Coupling caps- can you explain them to me?

[flatfive]

So the higher the value of a coupling cap, more low frequencies get through therefore more bass.

But the higher the value of a bypass cap, fewer low frequencies get through, therefore more treble.

If lowering the value of bypass caps bleeds off more low frequencies to ground, one would assume that a lower value coupling cap would allow more low frequencies through. But this is not the case. Why?

[tubeswell]

But the higher the value of a bypass cap, fewer low frequencies get through, therefore more treble.

No its the other way around - the higher the value of a bypass cap, the more gain there is for more frequencies to get amplified - so a 22uF cap will add overall gain to all the signal (resulting in treble, middle and bass frequencies all getting amplified), whereas a 0.47uF cap will only add gain to the higher frequencies.

[Chuck H]

The reason coupling caps and bypass caps seem to work inversely is that they are not performing in the same circuit. In a plate coupling circuit a cap works as you would expect. Small caps only pass high frequencies and as you increase the cap value you accesss a correspondingly lower frequency. A cathode bypass cap is there to REMOVE AC from the cathode circuit. The presence of AC on the cathode creates local negative feedback. This can improve liearity but at the expense of gain. A bypass cap bleeds AC off of the cathode to ground but leaves the biasing DC on the cathode. So, used in a cathode bypass circuit small values remove high frequencies from the local negative feedback, thus increasing the gain for those frequencies. Increasing bypass cap value produces a correspondingly lower frequency gain increase. The reason the values seem disproportionate (ie: .047uf is a typical coupling cap size but 22uf would be a typical bypass cap size) is because of the two circuits different impedances. The impedance of a circuit is one of the dictating factors for determining a capacitors rolloff frequency.

Chuck

[cbarrow7625]

Is everybody talking with the same terms here? I don't think that "higher" and "lower" values are the best descriptors for caps. "Larger" and "smaller" VALUES for caps is much clearer. Where "larger" = increased capacitance and "smaller" = decreased capacitance.

I have seen a number of times when people say "higher" value, meaning a smaller cap that passes only higher frequencies. High & low are proper for frequencies. To avoid future confusion may I suggest that we standardize on "larger" and "smaller" capacitance values when referring to caps?

[flatfive]

When I say "higher" cap, I mean higher capacitance. But I'm ok with larger and smaller.

[Enzo]

A coupling cap passes the signal along to the next stage. The larger value that cap is, the more lower freqs can pass through it.

In a tube gain stage, the current through the tube causes a voltage drop across the cathode resistor. As the current increases, so does this voltage. That increase in cathode voltage tends to bias the tube to less conductivity. A sort of natural negative feedback. Increasing the currrent through the tube tends to reduce the gain of the tube at the same time.

A bypass cap lets some of the signal bypass that resistor. That means that any signal bypassing it won;t be contributing to that voltage drop, and won't be negating any gain.
So a smaller cap tends to let only highs bypass the resistor, so only highs don;t have their gain negated. The larger the bypass cap, the lower the freq that gets bypassed around that resistor. And thus the lower the freqs that don;t have their gain negated. So the larger the bypass cap, the lower the freqs that get the gain. SO more and more bottom as the value increases.


This may not be a very rigorous description technically, but I hope it makes the concept more understandable.

[EFK]

Enzo thank you for a very clear and simple description!

[flatfive]

Enzo, the only wire that comes out of this cap/resistor goes to ground.

It would seem that any signal (even if it bypassed the resistor) would end up going to ground also. But somehow I don't think this is right.

So this signal that is bypassing the resistor, where is it going?

[Enzo]

Not sure what you mean. The cathode resistor and bypass cap are wired to ground at one end and the cathode of the tube at the other.

The DC tube current flows through the resistor to provide a bias voltage for the tube. The signal would normally do that as well. Adding the caps allows signal of any freq over whatever the cap dictates to be shorted to ground. Any signal thus shorted will not contribute to any voltage forming across the cathode resistor. Only the unbypassed freqs will do that. If they don;t contribute to the cathode voltage, then they don;t participate in gain reduction. SO those freqs are amplified more than the lower freqs which have some of their gain negated by the action of the cathode resistor.

