Video Jon Oxer Talks About the ArduSats That are On the Way to ISS (Video) 17
Oxer : I am here to talk about ArduSat which is an open source satellite project that is going to be launched in about a week and a half. Satellite technology is something that is normally considered to be way out there, it is something that you can’t participate in, it is something that NASA does, big budgets and all that sort of thing. And the ArduSat project aims to bring that down to a hobbyist level, make it accessible even to students, and to make the design of a satellite as open source as possible. And that is not just so that people can see how the satellite works and understand it which is obviously a really important element of open source, but also to collaborate and participate in it, and get their hands on the hardware themselves, and actually do experiments themselves at home or in the school classroom.
Tim: Now, ArduSat is based on the CubeSat standards. And the CubeSat is a standard that anybody can look up, it is freely available, but the guts of it, it differs by being open source in what sort of way?
Jon: Okay. The CubeSat standard is defined to try to reduce a reinvention of the wheel basically. Rather than having every satellite as a one-off design with a whole lot of supporting infrastructure around it, not just the satellite but the deployment mechanism and everything else to try to allow people to focus on just the particular requirements for their mission. So the payload that goes within it. So a lot of CubeSats use things like commercial off the shelf flight computers, power supply and various other things. And then just focus on having the particular experiment payload as open source or a design for that particular mission. For ArduSat-1 in order to reduce risk most of the infrastructure is using proprietary parts, things like the flight computer and the power supply. My objective over time is to progressively replace all of those subsystems so that the entire satellite stack becomes open, both in terms of the hardware design and also in terms of the software that runs on it.
Tim: So for this first launch, what have you been able to actually effect that replacement?
Jon: At the moment, the payload itself that runs the experiments is going to be an open design. We have a couple of parts to it - there is the main processor module which has essentially 16 Arduinos on it, and each one of those can be running experiments in parallel. There are a whole bunch of sensors on the satellite and then each of those nodes can be running an experiment defined by a hobbyist or whoever wants to get involved, and talking to those sensors and storing the data that comes back from them. So at the moment, it is just the part that executes the experiments – it is not the rest of the satellite.
Tim: Talk about your background in hardware that allows you to actually execute this type of project.
Jon: Yeah sure. My background way back in school and just after school was actually in hardware. Then I ended up going off and spending about 20 odd years working purely in software. It was about seven or eight years ago that I became more interested in hardware again. I started playing around at the hobbyist level, came along to OSCON and did a couple of talks about hardware hacking for software developers, really trying to show software people that it is not that hard to play around with hardware. I did some demonstrations using Arduino, and it really took off from there.
Tim: Speaking of the hardware, you got some on the table. Would you mind picking up a few and show us what you’ve got there?
Jon: Yeah sure. Absolutely. I’ll start with showing you some crude things just to show what can be done relatively low tech. Now what this does is illustrate the format of a CubeSat. It is a fairly standardized format. It has to be within a 10 cm cube and has to weigh less than 1.3 kilos. So there are standardized mounting holes that are used and then boards that stack on top of each other. So this is a fairly typical system. This is the CubeSat bus and it is a stacking connector so you can have multiple modules sitting on top of each other.
What I’ve done here is designed a prototyping module – so rather than being for a specific purpose it is just a big open grid of holes with power rails and various breakouts and things. So you can put whatever parts you like on it. This is great for ground prototyping just testing out ideas on the workbench. So what I’ve done is used one prototyping module to make a really crude power supply - it is basically a battery pack and the regulator. And then another one that has some sensors and a little Arduino compatible board in it; so this is called a LeoStick, it is basically an Arduino Leonardo but in a memory stick form factor. So I just mounted one of those on there. There is a real time clock and some sensors, accelerometer and a couple of radios.
Tim: That’s a one-off?
Jon: So this is a one-off. This is a proof of concept, the sort of thing that hobbyists can put together for relatively little money just on the workbench. Obviously, you are not going to fly that in space. But then there are a number of other boards, for example, the next level of sophistication is a board which is the ArduSat Arduino Adapter Module – it allows you to take a regular Arduino and it is basically like an upside down shield so you can plug an Arduino into it and your Arduino can become part of the satellite stack. Which is a very neat way of doing development. Now beyond that, I’ve designed a couple of boards – this particular one is the first prototype of the payload processor module that is in ArduSat-1, and each one of these little rectangles is essentially a complete Arduino Uno. And then up here in the corner that is the equivalent of an Arduino Mega. So this is the supervisor node that controls all 16 of these other Arduinos.
Tim: How much memory does that have altogether?
