8
\$\begingroup\$

I have a low pass filter like this:

schematic

simulate this circuit – Schematic created using CircuitLab

\$V_{\text{out}}\$ is measured right after \$R_1\$, which I suppose means that it is measured over the parallel part.

\$R_2\$ is the load of the filter. When this circuit is measured with an oscilloscope it seems like it is not dependent on the frequency at all. I would like to investigate why.

I tried to calculate the transfer function for the filter, but I am not sure that it is right.

$$ H(j\omega) = \frac{1}{R_1\left(j\omega C+\frac{1}{R_2}\right)+1}$$

I'm using \$R_1 = 33\text{k}\Omega\$, \$R_2 = 1\text{k}\Omega\$, and \$C = 220\text{pF}\$.

If I plot the frequency response in Matlab with this, I just get a straight line going from the origin through (1,.5), (2,1) where (Hz, H(w)) and so on.

Is this correct?

\$\endgroup\$
5
  • \$\begingroup\$ Yes, a schematic would be helpful. Also, you said you measured this circuit. What values of R1, R2, and C1 did you use? And what frequencies did you use for Vin? \$\endgroup\$
    – Eric
    Nov 25, 2014 at 17:01
  • \$\begingroup\$ Welcome to EE community! I allowed myself to edit your question and draw a schematic for you. Please check if this is what you mean. \$\endgroup\$
    – Kamil
    Nov 25, 2014 at 17:09
  • \$\begingroup\$ Thanks! The circuit looks right! I used a sinus frequency of between 1kHz up to 100kHz as input. My components has the values R1=33kOhm, R2=1kOhm, C1=220pF \$\endgroup\$
    – rablentain
    Nov 25, 2014 at 17:11
  • \$\begingroup\$ I added "box" after I read what is filter and what is load. You can edit it and enter component values if you want. \$\endgroup\$
    – Kamil
    Nov 25, 2014 at 17:16
  • \$\begingroup\$ Try changing your load to 1 Meg ohm. (The 1 k is loading the entire circuit down. Without the capacitor R1 and R2 make a voltage divider that is cutting your input down by a factor of 33!) \$\endgroup\$ Nov 25, 2014 at 17:36

3 Answers 3

4
\$\begingroup\$

You can interpret this circuit as a voltage divider using $$R_2 \parallel \frac{1}{j\omega C} = \frac{R_2}{j\omega R_2 C + 1} $$ and \$R_1\$. The transfer function is therefore

$$H(j\omega) = \frac{R_2 \parallel \frac{1}{j\omega C}}{R_2 \parallel \frac{1}{j\omega C} + R_1} = \frac{\frac{R_2}{j\omega R_2 C + 1}}{\frac{R_2}{j\omega R_2 C + 1} + R_1} = \frac{R_2}{R_2 + R_1(j\omega R_2 C + 1)}$$

If you divide numerator and denominator by \$R_2\$ this is the same expression you calculated, but I think it's easier to understand the filter using my result. As \$\omega \to 0\$ $$H(j\omega) = H(0) = \frac{R_2}{R_2 + R_1}$$ which is what you would expect for a simple voltage divider using \$R_1\$ and \$R_2\$. As \$\omega \to \infty\$ the denominator dominates and \$|H(j\omega)| \to 0\$. This is a low pass filter so the output should depend on the frequency (provided you sweep to a high enough frequency).

Here is your circuit in CircuitLab setup so that you can simulate it within CircuitLab:

schematic

simulate this circuit – Schematic created using CircuitLab

And here is the frequency sweep on the circuit as reported by CircuitLab (click to make it larger):

enter image description here

You can use this to verify your Matlab code. If you post your Matlab code we might also be able to help you find a problem with it.

\$\endgroup\$
2
  • \$\begingroup\$ I did actually tried to run my code again, with a higher frequency span this time. It now looks like the graphs you posted so I think its correct! :) What does the || notation mean in your calculations? \$\endgroup\$
    – rablentain
    Nov 26, 2014 at 7:43
  • \$\begingroup\$ Great! The || symbol denotes the parallel combination of two impedances. \$\endgroup\$
    – Null
    Nov 26, 2014 at 13:35
6
\$\begingroup\$

You can rearrange the circuit:

schematic

simulate this circuit – Schematic created using CircuitLab

But using Thévenin equivalence, you have this:

schematic

simulate this circuit

Which is a standard RC circuit with $$ V'_{in} = {R_2\over{R_1+R_2}}V_{in} $$ and $$ R_p = R_1||R_2 = {R_1R_2\over{R_1+R_2}}$$

\$\endgroup\$
5
\$\begingroup\$

Your equation is correct but written a bit unusually.

You can re-arrange to get:

$$H(\omega) = \frac{1}{C_1 R_1} \times \frac{1}{j\omega+\frac{R_1+R_2}{C_1R_1 R_2}}$$

so your pole frequency (or cutoff of your lowpass filter) is $$f=\frac{\frac{R_1+R_2}{C_1 R_1 R_2}}{2\pi}$$

For your values that works out to be \$745\$kHz which is well above your test signal's highest frequency (\$100\$kHz). So you won't see any rolloff.

Here's what the mag/phase versus frequency looks like:

enter image description here

\$\endgroup\$
3
  • \$\begingroup\$ akellyirl - please recalculate the given cut-off frequency. It should simply be wc=1/RpC (Rp=R1||R2) and fc=wc/2Pi \$\endgroup\$
    – LvW
    Nov 25, 2014 at 18:23
  • \$\begingroup\$ That's the same thing..... \$\endgroup\$
    – akellyirl
    Nov 26, 2014 at 9:26
  • \$\begingroup\$ I don`t think so: f=2*pi....? \$\endgroup\$
    – LvW
    Nov 26, 2014 at 13:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.