Okay. Our next talk will be from Ilyas and this is the first of a dual of talks that
we'll have on a similar topic, so you are in for a treat on the future of space flight
in open source. Thank you.
Thank you Ilyas.
So my name is Ilyas. I am a core contributor at the Libre Space Foundation and my job
title is doing space stuff. So before we move on, let's get the audience up to speed.
In the title, I think there are like two words that may be unfamiliar. So does anyone know
what a pocket cube is? Okay. A pocket cube is a really small satellite. It's like five
centimeters, a cube of five centimeters size. So it's this. It's really small. And the one
P, the one P is the one unit for pocket cube. So this is one P. If you go to two P, you extend
one time the size three P, you extend two times the size four P, it makes no sense. You just go to
a bigger satellite. Do you know what a deployer is? Most of you. Okay. So rocket people do not like
to put your satellite with duct tape on the rocket. So they want a fancy box to put everything in.
That's a deployer. And the last term, do you know what Leo is? Leo is low earth orbit. Start from
around 200, 300 kilometers and go up to like 1200. So this is the terminology. Let's start our story.
So this is the story of cubic. So it was, I think, three years ago, summer time. And we got a phone
call with, it was actually an email. And they told us, you know, we have a spot on the rocket in a
deployer for one P satellite. And it's free. And this matters because the cost for launching one of
these could be like 15,000 years, maybe more. So that's a great opportunity. And we said yes. And
they said, okay, you have a few months to get it ready to go. Okay. So let's see what we have. We
looked around on the space staff we have. And we have a comms board. That's important. Because
like a basic satellite, you need something to communicate with. And you need someone, some
source of power. And that's what we had. So it was quite capable. It can do all this modulation
and stuff. I won't bore you with it. But it was a starting point. It needed a bit of tweaking.
And that's it. So where are we going? They told us you're going to an orbit of 300 kilometers.
That's nice. For the weight of these satellites, you expect to be up there for one to two weeks. Good
enough. And it was the first flight of the rocket. Not good enough. So we came up with the plan. There
was lots of hardware that had to be done. So we had to test the comms board we had. And then we
needed a way to add power to it. And even it would be even nicer to be able to not just have a
battery that will die and have some solar panels to get extra power. And then all these had to be
somehow bind together to form a satellite. And then put all this on a plate that can go into the
deployer because there are some kind of specifications that require you to have like a specific
shape that goes in there. And then there were some testings where you put this in an oven and some
materials that may evaporate. They do that in your oven and not on the lens of the satellite next to
you that messes everything up. And then there is some campaigns where you torture your satellite and
shake it and hit it. So you make sure that nothing will come loose during the launch. And then you
just send it to the rocket people and you have pizza. On the software side there was some software
that could do communications but it needed to be tested and evolve a bit. And then once we were
happy with what we have we flashed the final firmware in the satellite and then we have pizza. And
on the bureaucracy department we had to coordinate frequencies. So what's that? Let's say you have a
radio and you decide that you duck deep push the button and then you put a tape recorder next to it
and say I am Joe. I am Joe every second. And then you just pick a frequency. Let's say I just put my
birthday in. And you just start transmitting. And then you get some kind of military band and local
people around the country don't like it. Imagine that but the whole planet will not like it. So you
need to get a frequency allocated to you. And then there are some managerial staff for exporting
this and what is this and what is that. And customs people are not happy and then you have to make
them happy with lots of paper work. They love paper work. And then have pizza. So all this is good. We
can do it or we think we can do it. But why? So what should this satellite do? And we realize that
there is some kind of a problem that we could try to address. So imagine this. When you, your
satellite and many other satellites and for this size it could be like 50 or 100 go out in space. It's
the equivalent of going on the top of a hill, have like a jar of marbles, just throw it down hills, go
to the next city and try to pick one which is yours. So there are raiders for this thing and there are
like military services that do track all this stuff. But during the first weeks it's like a blob of
stuff. So in order to be able to communicate you need to think. You need to know which one is yours
and you need to know where is it. So the first part is quite easy. Because you just wait for the
lands, satellites go out and then you wait, you look up and at some point you say, you pick. Hey,
it's me. Nice. Can you hear me? Hey, it's me. Okay, where are you? Hey, it's me. Can you tell me your
position? Hey, it's me. And then no contact. Luckily the play me part, it's identifiable. Because you
would have some kind of ID beacon, whatever that you design. So you know that's, that's yours. The other
part not so easy. But we thought that if we exploit the Doppler effect, because these things go really
fast and really fast is like seven to eight kilometers per second and to give you like a sense of how
fast this is, I will go to the center of the city and come back. So I'm going and back that fast. So the
Doppler shift that happens to the transmission from the satellite can help you start identifying the
orbit. There are some, we call them orbital elements which define the orbit. I won't bore you with that. But
there is a way that you can do that. So that's the experiment. So on the hardware, we started designing a
power system. And so we got really popular solar harvesting chip that doesn't PPT that's the SPV 1040 with
some solar panels. We also added the battery management chip. So it actually it's a battery goes so knows how
many. Can't go in goes out. And you can get a good sense of what's going on your power system. Some minor
modifications to the comms, the communication board. And then we we designed all the structure that will keep
everything together. So again, the power budget. We wrote a tool to do that because commercial ones are not open
and they're very expensive. So this is what this tool produces. You just enter how much solar panels you have. And
then you get a really nice plot, which is for each side of the satellite, how much power you're getting. And this is
awesome. But the thing is when your satellite is kicked out of the box, because it is kind of mounted on the bottom,
just flips around. So the reality is something like that.
