0
\$\begingroup\$

In a PC power supply I have, it implements a synchronous buck converter that uses two active switches without a diode. I found a picture of the schematic online, whose link is below: https://e2e.ti.com/cfs-file/__key/communityserver-blogs-components-weblogfiles/00-00-00-03-59/1411.fig1.png

I can see why they use a MOSFET for a lower switch: it may have a smaller loss than a diode.

This circuit reminds me of the regenerative control for DC motors, where the two buck switches can be used as a boost converter when treating the output as input. I am wondering if regeneration is relevant for this power supply, or it is not one of the reasons they choose a synchronous buck converter.

\$\endgroup\$
2
  • \$\begingroup\$ What energy do you expect could be regenerated? \$\endgroup\$
    – Hearth
    Commented Oct 4, 2022 at 1:37
  • \$\begingroup\$ A synchronous buck converter forced to operate in constant conduction mode may very well be capable of regen. However, there would have to be a power source connected to the output. \$\endgroup\$
    – user57037
    Commented Oct 4, 2022 at 2:19

3 Answers 3

1
\$\begingroup\$

As far as I'm aware the answer is no, and it wouldn't make much practical sense.

A vehicle takes energy from some power source (electric, chemical) and turns it into kinetic energy as its velocity increases. When you want to slow down the vehicle, you must convert some of that kinetic energy into another form. Usually, with brakes, you're turning that kinetic energy into thermal energy. Since a motor also happens to function as a generator if you rotate the shaft, you can instead translate some of that kinetic energy into electrical energy to recharge the battery, i.e. regenerative braking. The DC-DC converter design has to be built to support this, but it's not really the key factor behind why this kind of regeneration works in cars.

In a computer there is no build up of energy. The computer isn't gaining kinetic energy as it computes things (or at least I hope it isn't), nor is it storing the energy in some other form like chemical potential energy. The electrical energy is turned into thermal energy, which we have to dissipate in order to keep the temperature of the components at a reasonable level. That thermal energy is dissipated out into the environment and cannot be recovered by the power supply.

That isn't to say that datacenters don't sometimes use that waste heat in a useful way. Air conditioning also uses energy, so for every watt of thermal energy that the equipment in the DC dissipates you have to use even more power to pump that heat out of the building. This turns out to be really expensive. In some cases datacenters will use heat exchangers to heat up their building's hot water, to make some use out of that heat.

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

I seriously doubt it.

For motor-based regeneration, you take the kinetic and magnetic energy in the motor (and whatever is attached to it) and put it back into the power source. This could be quite substantial amount of energy and may be worth reclaiming.

For a computer system, what energy exists that could be put back into the power source?

  • Power supply capacitors have energy stored in them, but it if you take that energy out, then the voltages would drop, and the computer would turn off.

  • Magnetic components in power supply filters or switching converters have energy stored in them. But it's typically an insignificant amount of energy compared to how much that kind of system normally draws.

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

Regenerative control uses an h-bridge with four transistors (FETS or BJTs) so current can be controlled in both directions.

enter image description here Source: https://www.roboteq.com/applications/all-blogs/22-understanding-regeneration

In a Half bridge circuit as shown in the OP the current can only flow through the switches in one direction. A Buck converter can use a diode for the lower switch or a mosfet which is more efficient because it has very small voltage drop across it (compared to a diode)

\$\endgroup\$
2
  • \$\begingroup\$ Well, a synchronous buck in continuous conduction mode can maintain regulation at the output even if current is injected at the output. Of course, this will deliver energy to the input side, which may result in increased bus voltage unless the input side can somehow utilize the energy. \$\endgroup\$
    – user57037
    Commented Oct 4, 2022 at 4:40
  • 1
    \$\begingroup\$ But there is no motor in the OP's question so... \$\endgroup\$
    – user57037
    Commented Oct 4, 2022 at 4:42

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.