Working towards v2 of the Ground Station Gear Assembly, the SatNOGS hardware team has experimented with various gear designs. All of them where designed and 3D printed from scratch focusing on ease of reproduction and excellent mechanical operation and properties.
The new gear assembly (which is almost identical for both Azimuth and Altitude) sports a much thinner and compact design, driven by a much more powerful NEMA 17 Stepper motor.
Failing was part of the process. As you can see in the pics above, 3D printing a worm gear can be a monster of its own. After careful calibration and fine-tuning (layer heights, cooling, leveling, infill settings etc) on our 3D printer we were able to consistently reproduce worm and matching gears. We even experimented with acetone abs curing, which (based on the results we are seeing) we believe should be the default for our gears moving forward.
v2 of Ground Station is around the corner, and we will be posting later this week with full documentation, designs and instructions on how to build. Till then, back to gear-craze!
SatNOGS is focused on TX operations for Low Earth Orbit (LEO) satellites. Most satellites in those orbits transmit signals in relatively low power (compared to GSO satellites). 100mW to 2W is a typical range for LEO. Given that power output and our RX assembly (yagi + dvb dongle for reception) a Low Noise Amplifier can really improve our RX capabilities.
We decided that LNA would be an integral part of our RX assembly early on, but we could not easily find something that would meet our requirements (bands, noise figure, cost, size etc). After browsing and researching a lot of different options out there, we stumbled upon LNA4ALL.
LNA4ALL  is a great project by 9A4QV. The amplifier is built around Mini-Circuits PSA4-5043+ E-PHEMT Ultra Low noise MMIC amplifier operating from 50 MHz to 4 GHz. Small SOT-343 package combine low noise and high IP3 performance with internal match to 50 ohms. With 20 Euros as a pricetag, this was the LNA we were looking for.
We are using our 5V output from the regulator we have inside SatNOGS as VCC for the LNA (required a tiny bit of modification) and the LNA is connected in line between our RX dongle and the antenna (using SMA connectors)
Initial tests are showing great improvement in our reception (tested against a variety of bands, encoding and satellites) making this LNA an irreplaceable part of our project.
Proper shielding and housing should be in place, so we designed  and 3d printed quickly a housing for this LNA. Some grounded aluminum tape makes up for a shield.
More tests will follow, and we are designing new antennas and evaluating their matching.
SatNOGS as a project has been concieved as many Satellite Ground Station implementations (like the v0.1 which is ready) coupled together under a global Network that would enable anyone to utilize SatNOGS as a single platform for observations.
The UX idea is simple. An observer/astronomer/maker/hacker accesses our global Network via a web interface. Then she provides details about the observation that she would like to schedule like, which satellite, which band, what timeframe, which encoding etc. Then, having all the info , the system calculates the possible observation windows, on the available Ground Stations connected to the Network for the given timeframe (taking into account any tool, location and time constrains). Once the observer confirms the proposed “observation job” then it gets sent as a job to each Ground Station (GS) job queue to be executed when it has to.
GSs are gathering observations, decoding/recording them and sending them back to the Network, making them available to the observer (and the world!).
Here is a glimpse into the initial DB Schema of the Network (not all attributes are populated in the tables)
The general design of the Network part of SatNOGS constitutes a typical many-to-many scheduling problem (many observers to many ground stations) Interestingly enough projects have been developed for observations planning and scheduling on satellites like Hubble (HST) Projects like SPIKE  and ASPEN . Those are examples of many-to-one scheduling system s and we have been looking around for implementations that would be similar to our proposed one. Unfortunately we haven’t found anything closely related, thus we are building the Network part from scratch.
Given the expertise of our software people, we are building the application part on Django (python), and relying on rest APIs for communication with our ground stations. The first iteration of the network will be focused on a client-pull approach (versus a server-push) to eliminate any network restrictions.
Our challenges towards the global network are interesting. Given the data-heavy approach of the network (imagine all observations from all stations stored and indexed) we expect data storage to be a serious challenge. We are evaluating deep-storage as an approach to this, and we welcome any feedback or ideas around that.
For v0.2 of our ground station we have been working on consolidation of our electronics trying to minimize size and upgrade the components. Combined with the current minimized approach for the gear assemblies, this will enable us to deliver a smaller more rigid and reliable ground station.
The electronics that needed to be fitted in are the following:
* An Arduino. We selected Pico for the size advantage
* Two stepper drivers. We are using Pololu compatible A4983 based stepper drivers.
Agis (our electronics expert) designed an integrated board to fit everything in. Here is the result
And the schematic:
We are printing the PCB as we speak and will be posting a new log with the construction and testing.
Despite the fact that NEMA 17 stepper motors were considered our best choice, at the time of our first design some old NEMA 14, spare parts from an old Rep-Rap, where available so we stuck with them. Last week, we received our new motors and did some tests. The first test was to find the optimal torque without overheating the (also new) drivers. Although we have some results, tests will be continued and we will come up with some numbers probably by next week.
Additionally we are working on the new rotator design. We placed the motors on top. This way we managed to decrease the profile of the design a bit and also now the device is a lot more serviceable. Now you don’t need to remove everything if you want to access the motors or the worm gear. We are also trying different designs in order to improve the connection between the stepper shaft and the worm gear.