A theoretical perfect capacitor would have no leakage on DC and never lose its charge until a load was placed across it. But in the real world, there is always a slight leakage. The amount varies with the type of capacitor, and in many cases, its age. Old electrolytics tend to get leaky as they age. They also tend to lose their capacitance. But unless one has failed, the normal leakage current is in the microamp range.
The more modern they are, the better they behave.
Capacitor technology has possibly improved more in the last 40 or so years than just about anything else. I can remember walking into Central Electronics and buying a capacitor to replace one that went bad in some equipment.
The one I was replacing might be the size of my thumb, the new one, same value and often higher voltage, would be 1/4 the size.
On DC, when voltage is first applied, a capacitor is pretty much a dead short. You can measure that on an ohmmeter. As it charges up, the apparent resistance goes up until it shows as an open circuit. Infinite resistance. So, if a vehicle has several capacitors inside which have discharged, there will be a large inrush of current which will often cause a spark. But as soon as they are charged, the current drops to either zero, or whatever the various always on loads are.
Many years ago, I helped design and build a large piece of equipment. It had several power supplies and lots of wiring. When we went to power it up, one of the smaller branch fuses blew.
Since I had done the wiring, the rest of the crew had several comments about my losing my touch, etc.
But, no matter how I looked, I could find no short.
Finally I replaced the fuse and powered it up. No problems. We ran some tests, and shut down for the day. Next day, same fuse blew when we powered it up. Again, no short. And again, it ran fine when the fuse was replaced.
Turned out, we had used a switching power supply for one of the power units. They were fairly new then, and we had not realized that due to the large amount of capacitance in that type of supply, there is a large inrush of current when first powered on.
The power supply was getting about half charged before the fuse blew, and the remaining amount needed was not enough to blow the second one.
Due to the design of the system, there was not an easy solution. We needed that low value, fast blow, fuse for protection in other circuits.
We could redesign it so that the power supply was on its own circuit, but there was not a lot of time or room to do that.
Our solution was to add a delay-on-make 10 second time delay relay and put about a 100 ohm wirewound resistor in series with the input to the power supply.
The output of the power supply went through one set of Normally Open contacts and the other set of Normally Open contacts were across the resistor.
The resistor lowered the inrush current to something the fuse could handle. After 10 seconds, the relay closed, shorting across the resistor to supply full power to the system, and allowing the output of the supply to power it's circuits.
That way, there was no danger of the power supply not being able to give full power due to the resistor, and we added a small green LED to the output and labeled it "Machine Ready".
Told the customer that the machine had to do a short diagnostic when first powered up.
J.