simulate this circuit – Schematic created using CircuitLab

I recently put together a simple project that was three basic LED circuits all in parallel off of a AA battery pack. The interesting twist is that two of the LEDs are of the candle flicker variety and one was a regular white LED. What I noticed was that the white LED also flickers, presumably from noise generated by the flicker LEDs rapidly changing their current draw. In this particular circuit that's an unexpected benefit, but I had thought that the batteries would be able to react to changes in current like this when it's all hooked together with decent gauge wire and no other components. What I have are a few interrelated questions around this:

  1. Is this flicker caused by an inability of the batteries to rapidly react to rapid swings in current on the order of milliamps or by inductance on the wires? (They are some inches long.)
  2. Given that each LED is actually an LED combined with a microcontroller, the technically correct approach would be to drop a 104 capacitor next to each one. But, I sometimes have a bank of these things. Is it possible to (suboptimally, but practically) have them share a smaller number of capacitors and still eliminate enough noise? e.g. I have a bank of ten all in parallel and am pondering adding a capacitor to each end of the bank.
  3. Given that these are usually the load of a circuit and only really interact with the rest of the circuit by being hooked up to ground via a transistor (or mosfet) on the low side and directly to VCC on the high side, how much noise could meaningfully propagate to the rest of the circuit, presuming that the rest is properly designed with bypass capacitors, etc. around relevant parts?

Flicker LED datasheet: https://cdn.evilmadscientist.com/catalog/components/semi/LED/datasheets/BL-L513-B-S3.pdf

  • \$\begingroup\$ Show the schematics. Perhaps there is something important missing? \$\endgroup\$
    – Justme
    Commented Dec 31, 2020 at 17:58
  • \$\begingroup\$ This is an "in general" question. The specific circuit referenced as spawning this question is simply VCC -> LED(flicker) -> Resistor -> GND, VCC -> LED(flicker) -> Resistor -> GND, VCC -> LED(ordinary) -> Resistor -> GND as described in the question. \$\endgroup\$ Commented Dec 31, 2020 at 18:08
  • \$\begingroup\$ Include specific part numbers for the "batteries" and links to the datasheets for the LEDs and resistor values. Really, a schematic would be very helpful. \$\endgroup\$ Commented Dec 31, 2020 at 18:29
  • \$\begingroup\$ But I'm not asking about that circuit. All I'm asking about is what kind of bypassing is minimally necessary for the flicker LEDs in an arbitrary circuit where it appears as the load. Here's the datasheet for those (yellow) LEDs: cdn.evilmadscientist.com/catalog/components/semi/LED/datasheets/… The rest is standard stuff: AA batteries and generic through-hole resistors and LEDs. While admittedly text, the schematic is in my first comment. \$\endgroup\$ Commented Dec 31, 2020 at 18:37
  • 1
    \$\begingroup\$ Added. Really not sure how that adds any detail not already present. \$\endgroup\$ Commented Dec 31, 2020 at 18:49

1 Answer 1


All batteries have an internal resistance. That tends to increase as the battery runs down, especially for alkaline and zinc-carbon ones. So using thicker wires will make little difference.

A single large capacitor across the power lines is likely to be more effective than small decoupling capacitors. Try tens to hundreds of microfarads if you have the space in your project.


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.