Bandpass Filter

[neo]

I have been on this for a week and i am at the point i feel like i'm going . What i need is a bandpass filter (or 40) so that i can build a spectrum analyzer, i don't need precision as its just eye candy out an old vfd i know how to do everything except the bandpass filter. The closest i have come is the equation 1/(2*Pi*Freq*R*10)-12 which matched up to one of their results when given the same info but it repeats and is based on this equation 1/(2*Pi*Freq*R). What this all boils down to is i'm either missing something or trying to make it way too complex.

[trophosphere]

Where are you getting your information from? Where did you get "their results"?

[neo]

http://www.electronics-tutorials.ws/filter/filter_4.html
http://www.electronics-tutorials.ws/filter/filter_3.html
In the comments of the second one
"Question, for some outdoor hifi speakers (which do not have a built in high pass filter, so I get a lot of distortion from low frequencies) I’m planning om creating passive high pass filters do to the low costs.
Just to be sure, will this work:
Capacitor: 3900 pF
Resistor: 1M?
This should yield a high pass filter at 40,8 Hz right?

Posted on April 03rd 2016 | 10:47 am"
and i was able to reproduce the result

[trophosphere]

The capacitor and resistor values of 3900pF and 1M will indeed yield a -3dB cut-off of about 40.8Hz for a 1st order high-pass filter assuming a low impedance source and a high impedance load. The equation used to calculate the -3dB cut-off for both first degree high and low pass filters (using resistors and capacitors) is f=1/(2*pi*R*C).

As shown in the tutorials, you can certainly cascade them (with buffering in between each stage) to create band-pass filters. Their upper and lower -3dB cut-off frequencies will be your calculated values for each stage and the difference between the frequencies themselves will be your bandwidth. Note your pass-band won't be exactly flat as the attenuation starts quite a bit before the -3dB calculated cut-off particularly at the lower frequencies.

[neo]

But as useful as that is that is besides my point the math works for that but when i try to use it to get 40 different values it repeats.
1    1.6nf           10K  10hz
2   3.18nf       1k    50hz
3   1.6nf        1k     100hz
4      3.18nf       1k    500hz
those are my first four high pass calculations, so unless im misunderstanding a 50 hz is the same as a 500 hz high pass filter which is an invalid result

[LvW]

neo - you have mentioned your plan to use the bandpass for the purpose of spectrum analyzing, right?
Do you know that the passive filter as shown in the link is the worst bandpass available? It has a Q value below 0.5.
I think, you will need instead a high-Q bandpass filter which - in addition - is frequency-tunable. 

[danadak]

Google this, several solutions -

"lm3915 spectrum"


Regards, Dana.

[trophosphere]

But as useful as that is that is besides my point the math works for that but when i try to use it to get 40 different values it repeats.
1    1.6nf           10K  10hz
2   3.18nf       1k    50hz
3   1.6nf        1k     100hz
4      3.18nf       1k    500hz
those are my first four high pass calculations, so unless im misunderstanding a 50 hz is the same as a 500 hz high pass filter which is an invalid result

I think the problem is the decimal point in the math. The values of 3.18nF and 1K Ohm results in a cut-off frequency of 50KHz. For a 50Hz cut-off, you will need to either increase the value or your resistor or capacitor.

Edit: I found a website which can provide a bit more help when doing the calculations:
1st Order Low Pass Filter Calculator
1st Order High Pass Filter Calculator

[suicidaleggroll]

But as useful as that is that is besides my point the math works for that but when i try to use it to get 40 different values it repeats.
1    1.6nf           10K  10hz
2   3.18nf       1k    50hz
3   1.6nf        1k     100hz
4      3.18nf       1k    500hz
those are my first four high pass calculations, so unless im misunderstanding a 50 hz is the same as a 500 hz high pass filter which is an invalid result

You're messing up the math

1)    1/(2 * pi * 0.0000000016 * 10000) = 9947 Hz
2)    1/(2 * pi * 0.00000000318 * 1000) = 50048 Hz
3)    1/(2 * pi * 0.0000000016 * 1000) = 99471 Hz
4)    same as 2

Off from 10 Hz, 50 Hz, and 100 Hz by three orders of magnitude.  Make sure you're taking "nano" to mean 1e-9.  It looks like you're using micro farads instead on 1, 2, and 3, and some other units entirely on 4.

[neo]

Google this, several solutions -

"lm3915 spectrum"


Regards, Dana.

I tried that before but apparently dropping analyzer off the end netted more results, i found this  https://ubiyubix.wordpress.com/2015/06/27/7-band-audio-spectrum-analyzer-w-msgeq7-lm3915-attiny2313/   which while not 40 channels would work well enough

But as useful as that is that is besides my point the math works for that but when i try to use it to get 40 different values it repeats.
1    1.6nf           10K  10hz
2   3.18nf       1k    50hz
3   1.6nf        1k     100hz
4      3.18nf       1k    500hz
those are my first four high pass calculations, so unless im misunderstanding a 50 hz is the same as a 500 hz high pass filter which is an invalid result

You're messing up the math

1)    1/(2 * pi * 0.0000000016 * 10000) = 9947 Hz
2)    1/(2 * pi * 0.00000000318 * 1000) = 50048 Hz
3)    1/(2 * pi * 0.0000000016 * 1000) = 99471 Hz
4)    same as 2

Off from 10 Hz, 50 Hz, and 100 Hz by three orders of magnitude.  Make sure you're taking "nano" to mean 1e-9.  It looks like you're using micro farads instead on 1, 2, and 3, and some other units entirely on 4.

what can i say, i suck at math 


Thanks everyone.

[Audioguru]

Your 1st order filters are too simple then the slopes of the highpass and lowpass are very gradual. A spectrum analyser needs filters with very steep slopes created with a 4th to 8th order bandpass filter, not a 1st order filter. The tutorial you posted wrongly called it a 2nd order filter.

[suicidaleggroll]

The tutorial you posted wrongly called it a 2nd order filter.

2nd order bandpass = 1st order HP + 1st order LP

[Fulcrum]

Would it not make more sense to make the bandpass filters out of RLC filters? Here's a calculator that will let you choose a desired pass band frequency, quality factor, and a few other parameters.
http://sim.okawa-denshi.jp/en/RLCtool.php