When using batteries to drive a load via a voltage regulator, is it better to step the voltage up from two batteries in parallel, or step the voltage down from two batteries in series? If there is no absolute answer, is there a general rule (or set of rules) for deciding which is best for a particular situation?

Are there any interesting cases where a particular circuit type bucks the trend? (pun not intended)


3 Answers 3


Using a step-up converter offers the advantage of being able to use partially-depleted batteries whose output voltage falls below the operating voltage of your device, but offers the disadvantages of exposing the device to harm if the battery voltage exceeds the expected level (e.g. if your 3.3-volt device operates off two AA batteries and have a 3-volt power-in jack into which someone plugs a 5-volt supply), and of increasing battery draw as the batteries get depleted (which in turn increases the risks that rechargeable batteries will be damaged, or non-rechargeable batteries will "leak" (i.e. ooze corrosive chemicals). It also, for better or for worse, will generally cause devices to continue to work normally as the batteries age, until they suddenly reach a point where they quit altogether.

Using a step-down converter offers greater immunity to input overvoltage, and will cause the voltage which is fed to the main circuitry to drop once the battery voltage has sagged too much. In many cases, this will cause the device to start working less well as the batteries age--sometimes a good thing, and sometimes a bad thing. Such behavior may be a bad thing if the device becomes useless as soon as its performance starts to degrade, but may be a good thing if the device remains somewhat useful, and if a user might not want to have to change batteries at unexpected times.

  • 1
    \$\begingroup\$ If it is higher power a step down has the advantage of keeping the copper needed much lower until the last moment before load. \$\endgroup\$
    – Kortuk
    Nov 20, 2012 at 5:21
  • \$\begingroup\$ @Kortuk: At really large power levels, use of higher voltages can reduce resistive losses in cabling, but I doubt that would be a factor for applications powered off "consumer" batteries. I suppose my answer should perhaps have mentioned that connecting batteries directly in parallel is often not a good idea, even though I've seen consumer products do it (e.g. a small light which could use two or four AA batteries, which stated that using four new AA batteries would offer more than twice the life of two new AA batteries, but weigh more). I wonder if... \$\endgroup\$
    – supercat
    Nov 20, 2012 at 15:24
  • \$\begingroup\$ ...there are any buck-mode supply chips which would be designed to accept power from multiple inputs, in such fashion as to behave similar to "diode-or'ing" them, but without the extra diode drop? If one used a mosfet on each power source, such a supply could be just as efficient as a direct connection when two sets of batteries were matched, but wouldn't let current flow from one set to another (which could be harmful) when they weren't. Thinking about it, such a supply could be especially advantageous if it was designed to draw power from the worse battery, to a point, before using... \$\endgroup\$
    – supercat
    Nov 20, 2012 at 15:30
  • \$\begingroup\$ ...the better one. A device which used two sets of batteries in such fashion could inform the user when one set was depleted, allowing the user to replace that set at his leisure while the device remained functional. Further, while most devices give users a choice between replacing batteries while they still have useful life left in them, or having a device become non-functional due to dead batteries, a device which focused drain on the weaker batteries would allow users to get all the useful life out of batteries before replacement. \$\endgroup\$
    – supercat
    Nov 20, 2012 at 15:33

Do not put batteries in parallel. Their charge level and voltage behavior are not identical event if they are of the same model. This is due to the manufacturing process.

If you connect them in parallel, your design impose that that their voltages are equal. When the design is switched off, the battery with the higher voltage will discharge into the one with the lower voltage. If the chosen battery technology doesn't allow recharging, this energy is lost. After some time the voltage will be equal and the loss of energy will stop. But if an external disturbance appends, such as an increase of temperature, the voltage of the battery cells will differ again and some energy will be lost, again.

This is why you should never connect battery cells in parallel. And if you reverse-engineers battery powered devices, they are always connected in series.


No step conversion is best for heavy loads. Get a match for both battery and load for lower losses. Generally higher voltage loads save on copper losses and copper if long haul power distribution.

For redundancy parallel cell operation is better. The stronger battery delivers more power until it matches the weaker battery. However in series the weaker battery always limits the power.

Unless the battery cell voltage is too low to convert efficiently, consider two 7.6V cells with different characteristics shown by the ESR for each. Using matched power loads, notice that the series arrangment gives lower power to the load because the weaker cell ( higher ESR) reduces the power available to the load.

Also note this only applies to rechargeable cells which equalize the voltage during charge. Primary cells would not be used this way as they would discharge against each other. Batteries wearout with higher ESR and lower A-hr capacity. Where batteries have very high power and low ESR capability external resistors and for shunt arrangement and bypass for series arrangement are needed to protect them . Battery cellls are typically matched << 2% voltage for good operation in any array, series or parallel.

enter image description here

  • \$\begingroup\$ Despite being rather anaemic, this answer seems to make sense to me. Anyone care to weigh in as to why it was downvoted? \$\endgroup\$
    – Polynomial
    Nov 20, 2012 at 8:50
  • \$\begingroup\$ Perhaps if one was careful to match the batteries cell characteristics and age and prevent reverse cell voltage on heavy draining, series operation is acceptable. Some battery chemistry such as lead-acid will age faster when this occurs. Parallel operation offers redundancy then find a motor that runs on that voltage. If not possible then stepup using a LOW ESR choke (lower than MOSFET) which is again lower than motor winding ESR is a good criteria for design. \$\endgroup\$ Nov 20, 2012 at 14:04
  • \$\begingroup\$ In many cases it is difficult or impossible to match the battery voltage to the power supply requirements. For example, many chips require 5% or 10% tolerance on rails and batteries will vary more than that between fully charged and fully discharged. Also you are assuming a heavy load, while the OP didn't mention. I didn't do the downvote, but it is justified in this case. \$\endgroup\$
    – user3624
    Nov 22, 2012 at 5:17
  • \$\begingroup\$ My suggestion was for heavy loads to match cell voltage which comes in 1.2V incrments for NiMH. That should not be hard to match. If the regulation is that tight an LDO will give the lest noise immunity. If it not that tight, it doesn't matter. So getting random performance batteries in series is not a good as the weaker one governs the current, so putting in parallel is my 2nd choice. \$\endgroup\$ Nov 22, 2012 at 5:27
  • 1
    \$\begingroup\$ Tony, is that really you with a new loging "Richman"? \$\endgroup\$ Nov 22, 2012 at 15:35

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.