I am trying to design a four mode buck boost converter based on the LTC3789 controller such that I want to completely substitute all the PLL functions, clock and everything else internal to it so that I can exercise complete control using a PIC16F877 micro for my charge controller project. Please How do I do this?

enter image description here

  • \$\begingroup\$ I very much doubt that a PIC can do this because you will need fast analogue error amplifiers for accurately controlling the functionality equivalent to what an LTC3789 does. \$\endgroup\$ – Andy aka Sep 3 '15 at 9:47
  • \$\begingroup\$ Switched mode converters need dedicated controllers, using a generic micro is a recipe for failure. And a 4 mode buck-boost, that's one of the most complicated switching converter systems there is. ICs like the LTC3789 are designed by a team of experts with years of experience. Why do you think you can simply implement something similar using a PIC ? \$\endgroup\$ – Bimpelrekkie Sep 3 '15 at 9:53
  • \$\begingroup\$ @FakeMoustache I would beg to differ, as I have done it. Not sure what you mean by 'dedicated controller.' Sure, some micro's will have better off the shelf libraries for running certain control loops, but it's not too hard to code your own PID loops. I would say that starting out with this very complicated switcher would be a bit ambitious. Buck converters are sort of a 'darling' of the academic world. I'd start there as there are the most resources+linearized models for them. Maybe think about a dspic or a c2000 instead though. \$\endgroup\$ – RYS Sep 3 '15 at 10:03
  • \$\begingroup\$ @RYS There are controller ICs that are specifically designed for switched more power applications. These usually include some circuitry to do current monitoring for example. For simple applications your solution will work no doubt but how robust is it ? What happens if the output is shorted or the inductor saturates ? Dedicated controllers have circuits which respond quickly enough to prevent damage. For low power it is not much of an issue but for high power applications I would not want to risk that things blow up because the PIC program crashed. \$\endgroup\$ – Bimpelrekkie Sep 3 '15 at 12:48

Take a look at this circuit. It was developed to test a half bridge buck-boost controller using an ATmega: -

enter image description here

It might be worth considering using this approach rather than try and model exactly what might go on inside the Linear Tech chip. Here is the article. The circuit should be fairly easily converted to full H bridge by replacing D2 and D3 with mosfets and appropriate drivers.

If you think this is the way to go I would urge to to google "H bridge buck boost controllers" rather than "4 mode".

| improve this answer | |
  • \$\begingroup\$ Is it strange that this circuit shows 5V in to a 7805? \$\endgroup\$ – RYS Sep 3 '15 at 9:59
  • \$\begingroup\$ @RYS I believe they have the 7805 drawn the wrong way round. Input supply is on the right hand side, 5V output feeds the atmega. \$\endgroup\$ – Andy aka Sep 3 '15 at 10:03
  • \$\begingroup\$ that would make more sense! \$\endgroup\$ – RYS Sep 3 '15 at 10:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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