Formula to calculate capacitor value?

[klitgt]

If you want to calculate the value of a coupling cap (value = C) from a preamp to a poweramp with the input impedance of R Ohms we agree to use the formula Fc=1/(2 * 3,14 * C * R) to calculate the C value if the low frequency is to be Fc Hz.
In case R=10K and we want a flat reponse down to 16 Hz C is going to be at least 1 uF, see example A in attachment.

If you want a mild cut of the high frequencies from e.g. 8000 Hz and up you may do as shown in example B?

How do you calculate the value of C2, assuming that C1 is 1uF and R is 10K??

[sbrads]

It won't work. Results will depend on the source impedance, if it was zero there would be no HF cut to speak of anyway. You need a series resistor at least. It's best to use a network synthesis program anyway, as there's always some interaction between the 2 poles (LF & HF). I use an ancient DOS program called ECAP for simple RLC analysis, I find it easy & quick, and has a frequency response plot, but I'm sure there are Windows equivalents.

[I_Forgot]

I recommend that everyone who builds anything electronic learn how to run a spice simulator like SwitcherCAD. It is free. There's just no reason not to learn it!

I ran your circuit and here is the result... I included some series resistance and parallel capacitance in the source.

I_F

[I_Forgot]

Here is a more detailed rerun, this time stepping the capacitor value. I've also added some capacitance at the load for completeness.

First the circuit:

[I_Forgot]

And here is the result. I can't control the colors, so some of the traces are hard to see because I had to reduce the image size to <1000 pixels wide, but you should be able to pick out a value to use. The capacitance at the source and load don't have much effect on the overall response.

You can easily set up this circuit yourself and change component values. If you want frequency response, run an ac analysis as I did here. If you want time domain response, run a transient analysis.

You can DL SwitcherCAD for free from Linear Technology's web site. There are a bunch of tutorial circuits included so you can figure out how to do things like parameter sweeps.

I_F

[poobah]

If you want flat response down to 16 Hz... you should calculate C1 using 1.6 Hz (10 times less) as your frequency. The standard formula finds the point where reactance from the cap equals resistance... this is the "corner frequency" or "-3 dB" point Most coupling caps are designed using 2-3 Hz as the "corner frequency.

For C2, do the same thing; in the opposite way. Calculate C2, with the same formula using 160 kHz (ten times more) as your corner frequency.

Keep in mind, in both cases, R is not just R, there is some source resistance adding to the input of this thing, and there is some load resistance the output of this thing. So for C1, you should assume that R is larger because the source resistance ADDS to it. For C2, assume R is smaller because the load resistance appears in PARALLEL with R.

[klitgt]

Originally posted by I_Forgot
I recommend that everyone who builds anything electronic learn how to run a spice simulator like SwitcherCAD. It is free. There's just no reason not to learn it!

I ran your circuit and here is the result... I included some series resistance and parallel capacitance in the source.

I_F

Thank you I_F for inspiring lessons! I will try to learn to run the SwitcherCAD.

[AndrewT]

Hi,
remember to add the value of the DC blocker in the source component to the DC blocker in the amplifier. Could be anywhere in the range 680nF to 10uF.
Then calculate for an RC time constant of about80mS to 100mS.

[Christer]

Originally posted by I_Forgot
I can't control the colors, so some of the traces are hard to see ...

Yes you can. Choose "Color Preferences" in the Tools menu. You can change colour of traces, background and even more stuff.

[I_Forgot]

Oh good! I had not run into that before. Thanks!

I_F