i have been designing flyback converter of a 5Volt 5 ampere specifications output deriving from 220v ac mains. But i have been stuck at one point and can't seem to get out it. Efficiency is particularly very poor in this case getting around 49% when running at 25W output. Earlier, due to leakage inductance voltages were going around greater than 600v on the drain of the MOSFET(10n60) causing high stress on the FET.
So to clamp the voltage , introduced RCD clamp(50nF, 3-5k). I know its a very low value of resistance used causing excessive losses. But only at this value i was able to clamp the voltage down to 550v running at full load.(For stability purpose this is too high, but can't go down anymore without decreasing resistance). At this, MOSFET temperature does not increase to high levels.
Whether this problem is coming due to the leakage inductance too high for this circuit or some other problem?
Transformer specifications:
30swg, 61 turns primary, 1 strand(1.7mH)
22 swg, 3 turns secondary, 5 strands(4.6uH)
35 swg, 1 strand, auxillary, 10 turns( 100uH)
EE28 core used, interleaved windings.
Since, i am winding transformer by hand not by machine, already it is a problem to wind 22 swg wire that too 5 strands so that is why i am thinking it might be a problem of magnetizing inductance too high in this circuit.
Uc3844 ic with isolated feedback using 817 opto with tl431 reference.
Any help would be greatly appreciated.
Edit 1: Circuit for the following converter(FET used is 10n60 not IRF840)(earlier IRF840 was used)-
Edit 2: The solution I got was to increase the current sense resistor to limit the peak current thereby less energy stored per cycle causing resistance value used to almost negligible(50kohm,considering losses). Efficiency boosted up to 76% at 5A Load. Still trying to increase it up to 85 by optimizing the core and rectifier losses. Will write the results soon.