Laser soft soldering in the field of electronic assembly is a high-precision welding technology mainly used for the welding of small and precision components. The laser soft soldering system uses a semiconductor laser as the heat source to provide a highly concentrated laser energy beam to achieve high-quality welding with fast processing speed, a small heat-affected zone, and high precision. In the machining system, the performance of the semiconductor laser directly affects the welding effect with current and temperature being critical parameters to control. These parameters directly affect the focusing and energy distribution of the laser beam. Semiconductor lasers, also known as laser diodes, are electro-optical conversion devices. During processing, part of the electrical energy is converted into heat, causing the temperature to rise. This rise in temperature decreases the efficiency of the semiconductor material and the laser’s output power. Excessive temperature leads to laser wavelength drift, affecting the interaction between the laser beam and the material. Conversely, too low a temperature can make the laser difficult to start and result in unstable output power. Therefore, to ensure the performance and reliability of semiconductor lasers, it is crucial to develop a temperature control system to maintain temperature stability.

To address these issues, we propose a temperature control system for semiconductor lasers based on a model predictive controller (MPC). First, a mathematical model is established for the thermoelectric cooler (TEC) and other components of the thermal control apparatus. Subsequently, a predictive control model for the laser’s thermal control system is constructed using this mathematical model and software simulations. Finally, the feasibility of the design is validated through practical experimentation on an experimental platform specifically designed for laser soldering processes. The experimental verification involves actual laser soldering operations, confirming the practicality of the proposed design.

Temperature fluctuations in a semiconductor laser can lead to instability in its output power, affecting the quality of the laser and the effectiveness of soldering. An increase in temperature alters the physical properties of semiconductor materials, such as the band structure and carrier concentration, which in turn affect the wavelength and intensity of the laser. To verify the effectiveness of this design in controlling the laser’s temperature under varying input currents, we design an experiment to assess the thermal control capabilities under randomly changing laser input current conditions. Experimental results indicate that under MPC, the laser temperature converges rapidly and remains stable (Fig. 6), with energy consumption being about 42% lower than that under proportional-integral-derivative (PID) control (Table 1). In the laser center wavelength stability experiment, the center wavelength shift measurement experiment of the semiconductor laser is designed to indirectly evaluate the junction temperature control stability of the semiconductor laser under the control of MPC. The drift fluctuation range of the central wavelength of the laser under MPC thermal control system is 0.36 nm within 30 min, which is 52% less than that of PID control (Fig. 7). Experiments on laser soft soldering capabilities show that under MPC control, the laser can efficiently complete soft soldering tasks, achieving reliable and effective solder joints (Fig. 9).

To enhance the control performance of the thermal management system in the semiconductor laser of the laser soft soldering system, we conduct precise mathematical modeling based on thermoelectric cooling devices and non-equilibrium thermodynamics principles. This leads to the design of an MPC for the thermal control system. The effectiveness of the control algorithm and the actual laser soft soldering capabilities are validated through experiments. Under simulated conditions, a comparative study between the traditional PID control algorithm and the MPC control algorithm indicates that under random input current changes within the working current range of the laser, the MPC-based thermal control system offers more stable and rapid temperature control. The temperature overshoot is reduced to 1.12%, significantly shortening the time and range of temperature fluctuations of the semiconductor laser during operation. In addition, the energy consumption of the entire thermal control system is reduced by about 42%. Under real processing parameters and environment, the MPC algorithm-based thermal control system produces a solder joint temperature curve that closely matches the ideal temperature curve compared to the PID control algorithm. The average stabilization time is 42.53 ms, with better robustness and higher reliability and quality of solder joints, meeting the stringent thermal control requirements of the laser soft soldering field.