Battery and Supercap application

Hi, I'm trying to sort out a building block circuit out of this idea:

Given a battery, and a bank of supercaps (serialized and equalized) from where could I start looking up to find a building block to wire the two objects toghether so that, during steady time loads, I use power from the battery, and then during fast transient of high consumption I can switch over to the supercap storage?

_ I could hook two independant SEPIC converter to the two unit to derive the voltage I need for the load, but then how could I balance absorpion from the two DC bus line? it's never good to parallelize two power unit bare naked..

or

_ I could have a single SEPIC converter which is feeded by either the battery or supercap, but not just as you would do with a relay or a couple of diodes, I was thinking at some active switching schema, maybe some combination of mosfets with a asymmetrical turn-ON-OFF time?

I think overall I am missing a basic building block in electronic domain, sort of balancer, but before it, not just the plain circuit solution, I really mean how is possible in the world to even balance absorption from two identical batteries? right, starting from a basic exercize, give two AA 1.5v batteries, how would you accomodate them so that you could decide:

_ 100% power from battery 1 - no power from battery 2
_ 20% power from battery 1 - 80% power from battery 2

.. and so on? we could easily put 3 current monitor, 3 voltage monitor, a 16F690 pic (first to my mind) and write a quick looping C program and implementing a simple polling stepped response algoritm to trigger absorption from one or the other battery (outputting one or two reference pwm signals example), given sensor input and some kind of actuating signals.

But the question is.. I got all the base blocks (they are all self explanatory) but looking at the circuit, what should I enforce in the critical node to balance the flowing of currents? I simply cannot get an answer..

Are you thinking of playing with output voltages? wouldn't I risk of dissipating power if I mess with voltages outside the schema of a standard converter? again, many stardard brick power converter (traco, densei-lambda) prohibits from conneting them in parallel connections due to Vsense pins which would trigger some kind of self response..

I'm really baffle! Any ideas?
Tnx.

Just parallel the battery with the supercaps. When the current draw becomes high enough that the battery voltage starts to drop, the supercaps will start supplying extra current.

Or are you trying to protect the battery from high current?

Right! on one side I'm trying to put an upper limit to the current I will take from the battery - second, for sure I will scale up the system, and I really need to control currents frowling in the system.

Why there's this black hole in current control techniques, even at a small scale?

To explore the lack of solutions, I first wonder why?

The first problem, the battery with the super-caps. The battery has some internal resistance, and can't meet the peak requirement of the load. The resistance defines the problem and is also the solution. You might ask, what would happen to the poor battery if the resistance is too low? Won't the super-caps be deprived of their function? Yes they will, because without internal resistance there would be no problem to solve.

The second problem, I interpret as, how to exhaust the first of two batteries before beginning to use the other? I'm truly curious where and why for that one?

foxOnTheRun said:

Right! on one side I'm trying to put an upper limit to the current I will take from the battery - second, for sure I will scale up the system, and I really need to control currents frowling in the system.

Why there's this black hole in current control techniques, even at a small scale?

Click to expand...

Why are you concerned about limiting short-term current surges from the battery? Normally they do no harm. If it's a long term over-current then it would require a super-cap too large to be practical.

Current limiting is not normally done unless necessary. It usually reduces the efficiency of the systems since it requires a voltage drop to measure the current flow (unless you use hall effect devices) and the circuitry to measure the voltage drop and use that to control a current limiting device (typically a transistor).

A lot of "supercaps" are termed supercaps because they have exceptionally high capacitance for their size- however, many do not have an exceptionally low ESR. IIRC while an electrolytic might have a few ohms, and a ceramic might be low milliohms, I've seen supercaps with resistances in the tens of ohms. And since they were also low voltage caps (like 2.5v), the possible current was very low and series stacks would only worsen that situation.

That limits the useful applications. The high ESR ones can't help with surges or provide much power with the battery removed. They were only good for long-term, very low-current draws.

All right, these are good conservative suggestions, but I'm seeing we are going off topic, the question was about some building block about flow control of two, even simple, power source.

A naive idea would be that given a stabilized output voltage, eg 5v, parallel units would just cheat their outputs by few millivolts less/more individually, just enough so that having them in parallel would force them to pump less/more current (not interested in assorbing it back..), while keeping the bus voltage steady - this is just an idea, the problem is the wrong feedback from the output terminal that my units would receive back.

From what I can see, even a simple switching regulator has always (by design) a simple resistor network on the output, so if I parallelize multiple unit they would simple go crazy, becaus their feedback signal would be poisoned by the output of the other converter; what I'm feeling right now is to investigate in the building block of a more distributed system where each power source is outputting some current and it must cooperate (and know the existance of) other power modules.

It certainly sounds esotic and complicated.. that's why I started asking taking as base example a simple system..