Deciding cut off frequency for low/high pass filter

Go To Last Post
15 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

1. I have a application in which need to get 10Hz-240Hz signal. So first made a 240hz LFP & then connected 10Hz HPF after that. Ckt is attached. Res/cap values are taken from TI workbench, nearest available will add later.
2. I was reading that when designing a filter, at fc gain is at 3db lower. 
3. I need to measure 10Hz to 240Hz, so at my 10Hz signal with 3db lower & similarly 240Hz signal will be 3db lower.
4. So what is I want no gain loss at 10Hz or 240Hz.  How much lower frequency should I select than 10Hz, so that at 10hz there is no loss.
Similarly for 240Hz, how much higher frequency should i select so that there is no signal loss at 240Hz. 

 

5. is there any general rull of thumb for this?
6. Similarly for sample signal at 240Hz with MCU adc, what is frequency for adc sampling to be selected. Generally I have select 10 times atleast. But is there any general rule of thumb for this also. 

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1


Vindhyachal Takniki wrote:
Ckt is attached.
Life is a lot simpler if you simply do this...

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1


Opamps symbols look like this:

Please use standard schematic symbols!!!

 

Now, can you tell us more about the signal you want to filter,  why does it need filtering?   What is the source of the noise?

Can you add a gain stage (3db) after the filtering to boost the signal back up?

 

Jim

 

Click Link: Get Free Stock: Retire early! PM for strategy

share.robinhood.com/jamesc3274
get $5 free gold/silver https://www.onegold.com/join/713...

 

 

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 2

Did you hear about Nyquist theorem ? Aliasing...noise...sampling..etc ?

 

if you havent, please study these things well.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

 So what is I want no gain loss at 10Hz or 240Hz.  How much lower frequency should I select than 10Hz, so that at 10hz there is no loss

Well, that's not really the question...you can always increase the gains to give unity or more at any frequency.  The real question is how steep of a cutoff (or pass) do yoyu want, the flatness in the passband...essentially the boundary limits on the shape you desire.  There is also phase/group delay & distortion, if that is important to your needs.  Better take a look at some of the references for active filter types.   

 

Get a copy of LTspice & simulate your active network...you can plot the gain & phase vs frequency responses & tweak your circuit arrangement & values.

https://www.seas.upenn.edu/~jan/spice/LTSpiceQuickGuide.pdf

 

http://www.simonbramble.co.uk/techarticles/active_filters/active_filter_design.htm

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

These look like Salen-Key 6db/octave filters that probably have a flat response at the cut-off frequency.    They won't be cutting off the frequency pass band very sharply.  Plus the components (especially the caps ) will have about +/-10% range, even if you could get a 1.23456789 uF cap.

 

For sharp filters you could look into switched capacitor filter ICs.  You can get 12db+ cutoff at a multiple of the square wave frequency that you input into the the switch-cap IC.

Last Edited: Thu. Oct 24, 2019 - 08:21 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

In your case, for an acceptable low attenuation (surely not 0 dB) when using second order filters, I would use (if I remember well):

fcut_lo = f_lo/3 = 10/3 Hz

fcut_hi = f_hi*3 = 240*3 Hz

The suggested parameter '3' decreases if higher order filters are used.

 

Anyway, I usually simulate my filters, using 'real' available values of R and C, so that I can adjust these values to get the 'closest' response that the application requires.

 

Kerim

 

 

 

Last Edited: Fri. Oct 25, 2019 - 12:01 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

thanks for inout 5-400hz gives desired results in my case

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

clawson wrote:
Life is a lot simpler if you simply do this

 

For instructions on how to do that, see Tip #1.

 

Also see Tip #5.

Top Tips:

  1. How to properly post source code - see: https://www.avrfreaks.net/comment... - also how to properly include images/pictures
  2. "Garbage" characters on a serial terminal are (almost?) invariably due to wrong baud rate - see: https://learn.sparkfun.com/tutorials/serial-communication
  3. Wrong baud rate is usually due to not running at the speed you thought; check by blinking a LED to see if you get the speed you expected
  4. Difference between a crystal, and a crystal oscillatorhttps://www.avrfreaks.net/comment...
  5. When your question is resolved, mark the solution: https://www.avrfreaks.net/comment...
  6. Beginner's "Getting Started" tips: https://www.avrfreaks.net/comment...
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Good evening, some thoughts/guidelines regarding filters, but you will have to search/study a little

 

1. You are design a band-pass filter. If you construct the ''real'' filter, place High pass first since will have better adaptation

2. For Sallen-Key filters (like the ones in the design), the fc is your cut-off frequency (-3db frequency - here 10Hz and 240Hz respectively)

3. Since you have connected the Op-Amps as voltage followers, total gain of the filer will be K=1 (1*1 for the two stages)

    If you want any gain in your band-pass you must re-arrange the Op-Amp connections for gain (2 resistors in (-), you must already know)

    Then total gain will be the product of the two individual Ktotal=K1*K2 and better choose same gain for better adaptation again...

4. The response will be as expexted (what you write in 3) if component values are those found in simulation (or near that)
5. For no gain loss at cut-off frequencies, you must select lower cut-off for HPF and higher for LPF. The *3 mentioned from KerimF is such a choice
    but you can find the ''correct'' for your filter if you take some measurements with the various cut-off frequencies you might select.

6. For ''sample signal at 240Hz with MCU adc' i suppose you mean the sampling rate of the adc of MCU? Usually all Microchip MCUs mentioned here have at least 62500Hz and above clock for adc, so they will do the job. Nyquist (and many others) theorem instructs a sampling rate at least 2 times the highest frequency  found in the signal per second in order to have ''perfect reconstruction''

7. You can also use MF (multiple feedback) active filters or switched capacitors filters for higher attenuation of frequencies outside your band-pass.

    Alternatively you can increase the order of your existing filter by placing two of each (HPF-HFP , LPF-LPF) for 40db/decade or 12db/octave attenuation outside the band.

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi Ntano

Whats the techincal reason for this? " 1. You are design a band-pass filter. If you construct the ''real'' filter, place High pass first since will have better adaptation "

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hello Takniki,

You place the High Pass First in order to "pass" all lower frequencies without attenuation and then

the low pass will cut them in your high cut-off frequency (the 2 cut-off frequencies in the band pass filter

are: the higher for the low pass, the lower for the high pass). 
It is (or at least it was in "my times" a common practice in order to have better response because all lower

frequenices pass from the high pass without any attenuation (if you place them the other way around then you will notice
a slight attenuation in the area it shouldn't be (the "flat" region) and because Op-amps used are not ideal the

practical response will be a little different than the theoretical.

But as always, the best is to try it and see if this makes sense...
A signal generator and an oscilloscope will show you if there are any differences between the two connections.

As long as i recall this was the general guideline i was tought and using when i was an E.E. student (a long time ago...).
Hope this helps.

 

EDIT: "adaptation" was my quick word (as english is not my native language),

"matching" (or something like that in english) was the reason taught back then

Last Edited: Wed. Oct 30, 2019 - 08:37 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Ntano wrote:
place the High Pass First in order to "pass" all lower (sic) frequencies without attenuation

You mean pass all higher frequencies without attenuation?

Top Tips:

  1. How to properly post source code - see: https://www.avrfreaks.net/comment... - also how to properly include images/pictures
  2. "Garbage" characters on a serial terminal are (almost?) invariably due to wrong baud rate - see: https://learn.sparkfun.com/tutorials/serial-communication
  3. Wrong baud rate is usually due to not running at the speed you thought; check by blinking a LED to see if you get the speed you expected
  4. Difference between a crystal, and a crystal oscillatorhttps://www.avrfreaks.net/comment...
  5. When your question is resolved, mark the solution: https://www.avrfreaks.net/comment...
  6. Beginner's "Getting Started" tips: https://www.avrfreaks.net/comment...
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

yes obviously...

(i was intending to write ..... from your lower frequency cut-off but messed in editing...)

thank you for correcting.

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

For instance, also in general (in practice), the positions of filter sections (mainly of a high order) are chosen so that none of the opamps reaches saturation; by starting from the section of highest damping factor ξ and ending at the one of the lowest ξ.

 

Kerim

 

Last Edited: Thu. Oct 31, 2019 - 02:52 PM