4

What is the effect, downside or purpose, if any, of using buck converters in parallel? For example, to connect the Vout and Gnd together on two converters. Could this be used to increase the current capability? Assume that the converters are identical in design.

Pål Thingbø
  • 333
  • 1
  • 3
  • 12
  • related, if not duplicate: [Is it ok to connect the output of buck regulator in parallel?](http://electronics.stackexchange.com/questions/55270/is-it-ok-to-connect-the-output-of-buck-regulator-in-parallel) – Nick Alexeev May 19 '14 at 16:57

3 Answers3

5

No real benefit is achieved. The buck that regulates to a slightly higher output voltage wins and the other buck basically switches off. However, it's probably a bit worse with synchronous buck regulators because they have active high and low mosfets and can act as a shunt regulator.

This is likely to be a no win situation for either buck with one device trying to regulate downwards by shorting the output whilst the other will try and connect it's series pass transistor to the incoming supply rail far too often in order to fight the other device.

My advice is get chips that can work in master-slave mode or design a better buck regulator.

Andy aka
  • 434,556
  • 28
  • 351
  • 777
  • If you have a 10A 12V buck regulator, then once you hit the 10A, it is likely to drop it's voltage a bit, giving the other regulator a chance to make up. Depending on how, exactly, they drop voltage, you may be able to parallel them. If they drop output to 0, then it won't work, but if they droop a little first, it might actually be fine, and you may be able to get the current you need from both. But it's not an ideal situation, and you really should be using converters meant to run in parallel. You can also force this situation by using output resistors and making them share the load. – Adam Davis May 19 '14 at 18:57
  • Can you elaborate on using output resistors? Do you mean a simple output resistor (e.g. 1 ohm) on each output? – Pål Thingbø May 21 '14 at 09:09
  • Could you please learn about the difference between "its" and "it's"? Thanks. – Matthias Urlichs Dec 23 '15 at 17:19
  • @MatthiasUrlichs I cannot learn this because I am not consistant in its use - in my answer I use it correctly and incorrectly. In this comment I've used it correctly - have you seen me use it incorrectly (or correctly) elsewhere? – Andy aka Dec 23 '15 at 17:30
4

The purpose is generally to get higher power output capabilities. In general, it may be less expensive to design a single 20-30W converter and duplicate it than to design a single 100W converter.

To reduce loading on the input supply and improve noise characteristics, you can find multiphase buck controllers. These work by sending the same clock signal but phase-shifting between the multiple converters.

HL-SDK
  • 2,539
  • 1
  • 19
  • 26
0

This can be done to:

  • improve transient response: output inductors are seen by the load as one inductor with lower value (L_{tot} = L_{ph}/N). This improves converter's transient response (check out VRM converters)
  • improve efficiency: individual phases can be switched off at light load to maximize overall efficiency
  • decrease input/output current ripple: EMI benefits, in some cases this can lead to reduction of installed electrolytic capacitors and this can lead to improved reliability
  • increase current rating/decrease conduction losses: current is split between many phases. Good at very high currents.

Downsides:

  • increased active component use: pricey
  • needs additional control (1): if individual converters are voltage control they will not share current equally between phases. This can be mitigated a bit by special design. Current control or voltage droop characteristics is advised.
  • needs additional control (2): individual controller ICs needed (pricey) or controller has to have many PWM outputs (limited number of those in typical ucontrollers)
Szymon Bęczkowski
  • 2,595
  • 15
  • 21