About low pass filter and high pass filter

[winonet]

hi everybody, I am new in electronic and I want to know, What advantages/disadvantages of using a First order filter vs a third or higher order filter? and the last question: what is the highest order or analog filter that can be constructed?

Thank you for your help
have nice day

 

[flatulent]

Fewer poles cost less. More poles have higher attenuation out of band. As far as the maximum number of poles, there is a practical limit when the coupling through the air between the input and output is higher that the signal passed through the filter proper in the attenuation band. Usually, 60 dB is the practical limit of out of band attenuation.

 

[LvW]

I am not sure if the above reply answers the question completely. Therefore, some addtional remarks:
1.) Each lowpass filter has a pass region, a transition region and an attenuation region. The transition region of a third order filter is smaller than for a first order filter because the slope of the filter transfer function is steeper. That means, an attenuation region which is rather close to the end of the pass band requires a higher order transfer function.
2.) It is not possible to fix an upper limit for the order of a filter. There are many, many aspects which influence the decision regarding filter order (topology of the filter, cost, complexity, delay aspects, time response, exactness, power consumption...)

 

[winonet]

Thank you , you help me so much. I have a question : does the cut-off frequency change as we change the filter order? why?

 

[LvW]

This question cannot be answered because changing the filter order does require another filter structure with new dimensioning.
That means: YOU are responsible for the design. Perhaps you mean the following: Can two filters of different orders have the same cut-off frequency? The answer is: YES.

 

[ic98]

you can keep these in mind:
- the higher the filter order the greater the filter power consumption and silicon area (for integrated), at a given bandwidth
- high order filters cannot be used within feedback structures
however, achieving specifications of an ideal filter is desired, why?
coz you want to suppress higher frequencies greater than a specific, so another system requires to you deliver a signal with specific SNR..
in brief, the freq. specs of the systems you use filter between them results in the requirement for the filter design

 

[LvW]

...........
- high order filters cannot be used within feedback structures

IC98, please can you verify/explain this statement? I have some doubts about it.

 

[ic98]

LvW, it's simple; but i don't know what amount of description is needed.
if you have read some fundamental text from books on electronic circuit design by Gray & Meyer, Sedra & Smith, Razavi, and others, you might have found some material on feedback stability analysis of high-gain amplifiers (operational amplifiers). if so, generally a filter (with a given gain, lower and upper cut-off frequencies) is can be assumed as a general amplifier (say operational amplifier), and vice versa! and all your studies can be used to analyze a filter in feedback structure. OK?

 

[LvW]

Hi IC98,

perhaps, now I understand what you mean. Of course, there will be stability problems when a frequency dependent block of the order of 2 (or higher) is part of an overall feedback loop. But this problem - for my opinion - has nothing to do with the original question which more or less was a basic one regarding filter theory. Therefore the misunderstanding. (My first impression was that you don't allow higher filter orders realized by feedback topologies.)

 

[FvM]

- high order filters cannot be used within feedback structures

Of course, you can design a filter of order >= 2 that can be used in a feedback loop, it simply needs a dominant pole. In so far the statement is incorrect, strictly spoken. But mainly it's answering a question, that hasn't been asked. And in my view, it won't address a meaningful category in the classification of filters.

 

[ic98]

LvW, FvM
I think I answered the writer's question in my first reply to this post.
Of course that, "of course, you can design a fiter of order greater than whatever you desire in a feedback loop", but your loop is unstable. the problem is that, as far i know, after you model a system(circuit) with its dominant poles (say at most two), if there are more nondominant poles (say n poles) , thus you say our system has n+2(dominant) order..
Maybe it's enough for the author!
good luck