Mitigating non-"always-on" USB battery packs

[timbopoise]

I have built several IoT devices that can be powered by USB.  I also bought several USB battery banks/packs, e.g. 20Ah, but it turns out they are not one of the few products that support "always-on".  As a consequence, when powering my PCBs with these USB Batteries they shut off after a short time.  Apparently they draw too little power and the packs shut off.

Given that "always-on" battery packs seem to be much more expensive and that I have these other packs I want to use, I had an idea for mitigating this issue and wanted some constructive feedback/thoughts:

Develop and build a small USB "dongle", ideally with a Jack and Plug Type-A connector  (to start), in case the pack only has one Jack.  The idea is to build a "loading" circuit that periodically loads
the VBus with sufficient draw to reset the battery pack's built in timer.  I'm sure battery packs will have different timeout but let me set that issue aside for now. 

The goal is to trigger a minimum frequency load pulse on the VBus (~10-30sec?) where the draw is minimal power for it's short duration.

Maybe there's such a device/product already that I can buy?

Your constructive thought?

Thanks

[Gyro]

Welcome to the forum.

Forum search pulls up several threads with solutions to this problem. Eg...

https://www.eevblog.com/forum/beginners/need-cheap-usb-power-bank-with-always-on-instead-of-auto-off/

https://www.eevblog.com/forum/projects/usb-battery-bank-auto-shutoff/

https://www.eevblog.com/forum/projects/(yet-another)-power-bank-stay-on/

Irrc, there have been a few others along the same lines. A cmos 555 timer pulsing a load resistor is probably the simplest solution.

[Hawaka]

I had the same problem while I was doing a project. I had bought one of those "BMS" board (the one in the picture) and my device would always turn off. After some reverse engineering, I found out that the board is exactly the typical application from the datasheet of the chip that does the power management.

The current is sensed through a pair of mosfet. I removed one of the mosfet to make the sense-current double what it is, and luckily enough that was good enough and no more auto-shut off. Nice thing was also that the charge current was not impacted, as it used another set of mosfet.

If you are planning to make your own battery bank it's an easy trick and it is zero cost. Maybe that solution can be done on commercial bank?

[voltsandjolts]

I found out that the board is exactly the typical application from the datasheet of the chip that does the power management.

Interesting, could you please link the datasheet here?

[Hawaka]

This is the chip: SW6124. PDF is linked

[james_s]

It's a shame these things don't all have an always-on mode. Given they already have a microcontroller in most cases, it would cost $0 to have a mode enabled by say holding down or pressing the power button multiple times to defeat the auto power off.

[tooki]

It's a shame these things don't all have an always-on mode. Given they already have a microcontroller in most cases, it would cost $0 to have a mode enabled by say holding down or pressing the power button multiple times to defeat the auto power off.

Better models do exactly that.

[james_s]

My point is simply that they should *all* do that, or at least all of the ones that contain a microcontroller, which I think is most of them. Instead you have to do some significant digging in order to find one.

[voltsandjolts]

It's a shame these things don't all have an always-on mode. Given they already have a microcontroller in most cases, it would cost $0 to have a mode enabled by say holding down or pressing the power button multiple times to defeat the auto power off.

Better models do exactly that.

Can you link some?

[tooki]

Many Anker models have it, for example https://www.anker.com/uk/products/a1229?variant=37339926397092