We take on work that starts as a dynamics question and has to end as hardware someone can build, certify, and service. The useful thing we bring is that we do not hand the model over and stop.
Lyapunov-based and variable-structure design, state estimation, and loops that have to stay stable across the whole operating envelope rather than around one linearisation point.
MATLAB, Simulink and Simscape, plus Python where a script is the better tool. Plant models built to represent the real machine — component tolerances, sensor behaviour and failure modes, not an idealisation that flatters the controller. Runs headless, so a study is a batch job rather than an afternoon of clicking.
Analogue and mixed-signal circuit design taken through schematic capture and board layout in KiCad, with the noise and gain structure worked out in simulation before anything is routed.
Sensor and converter selection with the error budget written down: what the measurement actually resolves, and what it costs to resolve more.
Sensor voting, independent backup layers, and fail-safe behaviour chosen deliberately. Where a loop can hurt something, the architecture says how it degrades.
FreeCAD and Blender driven from Python: geometry generated from parameters, verified rather than eyeballed, and packaged as STEP plus 2D cut files a shop can quote without a phone call.
Design rationale, assembly manuals with rendered step figures, and bills of material. Written as a deliverable, not scraped together at the end.
Most work starts with a fixed-fee concept phase: we build enough of the model to find out whether the idea survives contact with the real machine, and we say so plainly if it does not. That phase produces a written design stance — what we recommend, what we ruled out, and what it will cost to be wrong.
Then build. Simulation, hardware selection, control implementation, and the drawings and documentation to hand to whoever fabricates and maintains it.
We are comfortable being the engineer of record on the dynamics and letting a specialist own the parts that should be specialised — mains power, certification, structural sign-off.
The studio's technical spine is nonlinear control theory — the same body of work behind the DS-C converter architecture and the drive systems in our kinetic installations. Peer-reviewed publication in IEEE and ACC venues, and delivery experience across motor drives, agricultural and industrial control, data converters, and instrumentation.
A short description of the system and the problem is enough to start. We will tell you quickly whether we are the right people for it.
takis.zourntos@emads.org