One 18650 cell with boost converter vs two 18650 cells with step down converter?

So I want to power a 5V Arduino and other 5V modules with 18650 cell(s). The problem is I don't know which version would be more efficient in power consuption:

One 18650 cell with a boost converter which makes 5V out of the cell's nominal 3.7V.

Or two 18650 cells with a step down converter to 5V for the modules.

Both options have about 90% conversion efficiency, but look at the individual converter efficiency specifications to be sure.

Pololu's converters are well characterized and high quality, unlike cheap eBay or Banggood.

1 Like

tosoki_tibor:
So I want to power a 5V Arduino and other 5V modules with 18650 cell(s). The problem is I don't know which version would be more efficient in power consuption:

One 18650 cell with a boost converter which makes 5V out of the cell's nominal 3.7V.

Or two 18650 cells with a step down converter to 5V for the modules.

You should use two 18650 cells and then can use LM7805 to step down voltage to 5V.

jackthomson42:
You should use two 18650 cells and then can use LM7805 to step down voltage to 5V.

If the OP is looking for power efficiency, then 2 batteries and a buck converter are probably the winner. A linear regulator would be the least efficient.

A buck/boost converter is more efficient because it is basically a switching power supply.

Pololu’s step-up or step-down regulators run at about 90% efficiency, for either 3.3 to 5V or 7.4 to 5V.

The 7805 7.4 to 5V option would be about 67% efficient (2.4V wasted as heat).

A buck convert can be marginally more efficient, but another consideration might be that charging a single cell is easier and has fewer ways to go badly wrong.

So no real difference in efficiency provided you ignore the 7805 suggestion. Your choice probably depends on whether you mind the extra size and weight of 2 cells. Obviously two 18650s will have twice the energy of one so will last longer between charges.

Steve

If you are also going to build a charger into your project, then a 1S setup might be easier to deal with. When you charge cells in series, things can get complicated (balancing, and all that).

If one cell doesn't give you enough time, an alternative might be two 18650s in parallel. I think that would preserve charging simplicity without any real loss in efficiency.

Here's an 18650 module widely available from the usual sources:

It has both a charger and a boost converter. And unlike most powerbanks, it doesn't shut off when the load current drops. Apparently there are also two-cell and even four-cell versions of this module. I would guess the cells are paralleled, but don't know for sure.

Andreas Spiess has a great video on this module:

And I'm working on adding a load sharing circuit to it to overcome some of the limitations Andreas talks about - so it can be a real always-on Arduino UPS.

That is a very interesting battery/charger/booster module!

Thanks for posting the links.

jremington:
That is a very interesting battery/charger/booster module!

Thanks for posting the links.

Yes, it has the advantage of having everything in one package. Well, basically it's a powerbank, but designed to work with Arduino, ESP8266, etc., projects that may go to sleep from time to time. Every other powerbank I've found just shuts down when it thinks the "battery" it's charging is fully charged.

I've ordered one from Banggood, which I hope has faster shipping than AliExpress.

Not an answer to your question, however have you considered a 3.3v Arduino board? Not only does the board require less power but the losses from going from 3.7 - 4V down to 3.3 to 3.4 volts can be pretty efficient.

If you stay with the 5V boards keep in mind the efficiency of switching supplies is measured at full power. At low power outputs the efficiency drops considerably.

Also some arduino boards run on 5V but drop the 5V down to 3.3V for the processor.

JohnRob:
Not an answer to your question, however have you considered a 3.3v Arduino board? Not only does the board require less power but the losses from going from 3.7 - 4V down to 3.3 to 3.4 volts can be pretty efficient.

If you stay with the 5V boards keep in mind the efficiency of switching supplies is measured at full power. At low power outputs the efficiency drops considerably.

Also some arduino boards run on 5V but drop the 5V down to 3.3V for the processor.

I personally feel its harder to use 3.3V boards, Using them with external components is a bit rough. Almost always you need a level shifter, plus the pins output a bit less than 3V that just adds complexity for bigger projects.
But they can be useful nonetheless.

JohnRob:
Not an answer to your question, however have you considered a 3.3v Arduino board? Not only does the board require less power but the losses from going from 3.7 - 4V down to 3.3 to 3.4 volts can be pretty efficient.

If you stay with the 5V boards keep in mind the efficiency of switching supplies is measured at full power. At low power outputs the efficiency drops considerably.

Also some arduino boards run on 5V but drop the 5V down to 3.3V for the processor.

I have been thinking of that, but the other components still need 5V. Plus for example the wind vane has a 0-5V output, with a 3.3V Arduino I would need a voltage divider. So having a 3.3V board would make the project more complicated.

Do give more info on those other components. There are very few components that really need 5V. Many will happily work on other voltages.

A 3.3V Pro Mini will be perfectly happy with 3.7V fed to the 3.3V pin, thus bypassing the regulator for best power efficiency (remove regulator and power LED and you have a really low power board, especially if it sleeps most of the time). It actually can take anything from 2.3V or so (the minimum spec for running at 8 MHz) up to 5.5V (the maximum spec of the ATmega328P) on the 3.3V/Vcc pin.

wvmarle:
A 3.3V Pro Mini will be perfectly happy with 3.7V fed to the 3.3V pin, thus bypassing the regulator for best power efficiency (remove regulator and power LED and you have a really low power board, especially if it sleeps most of the time). It actually can take anything from 2.3V or so (the minimum spec for running at 8 MHz) up to 5.5V (the maximum spec of the ATmega328P) on the 3.3V/Vcc pin.

My IR remote uses that exact setup - an 8MHz 3.3V Pro Mini, driven directly by an 18650 at the 3.3V pin, with regulator and LED removed. It is awakened by a keypress, so it can sleep with no oscillators running - just keeping ram alive and waiting for that pin change. I don't have anything that will measure it accurately, but the sleep current is some fraction of a microamp. No need for an On/Off switch.