Jon: It is the same basically as an Arduino. So it has only got 2.5 k of RAM in each of these MCUs, so they are very limited. And the limitations you work with are the same as if you had an Arduino sitting on your bench at home – it is just that this is 17 Arduinos on one board. So it is very compact. And the next generation beyond that there are a few flaws on that design I have here. And it is mounted to a power supply module so you can’t really see into this too well, but that is a power supply module which is not going to be used on ArduSat-1 but the idea is that this could be used in something like a high altitude balloon flight or in bench testing. It means that instead of having to put in a $7000 flight ready power supply into your little test stack, for a couple of hundred dollars you can have something that is functionally equivalent but not space rated. And you can build up a stack that is really neat and will behave just like a regular satellite but without the big dollars.
Tim: Will people be able to order this from your company Freetronics?
Jon: Yeah, that’s right. There are a couple of issues that we have which are very interesting in terms of the open source aspect of this. And one of the problems is regulations in relation to distribution of designs or hardware for space technology. There is a regulation called ITAR which is International Traffic in Arms Regulations which restrict our ability to share designs. So one of the challenges we have to overcome at the moment is how do we make this open source and publish the designs without running foul of ITAR regulations.
Tim: What do you do?
Jon: Well, there are a couple of things. One is to sell it as an educational device. Now, this board that is underneath, this is one of five boards in existence, and two of these are actually going into space in a couple of weeks’ time. So that is a fully space rated board. And that may have some issues in terms of being able to sell it directly. But what we can do is sell a variant of it as an educational board for ground use only. And that should get us around the ITAR regulations.
Tim: Where will the launch itself be from?
Jon: It is being done by JAXA which is the Japanese Space Agency, and it is being launched from the Tanegashima Launch Center which is in Japan. It will be going up on the HTV-4 mission which is being launched on August 4th.
Tim: Now you are in Australia. Is the entire effort an Australia based project?
Jon: No, in fact, I am the only Australian. It is a bit of an international effort. The main team is based in California, but they came together from various countries, because they made it at the International Space University while doing their master’s degree. And it is in California, where the main office is. The company behind it is NanoSatisfi.
Tim: It sounds like a lot of work.
Jon: Yeah, there is a lot of work. And I am only seeing a tiny fraction of it. It is not just the design of the hardware – there is a huge amount of infrastructure that goes in and around this. You have to obviously have the software to run on it, and you also need all of the infrastructure for managing things like uploads, so there’s communications with the satellite, how do you handle things like taking submissions from people about experiments to run on it and vet them, and then execute them on the satellite. So there are a lot of people doing a lot of work.
Tim: If people are interested and want to see what the open source specifications of these boards are, obviously it is hard to make one in your basement, where would people go to find out more?
Jon: Yeah, okay. If you go to ardusat.org, there is general information about the ArduSat project. If you go to freetronics.com/ardusat, you will find links to a couple of these boards, including downloadable files for the EAGLE projects that are used to design the PCBs. So you can see the design yourself.
Friday, August 8?? (Score:2)
Re: (Score:3)
Thanks for noticing the typo. Aug. 9.
Re: (Score:1)
70 second advertisement???? (Score:2)
No fscking way am I going to sit through that to watch -- if the past indicates the future -- a lame interview.
Re: (Score:2)
Re: (Score:2)
Eh? When I tried to watch, there was a 70 second YouTube-style advertisement, but without the ability to skip after 5 seconds.
Missed opportunity? (Score:4, Insightful)
But it's hard to understand why a 17-wide parallel configuration of 8-bit microprocessors each having just 2.5KB of RAM makes for a sensible satellite payload processor. Why not something with an architecture more like a Raspberry Pi or BeagleBoard? Not those specific boards necessarily, but a similar, simple one-chip SoC approach and a decent amount of memory. A processor like that could drive a bunch of experiments (more than you can fit in a Cubesat), and have enough room for the software to be comfortable and maybe even maintainable on-orbit.
A SoC-based system would fit in the same low cost profile but could run much more interesting applications. Ardusat feels like a missed opportunity, because it has lots of other things going for it: open source, submission process, international coalition, hobbyist/student focus, etc.
Re: (Score:2)
Simple microprocessors like this have an advantage of having large gates and therefore being more resistant to cosmic radiation. I'm actually pretty curious about how big the effect of this would be when jumping up to ARM Cortex-M3 and then to ARM11. Surely state-of-art cosmic technology uses more advanced CPUs than ATMega328, is it fine to just pop it in a better case, or is the effect not so strong in practice?
Regardless, AIUI, with something like ATMega328, you don't really need to worry much, maybe have
Re: (Score:2)
Failover is rarely "simple". There's a lot of code and mechanism, somewhere, to decide when a
How it works and what it is (Score:2)
Skeptical about radiation performance (Score:4, Interesting)
I have put an Arduino in space, btw, but only on a suborbital flight lasting about 20 minutes. Also, for what it's worth, IAAREE (I am a radiation effects engineer).
Re: (Score:2)