 Welcome to stage C. Before we start, I just would like to remind you this is entirely run by volunteers and we need more volunteers, most notably to run the audio and video system, to run the bar, to run the car park. There are plenty of things that needs to be done and we need more people for this. So you should really go to volunteer.emfchem.org or to the volunteer tent behind you. But for now, I will leave you with Strato-Binson who is going to talk to you about satellites. Good morning. This is the story behind $50 sat, which became the world's smallest satellite and by quite a long margin the world's cheapest satellite. November 21st 2013, Dombrovsky in Russia. That's a Dnieper ballistic missile. Apparently after the Cold War the Russians found they had plenty of these spare and they've been repurposing them for launching small satellites. They're proving alarmingly reliable. Right, that was launched at 7.10 in the morning after two orbits and about two hours later in my garden shed. That was the first message picked up. That's the telemetry from the satellite and that is basically telling me that it's all okay. The battery is okay. That's the figure at the end, three, seven, two, three. And it was all running well. And at that point I'd have to say that the entire team were quite amazed. This was really a very basic satellite. The original title, one of the titles I was going to give for this talk was building complex satellites is easy but building simple ones is not. And there's an element of truth in that because if you have to build a satellite there is a tendency to go overboard and introduce too much complex tech. There are surprisingly quite a lot of off the shelf components for small satellites you can buy. Solar panels, power units, radio units and all this. But this satellite was very much smaller than anything that had been launched before. So there was nothing available and we had to start from scratch. And rather controversially at the time, certainly in the amateur radio world we decided to go really basic. What was the minimum we could get away with and prove that there's such a small satellite would actually work? Now I have a working model here. That is how big it is. That operated in orbit far longer than we expected. It operated for 20 months. We only expected it to work about one month, the main issue being the battery. Right. And as an example, one of the reasons for giving this talk was I wanted to explain that even something like a satellite, you don't need a lot of complex stuff, equipment to build, design and make it work. It was done in a garden shed. That's quite true. And the only complex bit of equipment I used was an oscilloscope. All the rest was using the natural environment around us. Okay. This was the first of a new type of satellite called pocket cubes, pioneered by Professor Twiggs in America, Bob Twiggs. He'd also come up with the CubeSat standard some years ago. Those are 10 centimetre cubes. But he decided to do something smaller. And he called them pocket cubes. And these were based on a 50 mil cube size. So that's two inches. So he started a project to do this. And the idea was that they would be very simple to build. We'd use off the shelf components where possible. Perhaps the only exception would be the solar panels. And the construction had to be within the remit of your average school or college. And for that reason, they decided to use a pickaxe microcontroller. Some of you might be familiar with. They're very popular in schools in this country for teaching technology. And they're also very popular in America. And part of the reason why I became involved in this project as effectively as a consultant was I found myself as being the only person in the world who had actually connected a pickaxe to the radio module we were going to use. And the radio module was going to be an RFM 22B. Now, I'll show you one later. But these things are about the size of a 10 P piece. You can get them for about two pounds. But bear in mind that we're working with a 50 mil cube size. So it's got to be really small. Now, the minimum distance we needed out of this for a radio was 600 kilometers. Now, when we started, we said, well, if we can just hear a Morse code beacon, that would be fine. Now, we had a very good idea that that would work. Because there are satellites in orbit using such low powers, which you can hear. So we were fairly confident we would hear it. But we also needed there was another another requirement we put upon ourselves. We had to be able to communicate with it two way. And the reason for this is in theory, if you launch a satellite, you're supposed to be able to turn it off if you're asked. Now that requires two way comms. Now that was the tricky bit. Because the manufacturers of this device hope they reckon that it was good for a kilometer. We wanted 600. Eventually, we got about 2,200 out of it, various means. This device is frequency agile, essential for us because we would you would be allocated a frequency to run on in the case