Today I'm going to teach you about
feedback resistors in DC to DC converters. And I'm going to show you how to use this
knowledge to build an adjustable power supply with an output voltage between 2.5 volts
and 14 volts. So in a previous video I showed you how
to make a 5 ampere buck converter with a five volt output.
I told you that these resistors configure the LM2678 to have a 5 volt output. Let's talk more
about how this works. Most DC to DC converter controller
chips have a pin called the feedback pin. This is the part
of the chip that's used to monitor the output voltage. The controller basically looks at the
voltage on the feedback pin, and if the voltage is too high or too
low, it adjusts the pulse width of the switching waveform, which then gets
filtered, and the correct output voltage gets restored. In this example I have a 10 volt input
going to the supply and the load is changing between 0.5 amperes
and 5 amperes. The feedback mechanism takes care of
this, adjusts the duty cycle, and a perfect 5 volt output gets
maintained. Let's talk more about how this works, and how we can design our own feedback
resistor network. DC to DC converter controllers usually have a precise internal reference voltage called "VFB".
The exact value will depend on the chip you're using but it will always be given in the
datasheet. And it's usually around 1.2 volts. For our LM2678 it's 1.21 volts. If we removed the
feedback resistors and connected the output of the supply
directly to the feedback pin, the controller would look at the output
voltage, compare it to 1.21 volts, and then do whatever it has to do to
ensure that the output stays at 1.21 volts. But that's not very useful is it? Why
would you want to use a 1.21 volt supply? Okay let's add a 10 to 1 voltage
divider here, so whatever the output voltage is, it gets
divided by 10, and that's what the feedback pin on the
controller is receiving. This effectively multiplies the output voltage by 10 and you get 12.1 volts on the
output. So... we're dividing... but we're multiplying... which is a little weird... but
check this out. Let's say the output of the supply was
12.0 volts. This gets divided by 10, and the controller
would see 1.20 volts on the feedback pin. The controller would then say, "Hey! This
is too low! We need to increase the output voltage!" So it increases the pulse width and
raises the voltage to 12.1 volts again. The controller sees 1.21 volts on the
feedback pin, and now it's happy. Now let's say there's
a sudden drop in the output current, and the output voltage shoots up to
12.2 volts. The controller would see 1.22 volts. The negative feedback control loop
inside the chip would then reduce the duty cycle,
restoring the desired output voltage of 12.1 volts. By changing the
values of the resistors in the feedback resistor network here, we can set the output voltage to be
almost anything we want... assuming all the components can handle
the extra voltage! You can use these formulas to set the output voltage to whatever
you want it to be, within the limits of what the controller chip is capable of. You can also have a little bit of fun. If
you make these resistors fixed, and also add a variable resistor, you can
create a variable output voltage power supply. Now you have a step down power supply
that can output 2.5 volts to 14 volts DC. Right now I
have my power supply set to 13.8 volts and I am using it to charge a 12 volt
lead-acid battery. I can use the supply to dim LEDs, power amplifiers, or just see how much
voltage something can handle. Now if you remember my video about
voltage dividers I talked about how it's the ratio of
resistance values that determine the voltage. If that's the case, why not just use
these resistor values? If you think about it this would reduce
the power consumption of the circuit. But there is a trade-off! Our switch mode
power supply is switching high currents at high
frequencies. Whenever you do this your circuit will put out some
electromagnetic interference. You can see this for yourself with a
cheap AM radio. The electromagnetic interference is
inducing a small current into the antenna of my radio and it's
getting picked up as unwanted noise. Now there's a difference between how
electric and magnetic fields affect things but I'm just trying to keep things
simple here. Things get really interesting when you realize that the switch mode
power supply can actually interfere with itself! Let's say some interference from the
inductor reaches the feedback resistors. This will induce a tiny unwanted current
in the resistors. When you have current flowing through a
resistor, a difference in voltage gets created. Because volts = current multiplied by resistance, the higher the resistance, the higher the
unwanted voltage you get in the form of noise. And this can affect the controller's
ability to regulate the output voltage. In general you want to keep the total
resistance of your feedback resistors somewhere between a few kiloohms but
under 1 megaohm. This will minimize the amount of noise
in your power supply that's created by interference. This is also why I like to work on
high powered electronics with my oscilloscope probe set to X1
attenuation. The lower resistance makes them less
susceptible to interference. Alright, now you know what feedback
resistors are, and you can use this knowledge to change the output voltage
of almost any dc-dc converter! Just make sure you
double check the voltage limits of your capacitors, diodes, MOSFETs etc. according to
the design guidelines of your controller's datasheet. Sometimes overclockers use this trick on
their video cards and motherboards to change the power supply voltages to
achieve higher clock speeds. Or if you want to save power you can run
things at lower voltages. Thank you for watching and if you
enjoyed this video please check out the video description section to see how you
can support me. Make sure you check out Patreon which
is a way you can donate money per video to keep the channel going.
