If your 3.3V microcontroller doesn‘t use much power, consider a linear regulator from the 5V rail. The lower quiescent current of linear regulators might just make it more efficient that using a switched mode.
(You’ll have to work it out)
One thing that it definitely will give you is a cleaner and more accurate 3.3V supply. With something like a LP2951, as accurate as a voltage reference, and that may save money and energy on voltage references.

 

I just realized that this is a Homework Assignment.

What is the outline / limitations of the proposed Load ?
.
.
.

 

Well, the task for the HW Assignement right now is to design a converter (SEPIC or BuckBoost ) which converts the input voltage coming from a solar panel (ranging from 3V to 19V) to a standard voltage level (8.4V) at its output so as to charge a battery (at 8.4V) and supply the 3 DC-Dc converters i described at first. But i still can't find the components' values to make it operate properly.

 

Well, the task for the HW Assignement right now is to design a converter (SEPIC or BuckBoost ) which converts the input voltage coming from a solar panel (ranging from 3V to 19V) to a standard voltage level (8.4V) at its output so as to charge a battery (at 8.4V) and supply the 3 DC-Dc converters i described at first. But i still can't find the components' values to make it operate properly.

One little question i would ask is do you know how to analyze a circuit like this in the time domain? It helps a lot to be able to do that vs using a simulator because it points out various key points in a design that you dont ever see with a simulation. That's not to say that a simulation is bad, just that a direct mathematical analysis can help understand the key points that make a design better or worse than another or why certain parts are better than others. A careful study of the parts that go into switching converters is also a good idea because they have to be able to handle stresses not always seen in other types of circuits. This is especially true of the capacitors used they can be vulnerable to over heating due to high pulse currents. Inductors of course have to be free from saturation or the whole thing literally can blow up. Transistor voltage stress as well as peak current stresses.
Another concern of late is power line harmonics but because your project is going to be run from a solar panel you probably wont have to be concerned with this unless it connects to the power line for some backup reason or something. Solar panels should also be studied as to their vulnerabilities ... back some time ago i worked with a million dollar solar panel so we had to be very careful not to damage it. Solar panels are still a little pricey but new technologies are on the way that are going to change all that. Until then through it pays to be careful.

 

hi Peter,
What other sources of SEPIC information have you studied.?

E

 

hi Peter,
What other sources of SEPIC information have you studied.?

E

Hi,
not that many, 1-2 that provide very similar calculations to the source you attached. I think that the resistance that my SEPIC "sees" on its output is very low (i calculated it around 0.77 Ohm) and that's why it cannot reach the 8.4V for inputs from 3V to 8~9V.

 

hi Peter,
Do you have an updated circuit of your design to post, I could try it in LTSpice.
E

 

hi Peter,
Do you have an updated circuit of your design to post, I could try it in LTSpice.
E

Yes, of course.
Here is the SEPIC :

And here are the three DC-DC converters connected on its outputs :

And the whole circuit :

 

The dot on L9 should be at the other end.
Not a problem if you have two separate inductors, but if you want to try coupled inductors, it should be the other way round, so that it changes to a flyback circuit if you remove C10

 

hi Peter,
This is what LTS shows for a 4R Load, check my layout for accuracy.
E

 

Hi Peter,
Load currents from 0A step 0.5A to 2A
E
Updating plot
Corrected Plot 20.08.21

 

hi Peter,
This is what LTS shows for a 4R Load, check my layout for accuracy.
E
View attachment 246001

Hi,
why do you place a 4 Ohm resistance at output? I ask because i don't think that the equivalent resistance that SEPIC "sees" on its output is 4 Ohms

 

Hi Peter,
Load currents from 0A step 0.5A to 2A
E
Updating plot

Could you explain what is this symbol at the output of the SEPIC?

 

hi Peter,
Do you mean the 'I1' , if yes, thats a dummy test load, only for the simulation.

Its possible in a sim to step that I1 component, in order to show the effect of different load currents, OK.?
E

Update: Corrected image in post #31.

 

hi Peter,
Do you mean the 'I1' , if yes, thats a dummy test load, only for the simulation.

Its possible in a sim to step that I1 component, in order to show the effect of different load currents, OK.?
E

Yes i mean I1. And in what was the range for it?

 

hi,
0A thru to 2A, in 0.5A steps
Note the image correction in post #31, sorry about that, posted the wrong clip earlier.
E

 

hi,
0A thru to 2A, in 0.5A steps
Note the image correction in post #31, sorry about that, posted the wrong clip earlier.
E
View attachment 246092

Ok now clear, and could you answer me in post #32?

 

hi P,
OK, when running a Sepic psu simulation it us usual to add some load on the output.
otherwise it would not give a meaningful result, I chose 4R as an arbitrary value.

What is the actual value seen by the Sepic output.?

BTW: in the sim the PWM is a fixed value.

E
Example:
This image for sim, shows the Vout with no load.

 

hi P,
OK, when running a Sepic psu simulation it us usual to add some load on the output.
otherwise it would not give a meaningful result, I chose 4R as an arbitrary value.

What is the actual value seen by the Sepic output.?

BTW: in the sim the PWM is a fixed value.

E
Example:
This image for sim, shows the Vout with no load.

The resistance seen by SEPIC is about 0.77 Ohms. It is the equivalent of the 3 DC-DC converters connected on its output (if you design the whole circuit as i posted it). And my assumption is that it is too low to reach 8.4V at output (for inputs 3-8V)

 

hi P,
Using 0.77R load, this is what I see.
Note: mutual inductance coupling of L1 & L2 =1
E