Quick Question: Function of a Capacitor

[ryanscissorhands]

I think I just came to a realization. . . or I'm totally wrong. Let's see if I have this right: when a capacitor is placed int the signal path, it "absorbs" or rolls off different frequencies, depending on the value of the capacitor. When is is placed from signal to ground, however, it has the opposite effect, because the frequencies that it would normally block in the signal path, it prevents from bleeding to ground. . .right?

Or am I totally off?

[aron]

Check out these cool links:

http://www.ethanwiner.com/filters.html
http://www.physicsenterprises.andrews.edu/diy_archive/references/capacitors.html
http://www.capacitors.com/picking_capacitors/pickcap.htm

[gaussmarkov]

I think I just came to a realization. . . or I'm totally wrong. Let's see if I have this right: when a capacitor is placed int the signal path, it "absorbs" or rolls off different frequencies, depending on the value of the capacitor. When is is placed from signal to ground, however, it has the opposite effect, because the frequencies that it would normally block in the signal path, it prevents from bleeding to ground. . .right?

hi ryan!

as i understand it, when a capacitor is in the signal path it blocks dc but lets through the (ac) signal.  in this configuration, it is used to isolate different stages of a circuit in terms of the dc power supply.

capacitors also have a lower reactance ("resistance") for high frequencies than low frequencies.  (dc is zero frequency and that's why it is completely blocked above.)  so when a capacitor goes from the signal path to ground it bleeds off to ground more high frequency signal than low frequency signal.  it creates a low pass filter. 

if you place a resistor to ground after a capacitor in the signal path then you create a high pass filter. 

so there are 3 cases to consider:  a capacitor in the signal path, a capacitor from signal path to ground, and a capacitor in the signal path followed by a resistor from signal path to ground.  the first is a simple coupling capacitor blocking dc.  the second is a low pass filter. and the third is a high pass filter.

you can change the shape of the low pass filter by putting a resistor in front of the capacitor going to ground.  that makes the two filters look symmetric.  they are both called RC filters, too.  they are basically more complicated versions of the voltage dividers you see in our circuits to create 4.5V (so-called reference voltage) along with the 9V supply.  it's this voltage divider action that makes the high pass filter thing work opposite to the low pass filter configuration for ac/signal.

hope this helps!

p.s. i will make edits in green.

[niftydog]

the impedance of a capacitor is inversely proportional to the frequency of the signal.

That is to say that at very low frequencys a capacitor has a very high impedance (approaching an open circuit) and at higher frequencys the impedance drops. [I learnt this by keeping in mind the basic concept that a cap is an AC short circuit and a DC open circuit!]

So, placed in series it will block low frequencys (and DC) while passing higher frequencys.

Placed in parallel (ie; to ground) it still does the same thing but the effect is opposite. It shunts the higher frequencys to ground whilst leaving the lower frequencys (and DC) less affected.

[cab42]

This is a great thread!

Thanks to Ryan for asking a question that I have thought of, but not been able to formulate and thanks for the replies. I feel very enlightened 

Regards

Carsten

[brett]

Hi.
Understanding interactions between caps and resistors is fundamental to "seeing" what's going on in schematics.
Stay with me for a second while I do some ugly maths.....The impedance (Z) of a capacitor (in ohms) = 1/(2.pi.f.C)  where dot (.) means multiply, pi is greek for 3 and a bit, f is the frequency and C is the capacitance (in Farads). 
So at 1 kHz (right where you solo a lot - A above middle A), 22uF has about 7 ohms of impedance.  Not much at all.  Which is why you see big caps: 10, 22, 47uF used to "bypass" larger resistors.  A classic example is the fuzz control on a fuzzface, where a 22 uF cap is used to bypass (to ground) a variable amount of the resistance in the 1k potentiometer.
cheers

[KORGULL]

*edit* -Had question - found answer in a book- never mind.

[ryanscissorhands]

Thanks for the responses! So I almost had it right. . . except totally backwards. I thought that caps passed low frequencies better than highs. That's good to know, for when I decide to mod effects by swapping caps. Saves me a lot of frustration due to a bad mentalization.

So had I understood correctly the impedance vs. frequency, my epiphany would have been correct. I was staring at a schem--someone's new fuzz based on the fetzer w/diode clipping--i was staring at the cap to ground, and it hit me.

So it's good to know that I realized something, and made a big leap in my surface-level knowledge of this stuff. And it's good to be corrected. I feel that I just made a big step.

