zaterdag 25 oktober 2014

The start of construction of the TUDelft satellite tracking station: DopTrack

Monday 20 October 2014, we started the construction of DopTrack, a satellite tracking station on top of the tallest faculty of Delft University of Technology for students to use in satellite orbit determination. I will try to report on our findings of building it in this blogpost. We are in the process of setting up a dedicated website, but for now you have to do my blog ;)

Project DopTrack
As a master student, I followed several courses on orbit determination and satellite data analyses in the Astrodynamics & Space Missions group. During this time, I also observed the design and launch of two cube-sats of the space system engineering group. I was inspired by what students and university staff could accomplish. Come on, they launched two satellites in space. SPACE!!! However, after launch and succes of the missions the satellites were left alone. It would be cool to combine the use of these satellites and the knowledge from satellite data analyses courses. So I decided, when I became a PhD candidate, to construct a space operations practical that allows bachelor and master students to get hands-on experience with satellite operations and orbit determination. The idea for DopTrack was born!

I found another researcher and a bachelor student mad enough to help me in my adventure. Together, we started to design the tracking station and wrote a proposal for some budget to buy equipment. Because both professors saw this project as a good opportunity to improve cooperation between the groups, the proposal was accepted and we could start ordering equipment.

And before we knew it, the week of construction was showing up in our agenda.

Day 1: Installation of UHF antenna and ground equipment
Monday the weather enabled us to work on the roof. The weather forecast of the rest of the week looked quite bad, with a fall storm coming our way. So, we decided to install all the equipment on the roof and make it waterproof. This meant to check the already installed VHF antenna, placing the (bit smaller) UHF antenna and GPS antennas. we need the GPS antenna for our clock, such that the radio, SDR and computer all have similar frequency stability. We saw in the manual that the clock also calculates the position of the antenna (that is what normal GPS receivers do), so we placed the GPS antenna in between the two radio antennas. The sight was spectacular:

Antenna set-up (from left to right): omni-UHF, GPS, and omni-VHF antennas
Down below, we had managed to place an old network-server cabinet (it was very heavy, and there is no elevator between the 21ste and 22nd level!) for our electronic equipment. We ordered rack mounts for our radio and SDR, while the GPS receiver clock was already in rack mount geometry. After putting the rack mounts together, we placed the equipment in the cabinet. The computer, we have now, is not yet a server rack mount one, but we are trying to get this in the coming months. For now we use a common desktop. It looks like this:

Cabinet set-up with SDR, GPS clock, radio, and computer (top to bottom)
All in all, a good day, but fall was coming!

Day 2: Constructing cables and stormy weather
A big storm had arrived in Holland and at the top floor of this building we could feel this. Worried for the antennas on the roof, we checked their status with a remote video camera pointed towards the antennas. Luckily, everything was fine, we had installed them properly! Despite the horizontal rain fall:

See the movie with HD settings, otherwise it is difficult to see. During this stormy weather, we continued the construction of the set-up. This meant making a lot of cables and soldering the connections. I levelled my soldering skills a lot ;).

Day 3: Making more cables and connecting to the radio
The next day, even more soldering of cable connectors. We also checked if it was possible to remotely operate the radio. And after some internet coding and scripting, we were able to remotely turn the radio on and off again (this is especially useful for trouble shooting, somehow this is important for engineers). 

To be even more in control of the electronic devices, we had installed a power switch, which made it possible for us to remotely turn the power on/off of 8 different devices. This worked and gave use nerdy feelings, when we were turning on/off equipment with our laptop (a bit like the first episode of "The Big Bang Theory"). 

Day 4: Setting up local network and finishing installation of equipment

Finally, all the different cables were made and we could hook up the different components. The radio to the antennas and SDR, and the GPS clock to all the other equipment. In between the antenna and radio we had to put a splitter-board. We ordered antenna splitters, because the antennas would also be used by other setups in the ground station. The initial setup looks a bit like this (we still need to fix it to the wall):

Antenna splitter panels for the VHF, UHF, and S-band antennas
In front of the splitters we placed bias-T amplification, which puts power in the cables to the antennas. This results in less loss in the long (25 meter) cables that run on the floor of the roof. After hooking it all up, we wanted to see if things worked.

First recording of a satellite was a fact!!! You can clearly hear that this signal is moving due to the continuous shifting of the frequency (the tone of the signal constantly changes). A good result of day four.

Day 5: Getting the SDR and GPS clock working

Friday was the last planned day for work on the ground station. This was mainly trying to get remote acces to all the different equipment components. Hours of reading operating manuals, setting up connections from the router/computer to the different ports of the equipment and testing the remote connection on a different laptop. This will we be doing for some more days, but we got connection to the radio and GPS clock:

Remote connection to the GPS clock is established
The GPS clock is up and running and locked to several GPS satellites. The hardware is installed and now we need to fix all the software, before we can test the tracking capabilities of our station. And hopefully we get figures like this:

The Doppler curves in the radio frequency domain of the Delfi-C3 and Delfi-n3Xt (courtesy of Nils von Storch, operator Delfi ground station)
In the end, we will convert this to actual information about the satellite's velocity and position. When you, as a student, are interested in working with this equipment for your master thesis, please send us an email (doptrack[at] Furthermore, this facility will be used in the new Space minor of the Aerospace Engineering faculty of the TUDelft. From a student idea made to an actual education application. 

zaterdag 11 oktober 2014

My first estimates of the gravity field of the Rosetta comet, 67P/Churyumov-Gerasimenko

Last week, I saw a very incredible photo compilation. The Rosetta ESA satellite is currently orbiting a comet in outer space. Onboard is a camera, that is constantly taking pictures of the object. On thursday 2 October, the ESA Rosetta blog showed active geysers, shooting water of the comet out into space. What we learned as a kid, we can now see with our own eyes (well via the camera lens, a high-data space downlink, and a remarkable team in the ESA Operations & Science). Active geysers on a COMET!!!!!! Amazing.

