# Intermittently supply a micro controller from a weak DC source

#### sciortapiecoro

Joined Sep 9, 2022
3
Hi everyone! I have the following problem at hand, which is probably trivial but way beyond my ability to design circuits.

I have a very low current DC voltage coming from a Cockcroft-Walton Multiplier, that I would like to use to intermittently supply at least T seconds of 3.3±0.2 V DC to a micro-controller while guaranteeing its requirements for peak current.

It feels to me that, at least in theory, I should be able to charge a (big enough) capacitor for long enough and then discharge it in a controlled way such that I can guarantee T seconds of operations to the microcontroller.
I'm happy to wait indefinetely between each discharge to make sure that the capacitor charges despite the weak input current, but I'm by no means able to understand how to design a circuit like that.

Any hint would be appreciated!

#### Papabravo

Joined Feb 24, 2006
19,288
The central feature of the circuit is the capacitor. For a given current the voltage will rise exponentially and you can tell how long that will take. In order to speed the process up as much as possible you want the smallest possible resistance in the charging path. For a given load current the voltage will discharge exponentially as well. In this case you want a large resistance to slow the discharge down, so you get at least T seconds at or above a certain voltage.

I think this quest is unlikely to produce the results that you seek, but you are free to run the experiment.

#### BobTPH

Joined Jun 5, 2013
5,762
The fact that you have a multiplier implies you have a lower AC voltage coming in. can you not use that source to either charge the capacitor or run the micro directly?

#### sciortapiecoro

Joined Sep 9, 2022
3
Thank you both for your replies!

The central feature of the circuit is the capacitor. For a given current the voltage will rise exponentially and you can tell how long that will take. In order to speed the process up as much as possible you want the smallest possible resistance in the charging path. For a given load current the voltage will discharge exponentially as well. In this case you want a large resistance to slow the discharge down, so you get at least T seconds at or above a certain voltage.

I think this quest is unlikely to produce the results that you seek, but you are free to run the experiment.
That makes sense to me, what I don't understand is how to automatically switch from the charging path to the load path when the capacitor is enough charged.
Normally I would do it with a switch, either manual or controlled by some external signal, but I'd like to have some circuit that does it by itself.

In practice, it should do something like

1. Charge the capacitor until reaching some voltace V_max
2. When V_max is reached, disconnect the charging circuit and switch on the load circuit
3. When the voltage across the capacitor goes down to V_min, switch off the load circuit and switch back to the charging circuit.
4. Back to 1.

The fact that you have a multiplier implies you have a lower AC voltage coming in. can you not use that source to either charge the capacitor or run the micro directly?
Yes I can definetely use the source to charge the capacitor, it is just weaker than the requirements for the microcontroller both in terms of voltage and maximum current.

Also, I'm also quite curious now to understand if this problem is solvable with those constraints.

#### sciortapiecoro

Joined Sep 9, 2022
3
To make an analogy: I'm looking for an electronic circuit equivalent to a toilet flush
It charges slowly but then discharges completely with a shorter burst of higher power

#### DickCappels

Joined Aug 21, 2008
9,317
Can you explain more about how the multiplier is being fed? The output voltage is very low. Are there active switches in the multiplier? May as well take care of any potential problems with that before getting to deeply into how to pay out the energy.

#### BobTPH

Joined Jun 5, 2013
5,762
What is the input source? perhaps a voltage multiplier is not the best solution. They are not efficient at low input voltages due to the diode drops.

#### Papabravo

Joined Feb 24, 2006
19,288
Since you know the rate of charging and the rate of discharging you can do it all with a time base circuit. So let us say that you have 100 units of time, and you wish to charge for 93 units of time and discharge for 7 units of time. A little bit of circuitry to provide some deadtime between each of the two time intervals and you should be set to go. The unsophisticated experiment, after reaching steady state, looks something like this:

A more sophisticated circuit would involve actual switches.

Last edited: