0
\$\begingroup\$

The LT3080 variable voltage regulator's datasheet shows a circuit design in which two regulators are placed in parallel to share the power load, resulting in double the maximum current capability of a single regulator. Essentially they just connect the outputs and inputs, and use the same control resistor.

The internal design of the regulator seems to be particularly ideal for this kind of application, due to the configuration of the output transistor, though I'm certainly no expert on this.

Are there any considerations I should make when setting up a supply in such a parallel configuration? Is it safe to assume that only regulators with explicit mention of this capability in the datasheet should be used in this way?

\$\endgroup\$
  • \$\begingroup\$ At best.. be sure to read any caveats carefully even if it is explicitly mentioned. Not everything in a data sheet is ready for prime time, and maybe not in your application. \$\endgroup\$ – Spehro Pefhany Mar 17 '16 at 13:45
3
\$\begingroup\$

Is it safe to assume that only regulators with explicit mention of this capability in the datasheet should be used in this way?

Yes it is !

In practice, you would need such regulators only in very rare conditions (and I cannot even think of one).

There are many regulators available with different output currents. In practice there is almost always a regulator that will fit your needs.

There are tricks to use standard regulators in parallel but these remain tricks. One of the challenges will be: how do you make sure each regulator takes half the load (share the load equally).

\$\endgroup\$
  • \$\begingroup\$ Regarding the last part of your answer, I presume you're hinting at equidistant physical placement to a single tap to the power plane on the board, to ensure equal impedance? \$\endgroup\$ – Polynomial Mar 17 '16 at 13:49
  • \$\begingroup\$ Where you tap the power plane is a much smaller effect than the possible imbalance in output voltage of two regulators. Any 'parallelable' regulators will need to have an active balance scheme that takes care of their imbalance, which will automatically sort out power plane issues. \$\endgroup\$ – Neil_UK Mar 17 '16 at 13:57
  • 1
    \$\begingroup\$ @Polynomial No, that will not help because no two regulators are 100% identical, even when the chips come from the same batch there will be differences. That's why the LT3080 needs a small output resistor. If 2 regulators for example have a 10 mV output voltage difference then you would need to drop like 100mV across a resistor to balance that difference out. That is inefficient. \$\endgroup\$ – Bimpelrekkie Mar 17 '16 at 13:58
  • 1
    \$\begingroup\$ An additional complication is the need for suitable heatsinking. If both go on the same heatsink, be aware that it has to be sized for the total heat, not just that of a single device - obvious, but some get it wrong. \$\endgroup\$ – Neil_UK Mar 17 '16 at 13:59
  • \$\begingroup\$ Very good points. I'll keep them in mind if I ever run into the need to run two regulators in parallel! \$\endgroup\$ – Polynomial Mar 17 '16 at 14:03
1
\$\begingroup\$

Regulators which have output stages that can only supply current one way (either source only or sink only) are generally safe to parallel. This includes most bipolar Linear regulators and shunt regulators.

Load sharing can be equalized by adding a small amount of resistance (say 0.1 ohms) in series with the output of each regulator.

Schottky diodes may also be used at the output of each supply when paralleling them if you are willing to have a few tenths of a volt of drop.

\$\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.