Circuit design for 5V USB powered Ultra/Super Capacitor 'Booster Box'

[AlexPerth]

Managed to pick up a few 5V 3 Farad PowerStor PM Series supercaps - http://www1.cooperbussmann.com/pdf/e24b7577-c234-4ed3-a0cb-d4f1d4122374.pdf

I'd like to build a small USB 'Booster Box'. Basically a USB passthrough that would work on low power 500ma USB ports whilst still delivering a big 2-3amp punch when vaping thanks to the supercap.

With my limited knowledge these PM series supercaps appear to be an excellent candidate for this application, they have a low ESR and are designed for high current pulse applications.

I'm wondering if anyone has built anything like this before or had any thoughts on the circuit design? I know I'd have to use a resistor to limit the USB current draw for charging the cap(s) but I'm not sure if there are other considerations I have to keep in mind. Can I stack these in parallel across the power rail if I want to increase the energy storage capacity?

I'm also a bit wary of the safety factors involved and how best to fuse a device like this...I havent done the maths but I imagine discharging 3 farads via a short circuit would result in unpleasant side effects...

 

[P1NkY]

That's actually a very interesting idea, Alex! Sounds to me like it could work, also.

In my past experience with supercaps at my old job, I'd charge it up, solder an LED to one and throw it around. The damn thing would still be lit the whole 8 hour shift! And those were 1F caps! (granted, an LED is not much of a load, but, yeah_.
Stacking them in parallel will increase the capacitance and shorting them out will produce a spark, but not a life-threatening one, given the voltage (5V).

Imagine the 500ma as a trickle of water. The supercap would be a glass of water the 500ma would trickle into. The more capacitance, or stacked caps, the bigger the glass. Drinking the water out of the glass would be vaping (or the output). The larger the load (or the lower the resistance) the bigger the gulp of water. It will take time to charge said caps, and vaping could take big gulps from said glass. What's left to see is if the 500ma can fill the glass in a reasonable time frame to allow for "normal" vaping.

BTW, I hope you attempt this via a powered hub of some sort; you don't want to risk your motherboard on an experiment!

Let us know if you try this and the result!

 

[CraigHB]

Following is a dishcarge curve for a 6 Farad capacitor initially charged to 5 Volts with a 3 Ohm load. That would be two 3F caps in parallel. Charge current is not limited for this SPICE simulation. Supply power is disengaged when the capacitor is loaded.

You can see voltage falls off rather quickly, but provides a useable voltage for about 10 seconds.

It will take an amount of time to charge the capcitor at 500mA. The simulation charges the cap with a high current limit. You would need to limit USB supply current into the capacitor with a 10Ω 2.5W resistor to keep it under 500mA at all times. With the required current limiting, leaving the supply engaged while the atomizer is energized would flatten out the discharge curve to some extent, but it's not going to be a huge difference, about 30% more run time.

 

[P1NkY]

Looks like it's not gonna work, Alex. As CraigHB showed, you'd get ~10 seconds of usable voltage. I just simulated it with Electronics Workbench (yeah, I'm THAT old, lol!) and that's what I got, also.
That's not the worst part of it, though. A 10 second drag @5V is pretty harsh, indeed, but it does end up at less than 3V. The bad part is the amount of time required to get the caps charged up again.
I was able to simulate the charging time with EWB and as seen above, it charges along an exponential curve; meaning it starts charging up relatively quickly, but as it charges it pulls less and less current from the power supply, so it charges slower and slower:

1V in 17 sec
3V in 60 sec
3.6V in 79 sec
4V in 100 sec
4.3V in 121 sec
4.5 in 141 sec
4.9 in 240 sec
At the 5 minute mark, it was still at 4.96V

I then proceeded to simulate the circuit with the caps fully charged, taking a quick 3 second drag and timing how long to get back up to 5.00V. It took 7.6 minutes!!! Not good at all...

Craig was right about the 10 ohm resistor needed to keep the power supply happy. Without it I saw the current spike briefly to 17A with the caps fully discharged! The power supply would not be able to take that kind of abuse; it'd look like a dead short to it!