[ryanscissorhands]

Okay, just a little thinking outloud (err, on my keyboard. . .).

The impedance (Z) of a capacitor (in ohms) = 1/(2.pi.f.C)  where dot (.) means multiply, pi is greek for 3 and a bit, f is the frequency and C is the capacitance (in Farads). 

Z = 1/(2pi.f.c)
z.f.c = a constant, namely 1/2pi, right?
Simplified, z.f.c=K

So, let me see if this is right. For a given frequency, as capacitance goes up, impedance goes down. A larger cap will be less "resistive" of that frequency. Is this true? If so, again, I was backwards in my mental scheme of how a capacitor works, which is good to be set straight on.

Thanks for your help.

[niftydog]

So, let me see if this is right. For a given frequency, as capacitance goes up, impedance goes down.

yes, that too.

So, inversely proportional to both the frequency and the capacitance value.

[gaussmarkov]

this thread seems like a good place to ask this follow-up:  what about current?  in a voltage divider, or an RC (resistor-capacitor) filter, the division of voltage does not depend on the level of the current.  but suppose we are using a capacitor all by itself (if that makes any sense--maybe it doesn't?).  then the definition Z=V/I  where I is current and V is voltage only governs the ratio of voltage to current, not current itself.  so doesn't the actual filtering behavior of such a capacitor depend on the properties of the circuit?

here's an example of my thinking (just resistors).  if i am going to stick a resistor rated R ohms into the middle of a series of components, then i know that R=V/I ... but the voltage across the resistor isn't predictable from this information alone.  i see that the same thing is true for a capacitor, except voltage isn't the only thing affected:  as i understand it, the scale of the AC filter is affected.  i guess the "shape" of the filter is unchanged:  1/(2 pi f c) as ryan notes above.  so then relative frequency effects are constant but amplitude (volume) is variable?  maybe i have answered my own question. 

maybe you can see why i thought there were three cases, not just two as nifty explains. 

[niftydog]

hmm, I know you're just dumping your thoughts onto the screen, but I'm not sure I follow you. Are you still struggling with this or have you found your answer? Maybe you can rephrase? I'll have a hack, I might be on the right track...

so doesn't the actual filtering behavior of such a capacitor depend on the properties of the circuit?

Yes, because of the circuits affect on the impedance loading on the capacitor itself. Keep in mind that a capacitor has some resistance, and any other resistance in parallel with it will affect it's filtering characteristics.

When you talk of a DC blocking capacitor, it's really just an extreme high pass filter - so the three cases you talk of are really just two, but the "third" one is a special case where the impedance of the circuit acts as the resistor to ground following the capacitor.

[gaussmarkov]

hmm, I know you're just dumping your thoughts onto the screen, but I'm not sure I follow you. Are you still struggling with this or have you found your answer? Maybe you can rephrase? I'll have a hack, I might be on the right track...

so doesn't the actual filtering behavior of such a capacitor depend on the properties of the circuit?

Yes, because of the circuits affect on the impedance loading on the capacitor itself. Keep in mind that a capacitor has some resistance, and any other resistance in parallel with it will affect it's filtering characteristics.

When you talk of a DC blocking capacitor, it's really just an extreme high pass filter - so the three cases you talk of are really just two, but the "third" one is a special case where the impedance of the circuit acts as the resistor to ground following the capacitor.

niftydog, i really appreciate you taking a crack at my question.  this impedance stuff is tricky!!! 

everything that you say squares with my current understanding (i think), so i feel reassured.  except for the fact that i sound so confusing.     i had the effect of coupling capacitors inside of circuits over simplified, thinking they only blocked dc and not recognizing that they also attenuate lower frequencies.  those caps are almost always followed by resistors to ground, so i guess they are generally configured into RC filters.  i was imagining that there might be some cases where a cap would go straight (in series) into the base of a transistor (say).  now i see that even without the resistor to ground, the cap still attenuates lower frequencies with complete blocking at zero frequency (equals dc).

i think the reason i sound so confused is what's been tripping me up lately: how placing some new component into a circuit or changing the value of an existing component has consequences throughout the circuit (forwards and backwards so-to-speak).  it's like i wanted the coupling capacitor to just block dc at its location and have no other consequences.  but like you said, it has some resistance (different amounts at different frequencies even) and so (for example) it affects the current that runs through the whole dang circuit.

whatever my confusion is, thanks again.  i'm learning a lot.