Hi,
I'm new to buck regulators and would like to put together a very simple circuit to study the behavior. I'm using a bx-24 module to generate pwm logic-level signal at 14Khz. I don't want synchronous buck yet, but will probably do that next. The input voltage is 13V and I want to step it down to around 4V at 500mA. I'm not concerned about details like output ripple yet. I just want to get a basic circuit working with discrete components. All the circuits I can find on the internet don't show discrete solutions....but rather chips with integrated fets. I've looked everywhere for example showing discrete fets, etc.. and can't find it.

Thanks deeply for your help.

thundar

 

very simple circuits for the common topologies here:
http://www.powerdesigners.com/InfoWe...converter.shtm

building a circuit to understand the fundamentals is fine, but you're much better off with an integrated solution for any projects. the hard part is the feedback circuit and the switch driver circuit, as well as various forms of compensation that most of these chips perform.

__________________
If you don't have a planet, what good are gold bars?

want to contact me directly? gmail gordonthree
check out my project website: http://projects.dimension-x.net
Favorite numbers:
09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0

 

The "hard stuff" you mentioned is the fun stuff that I want to attempt using the bx-24 module. The primary goal is learning, not a robust functional product.

Thanks!

 

Thanks for the diagram. I've seen that one before. I'm looking for a diagram that shows the gate drive circuitry. That is the "black box" that I can't see to find anywhere, and which I want to replace with my uP and any other necessary components.

 

there's nothing special about the gate drive ... usually just a totem-pole (push pull) configuration of transistors, a lot of it depends on how much current you intend to switch, how fast you want to switch it, then you get into the nitty-gritty like slew-rate control, pulse skipping, etc.

the output stage of a 555 is a totem pole I believe, so are most microcontrollers, they should be able to drive a logic level fet (or darlington bjt) without any additional circuitry, at least at low current level (not more than few amps)

a recent thread in this forum was discussing various methods to drive a fet's gate.

oh btw, here's some of my fumblings regarding "digital smps" (microcontroller powered switcher):

http://forums.linear1.org/index.php/topic,604.0.html
http://projects.dimension-x.net/archives/54
http://projects.dimension-x.net/archives/55

__________________
If you don't have a planet, what good are gold bars?

want to contact me directly? gmail gordonthree
check out my project website: http://projects.dimension-x.net
Favorite numbers:
09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0

 

This is my schematic so far. N-channel fet. I think I need to boost the gate drive voltage (current 5V pulsed) above the input voltage (13Vdc). Do I have the mosfet setup properly (drain, source) ?

Attached Files

Bills_switcher.pdf (13.3 KB, 20 views)

 

just how much voltage do you intend to produce with your switcher? my opinion is your choice of fet is not suited for this application.

pick a vendor (fairchild, st, irf), and then browse their application catalog for switch-mode power supply transistors. the bss123 has a current rating of 170mA at 100v, I'd go for something rated 3 or more amps at 30v or less.

gate voltage is not as big an issue with mosfets, gate current is. voltage is a bjt requirement. your driver needs to source sufficient current to quickly charge and discharge the gate capacitor inside the fet. Most fets specify a rated voltage "Vgs" ... good switches are 2.8v, acceptable is 4.5v, anything above that will need more voltage, hence a special drive circuit (like a charge pump).

__________________
If you don't have a planet, what good are gold bars?

want to contact me directly? gmail gordonthree
check out my project website: http://projects.dimension-x.net
Favorite numbers:
09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0

 

Thanks for the tips about the mosfet selection. Since my input is 13V, and my gate is 5v max pulse....don't I need to get the gate voltage above the 13V input in order to turn on the fet? I'm still a little fuzzy on which terminal (D or S) should be connected to the 13V input voltage. As you can see, I'm trying to learn about mosfet operation in the process of doing this. Thanks so much for your help.

Is a resistor necessary between the pulse output of the uP and the gate of mosfet? The impedance of both devices should limit the current right?

 

Forgot to answer your questions. I'm stepping the voltage down to around 4~6 volts. The goals is to eventually turn this into a 4.1V li-ion battery charger with current sense feedback, etc.. I know there are easier solutions to do this, but the overall goal for me is to thoroughly explore all the variables involved in designing a li-ion charger.

 

if you were switching the high side, as in synchronous rectification, you'd need a higher voltage gate drive for the P channel device. but a simple switcher just switches the low side to charge the inductor, and uses a catch diode for the return current path. select a logic level mosfet with as low a Vgs threshold as you can find (within reason of course, no need to pay big bucks for the ultra low voltage models), and it will switch fine with 5v

__________________
If you don't have a planet, what good are gold bars?

want to contact me directly? gmail gordonthree
check out my project website: http://projects.dimension-x.net
Favorite numbers:
09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0

 

I have another amateurish question. What is the definition of high-side vs low-side. I thought that since the MOSFET was between the input source voltage and the load, that was considered "high side".

 

Will do what you want, is 90% efficient and will do current limiting.
http://www.romanblack.com/smps/smps.htm

__________________
I also post at the following sites:
http://www.stop-microsoft.org http://www.heated-debates.com
Screen name: Aloone_Jonez
And http://www.silicontronics.com, same screen name as here.

 

Victory is mine! I got my circuit to work (well, theoretically at least) using the attached schematic. My uP will replace the square pulse voltage source. I smoked my first bx-24 by not limiting the current from the pwm pin. Amateur mistake!

Attached Files

	Bills_switcher.pdf (14.0 KB, 13 views)
	Bills_switcher_plot.pdf (61.3 KB, 8 views)

 

That would work but the problem is it's ipen loop so you don't get any regulation.

__________________
I also post at the following sites:
http://www.stop-microsoft.org http://www.heated-debates.com
Screen name: Aloone_Jonez
And http://www.silicontronics.com, same screen name as here.

 

Yes, I'll be implementing current and voltage feedback to the uP. I just wanted to simulate a working buck reg first.