So the structure, that's a structure. We designed this
in the concept of having the PCBs that mount the solar panels and everything to be structural element. And this is an
exploded view. And this is how the systems fit in there.
There was a lot of room because we can have like four PCBs in there based based on the standard we tried to follow. But we
only had like a power board and the comms board. So we had two batteries just to be on the safe side. And then just to
add weight. So the more heavy you are, the longer you stay up there. And the more you pay to go over there.
So there's a ballast board, which has some weight to like reach the top weight you can have on these things. The
antenna on the bottom, it's a measuring tape. It's really good on unfolding itself back to shape, whatever you do to it.
And there are actual numbers on that, as you can see. Actually, let me pass this around.
So the antenna, we did some simulation for the antenna and the radiation pattern. On the right hand side, you can see how
this thing is tied down to the satellite in the stored configuration.
And then we get another goal. And they say, you know what? We have an extra slot for you and say, okay,
instead of building one, we just build two. That's great. And then there's the thing with a deployer where you were going in.
So can you build it for us?
So we said yes for some reason and we came up with the revised plan. So the revised plan also includes the deployer. I
won't go into that because that's the next talk. But this was the birth of PicoBus, the deployer we built. So you have to wait
15 minutes to hear about this.
Moving on. We had all the circuits and the PCBs ready. You need to add some kind of conformal coating to protect the stuff.
So you spray this and then you inspect it with UV light and it glows and where it doesn't glow, you have to apply more.
Quite simply. This is an almost finished structure. It's a photo from the, during the assembly.
And these are the two assembled satellites ready to go. So in the ballast board where it's nothing, we thought we could go like,
let's put some ideas in there instead. So this is a small board with like the four principles that we believe in regarding space and the openness in space.
We moved on to the bakeout procedure. So that's like
upside down jar that goes on the vacuum machine thingy and it goes to really, really big, low vacuum.
And then there's like a flat light on the right that does infrared
heating of everything there. And what you do is you measure the mass before you bake it.
You measure the mass after and you have to be within some specs or something really important evaporated in there. So it's an ago.
And the final step was to do the proto flat campaign. This is where you put everything in the deployer and then you just torture it.
And it hope, you hope that when you open the deploy again, you don't get sand coming out.
On the software side, we built some drivers for the hardware as a standalone project so it can be used for other things.
We built the elementary and telecommand. There was a finite state machine for orchestrating of what the satellite should do.
And then because there was a delay on the launch, the software people, we had like more time in our hands. So we decided, okay, let's add another project.
So there was a new project called open space data link protocol. And it's a way to structure your data and telemetry that goes up and out to the satellite.
And we also built some kind of ground station telecommand software to operate the whole thing.
Another interesting aspect of the development was that while the hardware was ready and the software was written,
we were using the actual ground station software that was going to be used during the mission, which was part of satnav.
So during development, we had like a satnav dashboard that would give us the state of the satellite. So everything is good and nice because it's next to us.
But the good thing is that this is ready. So when satellite goes up, you just reopen this and you have your data.
So these are like the final steps, final firmware is going in. Everything goes into the deploy for the last time.
And then you wait for the launch. The launch provider was firefly aerospace.
Like the dream payload was like the mission we were invited to join.
And a few months has passed and there's the moment that you actually wait to happen.
And it's a very stressful moment because sometimes from firefly things can go to fireball.
And that's what happened.
As you can imagine, there were feelings.
And the only thing you can do since you know that since the thing blew up, they will probably going to build another.
So we build more cubics.
And a few months later or almost a year later, here we go again, biting nails, hadging teddy bears.
And then you get this picture.
So that's our deployer in the other satellites.
Firefly did a good job to interrupt the stream just before we were deployed and then do a playback but not say that it is a playback.
So they said all payloads are deployed and you get a video with this thing closed.
And they said, okay, what happened? But there's a switch there that says, you know, the door has opened.
So hopefully things go good and they did.
So we started receiving telemetry from the satellites.
It's cubic three and four because one and two kind of disintegrated or something.
There was a minor issue with the A2C bus on one of the satellites, but the reset solved this.
