Charging Super Cap w/ a battery

[search4aztlan]

I'm working on a pulse power design. The duty cycle is 10%: .1sec ON @ 100mA, 1sec OFF(sleep) @ 0.1mA. The super cap (.47F) will be discharged at startup. What is the best way to combat 'inrush' at startup? My goal in this design is to power the circuit with a 3.6V battery and when the pulses (transients) occur, to power the circuit with the super cap.

So far I have the following options:
1.) using a NTC thermistor in line with a diode
2.) using a resistor in parallel with a schotky
3.) using a current limited voltage reference.

Any suggestions? The simpler the better.

 

[VEBill]

There is a single transistor plus resistor circuit that provides a current limit.

Another option (if the math works out at all the extremes) is just a simple resistor.

It is possible that the battary could handle the charging surge - if so, then just connect with wire. The user might be offended by the sparks when changing the battery...

All these assume that you want the cap to charge to the battery voltage (less the inevitable drops).

 

[search4aztlan]

I forgot to mention something very important. The voltage cannot drop below 3.1V. With a series resistor that limits the current to about 500 mA the Vdrop will be about .7V which will drop the supply voltage from 3.6 to 2.9, which will not allow the circuit(load) to run.

 

[IRstuff]

How do you figure that? Once the cap is charged, you claim that the current draw is only 0.1mA, and unless you have a 7Kohm resistor, the voltage drop should only be 0.72mV, assuming that you calculated a 7.2ohm resistor.

Your bigger problem might be that the series resistor forces the initial charge time to be over 10 seconds and that the 3 time-constants apply to the 1 second available to recharge the capacitor. The largest resistor you can use to met your other requirements is less than 0.7ohm.

This would limit your inrush to about 5A and either the battery can handle it or it'll simply be limited by its ESR.

TTFN

 

[VEBill]

I think he means that the voltage cannot be allowed to drop below 3.1v during the worst part of the recharging cycle.

Maybe he needs two super caps - one for the Incredible Pulsating Load From Hell, and another for the rest of the circuit that wants 3.1v min.

...or, maybe just split the circuit and connect the fussy part directly to the battery and the IPLFH part to the supercap side.

 

[Warpspeed]

Another way to do it would be with some sort of inductive charge pump.

It would completely eliminate the inrush problem from killing the battery, and the super-cap could be charged to a regulated voltage independent of battery voltage. It would probably be a lot more energy efficient too.

There are simple switched mode voltage converter chips that contain everything you would need to do this with just a few extra passive components. The average power is only 3.6v at 10mA, not very high.

 

[sdmays]

I'm with Warpspeed. The best bet is to use a DC-DC converter as a constant current source to charge the super cap to an appropriate voltage within the 0.9 sec before the heavy discharge. Using the battery, charge the capacitor about 9 times less than the rate at which you discharge the cap. The discharge from the battery will, likewise, be 1/9 the cap discharge, thereby easing the load on the battery. And, by using a DC-DC converter there will be very little battery power wasted.