Kinetic Ships Servo Drive 3 Months Ahead of Schedule
By replacing their legacy SPICE workflow with AUV, Kinetic's motor drive team eliminated convergence issues and accelerated their development timeline by a full quarter.
“Zero convergence failures in six months of daily use. That alone saved us hundreds of engineering hours.”
The Challenge
Kinetic Power manufactures servo drives for industrial robotics and CNC machinery. Their R&D team had been using a combination of MATLAB/Simulink for control design and a SPICE-based tool for power stage simulation. The disconnect between these two environments created constant iteration delays and model synchronization issues.
The most frustrating problem was convergence failures. Their SPICE simulations of the SiC-based power stage would fail to converge approximately 30% of the time, requiring manual intervention to adjust solver settings, initial conditions, or step sizes. Each failure wasted 15-30 minutes of engineering time.
The Solution
Kinetic evaluated AUV based on three criteria: convergence reliability, simulation speed, and the ability to co-simulate control algorithms with the power stage. AUV satisfied all three — the DSED solver eliminated convergence issues entirely, simulation speed was 50x faster than their SPICE tool, and AUV's Python API enabled seamless integration with their MATLAB-based control models.
The Results
In the first six months after migration, Kinetic's team ran over 4,000 simulations without a single convergence failure. The time savings allowed them to ship their next-generation servo drive three months ahead of the original schedule.
The faster iteration cycle also enabled more thorough design optimization. The team explored 3x more design variants than in previous projects, ultimately finding a topology modification that reduced power stage losses by 12% — a competitive advantage in the efficiency-sensitive servo drive market.
Perhaps most significantly, Kinetic estimated that eliminating prototype iterations (enabled by simulation accuracy within 3% of hardware) saved approximately $400,000 in prototype fabrication and testing costs across the project.
Zero convergence failures in six months of daily use. That alone saved us hundreds of engineering hours.