This datasheet have the formula to calculate oscillator frequency. I reckon no one have this so posting. Might br helpful

 

I have made this,

The schematic for the same is attached below.

I don't have the 330pf cap so i have used 100pf one with the resistor value of 180k (R3 and C3 in schematic) and that gives the frequency of 39.6 khz according to the formula.

I want to ask what must be the value of R9, as someone said earlier that there must always be some load.

 

I have made this,View attachment 185673

The schematic for the same is attached below.

I don't have the 330pf cap so i have used 100pf one with the resistor value of 180k (R3 and C3 in schematic) and that gives the frequency of 39.6 khz according to the formula.

I want to ask what must be the value of R9, as someone said earlier that there must always be some load.

The 1000 ohm resistor will present a load of 24mA with an output of 24 volts. That ought to be enough, although you could add another one in parallel and have a higher load of 48mA. That should certainly be enough to keep things stable.

 

This too doesn't work

I was trying to simulate it on ltspice but got and error while running the simulation
Could not open library file "IR2153.sub"
I'm using bordodynovs' library (he gave me the link, not sure whether he made it or not)

Saw similar problems on other forums but that didn't worked.

 

Please correct me on this one.
The oscillator frequency that we are setting using a resistor and a capacitor will be the switching speed of the mosfets, and if i lower the frequency, say 10Hz, will i be able to see them switching alternatively? And if yes then how the LEDs be connected, anodes of both on pins 5 and 7 respectively and cathodes to ground with series resistor? Is this right.

 

Could not open library file "IR2153.sub"

Where have you saved the file? If you have Win 10 and LTspiceXV11 then LTS expects the file to be in your ...User\Documents\LTspiceXV11\lib\sub folder.
Although LTS can probably simulate the IR2153 with an oscillator frequency of 10Hz, the real-world IC may not work at such a low frequency.

 

I was able to run the simulation, and here is the result

Why is the amplitude of Voh is 240V and that of Vol is 10V. That much voltage will damage the MOSFETs Gate, isn't it?

 

Why is the amplitude of Voh is 240V

The gate voltage Voh gets boosted to ~Vcc above the source voltage of the high-side FET by the bootstrap arrangement on the driver IC. When the FET switches on, its source gets pulled up to the +ve rail, i.e. very close to the mains peak voltage. But what matters is Vgs, the gate voltage relative to the source voltage.

 

I'm doing a spice simulation of the circuit, but it is taking forever to run after i put inductors to make transformer. Before adding the inductor i was able to do transient response of circuit but now I'm not. Is there something wrong with transformer setup? Because I read it that the inductance of windings must be square of the turns ratio and my transformer have 41 turns in primary and 5x2 turns in secondary.

 

Not the inductance ratio, it is the impedance ratio that is the square of the turns ratio. and some simulators do not model transformers well, and just because you have the symbol does not mean that you have a good model of a transformer.

 

Not the inductance ratio, it is the impedance ratio that is the square of the turns ratio. and some simulators do not model transformers well, and just because you have the symbol does not mean that you have a good model of a transformer.

So how to make it work?

 

So how to make it work?

I do not use the spice simulator, so I am not able to help there. Transformer models are complex if they are accurate, since they have current controlled voltage sources as well as inductive reactances. And if your transformer is saturating then you need to include that in the model, because it is non-linear, and that leads to inaccurate results.
So really you need a spice expert.

 

@Bordodynov

 

For a simulation to be even close to meaningful, you need a model that closely matches your real-world magnetics. (transformer)
This will be very difficult to create unless you have deep knowledge of this subject.

When building the physical hardware, consider starting with a piece of all-copper PCB material as a ground plane.
A ground plane helps dramatically when dealing with fast high current transients that typically occur in a SMPS.
Keep wiring as short as possible, solder GND points directly to the ground plane via very short leads.
Make the FET's easy to replace - and stock up - you will be blowing lot's of them.

I admire your fortitude, I would be very hesitant to try building a SMPS myself.

 

The high frequency spikes and pulses in a switcher mean that every conductor is part of the circuit, and at that point I decide to use a ready made supply.And for short production runs and single unit projects it is far more cost effective that way. But to create a working supply will certainly be a big achievement and a useful learning experience. It would even look good on a resume as a testimony to your determination.

 

it is taking forever to run

Add a ground to the secondary side of the transformer (Spice doesn't like floating nodes).
Give the inductors some parasitic capacitance and resistance.
Some models of the IR2153 have problems.
BTW, your AC voltage source should have an amplitude of 325V if the rms voltage is 230V.

 

Add a ground to the secondary side of the transformer (Spice doesn't like floating nodes).
Give the inductors some parasitic capacitance and resistance.
Some models of the IR2153 have problems.
BTW, your AC voltage source should have an amplitude of 325V if the rms voltage is 230V.

Thank you

 

Add a ground to the secondary side of the transformer (Spice doesn't like floating nodes).
Give the inductors some parasitic capacitance and resistance.
Some models of the IR2153 have problems.
BTW, your AC voltage source should have an amplitude of 325V if the rms voltage is 230V.

I totally agree with you. But especially for letter readers (I'm not sure there is such a concept in English) I made a model of GNDFLOAT. It is actually a resistor with a capacitor connected to the GND (1GOhm and 1pF (can be changed) . If you specify in the UIC options, the potential of this floating earth will be equal to zero.

 

I totally agree with you. But especially for letter readers (I'm not sure there is such a concept in English) I made a model of GNDFLOAT. It is actually a resistor with a capacitor connected to the GND (1GOhm and 1pF (can be changed) . If you specify in the UIC options, the potential of this floating earth will be equal to zero.

Cannot find model GNDFLOAT in Sborka.lib

 

.subckt GNDFLOAT 1 Cp=1pF Rp=1G
c 1 0 {Cp} Rpar={Rp}
.ends GNDFLOAT
I update my libraries regularly. I checked the Sborka.lib file for GNDFLOAT and found it. At the moment the latest version is on my web page. Please update the file.