Help me understand capacitors

[Haunz 10+ year member]

Your conclusion is wrong.. by definition a 1F can will hold 1 coulomb of charge @ 1 volt of potential... at 2 volts it will hold 4... and at 12v it will hold 144...

50A current draw from a 14v source is equal to ~0.28 ohms of resistance.
The RC time constant is a fixed time interval equal to resistance times capacitance. It takes ~5 time intervals to fully discharge a capacitor.

280x10^-3 ohms x 1 F = .28 seconds. Five times this is ~1.4 seconds. This is from fully charged to fully discharged.

Since the discharge curve of a capacitor resembles exponential decay, it takes a very short time for the fully charged capacitor (sitting @ 14v) to reach 11v. This would happen within 1 time constant, or less than .28 seconds. (Edit: By looking at a discharge curve, 80% charge is remaining at just over 1/5th the RC time constant... in this example that is approximately 0.056 seconds)

It happens VERY quickly when using a capacitor as a voltage source.

You have to remember this is using the capacitor as the only source. Complimenting a 14vdc source will decrease the discharge rate.

Your post is fine but that is not the RC time constant for a typical 1F can...

if our charging system fluctuates between 12 and 14 volts a 1F can will hold a difference of 52 Joules... thats 52 amps for one second, 104 amps for half a second, 208 amps for one quarter of a second, and so on..... (what we are interested in is sustaining power for transients that may last for 10~500ms; so you can do the math)

Here is where the RC time constant comes into play... at 1Khz a typical 1F can will measure around .0018 ohms... that works out to a RC time constant of 18ms.... You EE wanabee's can check my math but since we are looking at the difference between 12 and 14 volts I see that 1F cap being able to drop 52 Joules in .00277seconds (2.8ms)

The catch is that we are not just dealing with an RC circuit.. but an RCL circuit... and although we may find an ESR of .0018 Ohms at 1khz it will be significantly higher at a 20Hz sample rate due to the cap's ESL.... plus you can add to that the inductance/resistance of any additional wiring that's needed to hook the cap up....

Taking everything into account, the best way to characterize the situation is to actually test it out in the real world... I've looked at several real graphs of voltage fluctuation with and without caps installed.... and although it is basically meaningless, the graph ngsm13 posted attempts to show that caps work as intended... (even if he dosen't know it)

Anyone with some old carsound mags might be able to find some real world test results that showed how 1F cans stabilize voltage... FTMFW...

 

[helotaxi 5,000+ posts]

Your conclusion is wrong.. by definition a 1F can will hold 1 coulomb of charge @ 1 volt of potential... at 2 volts it will hold 4... and at 12v it will hold 144...

Last I checked a Farad was 1 coulomb per volt (F=Q/V) which is a linear relation rather than the exponential one that you are giving.

For an example, if our charging system fluctuates between 12 and 14 volts a 1F can will hold a difference of 52 Joules... thats 52 amps for one second, 104 amps for half a second, 208 amps for one quarter of a second, and so on..... (what we are interested in is sustaining power for transients that may last for 10~500ms; so you can do the math)

And since the relation between charge and voltage is linear, this is out the window as well. The difference is simply 2A per second disregarding time constant.

 

[Lakota 5,000+ posts]

Add me to the list of EE majors.

 

[Haunz 10+ year member]

Last I checked a Farad was 1 coulomb per volt (F=Q/V) which is a linear relation rather than the exponential one that you are giving.

And since the relation between charge and voltage is linear, this is out the window as well. The difference is simply 2A per second disregarding time constant.

yea, you're right.. guess I didn't remember as much about the topic as I thought I did...

at least I got the time constant right... //content.invisioncic.com/y282845/emoticons/redface.gif.62fdbfe1a101588a808c4cff71bcb942.gif

 

[Jared530 10+ year member]

Why wouldnt you tell us the names of the bad cap companies? So we can buy them and be SOL i guess

 

[lostdaytomorrow 10+ year member]

IM a Dumbass