1
\$\begingroup\$

I have a circuit with several neopixels and an ESP32 MCU running with micropyhton. Under full load, the circuit including the ESP32 draws around 200mA at full load. Now I'd like to add another component, which is not controlled by the ESP32 but uses the same +5V net to be powered. The additional component is controlled by a switch. In normal state, it draws ~330mA. However, initially when I turn it on, it causes a current spike (I assume inrush current), which peaks at ~500mA. After the initial spike it goes back down to ~330mA. This causes the ESP32 to stall and not operate normally anymore until I take the load off.

For context, I'm running the circuit with +7.5V from a power supply capable of delivering 3A. The voltage is dropped to +5V via a LV7805CV 1.5A (including heatsink and decoupling capacitors). I'm looking to replace that with an LV78S05CV, which can handle 2A.

My question is how can I mitigate the initial inrush current of the component to prevent the ESP32 from being affected and continue normal operation. I understand that the current draw is very sudden and happens when the switch is pushed before the current draw levels off at ~330mA. Do I need inrush suppression?

Would an RC circuit before the additional component like this

How to limit inrush current?

make sense?

N.B. The additional component can be turned on/off anytime and multiple times during normal operation of the ESP32 causing a spike every time that happens.

Please let me know if something isn't clear enough.

\$\endgroup\$
7
  • \$\begingroup\$ You don't need the R in the RC Circuit, just add a large enough capacitor at your source. Values around 100u sound reasonable in your case. \$\endgroup\$
    – Adi
    Commented Jan 4, 2021 at 19:59
  • \$\begingroup\$ Yes that can work or an ICL chip on the LDO input of suitable size. \$\endgroup\$ Commented Jan 4, 2021 at 20:00
  • \$\begingroup\$ @AdityaChavan I tried a 1000uF cap right after the LV7805CV but that didn't work unfortunately. The ESP32 stalled until I removed the load again. \$\endgroup\$
    – Primesty
    Commented Jan 4, 2021 at 20:59
  • \$\begingroup\$ @TonyStewartSunnyskyguyEE75 do you have a suggestions in terms of an ICL chip? \$\endgroup\$
    – Primesty
    Commented Jan 4, 2021 at 21:00
  • \$\begingroup\$ digikey.ca/en/products/filter/inrush-current-limiters-icl/151 \$\endgroup\$ Commented Jan 4, 2021 at 22:10

1 Answer 1

1
\$\begingroup\$

I'm looking to replace that with an LV78S05CV, which can handle 2A.

Instead of replacing the original regulator, add another one to power the problematic 'component'. This keeps the surge current away from the original regulator, which then only has to handle voltage sag on its input. A 78M05 wiould be a good choice for the second regulator because it limits current to ~0.8 amps maximum.

If your power supply can deliver this current without sagging below 7 V the regulator powering the ESP32 should have no trouble maintaining a stable output when the 'component' is switched on.

If your power supply can't handle the surge current without dropping below 7 V then you can can add a diode and capacitor to the ESP32 regulator input to hold its input voltage up during the sag. The circuit looks like this:-

schematic

simulate this circuit – Schematic created using CircuitLab

Diode D1 prevents the 'component' from drawing current out of reservoir capacitor C1 when the power supply voltage sags.

Use a high current Schottky diode for lowest voltage drop (important because the 7805 has a 'dropout' voltage of ~2 V). Using an LDO (low dropout) regulator such as the L4941 to power the ESP32 could also help, as the reservoir capacitor can then discharge a bit more before the regulator drops out of regulation.

\$\endgroup\$
2
  • \$\begingroup\$ Bruce Abbot, thanks a lot for the detailed explanation! I'll definitely give this a shot! \$\endgroup\$
    – Primesty
    Commented Jan 5, 2021 at 14:32
  • \$\begingroup\$ Works like a charm! \$\endgroup\$
    – Primesty
    Commented Jan 5, 2021 at 19:10

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.