Latching of the UC3844

uc3844

I've just build 2-transistor forward off-line converter with the UC3844. The output voltages are 25V 6,5A (main output, stabilized) and 15V 4A (non stabilized). The two transistors that switch primary winding are driven from small gate drive transformer (the gates must be galvanically separated). UC3844 is powered from an additional slave flyback converter that outupts 18V to power the chip. But there is a problem.
I try tu run the converter by slowly increasing input voltage from autotransformer.
When the output reaches 25V i can hear loud and irregular "clattering" in the transformer core (ETD39) which most probably is the result of saturating of the core. I think that UC3844 generates wafeform which ratio is higher than 50% (but i don't know why - this chip should not work this way). After several seconds UC3844 stops working suddenly and draws as much as 100 mA from its Vcc. At the same time, VREF drops to about 3V and output waveform at pin 6 disappears.
I enclosed part of the schematic.
What could be the cause of the problem ?

uc 3844

I assume insufficient UC3844 supply decoupling triggering the UV lockout and/or output overload. Specified 1A current could be easily exceeded by the circuit, I think. To my opion, a level-shifting IRF gate driver would be a better option.

flyback circuit uc3844

It seems i improved the circuit. I simply removed DC restoration circuit at the secondary side of the gate drive transformer and added small resistor in series with primary (10R).
Now the chip doesn't latch.

i found the solution here:

www.powersystemsdesign.com/design_tips_dec06.pdf

IRF gate driver is not suitable for this purpose as the two transistors must be turned on at the same time whereas IRF chips drive the transistors alternately.
However i have new problem. The output voltage drops too much (1,5V) when only 2A current is drawn from the output but the ratio is below 50%. It seems it can be problem with feedback compensation.

schematic circuit uc3844

IRF gate driver is not suitable for this purpose as the two transistors must be turned on at the same time whereas IRF chips drive the transistors alternately.

Click to expand...

Some do, others don't. You apparently didn't check for suitable parts.

However, there may be a problem with initial preloading a boostrap driver in this circuit. An auxilary load resistor is necessary.