1

So I recently bought a ESP8266 chip, it requires 3V3 and a large current( around 210mA max which is more than that can be supplied by an arduino uno at least). I managed to connect it to my arduino uno and got it to operate in standby mode, but Wifi mode wouldn't work because it requires more power.

After reading multiple online articles, I've decided to construct my own power supply. I'm planning on hooking 3 AAA batteries in series(4V5) and then use resistors(of small values to decrease power loss) to construct a voltage divide which powers the chip.

My concern is that I have not seen a single person doing this online and this seems like a straight forward thing to do, which makes me worried. Did I miss something? Is my plan feasible?

Improve this question
4
  • See items like “5V 3.3V DC-DC Step-Down Power Supply Buck Module AMS1117 LDO” and “Mini MP1584EN DC-DC BUCK Adjustable Step Down Module” on ebay. Also, two brand new alkaline AA batteries in series,will be adequately close to 3.3 V, or two C cells if available to you, or one 3.7V 18650 LIPO cell by itself
    – James Waldby - jwpat7
    Feb 18, 2017 at 1:45
  • 1
    I wrote a whole blog article about why you can't do it: hackingmajenkoblog.wordpress.com/2016/08/30/…
    – Majenko
    Feb 18, 2017 at 10:11
  • if you want to implement jank, a couple inline diodes between +5v and Vin should drop the voltage to within the ESP's limits of 3.8-2.8v without a lot of heat and wasted current. A "buck converter" based on a switch-mode regulator would be far more ideal.
    – dandavis
    Feb 19, 2017 at 9:01
  • I've run an ESP01 off of 2 fresh AA batteries just fine; (1.6X2=3.2v; and even 2.8 is ok)
    – dandavis
    Feb 19, 2017 at 9:13

1 Answer 1

Reset to default
2

There's a good reason why you can't do this.

When you power something like an ESP, the current it consumes will vary over time, as it does different things. For the purposes of current consumption, you could model the device as a resistor whose resistance varies (greatly) over time. When it's in sleep mode with the radio off, it might be a 10k resistor. When it's active with the radio on full power, maybe a 30ohm resistor.

A potential divider divides voltage according to the ratio of the resistances. You are talking about putting your ESP in parallel with one of the resistors. Putting two resistors in parallel creates a new resistor with a resistance lower than either of the two.

If you want to calculate the resistance of two parallel resistors, it's R = (R1 x R2) / (R1 + R2).

What values would you choose for the resistors in the potential divider? You'd have to go for values much lower than 15ohm in order for the 'resistance' of the ESP not to greatly affect the behavior of the divider - say 5ohm in total. This would mean your divider alone would draw a current of 1A from a 5v supply.

Fortunately, voltage regulators and buck converters are easily and very cheaply available. I recently bought a pack of five 3.3v regulator modules from China, delivered for less than £2.

Alternatively, you can create a reasonably stable supply from a reference voltage with just a single NPN transistor, but I don't know offhand whether it would be good enough for this application.

Improve this answer
6
  • An esp presents 16-65 ohms (4X) under normal conditions, but the allowed input swing is only 0.3X...
    – dandavis
    Feb 19, 2017 at 9:11
  • no, the swing relative from 3.3 to 3.8v and 2.8v (1v/3.3v); the official range of the ESP
    – dandavis
    Feb 19, 2017 at 9:14
  • i was just adding numbers to detail why a voltage divider won't work because the highly variable current draw would force a static divider to provide volts outside the acceptable range. If the ESP always used, say, 85-110ma, one could likely get away with a naive divider (heat excepted)
    – dandavis
    Feb 19, 2017 at 9:18
  • if means that the divider could only swing 30% of it's pre-programmed 3.3 voltage without issue; but how can one accomplish that when the resistance varies 400%? (one can't)
    – dandavis
    Feb 19, 2017 at 9:21
  • or, put another way, ~15% from the midpoint for a total range of 30% of the target. you can go 15% down or 15% up, but no more. 0.99v is the spread from 3.8 to 2.8... let me rewrite my comments to clarify, can you delete your responses and i'll follow suite with a to-the-point comment? ( i hate to gunk up a good answer)
    – dandavis
    Feb 19, 2017 at 9:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.