he's building to handle case when cubesats aren't big enough, which can happen more ofen than you think
    • new standard for larger-than-cubesat: 350mm x 350mm x (100-600)mm
    • what does this do to launch costs?
      • don't forget, launch cost will include PPOD as well -- so yes, more mass, but total cost ~ the same
      • not that much bigger (as far as volume goes); yes, will still fit on RocketLabs' Electron
      • a lot of engineering effort can go into miniaturizing instruments so that they'll fit into a cubesat; this is a standard that has more space, so you don't have to spend that time/money
      • aimed at startups; "here's 10 million, show me your proof of concept in a year."
    • why proprietary launch ring?
      • there were standard options but they're more expensive; will need diff adapters for diff rockets
    • about two years for docs; why?
      • work still in progress, heavy iteration. next step is to build test version, and will probably adjust standard when that happens (lessons learned); can email him for standards in meantime
  • Daina Bourquin, Harvard Library
    • talked to her about metadata, citations for software
    • tools:
      • zenudo: cern project to archive (and create DOI for) github repo via webhooks when you cut a rlease
      • codemeta
      • reprozip
      • software citation implementation working group
      • software presentation network
    • Her challenge: make it easy for teams to do the right thing re: archiving
    • why satnogs? because it's a challenging audience! if i can serve & make things easy for a dynamic community that does not have this in mind, then the rest is easy.
  • pc104 workshop/librecube:
    • organizer's intention was to keep this focused on certain pin arrangements in header; that didn't happen
    • problems with pc104:
      • how to make flatsat with pc104? have to make daughterboard; this is not how it will fly; assembly is a big part of testing how you fly, and can't do that & still have easy access to boards. switching to backplane would fix this
        • yes, would still have to test for insert cycles, thermal, vibration -- but this would be a great place to start.
      • know what else sucks about pc104? slow, annoying serial communications. ethernet would be great for this.
        • need higher data rates: 10mbit from an instrument, 2Gbit download (these are real examples); well beyond what pc104 can support.
      • pc104 takes up a lot of vertical room; going custom also allows you to save a lot of space.
      • what we have, has lasted us 10 years. we should be developing for next ten years, not optimizing a very old standard.
    • but pc104 is what people are using and selling, it's not obsolete. and physical support from header is an important benefit.
    • failure rate on cubesats. any ideas?
      • radios are a problem -- ask any ham why using power meter chips as alternative to AX.25 is a good thing (not sure i understood this point; Bruce was the one who made it, so it might be good to ask him)
      • also, i2c can lock bus -- and if you don't have good, well-tested watchdog, you have now just frozen your sat.