Remember, the tube is not some solid block inside. We can ground the cathode completely with a wire, if we want. The tube will still function. it is not like grounding the cathode - directly or only at certain freqs - is grounding the signal path through the amp. perhaps we should make the understanding that the signal path through the amp is not the same thing as signal passing through a tube. The signal path through a stage is usually in the grid, out the plate, and on to the next stage. But in that process, we put the signal on the grid. The grid is not connected to anything inside the tube, it just exercises its influence over other things to control what comes out of the tube. Just as you can use your fingers to turn the knob on a light dimmer and cause the bulb to change brightness, but your fingers are not connected to the bulb.

The grid affects the voltage at the plate. From that plate we take the varying voltage - which is our ongoing signal - and send it along to the next stage. The signal appears also at the cathode. But it is not conected to either grid or cathode. What we do with the signal that appears at the cathode does however affect what goes on at the plate.

Of course in other circuits like a cathode follower or a phase splitter, we can make use of the signal at the cathode. But in those applications it would generally be counter prodcutive to bypass the cathode then.

[Old Tele man]

...a COUPLING capacitor passes AC-signals through to the next stage.

...a DE-COUPLING or BY-PASS capacitor passes AC-signals through to ground (typically).

[Enzo]

Tele, I could be wrong, but I get the impression he is wrestling with the idea of grounding the AC signal. As if grounding the AC at the cathode would somehow ground off the signal path.

[flatfive]

Thanks for all your help, but I'm just not getting it.

[Enzo]

Well, where are we leaving you in the dust? Are you not getting why it works, or not getting that it works at all? Not getting the distinction that coupling caps (plate to next grid) do something different from bypass caps? (Cathode to ground) Does any part of it make sense yet?

In your post, you used the term "get through" for both situations. Maybe I don;t understand the question.

A coupling cap passes the signal along. Imagine a loudspeaker aiming out the window. We are sending our signal out the window. The larger the speaker the more bottom end we can send out the window. And of course the smaller the speaker the more it sounds like a tweeter.

The bypass cap is not passing the signal along, it is controlling the particular stage response. In post 2 tubeswell mentioned you have the bypass cap thing reversed. The larger the bypass cap value, the more low freq will pass through the stage. More passes through the bypass cap as well, but that doesn;t take away from what passes through the tube.

I tried to explain why a larger cathode cap will make more lows come through the stage than with a smaller cap.

A lower value of bypass cap doesn't bleed off low freqs to ground. Not from your signal path anyway. Without a bypass cap, the cathode resistor all by itself generates some negative feedback. To some extent, the tube fights its own gain. A bypass cap prevents that resistor from developing this gain reducing negative feedback effect, by bypassing the signal influence around the resistor. SO the bypass cap is not bleeding low signal to ground, it is bleeding to ground the cathode resistors ability to fight the tube's gain. A small cap passes only high freqs, so only high freqs get ignored by this natural gain reduction. The larger cap allows lower freqs to also get ignored by the natural gain reduction. The larger the cap, the lower this ignoring goes.

SO you could think of the bypass cap as coupling the signal to the "don;t turn me down" department of the tube. The larger the caps, the lower the freq that won;t get turned down.

[Robert M. Martinelli]

Since My English is rather poor, I'll try to make things clearer with an image...See the ( badly drawn ) attached file.

A capacitor's ability to behave differently at different frequencies is called Capacitive Reactance, indicated on theory books with the term Xc.

You can think a capacitor's Xc as the different resistance the cap opposes to the passage of different frequencies.

The formula to calculate Xc is rather simple : Xc= 1/(2*Pi*F*C) Where Xc is a cap's Capacitive Reactance in Ohms, Pi is 3.1415926..... F is the Frequency in Hertz and C is the cap's capacitance in Farads.

You can clearly see that, being the F factor at the equation divider, the higher the frequency, the lower the capacitor's Xc.

This said, you can have a cap used as an High Pass Filter, HPF ( in series with the signal ) or a Low Pass Filter, LPF ( in parallel with the signal ). In the first case the cap lets all the frequencies above the cutoff frequency ( the frequency at which the signal is attenuated by -3dB, or 0,707 times ) pass through, in the second case the caps shunts to ground all the frequencies above the cutoff frequency ( thus letting the frequencies below the cutoff pass through, hence its name ). With a single cap the HPF/LPF filter slope is -6 dB/Octave ( doubling the frequency the signal's amplitude gets halved ).

Hope this helps

Best regards

Bob

[tubeswell]

WTF - I am equally comfortable with the terms bigger and smaller, and higher and lower, recognising that some people refer to use one or the other - but I get what the OP meant (which is what counts).

Anyhow here is a handy page of a triode gain stage explaining the various bits

Tube Terminology for Dummies