A
A
Andrey2020-05-14 18:57:59
Electronics
Andrey, 2020-05-14 18:57:59

How to properly power STM32 with 18650 battery?

I want to make a device on STM32 with self-powered batteries (one 18650 is enough). The circuit will have peripherals, also designed for 3.3V.
The question is how to feed the circuit correctly?
The battery has a range of 3-4V, for STM32 and other peripherals 2.7-3.6V. Accordingly, there is a range in which Akum gives more voltage than necessary. In addition, I want stability in nutrition. Ideally, those same 3.3V, but the Akum range crosses this value, i.e. just the stabilizer will not work (as I think).
Thinking a lot about this (as well as browsing the Internet) I found three options for solving this issue:

  1. leave everything as it is, adding a limit to the battery charge (so that it does not charge more than 3.6V)
  2. use two batteries, thereby getting a range of 6-8V and lowering it to 3.3V, but this is a so-so idea. too many cans
  3. after the battery, use a step-up converter (for example, up to 5V), and after it, a step-down converter to 3.3V, but there is already an excess of converters in which the charge will be lost

So far, the only option that seems not so bad to me is 2 (with two Akum). I would like to know the opinion of other people. Or other solutions to this problem.

Answer the question

In order to leave comments, you need to log in

6 answer(s)
L
lonelymyp, 2020-05-15
@andreybold

take a ready-made converter like https://aliexpress.ru/item/32894217471.html or https://aliexpress.ru/item/33029072332.html
, depending on which parameters are better suited.

P
pfg21, 2020-05-14
@pfg21

put any LDO voltage regulator, it drops about 0.45 volts or less, that's enough for you.
and monitor the supply voltage on stm32, as soon as it approaches the lower limit - signal or xs.
you indicated the minimum operating voltage of stm32 at 2.7v. in this scenario, the battery will be discharged to 2.7 + 0.45 = 3.25 volts. lithium will survive such a discharge perfectly.
2. Do not be wise with two batteries, there is no point in this. the range of an ordinary lipo with ldo stabilizer is enough for you.
3. even more stupidity, in this case, neither to the village nor to the city. if it is interesting to draw all the energy from the battery, then look towards buck-boost converters.
again, at low current consumption, the total efficiency of the pulse converter may not be better than the linear one.

G
Grigory Boev, 2020-05-14
@ProgrammerForever

Try to use a tool from TI , you can choose according to the parameters. In this case, the power supply is from 3 to 5 Volts, the current is a couple of amperes and the emphasis is on high-performance converters.
Now I tried it - I gave out 21 options, each has a diagram, BOM and the ability to download pdf with details

V
VT100, 2020-05-14
@VT100

Not enough introductory energy savings.
If the requirements are not strict and there is a power switch, recalculate the circuit for nominal power, for example, 3.0 V and put any step-down converter to the required current.
If they are not hard, but there is no power switch (only by software) - look for a converter with "true shutdown".
If hard - look for the converter by the minimum current of its own consumption at low load. For example - converters switching from PWM to PFM mode (with increased short-term voltage instability). Remember to check the possible effect of this mode on analog nodes.
Etc.

V
Viktor, 2020-05-14
@nehrung

I liked your third option more, and I vote for it. Modern RF converters are close enough to 100% in efficiency to ignore the small reduction in efficiency due to the second conversion: well, instead of 98% it will be 96%, so what? In my opinion, this is not fundamental - until recently, such equipment had an efficiency of 80 ... 85%, and this suited everyone.
But in order for the efficiency to be truly high, the intermediate voltage should not be 5 volts, but much higher (at least 12) - the higher it is, the less against its background the loss from any voltage drops on the diodes and keys. Well, the frequency is higher - up to 1 MHz or more, with the appropriate design.

V
veydlin, 2020-05-19
@veydlin

If you need to save energy, then set LDO to 3.7V, as soon as lithium falls below this value, then the stub will, as it were, "pass through" it.
Your range goes from 3V - 4.2V to 3V - 3.7V, everything will work from this range if there is nothing critical to this, then everything is fine

Didn't find what you were looking for?

Ask your question

Ask a Question

731 491 924 answers to any question