I recently got very interested in electrical engineering. After watching a bunch of YouTube videos on circuit and PCB design, I figured it's finally the time to get a bit hands on. I decided to start by designing a relatively simple variable DC voltage supply, as I might actually use it in the near future.

Here are some design goals:

  1. VIN can vary from 5V to 24V and the range VOUT can be set to is 1.5V-24V. VOUT can be either greater or smaller than VIN.
  2. VOUT can be set programmatically via SPI from a micro controller.
  3. I would like to keep the parts relatively cheap, so that if I accidentally blow up something it won't be too costly.

Below is the schematic:

enter image description here

Apologize if my schematic is not very professional. I'm very new to this. I would be really grateful if someone experienced can give some comments on circuit design and schematic drawing.

Here are some reasonings behind it:

  1. I chose the XL6009, as it's the IC used by the buck-boost converter I already have, and it seems to support the desirable voltage range. Further it has internal compensation to make things simpler.
  2. The inductor L1 and L2 are uncoupled, as coupled inductors are kind of difficult to find (or am I looking at the wrong places).
  3. The value of L1 and L2 are adopted from this article.
  4. The resisters of the voltage divider network are calculated following this article

I also have two specific questions:

  1. What is the general process for rating the inductors and coupling capacitors for variable voltage SEPIC? All the guides I found (such as this) are for fixed voltage SEPICs.
  2. How to dynamically limit the output current so that no component (especially L1, L2, C4 and XL6009) is fried under any possible VIN/VOUT combinations?

And finally, two meta questions (not sure if this is the right place, but I'll ask anyway. If inappropriate, feel free to ignore):

  1. Is there any way to simulate the whole circuit (with ICs) so that I debug it without physically wiring it up?
  2. Is there any dedicated places to get feedback on circuit design?

Thanks in advance!

Edit: managed to find some SEPIC design guide for XL60XX chips. Put it here in case anyone needs it: link (It's in Chinese, but the formulas are still pretty readable).

  • \$\begingroup\$ The datasheet says the XL6009 has overcurrent protection, so provided you size all the components correctly for the XL6009 MOSFET, you probably don't need an external short circuit protector. You should simulate in ltspice or similar. \$\endgroup\$ Commented Dec 13, 2021 at 3:02
  • 1
    \$\begingroup\$ @Jasen, I got 10 from ebay for about 10 bucks, including shipping. \$\endgroup\$
    – NoSegfault
    Commented Dec 13, 2021 at 3:24
  • 4
    \$\begingroup\$ If you're asking questions like this, paying slightly more for a TI or LT part that comes with an 80 page manual full of design examples and spice simulation files might be a better idea. The idea behind using these cheaper vendors is usually that you know enough about design that you save by not paying for support. It sounds like you may benefit a lot from that support though. \$\endgroup\$ Commented Dec 13, 2021 at 3:50
  • 3
    \$\begingroup\$ If you ever need a coupled inductor, the easiest way to get one is to get a ferrite core and wind it yourself. Winding a few hundred turns of magnet wire by hand shouldn't be too big a deal for a one-off. \$\endgroup\$
    – Hearth
    Commented Dec 13, 2021 at 3:56
  • 1
    \$\begingroup\$ C2,c4,c5 component selection is critical. I don’t think you’ll be able to meet the ESR requirements for C5 with one component. Besides, 2200uF seems a bit large. The circuit might appear simple, but it is a very demanding configuration where every detail is important and the pcb layout critical. \$\endgroup\$
    – Kartman
    Commented Dec 13, 2021 at 4:29

1 Answer 1


Welcome to EE.SE!

Last year I did something very similar to your project, except it didn't require truly variable output voltage, but rather changeable to two different output levels. So the feedback network is conceptually simpler, just a resistor and a MOSFET. Like your project, it was a SEPIC topology.

However, I approached the design like this:

  • I used the LT3758 chip, which (as other commenters have noted) comes with a 38 page datasheet, including a layout example, and the chip can be simulated in LTspice
  • selected the components I intend to use, and wired up a simulation in LTspice
  • verified the resulting efficiency, power losses, and expected heating of the various components, and also verified I won't exceed their maximum ratings, at various input voltages
  • I was using coupled inductors, as this increases efficiency, and this was an important point for my project (it may be for yours, too - please consider it)

After the schematic was confirmed, then it came the very important step of laying it on a PCB, keeping close adherence to the Suggested Layout in the datasheet. You don't have much leeway there - just do what the people with dozens of years of experience of designing these circuits have to give you.

The manufactured prototype then came out just as expected, within the expectations for efficiency and thermal dissipation, and even better in some regards than the simulation suggested.

In conclusion, I'd not recommend trying to implement a complex design like this through trial and error. Simulate it first!

  • \$\begingroup\$ Thanks for the answer! I took a look at the LT chips, and LT1370 seems quite promising. It even has a through hole version which makes soldering much easier. I'll probably get both XL6009 and LT1370 in the end. I also managed to find some SEPIC design guides for XL6009 and added it to the question. \$\endgroup\$
    – NoSegfault
    Commented Dec 14, 2021 at 8:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.