How Is a Second Order Chebyshev Passband Filter Designed?

  • Thread starter Thread starter cathode-ray
  • Start date Start date
  • Tags Tags
    Filter
Join the discussion
Ask a follow-up here, or get your own question answered by working scientists, mathematicians and engineers — people, not an autocomplete.
Real named experts · corrections over time · the nuance an AI answer skips
7 replies · 5K views
cathode-ray
Messages
50
Reaction score
0

Homework Statement


Get the transfer function of a second order Chebyshev passband filter, with central frequency f0 = 1 [kHz], lower cutoff frequency fc=670 [Hz], 3dB ripple in the pass-band and 30dB of gain in the central frequency.

Homework Equations



Maximum allowed variation in passband transmission [itex]A_{max}=10log(1+ε^2)[/itex]
Transfer function of Chebyshev filter [itex]T(s)=\frac{Kw_{p}^N}{ε2^{N-1}(s-p_{1})...(s-p_{N})}[/itex]
Chebyshev filter poles [itex]p_{k}=-w_{p}sin(\frac{(2k-1)\pi}{2N})sinh(\frac{1}{N}sinh^{-1}(\frac{1}{ε}))+jw_{p}cos(\frac{(2k-1)\pi}{2N})cosh(\frac{1}{N}sinh^{-1}(\frac{1}{ε})), k=1,2,...,N[/itex]
N is the order of the filter
K is the gain
(expressions taken from "Microelectronic Circuits", Sedra, 5th edition)

The Attempt at a Solution



Hi,

My first attempt, to this problem was to calculate ε through [itex]A_{max}[/itex] and get the two poles(which are conjugated) through the expression given, considering [itex]w_{p}[/itex] as [itex]670\times2\pi[/itex]. I then had all that was needed to build the transfer function. The problem is that this is the transfer function for a low pass filter. I have no idea how to get the pass band filter at this point, and I am also not sure if what i did is correct.
 
Physics news on Phys.org
I made a search in the web before coming here and also found that document, but it didnt helped :S .
 
cathode-ray said:
The problem is that this is the transfer function for a low pass filter. I have no idea how to get the pass band filter at this point, and I am also not sure if what i did is correct.
I recall from my uni days that there are transforms to convert the LP to a HP, and also to a BP. That's why design equations deal only with the LP. Searching, I found mention on wikipedia. http://en.wikipedia.org/wiki/Frequency_transformations#Bandform_transformation

I think you are just looking for a second order system, of the form A.s / (s² + bs +c)
so it won't have ripple as such, there's only the one peak in the response, it's bandwidth being measured between the pair of -3dB points.

Beyond this, I'm of no help here, sorry.
 
rude man said:
I recommend http://www.analog.com/library/analogDialogue/archives/43-09/EDCh 8 filter.pdf

pp. 8.21 ff.

If you want the correct answer without the details, go to www.filterfree.com. Stipulate chebyshev bandpass type 1, 2nd order, input your upper & lower cutoff frequencies and midband gain, 3dB passband ripple, and get the answer.

SORRY, N/O. I'll try to figure out how I got their site. definitely did, only yesterday. Stay tuned, it's worth it.

OK, found the problem - the site is www.nuhertz.com. Scroll down to the "Filter Free" downlink.
 
I finally found the answer and the mistake i was making :D. To design the filter i was considering initially a second order low pass Chebyshev filter, with the specified characteristics, and then i applied the frequency transformation to make it a band pass. However to make the transformation what i needed was a first order low pass Chebyshev filter that as a transfer function of the form:

[itex]T(S)=\frac{1}{S-p_{1}}[/itex]​

where S is the normalized frequency with respect to the central frequency:

[itex]S=\frac{s}{w_{p}}[/itex]​

I then calculated the pole [itex]p_{1}[/itex] of the filter, using the expression i took from the "Microelectronic Circuits", which gave me [itex]p_{1}=-1[/itex]. Thus the transfer function for the low pass filter is:

[itex]T(S)=\frac{1}{S+1}[/itex]​

To make it a band pass filter i used a frequency transformation:

[itex]S\rightarrow\frac{S^{2}+1}{2\xi S}[/itex]​

After making the transformation i ended up with a band pass second order Chebyshev filter , with the desired characteristics, of the form:

[itex]T(S)=\frac{s}{\frac{s^{2}}{B}+s+\frac{w_{p}^{2}}{B}}[/itex]​
B is the bandwidth

In order to have the gain of 30dB at the central frequency i multiplied the transfer function for a constant k and calculated what value it had to have at [itex]s=jw_{p}[/itex].Then i finally got the desired result:

[itex]T(S)=\frac{s}{\frac{s^{2}}{Bk}+\frac{s}{k}+\frac{w_{p}^{2}}{Bk}}[/itex]​

Thanks for all the replies!

Note: Thanks for the link for that software rude man :) it will be useful