of $50 at 437, 437.505 megahertz. We could set it to any frequency light. There are very few other radios that will fit in a 50 mil cube. There is one now, somebody has built a radio for a pocket cube, and it costs about one and a half grand. This costs about two pounds. Right, that's the development model. I had to put together a working example of this using a code, using a code I developed. And that was how it was proved, solar panel, et cetera, et cetera. And that sat in my garden for quite some time. But what we needed to do was prove that the radio would work, that we would get two way communications. Now you can read the data sheet for as long as you like on this device, but it won't tell you anything really. Because those are based, the figures they give are lab conditions. We wanted to know how it performed in practice in the real world. So I live in Cardiff, South Wales, lots of mountains and hills. So we started using the hills in the area to do radio tests. That is in the far distance in the middle of the picture, it's Cofillie common, just between Cardiff and Cofillie. And I'm on Machen mountain. And one day I left my brother on the farm mountain with a pole and a radio transmitter on the top. And I measured the signals coming back from eight kilometers. And surprise, surprise, it worked. We got about eight kilometers and we were only using about 10 milliwatts. But still, we needed 6700. That's the route there on the map, Cofillie common, the green at the top of the picture. Very fortunate in Cardiff in that there is a hill to the north of Cardiff. And across the channel, there is another hill in the mendips. And this forms a nice 40 kilometer path where there is almost nothing in the way. So it's true line of sight. And we set about measuring what would happen. That was the route. It was a long drive for me. But we were able to pick up the signals from 40 kilometers away at 100 milliwatts. Now still, we're nowhere near where we want to be. Because we said we needed 600 kilometers. But there is something else involved. For anyone who knows the locality, that little blue line there, that is where some guy called Marconi was doing some radio testing in 1897. Quite famous, I believe. Right. I fed the results of that test hilltop to hilltop, back to Howie, colleague in America. And I don't expect you to read all that. I read it and it gave me brain hurt. But that's what he came up with. But the net result was that for the uplink, if we wanted to talk to the satellite in bold at the bottom, we needed 36 watts of power into a 10 dB beam. Now a 10 dB beam is something like one of these. This is an arrow antenna, satellite antenna. So using that and a 36 watt amplifier, I built my own. This thing here cost about 50 or 60 quid to build. We would be able to talk to our satellite. And that's all we wanted to know, really. As it happens, in the end, we found that coms was better than expected. And I once was able to send a message to a $50 satin that responded using 25 watts. And it was 2,200 kilometers away. Right. Now the team involved at this point, I was still asking, I was still being involved as a consultant. Effectively advising the students over there on the radio and the pickaxe and what have you, and doing the radio testing. About, I can't remember the exact date, somewhere around September 2012, Professor Twig said, well, do you want a launch slot? And I know that there is the famous saying, there's no such thing as a free launch. But in this case, it was true. And he said, well, why don't you build one? A pocket cube. Oh, yeah, of course. So the three of us involved got together and we decided to very quickly come up with the design for three people. I'm Michael Kirkhart. He's an electronics engineer. I'm the radio operator for many years in America. And the other one is how he did police. I'm not entirely sure what he does. He was prepared to go on up to working on a very famous airplane in America, perhaps the most famous jet airline jumbo there is. He did the comms on that. But the rest of the stuff he gets up to, he said he couldn't talk about. There was me. I notionally work in computer technical support these days. But I am an electronic engineer by background. I was involved in the manufacture of things like the micros and spectrums and Archimedes machines in the 80s. So I kept a very interest. And of course, there's Professor Robert Twiggs. He's the Professor of Space Science at Moehead University in the States. Right, design of $50 that. Having been given a free launch, we said, what are we going to do? I mentioned earlier, we could have gone nuts and put lots of fancy tech in. That's sort of satisfying in a way. But we took exactly the opposite view. We wanted to do the absolute minimum possible to the satellite. The minimum amount of components, less components, less to go wrong. And we had a philosophy. You simply cannot add simple. The more you add, the more complex it becomes. So we kept it really, really basic. We wanted to do the minimum possible to prove