PWM for buck converter

[johnsorious]

Is there a certain equation that I should use to size up my filtering capacitor? I'm seeing multiple equations online.

[johnsorious]

Here is my circuit. 

[johnsold]

I do not understand why you have C1 and C3 in parallel? Regardless, they are too small. Consider that the 5 V load current is 2 A. If the regulator was perfectly efficient, the current at the 12 V point would be 0.83 A. With a realistic 80% efficiency the current there will be about 1 A.  That current will completely discharge 500 uF in 6 milliseconds.  So your regulator will shut down on every half cycle when the input votlage goes to zero.

I do not have Multisim so I cannot see the parameters for the devices. If T1 is an ideal transformer, consider adding about 0.1 ohm in series with the secondary to represent the internal resistance of the windings.

Try putting a resistor in parallel with C1. SPICE simulators do not like nodes with no DC path to ground. Depending on how the diode and transistor models are built, you might get better convergence with a resistor there. The value is not very important as long as it does not interfere with operation of the circuit. Start with 1 megohm.

Lynn

[OStep]

It is indeed strange that this circuit fails to converge with default settings. We'll take a look at this circuit in greater detail.

For now, you can loosen the convergence criteria and it will run. If you're running interactive simulation, go to Simulate->Intreractive Simulation Settings->Analysis Options->Use Custom Options and set the following:

ABSTOL: 1e-8A

VNTOL: 1e-4V

These are completely acceptable for the voltage and current levels you are working with.

Also, I suggest that you study your rectifier circuit in a completely different simulation from the buck converter. Just replace the buck converter with some resistor that gives a similar current draw, e.g 6ohms.  Because you don't have the 400Khz switching signal from the buck converter circuit, your rectifier circuit can be simulated a lot more quickly. Once you've chosen your capacitor, then attach the buck converter circuit. 

[johnsorious]

Is there any specific equation that i should use to size my filtering cap? I'm seeing multiple ones online.

[OStep]

That is for you to experiement with. You have the theoretical equations and you have simulation tool. So why not study the simulation results and try to reconcile the equations. Try different loads (resistors) and try different capacitors and monitor the voltages and currents to understand cause and effect. 

[johnsold]

You probably see multiple equations because they are based on different assumptions about the circuit.

Just to explain my earlier comments in more detail I will show how I came up with the complete discharge comment. The charge in a capacitor Q = C*V. The charge transfered by a constant current over time T is Q = I*T. If you set those equal Q = C*V = I*T, you can make a quick and dirty estimate of the behavior of simple circuits involving energy storage in a capacitor. I used 500 uF because it is close to the capacitance in your circuit and easy to calculate in my head. For a rectifier output voltage of  12 volts Q = 12*500E-6 = 6E-3 coulombs.  Then calculate the time T it takes to remove that charge by a current of 1 A. T = Q/I = 6 ms. At 60 Hz one half cycle is 8.3 ms so the capacitor will completely discharge on each half cycle. 

Is this what you would get when you run a simulation or measure it with an oscilloscope? No. The charging is not instantaneous at the peak of the AC cycle. The discharge will not be a constant current down to zero. But it allows you to make a quick evaluation of the validity of your circuit. 

Simulations are useful for fine tuning circuits and for exploring behavior - under some circumstances. You need to understand the limitations of the circuit models and of the simulation process. You also need to have a good general understanding of how the circuit is supposed to work.

Lynn

[johnsorious]

Could you take a look and tell me what you think? I think I have it figured out. I'm looking to order parts soon so I can start building it asap for my capstone project.

[johnsorious]

I'm looking to order the parts but can't find a gate driver IC to use. I want to be able to test everything on a breadboard but the one that was recommended previously is a surface mount. Are there any through hole gate driver IC's that I can use for the buck circuit above?

[johnsorious]

Bump?