# Do capacitors automatically release their energy over time?

Will a capacitor automatically release its energy over time? Or will it stay in there until manually discharged?

So let's say I've had an old computer sitting around for a year and decide to take every piece apart... am I in danger of being shocked by the capacitors?

• I just wanted to add, BE CAREFUL when shorting capacitors to discharge them. Small caps you might get away with a dead short, but you're usually better off with a resistance or specific tool. Large caps, a dead short can cause a significant flash-bang, or even throw the remains of your screwdriver across the room (and into someone's face). One trick I've used is to use a wire to short the cap, but coil it several times before making contact. The added inductance limits the spark. – Stephen Collings Sep 14 '12 at 13:00

In theory it will. If an ideal capacitor is charged to a voltage and is disconnected it will hold it's charge.

In practice a capacitor has all kinds of non-ideal properties. Capacitors have 'leakage resistors'; you can picture them as a very high ohmic resistor (mega ohm's) parallel to the capacitor. When you disconnect a capacitor, it will be discharged via this parasitic resistor.

A big capacitor may hold a charge for some time, but I don't think you will ever get much further than 1 day in ideal circumstances. You should watch out if you have turned on the PC just 'a moment ago', but if you let it unplugged for a couple of hours and it will be fine.

The capacitors in the mains power supply are the most suspicious, these contain high voltages and high capacitance. If you don't know for sure, measure them. You can short them out if you find something, like the device Nick shows.. (it's probably a high voltage 1 kilo ohm resistor or something with some wires and isolation). But I suspect those are quite expensive and more designed for really high voltage situations (like kV's).

Or if you dare an old isolated screw driver (watch out for sparks though! :-) ). But I think it's obvious a very direct short will not enlarge the life of the components.

• The trick they tell you at any Tech Support/IT place is to hold down the computer's power button while the computer is unplugged. What they don't tell you is that this is draining any capacitors that might be malfunctioning and causing PC problems. – kevlar1818 May 24 '12 at 20:37
• There are NO UNSAFE voltages on a motherboard. But if you want to reset the card in case of CMOS latchup effects on the USB port or other PS2 port from glitches. Pressing the power switch will discharge the caps quickly with 1KΩ or so. – Tony Stewart Sunnyskyguy EE75 May 24 '12 at 23:38
• The item of significance on my equivalent circuit is C2 and ignoring that fact may give you a jolt. They don't make PC power supplies the way they made old TV sets so the storage time on caps is relatively short (minutes at best) – Tony Stewart Sunnyskyguy EE75 May 25 '12 at 0:10
• FYI be vary careful with capacitors in other devices. I one gave myself quite the fireworks show when I shorted a small camera flash capacitor with a metal rod (it melted): lifeisaprayer.com/blog/2007/rip-canon-powershot-s2 – geerlingguy May 25 '12 at 4:15

Discharge the capacitors yourself. This is a common procedure. There is even a tool for that, although you can make an improvised one.

from this post. Good discussion there too.

Well-designed high voltage circuits have bleed resistors for discharging high voltage capacitors.

Real (as opposed to ideal) capacitor has leakage resistance. It can be viewed as a large resistance in parallel with capacitor. There is a leakage current, which could be on the order of 1uA in large electrolytic capacitors.

Short it out for a few seconds.... On large electrolytic caps, like "main-frame" computer grade 100,000uF and TV HV 10uF 25KV doubler Caps, power supplies there is a phenomena like in batteries, known as memory. After you short it out the voltage creeps back. That's all you need to know. Short it long enough to discharge the memory effect.

Actually the capacitor has a few more non-ideal characteristics that can be put into the schematic. So the rest of this is for educational, technical, and factual values.

What's All this Stuff About Capacitors.

In fact some of my grey beard colleagues will remember that caps in storage like this required slow "conditioning" to prevent puncturing the insulation inside so a slow charge up for an hour was recommended before use. Thats the C2 cap physical property. It can short out.

The MAIN Capacitance is C1, the memory cap is effectively 5x to 10x bigger in electrolytics. However neglect ( <<1x C1 ) in ceramic/plastic caps . This Memory capacitance C2 may be the smaller or much higher so the original voltage is restore but the series resistance R3 is enough so you cant get much current from it but it can give you a jolt if you only short the cap for a zap or split second.

