These are some of the things I've done outside my main line of work or studies, just for fun.

Computational physics and math

  • Particle collectives in arbitrary force fields
    Take a force field with some arbitrary distribution (e.g. gravity from a group of stars). Throw a collection of particles in it, with initial positions and velocities. Calculate the trajectory for each particle.
    I used linear approximations of the field equations for very small domains. The equations of motion were also linearized, iterating over a very large number of very small steps. Therefore, any trajectory was simulated as a succession of infinitesimal parabolic arches. It was great to see the code converge to known analytical solutions (e.g. a single particle around a single star would move on an ellipse).
  • Volume of simple bodies in N-dimensional space
    In an N-dimensional space, consider an N-sphere circumscribed by an N-cube. Calculate the ratio between the N-volume of the sphere and the N-volume of the cube containing it, as a function of N.
    I'm sure there's a nice analytical solution somewhere. But if you're like me, you'll just do Monte Carlo with a ton of samples, and derive an approximate result. As N tends to infinity, the ratio tends to zero. That was very surprising.


  • 1-bit ADC/DAC (analog-to-digital / digital-to-analog) converter
    Take an analog signal (audio), convert it to digital via a 1-bit ADC (actually, a delta modulator). Send it via a single wire as a digital signal, convert it back to analog via a DAC at the destination. I designed the ADC/DAC circuits from first principles, based only on the theory of Delta modulation. Main design constraints were simplicity and low cost.
    The point was to show that a 1-bit digital line could carry sound just as well as an 8-bit or 16-bit line, provided that the switching frequency is high enough. It worked pretty well.
  • Wavetable synthesis sound card for 8-bit computers
    An offshoot of a larger project (polyphonic digital synthesizer) that never materialized fully. I designed and built the card from scratch. RAM banks were used to store the waveform and the envelope, DAC circuits and digitally controlled amplifiers instantiated the sound. I wrote drivers for it in Z80 assembly, and then I wrote a software synthesizer that could generate a variety of sounds.


  • Telemetry
    For a high altitude balloon, collect data such as position, altitude (via GPS), temperature, humidity (via appropriate sensors). Record data locally to an SD card. Send data to ground station via APRS in the 144 MHz band. I've used an Arduino clone for the microcontroller, and commonly available sensors. Ground tests went fine. I've not launched the balloon yet, I'm waiting for a good deal on helium.
  • Tachometer + fault protection
    For the telescope mirror grinding machine, I've built a device that can measure the rotation speed of the main rods, display it on an OLED screen, and cut AC power to the motor when the rods get stuck. Again an Arduino project, home-made tachometers, solid state relay on AC.
  • Random projects
    A small project for Halloween using an ATmel AVR micro, coded directly in C.
    A clock with a large LED display, mounted in the CD-ROM bay on a PC. It's synchronized to the PC.


  • Romanian Linux Users Group
    Founding member of RLUG (the Romanian Linux Users Group). I was part of the small initial group who created the organization. Educating the public at large, helping technicians, engineers and administrators working in the industry, providing resources (documentation, free software) and free support online - these were our main focuses initially.
    I believe we were very successful, locally, in helping the expansion of free software in general, and Linux in particular, which became a major component of the infrastructure in the region.
  • Computer technology summer camp
    I led a small group of technology enthusiasts and local teachers who organized a 1 week retreat, at a beautiful location out of town, during the summer vacation, for students who were interested to learn more about computers. We managed to pool resources, equipment and services borrowed from or donated by local businesses, schools, the Department of Telecommunications, and local authorities. The schedule included classes on programming, networking, systems administration, etc., but also unstructured time for kids to improvise, be creative, and apply what they had learned.
    The energy and enthusiasm were unbelievable. The kids learned a lot, but I'm sure the adults did, too - whenever we could sneak into someone else's class. This whole experience was its own reward. I would do this again any time.
  • Science for kids
    Literally a one-in-a-century event? Involving astronomy? Ocurring in the middle of the day? Perfect as a science event for middle school kids. It was the solar transit of Venus in June 2012. The turnout was bigger than expected, as you can see in this photo album.
  • Sidewalk astronomy
    Like this:

Astronomy and optics

  • It turns out, making your own telescope mirror, starting with an ordinary lump of glass, is not hard. It just takes a lot of persistence. Here's the polishing and figuring log for a mirror.
    I made my own optical testing equipment (Foucault tester).
    The telescope that uses that mirror can be seen at the very bottom of that very long polishing log.

Current interests

  • FPGA programming
    Specifically, its applications in computer vision.
  • Multirotors
    ...also known as "quadcopters" or "drones". Making progress already.
  • Autonomous flight
    Basically a combination of the previous two items. I want to make a drone that follows a red cue ball. Work in progress. Stay tuned.
  • Grinding / polishing machine for telescope mirrors
    If you want to make very large mirrors, you can't grind and polish by hand. You need a machine. Work in progress, as you can see in this video.
  • Very large telescope
    I'm aiming for an aperture of 1 meter or larger. This will likely require several successive steps, each one larger than the previous one. I'll have to use FEM to design the large but lightweight glass substrate; Elmer appears to work well for this purpose and it's free.
    Requirement: The viewport needs to be near ground level, which means I'll have to use a different design (not newtonian); I have several designs tested with OSLO that meet all conditions, are diffraction-limited, and do not require grinding impossible curves in glass.
    This project is likely to take a long time, so do NOT stay tuned.