After re-finding myself from being blown away by amazement, I continued browsing this blog. There, I found that they had build a 3D model of the comet and you (as an internet user) could download it to built your own comet. Totally awesome, unfortunately, I do not have a 3D printer. Nevertheless, I downloaded the files from the blogpost (see here). You can download .wrl and .obj files. Clicking on the .obj file, opened the comet in Adobe Photoshop. Look at that, I could rotate and play with my own comet:
This is what you see when opening the .obj file
After rotating the comet for several minutes, my nerdy brain started to pinch me: "Dude, if we had the coordinates of the surface, we could do some awesome stuff!". So I looked at the .wrl file with TextEdit (or Notepad) and found the following code.

#VRML V2.0 utf8

# Generated by VCGLIB, (C)Copyright 1999-2001 VCG, IEI-CNR

NavigationInfo {
type [ "EXAMINE", "ANY" ]
Transform {
  scale 1 1 1
  translation 0 0 0
      geometry IndexedFaceSet
        creaseAngle .5
        solid FALSE
        coord Coordinate
            -0.393756 0.401856 0.442509, -0.163294 0.491935 -0.000659, -0.515386 -0.259898 -0.343331, -0.277434 -0.260428 0.279815, 
            -0.55193 0.159748 0.155219, 0.091377 -0.282348 -0.211614, -0.792302 -0.192477 -0.092299, 0.785257 0.323246 -0.017733, 
            -0.093053 0.512756 -0.20542, 0.83051 0.3069 -0.045469, 0.795553 0.336202 0.063932, 0.168434 0.640004 -0.350024, 
            0.241654 -0.381833 0.019116, 0.200554 -0.206936 -0.422803, -0.215483 -0.128884 0.356076, -0.41427 -0.434269 -0.193254, 
            -0.687416 -0.006455 0.036488, 0.920413 0.222731 0.137458, -0.511265 -0.217672 -0.356012, 0.835122 0.2753 -0.114706, 
            0.724787 -0.241446 -0.305928, -0.574952 0.063756 -0.033269, 0.509054 -0.535483 -0.079728, -0.070761 0.50456 0.425794, ...

After the line "point" , a lot of numbers were printed, sets of three numbers separated by comma. These must be coordinates (x, y, z). They were normalised because the value of the numbers were between -1 and 1. So, I wrote a little code to separate these numbers and obtain three vectors, with the x, y, and z coordinates. To check if these were the surface-data points, I made a scatter plot.

The point cloud of the comet
Yes!!! I had obtained a data set of the surface of the Rosetta Comet (I call it the Rosetta comet, because its real name is very long (see title)). Now we could do some cool stuff. In previous blogposts, it might have become clear that I am a gravity scientist. My job is to look at objects (usually planets) and say something about their gravity field. In the beginning of my PhD research I had written software that is able to transform geometries + densities into gravity potential fields. I wanted to see if my software was able to generate a Rosetta comet gravity field. I already had the geometry of the object and an estimate of the density could be found on the ESA blog (do not use the English wikipedia page, because they think it is made out of material that is more than twice as light as water [400 kg/m^3], or they made a typo!), 4000 kg/m^3, which is typical for a stoney/ice comet. 

However, when I continued I run into a few problems. The .wrl file was normalised, so I needed to find a scale factor. On the blog I found a volume (25 km^3) that was given and using that value, I found that the scale value must be around 2500 m. Another issue was related to my software, which made me reorientate the origin of the comet (its one of those boring details, but if you really want to know, make a comment to this post). I translated the origin with the following values:

x = x - 0.2;
z = z + 0.25;

These modifications are a bit "sticky finger" work, but in the small time I had, I could not find better values and I really wanted to write this blog (maybe if I have some more time, or information). But now my software was able to make a gravity field model, assuming that the comet had a homogenous density distribution, which is not the case, but a good first estimate. Hopefully, after 12 November more information about that will be available (They are going to land a robot on the comet!!!!). Putting the geometry and density in my software, enabled me to make a model of the gravitational attraction of the comet.

The gravitational attraction in the radial direction at an altitude of 500 m above the highest point of the comet (in my reference frame), which is located around -120 degree longitude and +40 degree latitude.
I calculated the radial attraction (mGal is 1/100000 m/s^2, so very small) at 500 meters altitude above the comets highest point (in my reference frame). On Earth the radial gravity is dominant and doesn't vary that much, because of its spheroid shape, but here you even got negative radial gravity field, which is quite fascinating, or I am doing something wrong. Doubting my quick computing skills, I also plotted the magnitude of the total gravity vector.

Magnitude of the gravity vector felt at 500 m altitude.
Luckily there was no negative gravity, but in the southern area of the comet the gravity was quite low. At close distance the gravity field of a comet fluctuates very rapidly. You have to stop thinking in Earth or planet terms. To fly a satellite in this kind of environment is very tricky. I wish the Operations team at ESA a lot of success and hopefully they have a better representation of the gravity field (in a better reference frame) than I have.