C1 = Main cap C2 = memory cap in electrolytics C3 = vibrational cap in ceramic caps ( like piezo or crystals) (tiny but can cause noise)

D1 = in Polar Caps this reverse limit is usually > 15% of the rated voltage, which means you an use a Polar cap as a Non-Polar cap if you promise to use it only for small signals < 10% of rated V, such as undershoot. D2 = n.a. D3 = the forward voltage rating of the cap . D4,D5 = diodes for voltage steering behaviours and Drop > 10% of rated voltage

R1 = the main ESR of the Cap R2 = the self-leakage of the cap Very high in certain electrolytes 10^8 and plastic caps 10^10Ω , so the Effective Series Resistance of the Cap (ESR) is R1 and is temperature sensitive. R3 = the ESR of the memory cap .. >> 100x the ESR R1 R4 = the resistance of the Forward voltage rating in Polar caps is nonlinear and can be a negative resistance and cause fire-cracker like explosions in tantalum caps since it is also negative temp coefficient so self heating draws more current when at least 10% greater than rated voltage. and also when self-heating

L1 = self inductance of foil and/or leads. Monolithic caps are almost uncommon these days but bigger, more reliable but these days multi-layer metallized caps are most common.

The importance of each value depends on whether it is Polar or not, ceramic or not (C2).

The most ideal caps in electronics are also the most expensive. (we're not talking power transmission line PFC caps) When it comes to least leakage, low ESR, most stable with temperature, self-healing from spiked over-voltage , most reliable. I am referring to plastic Caps Teflon then Polyurethane, Mylar. (Mylar were default used in old telephones) If you want a time constant in minutes or maybe hours in some cases, that is possible. There a dozens of other materials including silver mica, and a few more exotic materials.

But to answer your question "Do not Forget", C2, the memory cap when discharging TV flyback triplers on old TV sets. Not an issue at all on PC's as there are only LOW VOLTAGE caps on the motherboard since all the HV is well protected inside the PSU case. I recommend shorting out C1 and count to 5 seconds, but dont take my word, zap one and then measure it.. If you have a 10MΩ DMM it would show slowly rising voltage. The resulting voltage indicates the Cap Ratio. Equal values would return to 50% voltage.

Just a heads up from 35 years of experience on caps, from all sides of the fence.

p.s. You will not likely find simulators use my schematic, but it is accurate. There are some variations and you can neglect most of them if you use it within guidelines of the component.

*p.p.s. If you have any Ultra-caps or just plain SuperCaps, you can measure these values. Ultra_caps are distinguished by remarkably low ESR. Supercaps were great for Car Bass boosters and  Standby power for embedded products with RAM where Lithium is not allowed. etc.*


Some thin film Polyurethane Caps are good for hundreds of Amps in small packages.and only \$1

Capacitors will lose their charge over time, and especially aluminium electrolyts do have some leakage. Even a low-leakage type, like this one will lose 1V in just 20s (1000$\mu$F/25V). Nevertheless, YMMV, and you will see capacitors which can hold their charge for several months.

It's wise to discharge them. Don't short-circuit them right away, they don't like that. Discharge them across a resistor. The voltage will drop quickly at the begin, and then slower and slower. If the voltage has been reduced to a few tens of percent of the rated voltage you can short them to speed up the process. Short for some seconds, if you only short briefly the voltage will rise again if you remove the short.

The ideal discharge procedure is through a constant current, so that the voltage drops at a constant rate and the total discharge will end quickly. Discharging via a resistor is exponential and theoretically takes forever.

What's All This Capacitor Leakage Stuff, Anyhow? (by Bob Pease)

• I know I suggest a silly idea. But can't we create a constant current drain using a OP AMP circuit? – Standard Sandun May 25 '12 at 18:27
• @sandun - yes. An opamp, a resistor and a transistor. Or just a transistor and a diode. It's not that hard. – stevenvh May 26 '12 at 3:50

The capacitors on your PC are unlikely to be able to harm you simply because the voltages are so low.

In the past, when vacuum tubes were common, DC power supplies at dangerous and lethal voltages were used. These power supplies were bypassed (filtered) with capacitors that could hold a charge for a very long time.

It became a common practice to always shunt these capacitors with a large resistor (1 M-ohm, for example) to discharge the capacitors when the equipment was turned off. This is the same idea as the discharge probe described in another answer to your question, but it's always there in the circuit. (By the way, the discharge probe uses resistors to limit the current of the discharge, which is a lot safer that just shorting the capacitor with a conductor.)

But if you ever come across an old radio or anything with a vacuum tube and high voltage supply (or an X-ray machine, if you're into that), be very careful. Especially immediately after the power has been turned off. But also, in the unlikely event that its designer forgot those shunt resistors (often called bleeder resistors, because they bleed the residual charge away), always keep one hand in your pocket when poking around.

• Yes Jim, the memory in those old caps is something we old timers dont forget ( pun intended) – Tony Stewart Sunnyskyguy EE75 May 25 '12 at 18:21