TOP256PN

It sounds like your power supply is going into auto-restart at power up. The control pin current must reach 2.5 mA in 64 ms or the power supply will go into auto-restart. Since the power supply starts with a 3.3 Volt zener, it indicates that you are close to the timeout. I would suggest reducing the output capacitance from 2000 uF to 1000 uF. This should reduce the time for the output to reach regulation. The reason that the TL431 works is because it has more gain and will sink more current near the regulation point.

From your schematic, the external mosfet is biased by the high voltage zener. The source voltage will be the zener voltage minus the Vgs(on) voltage. This will be the drain voltage of the internal mosfet.

Z. Cochrane

Thanks for the reply sir.
As per PI spreadsheet the primary inductance of transformer is 1.3mH.
I have made two bobbins. one is wounded by myself with the primary inductance 6.3mH and another one i got from vendor with primary inductance 1.5mH. the bobbin which i wounded is working fine with TL431 feedback(the problem which i posted first). In that same PCB if i mount the vendor made bobbin it is not working. it is going to autorestart mode. I measured bias voltage in the vendor made bobbin it was 34V for 100V input voltage. but for 6.3mH bobbin the bias voltage is 16V.Both bobbins was made as per the PI transformer design. Now how should i reduce the bias voltage in vendor made bobbin?.
And one more thing even after reducing the output capacitor to 470uH also power supply is not starting with zener feedback.
expecting for your quick reply. I have attached PI spreadsheet and tranformer design.

There are a couple of things I am concerned about. You measure 6.3 mH for the primary inductance. It should be 1.36+/-10%. I suspect that the core is not gapped correctly. You may be saturating the transformer core.
The vendor's transformer should produce about 15 volts on the bias winding. Check to see if the winding polarity is correct (the output winding and bias winding signals should be the same phase). If it is correct, then check the waveform with a scope at the bias winding. Either there is a large leakage spike that the rectifier is peak detecting or the flyback pulse is too large. If the flyback pulse is too large, then the number of turns is incorrect. If there is a large spike, then a resistor (start with 100 Ohms) in series with the bias rectifier. This will help filter the spike.
After you fixed the above problems, check to see if the output voltage gets to 5 Volts before 50 ms (the actual number is 64, but I want to see some margin). If it does not get to 5 Volts, then continue to reduce the output capacitor value. It it does reach 5 Volts but goes into auto-restart, check the feedback signal to the control pin.

Z. Cochrane

Thanks for your interest in this problem.
Actually i tried out all the method what you have suggested but nothing works out in vendor made bobbin. The instead of main output 5V it is only 1.3V and it reaches regulation around 32ms. but the other 5V output is 5.08V and the entire power supply is in autorestart mode.
When i reversed the bias winding the voltage in bias reduces to 5V instead before 34V.
I started testing in 6.3mH inductance bobbin. Till 50% load there is no problem. when i loaded 100% the clamp diode P6KE200A solder melts and the diode comes out from the board. the diode is too much heating. even the output diode for 5V main winding is also heating too much.pls suggest solution for this problem.Expecting for your reply.

You will need to look at the drain current with a current probe and oscilloscope. Also look at the drain voltage at the same time. This will help determine what is happening with your power supply. I believe that the transformer is saturating (6.3 mH transformer).

Z. Cochrane

Thanks for your reply sir.
I got transformer with 1.3mH itself and is working fine, taking load till 240 VAC. Beyond that it is going to autorestart for every 30 seconds itself. with no load power supply works fine till 540 VAC but with load problem is happening. with 230 VAC input voltage in full load power supply was kept inside hot air oven for one day, there is no problem but beyond that even at room temperature it goes to autorestart. pls suggest me.

If the device is turning off and on in a 30 second period, it is not auto-restart mode. It is more likely thermal shutdown. You will need to look at the drain current and voltage with an oscilloscope to determine if the power supply is operating properly.

Z. Cochrane

Actually the problem is with p6KE440A diode.if am not connecting the diode at 265 VAC power supply is not going to autorestart mode. it is working fine. when i connected P6KE120A in series with the P6KE440A also power supply is working fine at 265 VAC. I loaded power supply with 25% load that is around 3 to 4W, around 300 VAC input the diode p6ke440a becomes very hot and solder melts and diode comes out from the PCB. I tried by replacing P6KE440A by 4 numbers of P6KE120A. still result remains the same. now pls suggest me the solution. and one more thing i dont have current probe to measure the drain current. suggest me another method to measure the drain current.

I have connected a RC snubber of value R=1K and C=330pF across P6KE440A + P6KE120A. Power supply is working fine till 415 VAC on full load. now i have a doubt. i have simply connected these snubber values. whether this is correct? If so pls tel me the formula or method to connect the snubber. awaiting for your reply. Thank you

Snubber values will be determined by your load conditions, drain node capacitance and transformer leakage inductance. You will need to look at the drain waveform with a oscilloscope. Also I am concerned about if the transformer is saturating. If you don't have a current probe, you can try using a 1 Ohm resistor in series with the drain and make a differential voltage measurement with the oscilloscope.

Z. Cochrane

The drain current measured around 275 VAC is 8A. I have inductance measurement meter in which i can measure only the frequencies 1Khz & 100 Hz. I measured primary inductance which was 1.57mH (across the primary with all other windings open) at 1Khz frequency and also measured leakage inductance at the same frequency which was 2.76mH (across the primary with all other windings short). Drain voltage is around 600 V. pls suggest me.

Refering to my previous post leakage inductance was 70uH only. pls ignore my previous values of leakage inductance. Now how to avoid the saturation of transformer? should i decrease the VOR to decrease the primary inductance and reduce the number of primary turns? pls give me the suggestion.

I do not know if the transformer is saturating. You will have to measure the drain current to determine if there is a problem. If it is saturating, adding more turns to the transformer and increasing the core's gap so that you end up with the same primary inductance will decrease the flux density.

Z. Cochrane

I measured the drain current. It is starts increasing after 280 VAC suddenly to 8A.Now how many turns i should increase?

Please send me a scope photo of the drain voltage and current waveforms. Use 2 us/division timebase.

Z. Cochrane

It is not possible for me to send the photo due to official reason. I have ordered new bobbin with increased number of turns. let me see how it behaves. Thanks a lot for your help sir.

Hello sir,
I have increased the primary turns from 100 to 150 and corresponding secondary turns also by maintaining the same primary inductance.still result remains the same. transformer becomes very hot during 300 VAC input. even after switching off the power supply also, transformer is hot for around 15 to 20 mins. my question is whether EFD25 is sufficient for 13W operation at 66kHz in triple wire insulated winding method. still i cannot go for next higher bobbin due to space constraint in my PCB. please review my circuit once again and give me the feedback as early as possible.

Please give me solution its an urgent requirement