that the concept worked. Because this was entirely new. These small satellites had never been tried before. So we took it upon ourselves, apart from what other people were doing, to try and prove that the idea would work. There were four other pocket cubes launched at the same time as us. T-Logo Cube, which survived until the following end of January. Cube Scout and one called Ren. Those two never went operational. The success rate of amateur satellites is quite poor. But we really, really tried to make sure that we didn't do anything that would compromise the satellite working. We chose to use 40 mill boards. That's quite small. The idea there was that there would be two boards inside $50 at. One the processor and radio. Another would control the solar panels. But we needed to have space inside to allow wires to be fitted and whatever. So we said from the start, 40 mill boards. We were using a Pickaxe 40x2. It's still the only one that's ever been in space. There's likely to be another one fairly shortly, but I can't say much about that. The RFM-22B is the radio chip. Again, this had never been in space before. We needed to add watchdog and latch up protection. The reason here is that allegedly satellites, especially ones that have got virtually no screening, which this doesn't, are affected by particles in space and it causes them to crash and corrupt things and all sorts of nasties. And there's another problem is that if a particle goes through a mosque at exactly the wrong point, it causes it to turn on and latch up in a high current state. And the only way out of that to rescue your satellite is to power it down. So we built in a very simple bit of circuitry, which every so often would power the satellite down, but every two hours or so. But we could also, if we detected an anomaly, like the current was higher than expected, more of that later, we could force a shutdown. And that would disconnect the power for about half a second. And that's enough to clear that problem. I have to say we didn't have any problem with that at all. Nothing happened. Solar power, four sets of tasks, solar cells. These, if you look at them very carefully, you'll see the little tiny triangles. These are offcuts from circular, silicon panels that are made. They cut off the edges and they send the square ones up into on real satellites, if you like. But they sell these little offcuts. Each of those little triangles is about three dollars each. The state of the art, solar panels, solar cells, at the time about 28% efficient, which is massively more efficient than the average thing you see on a house roof or whatever. But bear in mind the area we had for solar panels was very, very limited. It's important to track the voltage point on the solar cell to get the maximum amount of energy out of it. Because at high current, low voltage, high voltage, low current, somewhere in between there's a happy medium. And you can then match the solar cell to the load in an efficient way. So we had a couple of circuit LTC 3105s doing this for each bank of solar panels. We had a current monitor on there and that was part of the protection. We needed to know how much current was flowing, both to the solar panels and through to the $50 set itself. That's the battery and that's the radio. The battery is the sort of thing you'd buy in eBay. I think we did actually. There are about £2. A little tiny thing. Now I'd have to say we didn't expect it to last very long. One of the problems was it was going to be very cold, more of that later. But it was also, it would go through an enormous number of charge cycles, which is the problem. We'll work it out. One charge cycle every orbit, 13 orbits a day, 20 months. I did write it down. I forgot what the card was. So we wanted to know, we could prove it would work because when it was launched you could hear it. But we also wanted to know what was happening. So the things we were monitoring were the voltages, battery, solar cells. We wanted to know how well the battery was doing. The currents, we needed to know how much power it was consuming, how much charge current we were getting from the solar cells. EEPROM corruption. Again this was something we put in because we were told it would be a problem. So every sort of transmission cycle, which is about a minute, I put in a check, which did a check some on the contents of an area of EEPROM to see if there was ever any corruption. There wasn't in 20 months. We also did the same thing on the flash, which is the core, where the core program is kept on the TICACS. Again we were told there was possible potential for corruption there from part of some space. There wasn't. We didn't detect any. We also did, I'd also did the same thing on a spare area of memory. Now this we warned was perhaps the biggest problem that RAM could get corrupted because again very little shielding and that you know you could get a bit flip on an area of RAM and it would cause problems to your program. We thought we detected one in 20 months, but we were never sure. But it