Electrolytic Capacitor: "Fast" charge / discharge cycles

[kellogs]

Hi,

In my project I am looking to store a bit of energy inside one of these caps during the zero volt half-period of a rectangular 10 - 15 V 1 Hz wave; the wave will typically go on for less than 1 minute at a time with max ambient temperatures of 60 degrees C. I am looking at the 270 uF @ 35V and 330 uF @ 25V caps.

These are strong 10 000 hours caps @ 105 deg C, but the datasheet does mention they are not well suited for "use in repeated charge / discharge applications" (page 5, 1.3-(2)) and further down they advise not to exceed 100 A of inrush current.

Will I be OK with a limiting series resistor of the standard 0.25 or 0.5W variety and, say, 25 ohm ?
And a side question, will there be much (if any) benefit if I went with a 35V rated cap vs a 25V rated one ?

[Zero999]

It's safe to assume the capacitor will completely charge/discharge, as the RC time constant is 8.25ms, which is much less than the period of 1s, for a 1Hz squarewave.

The energy stored in a capacitor can be calculated using the following formula:
E = 0.5CV²

If you have a 330µF capacitor, charged to 15V.

E = 0.5*330*10^-6*15² = 165*10^-6*225 = 0.037125 = 37.125mJ

When the capacitor is charged, the same amount of energy going into the capacitor, will be dissipated in the resistor.

When it's discharged all of the energy will be dissipated in the resistor.

Therefore the energy dissipated in the resistor per cycle, is double that of the formula listed above. Power is Joules per second and as the frequency is 1Hz, so the power dissipation will be 2*37.125m = 74.25mW.

The next point is the peak power dissipated in the resistor. 
P = V²/R = 15²/25 = 225/25 = 9W

Your resistor will obviously be able to handle the average power, but it will need to be able to withstand 9W peak. A 250mW thin film resistor will probably smoke. I'd opt for a 1W wire wound resistor. 25R isn't  a standard E24 value. Go for 24R, or 27R.

I doubt increasing the voltage rating of the capacitor will gain you much.

[tszaboo]

Ripple current is defined in the datasheet.
Polymer caps are not great for low power electronics, they are very leaky. YMMV depending on the application.

[kellogs]

Ripple current is defined in the datasheet.
Polymer caps are not great for low power electronics, they are very leaky. YMMV depending on the application.

Ripple current is 2.5A @ 100 kHz, so I guess I am covered on this side, my question was more about the type and wattage of limiting resistor to use. Can you elaborate on the second statement -"Polymer caps are not great for low power electronics, they are very leaky" ?

Like, I need to  draw some 0.7 mA for half a second from them. Is the leakage going to prevent me from doing this ?

Your resistor will obviously be able to handle the average power, but it will need to be able to withstand 9W peak. A 250mW thin film resistor will probably smoke. I'd opt for a 1W wire wound resistor. 25R isn't  a standard E24 value. Go for 24R, or 27R.

I doubt increasing the voltage rating of the capacitor will gain you much.

I should have mention there is a diode just before the cap. So it will not discharge at the 0V of wave but in the opposite direction, powering a MCU for 0.5s @ an average of 0.7 mA. Does this change the power rating and type for the limiting resistor ? Will a 27 ohm thick film, 0.5W @ 1206 handle this regime ?

Thanks