5
\$\begingroup\$

I have been reading about the buck converter and have also referred to the various online resources like here.

Enter image description here

In the above circuit, as I understand, when the switch closes, current starts to increase across the inductor linearly with time. What is the voltage at the output?

Current increasing would mean that the current across resistor would also increase linearly, and hence the voltage should practically be increasing, and not be stable, so how does it become stable/constant?

What exactly happens when the switch is opened?

The capacitor would cause ripples to pass through, but what about the resistor? From what I understood, the current decreases linearly with time and is due to the inductor (energy stored previously). But since it's decreasing, wouldn't the voltage at the output also decrease linearly with it?

So how do we get a stable output?

I guess the timing /duty cycle of switch would play a major part, but I still couldn't figure that out.

\$\endgroup\$
3
\$\begingroup\$

Here's the picture: -

enter image description here

It's significantly more complex than what I'm going to say but bear with me: -

Firstly I'm going to ask you to imagine that D1 is a switch like SW1 but, it closes when SW1 opens. What you then get at point (2) on the circuit is a square wave; it has a peak of Vin (let's say 10V) and rapidly drops to 0V when SW1 opens (remember I've asked you to consider that D1 is also a switch). This repeats at some arbitrarily high frequency such as 100kHz.

Let's also say that Vd (your output) is desired to be 5V. Now if the 10V squarewave at (2) spent half of its time at 10V and half of its time at 0V then the average value would be 5V i.e. exactly what you want.

Should you in fact require Vd to be 3.3V then the squarewave at (2) would spend about a third of its time at 10V and about two-thirds of its time at 0V. (Remember I'm asking you to consider that D1 is a switch that closes when SW1 opens).

So you've got a squarewave at (2) that has a duty cycle of (say one-third) AND now you have a low pass filter formed by L1 and C1 - the output from this is pretty much a dc voltage at 3.3V.

Then, you put a load resistor on (\$R_L\$) - does this alter the output voltage average level? The answer is virtually "no" because you are using an inductor and capacitor to form a low pass circuit and providing the inductor's internal resistance isn't too big then there won't be too much of a dc voltdrop across L1 and you'll still get 3.3V at the output.

However, if you load the output too much, the 3.3V will start to droop and this is when the control circuit starts to take over and apply a little more than one-third duty cycle to SW1 being on. This control loop is fundamental in all buck regulator circuits but it isn't necessary to understand the ins and outs of this to comprehend the basic working.

So far I've assumed D1 is a switch (like SW1) and what 've described is called a synchronous buck regulator - it uses two MOSFETs; one for SW1 and one in place of D1. I think it's easier to approach synchronous buck regulators first then move on to understanding standard (but less efficient) buck regulators.

Standard buck regulators have D1 (not a switch) and they would like D1 to behave like a switch (as described above) but it doesn't always do this. For a start it drops 0.7V across it when it is acting like a switch (maybe a bit less if you use a schottky diode). It conducts like a switch but the 0.7V across it loses energy in the form of heat - it can never match the efficiency of the synchronous regulator.

How does D1 behave like a switch - when SW1 is closed, a ramping current runs through L1 and when SW1 eventually opens, the back emf from L1 drives (2) negative in order to keep the current through L1 still flowing. This is the nature of inductors and if this is a little alien to you go and study inductors. This negative voltage rapidly falls below 0V until D1 starts to conduct - now it is behaving like a synchronous regulator (albeit with 0.7V drop across it). SW1 eventually starts conducting again and the cycle repeats.

Previously the squarewave described at point (2) was 10V peak and 0V at the bottom - now it is 10V peak but -0.7V at the bottom. L1 and C1 are still a low pass filter (as mentioned previously and if the duty cycle of the square wave was one-third, the voltage at the output would be about 3.1V. The control system would take over and alter the duty cycle until the output was 3.3V.

But there's a further problem when D1 is just a diode (and not a switch) and this really does make non-synchronous buck regulators quite tricky to get to grips with. If the load is very light, D1 doesn't act like a switch (as per a synchronous regulator) and the output voltage rises and rises because the energy stored in L1 keeps getting pumped into C1 and of course the output voltage rises. It's not a big problem because the control loop keeps this in check by applying ever smaller duty cycles the the squarewave at point (2).

I've got the point in my explanation where I need to take a break so if you manage to wade through this and want more let me know. The upshot of what I would explain is surrounds storing energy in the inductor (when SW1 closes), transferring it to the capacitor and making sure that the transferred energy \$\times\$ frequency (cycles per second) matches the power needed by the load resistor at the voltage the regulator is intending to regulate at.

Synchronous regulators are far easier to explain!!

\$\endgroup\$
1
\$\begingroup\$

Your assumptions are indeed correct. The output of a buck converter is not 100% pure DC like a linear regulator. There will always be ripple there, directly proportional to the converter duty cycle. During the switch on-time, the increasing inductor current causes the capacitor voltage to increase, and during the switch off-time, the inductor current decreases, as does the capacitor voltage.

The reason the buck converter is a reasonably-good DC power supply depends on the size of the inductor and the switching frequency of the converter. A well-designed buck converter's output ripple due to switching will be around 1% of the overall DC level (50mV of peak-to-peak ripple for a 5VDC output, for instance).

Since the inductor is supplying current during both the on-time and off-time of the switch, output capacitor ESR is usually the constraining feature in terms of how much capacitor you need (in order to get good ripple performance, you need small ESR; to get small ESR, you usually put many caps in parallel and you end up using more than what you 'need' to support the output voltage during the switch off-time)

\$\endgroup\$

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.