carried on working anyway. There's a technical problem with the RFM-22. It goes into a, it's got a circular smart reset which is anything but and it causes the whole thing to latch up. So you have to detect that and correct for it. What communications did we have from $50 sat? Well most call signs, the whole, the thing was would start and it would transmit call signs in order. Mine first and then the other three people in the project and then $50 sat's own call sign. The Morse was encoded in a way that you could tell, you could tell how well the battery was charged in case we couldn't get any of the other comms back. We had fast Morse data. Now this is really old-fashioned stuff, but 120 words per minute Morse, fast Morse, it was decoded and it, we were sending the solar cell voltage and the battery voltage and that. So anyone with an average hand receiver could have picked that up. Then there was the data telemetry from the RFM-22 itself. Now that we were very dubious about whether we would get back. As it happens we did just about. The limit was about 800 kilometers. And for that I used an antenna like this. I'm sure you recognize that as a quadrophila helix. It's made from plumbing pipe and bits of barbecue stand wire and it works extremely well on satellites. Plus you have to use a fairly sophisticated low noise amplifier. There's one here when I built myself. You can get it from a kit and you build it in the box. But with that we were able to get the telemetry back from, but it was created in kilometers. But the main thing that made 50 last such a success was we used FSK Ritty. FSK Ritty is an old style comms method used by Hams from years back. 100 board, but that could be received with the right amateur radio receiver when as soon as the radio, 50 last that came over the horizon, the radio horizon, that's about two and a half thousand kilometers away. And that's what made the project such a success. Those are prototype PCBs. The radio, the radio's got the programming connector. The other radio with no connections is the solar charge board. This is the build of the box. Again the build had to be really simple. Something you could do in the average workshop. So it was basically just bits of machined aluminium. You know, cut aluminium, drill, whatever. The only critical part of the construction was the base. That had to be accurate, nothing else did. That's the boards just before they were connected in place. That's the antenna end. Yes we really did use tape measure. That's the inside. There's a very good reason for tape measure. I'll show you in a second. Those wires are PTFE. You can't use PVC wire in space. It falls apart. Those are solar panels. That's it sort of more or less. Yes there is really tape measure. That's the inside. A lot of space really. That's it in its launch tube being shaken to bits because obviously rockets vibrate a lot. And that's it in a vacuum chamber because vacuums do strange things to plastics. And that is not our satellite but T-logo cube being launched into the mother satellite, which was called Unisat 5. And it's from there that the satellites were launched. Now the reason for the tape measure was if I mentioned there was on the vibration test it was in a launch tube and the launch tube was quite simple. All that happened was a door opened at the right time and a spring at the bottom of the launch tube which was square pushed the satellites out and of course as soon as our satellite was pushed out, out springs the antenna. So simple but effective. The satellite behind us had a more complex deployment mechanism and we think it didn't work. That's why they never heard from it. Meanwhile I had to get I had to organize the uplink and it was basically all home-built kit. The aluminium box is an RFM-22 receiver I made. You've seen the amplifier. That's a funky dongle. I used that to listen to it and at the right point I could press a button on the receiver and it would send the telemetry message, the command, up to fifty dollars that to do whatever it was expected to do and that worked quite well. What we go wrong? Well we did get something wrong. I'm afraid. The battery. This was a major concern. One of the problems was that there are no, at the time, and I think it's still true, there were no space rated batteries. There isn't. There are larger ones for CubeSats, lithium polymer packs you can buy, but nothing in our size. So that was literally was an off-the-shelf camera battery. So we didn't expect it to last very long and part of the reason was that those types of lithium battery you can't charge below zero. Now the satellite, I'll mention it again later, was to spend a lot of time below zero. There were some software errors. Cockups are my part mainly. In particular we were unable to discern between, because of a mistake I'd made, we were unable to tell the difference between a planned power down and an accidental power down, although we were able to work it out eventually by looking at the occurrence of resets because the satellite was sending out