Research On Failure Suppression Technique For Constant Current Load Driver

This paper is a research primarily focusing on the failure suppression technique of constant current load drive circuit. Failure in the context of this paper consists of two aspects: functional failure caused by damage during application and performance degradation during application due to lack of performance index. Based on how failure comes into being, the failure suppression technique presents new methods and techniques to improve the performance of the circuit that causes the failure, so that the failure can be eliminated.As the 3rd generation solid-state light source, LED has many merits and various applications, it will be studied in this paper as a typical load in a direct current output drive circuit. The development of LED driver is already quite mature, but as application requirements constantly become higher and requirements of application systems constantly change, the technology of its driver circuit has also been constantly developing. Only by improving the performance index and reliability of its driver circuit to the level of similar international products, can our product be recognized by the market and terminal devices of international brands. Hence, the research in this article has apparent market prospect and is of great technical significance.By analyzing the operating principle of power MOSFET and its drive circuit, one method and two techniques have been studied. This article proposes 1. a Progressive-Trigger-Control(PTC) method based on channel width change by analyzing the operating principle of the driver circuit’s power device. 2. A Automatic-Voltage-Tracking(AVT) technique for dead-time control based on the timing of the gates of synchronous power MOSFETs. 3. A frequency conversion technology with the same duty cycle based on charge-storage of capacitor. The main contents and innovations in this paper are as follows.(1) By analyzing the relationship between the power MOSFET’S conduction current and its channel width, the paper propose a di/dt Progressive-Trigger-Control method based on channel width change. Establishes a control releationship between the MOSFET’s turn-on/off transition state and the channel time-varied function W(t). Completes the theoretical analysis of the progressive trigger method. By implementing discretization on W(t), this method can be used upon controling the power MOSFET’s turn-on/off process in order to achieve a progressive trigger control over the channel width. Consequently, the breakover current increases/decreases progressively and the di/dt at the moment of turn-on/off decreases. The testing result of a tape-out show that, under same testing circumstances, the turn-off di/dt decrease by 50% while the turn-on di/dt decreases by 79%.(2) On the basis of studying the principle of synchronous boost converter and the relationship between the turn-on/off transition and the voltage waveform on each node, this paper proposes an Automatic-Voltage-Tracking technique. This approach automatically controls the switch-on time of the synchronous power device by determining the working state of the SW voltage. It can effectively decrease voltage overshoot on the SW node and increase conversion efficiency. The power devices gate-driven circuits with this approach is applied in the design of synchronous boost converter, and experimental results shows that the decrease of SW peak voltage during dead-time is over 40%, and the maximum increase of efficiency under light load is approxiamately 3%.(3) By analyzing causes of error in conventional frequency conversion technology, this paper proposes a frequency conversion technology based on the charge storage of capacitor. This technology converts the error in a single counting period into charge and store it in a capacitor. It then uses superposition method to eliminate the error in the next counting period. We can apply this technology in the direct PWM dimming control circuit of the LED driver. The test result of a tape out shows that the input signal frequency lies between 200 Hz to 10 KHz, after conversion the output signal frequency is 1KHz and the error of the duty cycle is below 1%.(4) By analyzing causes of error of the LED driver circuit’s current regulation, it is determined error of analog regulation is caused by the amplifier offset voltage of the control loop. By analyzing the principle of analog regulation, the minimum regulation range of mode dimming can be decreased to the 1% objective for mass production using the digital-calibration technique. As for direct PWM dimming, a structure that combines auto-zero technique and double switch technology can be used to increase the LED dimming accuracy up to 1% while controlling the matching degree of multi-channel current below ±0.6%. These parameters are competitive enough even against products of international brands.