I'm testing a ir2153 based smps on breadboard. For the first time i switched on the circuit it turned on and gave 18.8V on output. Mains 220V ac is connected to the circuit using a 100W bulb in series. When i switched on the mains supply the bulb turned on for a fraction of a second and then turned off, this is for the first time I powered the circuit.
Immediately after turning off the supply, i touched the MOSFET and diodes to see if they are getting warm. Everything was normal. Then in the second run as i connected the mains supply the bulb firstly glow for a fraction of second and then started flashing and after that it remained turned on as if there is a short in the circuit.
I quickly disconnected the mains and check the MOSFETs and they both were gone. What is the reason behind this?
Circuit is attached below
C1 and C2 in the circuit is 330uF but i don't had those so i used 680uF ones, is this the reason?

 

I forgot to attach files.

 

The capacitors C1,2 are to provide an Half voltage supply for the push-pull transformer output, according to the datasheet the transformer is being pulsed across the lower mosfet and not as your diagram...

 

That transformer looks like a power-line type with standard laminated iron core and, if so, is not appropriate for operation at 47kHz switching.
It likely has a lot of leakage inductance creating large voltage spikes during the switching.
You need a transformer designed for high frequency switching, such as one with a ferrite core.

I suggest you test the circuit at a much lower voltage with an isolated supply so you can look at the signals with an oscilloscope to see any abnormal waveforms or spikes.

 

transformer looks like a power-line type with standard laminated iron core and, if so, is not appropriate for operation at 47kHz switching.
It likely has a lot of leakage inductance creating large voltage spikes during the switching.

It is a ferrite core transformer not iron core.

 

The capacitors C1,2 are to provide an Half voltage supply for the push-pull transformer output, according to the datasheet the transformer is being pulsed across the lower mosfet and not as your diagram...

So the capacitor c1 and c2 of 680uF have no adverse effect on the circuit?
And if i change the transformer primary as there it is in datasheet then how likely is it going to work?

 

I don't see the need for the half-supply-voltage arrangement involving C1/C2, since the transformer is AC coupled by C7?
With the stray inductance of all the straggly wiring on the breadboard it's possible the FETs were going into uncontrolled high frequency oscillation that could have caused their demise.

 

It is a ferrite core transformer not iron core.

Do you have any information/data sheet on it?

 

Where did the circuit come from? AND, switching supplies are very demanding of having the circuit pattern just right. But if it is built up on a breadboard setup even a good design will probably not work right, if it works at all. Switching supplies are a big challenge to get right, even with a known-good circuit.

 

Do you have any information/data sheet on it?

I got it from a old computer power supply. I don't have the datasheet of it but the power was rated for 500W, so it must've to be a good one and also it was from a well known brand.
I did the rewinding of the transformer. Primary have 41 turns of 18 gauge wire and secondary have 3 wires of 23 gauge with 5x2 turns. All as per the circuit says

 

Where did the circuit come from? AND, switching supplies are very demanding of having the circuit pattern just right. But if it is built up on a breadboard setup even a good design will probably not work right, if it works at all. Switching supplies are a big challenge to get right, even with a known-good circuit.

From a YouTube channel- kasyan tv

 

As MB2 noted, being built on a plugin breadboard means it has a lot of stray capacitive coupling with a high ground impedance, all of which make if difficult to get a switching converter operating properly.
It really should be built on a vector board that has a copper ground plane.

 

As MB2 noted, being built on a plugin breadboard means it has a lot of stray capacitive coupling with a high ground impedance, all of which make if difficult to get a switching converter operating properly.
It really should be built on a vector board that has a copper ground plane.

Yes. Noted

 

What is the formula to calculate the frequency of R C oscillator? 1/2piRC ?

 

What is the formula to calculate the frequency of R C oscillator? 1/2piRC ?

It depends upon the design of the oscillator.
There is no one-size-fits-all equation for that.
If you are referring to the ir2153, the data sheet shows the relation between R, C, and frequency.

 

I've modified the circuit little bit. Just compare it with the previous one and give suggestions.

Changes-
Changed the RC network with values 10k and 330pF as per the datasheet
Transformer is connected to the centre point of two in caps

 

D2 is shown as a 1N4007. Normally a fast-recovery diode is recommended for that position. I don't know if that has any bearing on why your FETs died, but if the bootstrap circuit wasn't operating properly then Q1 might not have been turning on fully so could have overheated.

 

So D2 can be replaced with 1n4148?

 

So D2 can be replaced with 1n4148?

No, because a 1N4148 is a signal diode. It should be selected for the associated voltages and currents. And still you will need to change the wiring from that prototype arrangement to an arrangement with much less wire inductance and resistance.

 

Then D3 also needs to be changed?