a constant count of the number of resets it had gone through. Spin fading. Anyone who ever listened to the FSK RIT you will have noticed it fading in and out. The reason for this is the satellite was spinning about once every two seconds. Now we hadn't thought too much about that. We did have a magnet inside the satellite so that it was actually orbiting at a constant angle but we didn't think to put one in to correct the spin. So but it wasn't a huge problem. Time below nought degree centigrade. This was a big surprise. Cube sats, the ones 10 centimeter cube things, tend to stabilize around zero because at one instance they're at minus 30 and then they're in minus plus 30 whatever. So they tend to stabilize internally at about nought degrees somewhere around there, five degrees something like that. Our satellite was so small at such little mass that it was recorded. Every orbit it went down to minus 30 in the dark and up to plus 30 in the daylight. And this actually meant that because of the battery charging problem we didn't have that long to be able to charge it. About half the time we really thought because we had to wait. There was a temperature sensor on the board and we had to wait for the battery to warm up before attempting to charge it. So in reality when it was coming north over the UK it was probably in sunlight for about 15 minutes or 20 minutes before it started to charge. We didn't point the solar panels at the Sun and this was something we worked out eventually while we were puzzling as to why we won't get as much charge current as we thought. And what was happening is although there was a magnet that was stabilizing it in attitude as it went around the Earth, what was happening is the Sun was over there say and we should have stabilized it so that it went around that way. You know rather than going around orbiting that way it should have been done that way. And we hadn't thought about it. What we got right. Well the down thing worked. We did keep it extremely simple. You really could build one of these in your own garden shed. All the stuff is open source and public. If you want to look at the hardware designs or the software or whatever that's all online. Just look up Google $50 app. The amount of testing. Extraordinary amounts of testing on the radio setup. You'd be amazed. The use of FSK RITI. That had a dramatic effect because that was so easy to use. And we proved that the digital telemetry from RFM 22 was viable. Thanks to some of the people involved. In particular Professor Twiggs and the Gauss Group University of Roma. At the bottom some of the amateur enthusiasts who filled in reception reports for the FSK RITI. Now this originally this slide was finished my presentation. And it was quite it's quite ironic that yes it is the moon. But what I wasn't to know is that you might be aware that there is now a radio, an amateur radio project to send a pocket cube, not a pocket cube, a CubeSat to the moon. So it's quite prescient. But after this I did get involved in something and this I wanted to take the time to mention it because it's worthwhile. And it's to do with Laura. Now Laura is a new radio technology released in about 2013. It's used as spread spectrum which is not new. But it's the first time it's been done in a low-cost module. It's about the same size as the radio we were using and about the same price. However it operates a 20 dB below noise. That doesn't mean a lot to most people I guess. But it means it's very long range. And I first found this out when because I'd done so much testing on IFM-22. I set up an IFM-22 with low power on Tenby Beach and I lost the signal 100 meters away. I lost the Laura signal, same power, a kilometer away. It was capable of covering 10 times the distance on the same power. So I repeated the test across the channel and that again proved that. What had taken the radio in $50 had taken 100 milliwatts to cover 40 kilometers. The Laura device covered in two milliwatts. And I then decided that the best way of testing the long distance comms is to put a little tiny tracker on a foil party balloon. There's foil party balloon. And this has shown that Laura produces, Laura will cover at 10 milliwatts 400 kilometers, 1500 bits per second. That's quite fast. You would get, if you use the full 100 milliwatts the device can produce. You get three times the range and you get nine times the range at 100 bits per second. Now the net effect of all that is someone's going to do this in space very shortly I believe. You're talking about being able to receive digital telemetry from a satellite in orbit using a little receiver like that you can build in a lunchbox. Because the range should be about 4,000 kilometers or so. That's all. We haven't got time for questions. If you want to find anything more about Laura you can look at that website at the bottom. If you want to have a look at the model I've got the satellite or ask me a question. I'm sure you'll see me walking around. I'm the guy with the satellite in his pocket. Thank you very much.