0
\$\begingroup\$

I just watched Bullzoid/Actually Hardcore Overclocking's video on how VRMs actually work, and I have a couple questions.

In the video, the Bullzoid mentions how the transistors are switched on and off. I'm not if I'm right, but are the high side and low side transistors turned on and off exactly 180 degrees out of phase of each other?

wikichip

Also, since the period of switching and on and off seems to be constant, could a constant PWM signal be used to to control the VRM transistors? Given the high switching frequencies for modern motherboards, how is this signal generated so quickly yet precisely? And finally, how is the PWM signal phase shifted when a motherboard has multiple VRMs phases?

\$\endgroup\$
2
  • \$\begingroup\$ There are a ton of details and logic behind the control. You might as well just read how buck converters actually work than asking about aspects of VRMs. No, you cannot use a constant PWM signal because that defies the definition of a regulator. "Phase shifting" is trivial when you have a clock faster than the one you need. You just time offset a bunch of independent signals rather than shift a single signal. \$\endgroup\$
    – DKNguyen
    Oct 2, 2022 at 0:45
  • \$\begingroup\$ While @DKNguyen is right that a constant duty cycle would make a poor regulator, it actually WOULD work to some extent. If the load was constant (for example a resistor) it would actually work pretty well. But for a load where the current changes suddenly it is better to use a regulator to control duty cycle. There is a lot more to DC-DC converters. It isn't possible to sum up everything in a short answer. \$\endgroup\$
    – user57037
    Oct 2, 2022 at 7:19

2 Answers 2

3
\$\begingroup\$

What you're asking about here has nothing to do with a motherboard; you're asking about a buck converter, of which there are several, perhaps dozens, on a motherboard, but also in countless other devices.

The high and low side FETs are not turned on and off exactly 180° out of phase. Rather, a small delay is added between turning one FET off and turning the other on, to avoid the situation where both FETs could be on at the same time (which would short out the power supply). There are also cases where both FETs may be turned off, like in light load conditions with certain types of controller.

A constant PWM signal is not suitable for controlling a buck converter except in the condition of constant load, and even then it's not ideal. More advanced control is required; most buck converters in common use today use variants of what's called current-mode control, where an inner feedback loop regulates the current through the inductor and an outer feedback loop regulates the voltage at the output. This is in contrast to the less common voltage-mode controllers, which have only the voltage loop. In both cases, the feedback loop is tuned to give stable and quick operation under the expected load conditions; I won't go into the details, but it's a fair bit more complicated than just constant PWM. Most of that is abstracted away into a controller IC from companies like TI or LT/Analog, so the system designer doesn't need to do all the work.

\$\endgroup\$
0
\$\begingroup\$

The FETs are controlled by the buck regulator controller chip.

It also uses feedback, so it measures the output and controls PWM duty and/or frequency to keep the output at certain voltage even if the load current changes.

\$\endgroup\$

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.