Power Bank Keep-Alive Load

[parkview]

It can be infuriating when trying to use a Power Bank battery pack to run small things that don't use much power. I have four phone charger power banks and they all need 40-50mA of load to keep them running, otherwise they turn off. The turn off times range from 5, 30 and 40 seconds. The 5 second one is the oldest.

The other day it occurred to me that perhaps I could make a small PCB that just incorporated an appropriate load resistor to keep the battery outputting power. Luckily I had a quick Google search to see if there where some other projects out there and sure enough there where a few. Some people used MCU's like a ATiny85, others used a 555 timer. I decided to test out a 555 timer as I have a few DIP and SMD parts in stock. I bread boarded up a circuit and yesterday afternoon I had a play with testing my three 30-40 second Power Banks:

ksnip_20231020-114814.png (617.44 KiB) Viewed 9427 times

Initially I tried using a plain 555 timer IC to drive a load resistor, however I wasn't happy with such a high turn on, so later I tried a N-MOSFET to invert the waveform. It never turned off properly. I might have been able to get around that, but I thought a PNP transistor might be an easier to work with and it was. This is the circuit I ended up using:

ksnip_20231020-120149.png (84.23 KiB) Viewed 9427 times

It took most of the afternoon to work all this out and to trial/tweak the capacitor and resistor values to get to a low current, but still keep the Power Banks alive. Interestingly, they all needed a pulse length of at least 420mSec at 40mA when using a pulse rate/period of 9 seconds. The shorter the pulse rate, the short the keep alive pulse could be - up to a point.

It didn't take long to design up a circuit for it in KiCAD and layout a PCB. To clean up the ripple noise a bit I added a two order Pi filter as well and this is what it will look like:

ksnip_20231020-105741.png (265.6 KiB) Viewed 9427 times

[parkview]

I submitted the design to JLCPCB while I was recently in Shenzhen, China and a few days later I received the blank PCB. Luckily I was I had purchased a new hot plate for the November Asian KiCad Conference, so I could quickly assemble the board. While it pulsed, it didn't keep the Power Bank alive. Turns out the circuit wasn't operating the same as the breadboard version. No idea why.

I didn't have time to dig any deeper and troubleshoot this till I arrived back home. I recently purchased a Rigol DHO 924S oscilloscope, so this would be my first go at using it to troubleshoot a project.

Power-Bank-Load.jpg (66.19 KiB) Viewed 7551 times

For some reason the NE555 IC was operating at half the designed rate. These IC's where purchased via Aliexpress, so maybe that's the issue?
Anyway, it didn't take long to adjust the two main resistors to something that worked. In the end I had a 50mA pulse with a 5 second period, with a 9.4% duty cycle which was good enough for all my current power banks I have at home. I tested the dual Pi-Power filter and found that it was having a substantial effect. Looking at the FFT of the 5V out line it had less noise and harmonics on the line.

I am calling this a finished project and one of the boards will live in my travel bag.