Electronic – MC34063 heating & low performance problem

You can use smaller capacitors, depending on your current draw and ripple requirements.

The capacitor you're probably talking about is \$C_O\$, the output smoothing capacitor. You ought to look into ripple smoothing capacitors to see why frequency has entered your 'DC power' situation.

(Image source. The page itself generates a lot of browser dialogs, beware.)

The output capacitor can be calculated with the equation from the datasheet -

$$ C_O = 9*\frac{I_{out}\;t_{on}}{V_{ripple(pp)}}$$

Here's what that means. You know \$I_{out}\$, it's 20mA. You should probably know \$t_{on}\$ as well, it's the frequency part you mentioned.

The engineering enters the playing field when you need to pick a value with \$V_{ripple(pp)}\$ that fits your application. There is going to be some ripple on the DC voltage, how much is ok? Enter that value in the equation.

That equation will give you the capacitor value you can have within the manufacturer's design constraints. If you perform the calculation and come up with too big of a value, then you need to weigh that against the ripple value that you chose.

If the wire shorting out R1 is less than 0.07 ohms, that defeats the current limit function and should not be the source of the problem described.

The peak current through the inductor would be around 3 times the output current. That would be 10.5A. You need an inductor with saturation current rating higher than that. An undersized inductor could be a source of your problem.

In your schematic, there is nothing at the base Q1 to turn off the transistor. That is definitely an issue. The simplest thing you can put there is a resistor to ground.

I am curious on what do you use to simulate the circuit and where do you get the MC34063 simulation model. Also, I wonder why doesn't the simulation catch the lack of turn off on Q1.

The 180uH 10.5A saturation inductor is quite big. You can try to reduce its size by increasing the switching frequency and therefore reducing the inductance required.