Hello.
I have the problem with UCC28950 controller.
It is used in 540V to 50V isolated DC-DC converter, 3kW, Voltage Control Mode, open-loop configuration (amplifier is configured as voltage follower with EA+ input). Controller is connected to the secondary side. Switching frequency is 75kHz.
The primary resonant inductor is not used, there is only DC-blocking capacitor in series with primary winding of power transformer. I use current-double rectifier, NOT synchronous (diodes). Currently I’m experimenting with reduced power supply voltage of inverter.
At full load everything is perfect. The problem appears at NO load, when converters goes to discontinuous conduction mode. In this mode the voltage oscillations at primary are quite high but the overvoltage on transistors is acceptable. When I increase converter input power supply to about 40 volts the controller goes to instability. The control pulses at outputs A,B,C,D have random width, there are sudden pauses and I hear low frequency noise.
I’ve done several tests in order to reveal the reason. Let me describe them.
When I increase the input voltage there is intermediate stage before the controller goes to instability. At this stage the phase shift between legs AB and CD is slightly modulated. Moreover, the voltage at VREF-pin changes it’s level from 5V to 7V (proportionally to input voltage of power inverter). When it reaches 7V instability begins.
In normal mode the voltage of pin RT is constant and equals to 2.5V. In instability it changes intervals discretely (if not consider the noise) between levels 2.5V and 7V. When V_RT=7V the output pulses on pins A,B,C,D are zero. Also in instability mode the voltage on SS-pin is about 6.4V, while in normal mode it is about 5V.
Another observation is that during instability the voltages on pins TMIN, DELAB, DELCD sometimes drop to zero, while in normal mode they are always keep to be about 1V.
I’ve done two changes in controller connection. Firstly, I’ve shorted RSUM-pin to VREF-pin (I know, that it is out of recommended by datasheet, but I decided to try). Surprisingly, the problem disappeared. But, I could not change phase shift between legs, it always equals to the value set by TMIN (no matter of EA+ voltage).
Secondly, I removed wire from VREF-RSUM (restored resistor) and added second capacitor between VREF and GND (C=330uF). As the result I was able to increase input voltage to much higher value than 40V before instability begins. Moreover, I’ve noticed, that during instability the audio noise changed it’s frequency to higher values (in comparison with original test). Also, I can clearly see, that when I increase input power voltage of inverter and VREF goes from 5V to 7V the frequency of output pulses proportionally decreases to about 50%.
Could you explain this? I understand, that most likely the fundamental reason is the power noise, but why controller behaves this way? Why VREF and other described voltage levels are changed?
The peace of circuit, which shows controller connection, is included here.7651.Connection.pdf
I’m looking forward to hearing from you soon,
Roman Gorbunov,
Power Electronics Engineer at “FOTON” R&D
