Electronics help... need to maintain power during source switch

[K_W]

I am trying to make a battery backup for my home alarm system using a 5V power bank.

My alarm system brain uses a maximum of 5V 1A. I have a 10,000Mah power bank that both gives and takes 2 amps, a 2.5 amp charger, and a high quality 2.1A capable USB cable.

If I connect the alarm to it's wall power adapter it works, which is no surprise.
If I connect the alarm to the power bank alone, the alarm works off of the power bank.
If I then plug the charger into the power bank it starts charging and the alarm keeps working.
If I kill power to the charger with an X10 there's a split second power drop and the alarm reboots.

The power drop is not enough to see from the alarm's indicator light, it just turns red and reboots. If I hook a voltmeter with battery tester LED's to the wires I can see the blip in the output when wall power is cut.

It takes about 60 seconds for the alarm to reboot and re-arm... ok in a power outage, but too much time if bad guy is the one who cut power before breaking in.

I tried adding a small, big, and giant capacitors in parallel on the cable between the power bank and the alarm.

The alarm works with a 470, 1k and 2.2k uf capacitor, but still reboots.
The alarm lights up but will not boot and flickers with any change with a 12k uf.

I am ready to abandon capacitors as an option, but figured I'd ask first.

Can this even work with just a capacitor or do I need to look at other more complicated circuits?

 

[shibumiseeker]

You probably need way more capacitance.

Here's an easy to digest tutorial:
Energy storage in capacitors

Ultracapacitor may work but you need 2 of them in series as they are limited to about 3v.

Though TBH, I'd probably build a NiMH or NiCd with a direct charging circuit to put inline. Many LiFePo banks charge each cell separately which results in that blip when it switches to power out.

 

[JettaKnight]

These things can be tricky to calculate... which why I end up over-engineering.

If you know how much current the device actually consumes, and how low the voltage can sag before it trips, you can calculate the correct capacitance.


Ideally, the power bank would be both supplying power and maintaining it's charge at the same time - this is how a UPS works. In this case, it sounds like it actually has a switch-over circuit. And if that's the case, my guess is that the input is tied directly to output, such that no matter what capacitance you use, it won't switch until that's drained enough!

EDIT: After thinking, I now believe my theory is dumb.

 

[JettaKnight]

You could add a schottky diode....


Wall wort ---- power pack ------ (diode) ----|>|---- a few caps ---- alarm

This would allwo the power back to lose it's voltage, thus switching over, the diode prevents the power pack from seeing the caps.



I dunno - this is one of those things where'd I use a scope.

 

[K_W]

You could add a schottky diode....


Wall wort ---- power pack ------ (diode) ----|>|---- a few caps ---- alarm

This would allwo the power back to lose it's voltage, thus switching over, the diode prevents the power pack from seeing the caps.



I dunno - this is one of those things where'd I use a scope.

This is one of those times I wish I had a scope.

I've plugged the alarm into my laptop for now since it can stay up for an hour or so at rest... I'll get a new 20AWG USB cable to sacrifice and keep trying. Keep the ideas flowing.

I will be making a WTB for a left over grey AT&T gateway battery for my internet backup.

 

[woowoo2]

Plug the charger for the power bank into a small UPS.

 

[K_W]

Plug the charger for the power bank into a small UPS.

May as well use a UPS with a USB built-in.

 

[K_W]

After talking with the wife I went ahead and got a regular 120v UPS.