It's not a really nice way to do that. It's kind of windows you like, but at least it worked.
We attempted some telecommand and control.
The command was received.
We wasn't able to get the reply.
And the telemetry was received by the satellite network.
So these are like real space data here.
The power system kind of overworked.
So the battery was like continuously full, which is good.
So what do we get out of this?
The platform, the cubic platform was a success.
It is considered now Terral 9.
Terral 9 is like the view of the space and it works.
So it's like the top level. There is no pen there.
So that's good.
Firefly did not reach the target orbit.
So this affected the mission life because it was reduced to three to four days.
We managed to do orbit determination.
But since the orbit was decaying really fast, it was unusable.
The information was there, but you could not verify it because the next orbit was totally messed up.
And this gave life to four more projects.
PicoBus.
See, look, you should listen about that in 1630, which is exactly what the experiment is,
but in a more commercial way.
And we have the simulator and the space data protocol.
So the platform is one P-Pocket Cube bus, the one that's going around.
If you use one battery and you do not use a ballast, you have room for one or two payloads to put in there.
You get 350 to 500 milliwatts of generation for your staff.
There's battery monitor and management.
We have a really good documentation.
It's currently coming together from the various wiki pages and issues, but it's really detailed.
And it's a cost-effective solution for research, education, radio amateur or whatever you can imagine.
So what's the future?
We need to move comps to like version one and call it that.
Create a better power board, finalize a standard and then create a document about it.
And go bigger and fly more cubics.
This is all the software that we used. It's all open software.
And these are the people that helped this thing to become reality.
Thank you very much.
So if there is time for questions.
Yeah, yeah, I have minutes for questions.
Five minutes for questions.
How many ground stations and passages do you need to like estimate the orbital elements of the satellite?
So how many ground stations and passages you need to determine an orbit?
Okay.
Obviously the more the better, but with two or three ground stations and a single passage, you get a really good estimate.
And then with the following ones, you just kind of nail it actually.
So the other results, actually you can get more info on the SID look, talk about it, which is quite extensive.
Yeah.
Since you're using CCSDS, you also know the pain of CCSDS, I hope.
Have you thought about looking at other like open source space protocols that are already out there instead of writing a new?
Actually it is based.
So the question is whether we thought for existing space protocols instead of implementing a new one.
The CCSDS provides some compatibility with commercial equipment.
So we thought it was a good idea to go that way.
So there is an open solution that exists that is compatible with this kind of protocol.
Okay. Thank you very much for an amazing talk.
Tiedelini.
Oh, you've got it. Nice.
If I can ask one last question, or at least one more in any case.
Did you have any considerations for radiation effects in your design of the hardware and any mitigations in that sense?
Okay. So the question is whether we had any consideration for the radiation effects on the hardware.
When you fly what we call commercial of the self components, you know that they're not space hardened.
So there is always a chance for something going wrong.
But you can go around by designing in a like more clever way.
For example, MOSFETs may latch themselves.
But if you have a monitoring circuit that detects that and your power cycle and I think most of the times you get you can get out of this situation.
For the software, which is also quite important.
There was like a triple storage technique implemented for all the variables and information in the satellite.
So also there was an ECC RAM used from the microcontroller.
And there was a polling.
So you read three values and you choose like the two that match.
And this is it.
And there is also scrubbing frequently scrubbing where you read everything and you rewrite it.
So even there is an error, you have corrected it and then put the correct information in all places.
Okay. So illness will be available for questions outside.
Of course.
We have three more minutes.
Great.
Excellent.
Then maybe maybe maybe two more questions we had.
We had one right before.
It's all the way.
Yeah, but the ground stations.
Did you build your own ground stations or there's like a network that you can use or how this works because it looks like the most expensive part of the project is like you have to grab, you know, whole planet.
Okay.
So the question is, did we build our own ground station or did we use the session network?
The simple answer is sat mugs.
The more detailed answer is that we actually our first really first project was a ground station network.
It now has around the five, five hundred stations around the world.
It's operated by contributors.
So we have like a really good coverage in the UJF band and also on upper bands too.
Hello.
Thank you for the talk.
Two brief questions.
Did you have an attitude control system for the cubic or it was not needed and same thing for thermal management was important or in the design.
Okay.
So the question is, did we have an attitude control system and was there any thermal management?
So the first question is no, we just toss it out there and let it spin.
It was not important for the experiment.
So and also there was no time as you might have realized for the thermal management.
There were some tests in vacuum chambers and we saw that it would survive without needing like an extra provision for that.
Of course, the PCB design is built in a way where you dissipate all the heat from the components to the PCB.
But that's kind of best practice anyway.
And the temperatures we got from the telemetry were actually well within the automotive limit.
So we were good.
Okay.
Thank you.
Any additional questions we can go on the hallway.
Thank you.