passive filters, attenuation, dB and.....everything else i dont get!!!

Thread Starter

ninjaman

Joined May 18, 2013
341
hello
im studying for an HNC in electronics at college. I have to design a passive filter for a speaker cabinet that holds a tweeter and woofer. I have to chose 2nd order or up. I chose third order filters for both. including the impedance matching parts for the drivers. Lpad and zobel.
I don't understand the dBs, the attenuation bit and the orders bit. please could someone point me in the direction of a site that explains this. I have read this site on dB and attenuation and filters. I still don't get it. I have to show that I understand what im talking about to get the distinction. I have used online calculators to get the values for the above components. I am reading sites recommended for the zobel circuits.
I know that the crossover point is -3dB and that butterworth has been standard for certain reasons, that Bessel is no good for certain reasons. I don't understand the 6dB/octave part. im guessing that an octave is a set of frequencies. that the octave is probably best used because it is a musical thing. I know that an octave contains eight notes. i have read that it is doubling or halving. im guessin this is why it goes 3dB 6dB 12dB 24dB and 48dB for orders of filters. i don't know what that means though.
does that mean it gets quieter or louder. i know there are drop offs after the cross over, the higher the order the steeper the curve. so 3dB would be gradual and 48dB would "nose dive" making it better than the 3dB. (guessing on that one)
any of this correct at all.
is there a math formula that i could use to help me understand. i have looked. i went to the library and found that the only books that would describe anything like this was car audio books and newns audio hifi handbook. though only limited information is available and no answer to my question.
any help would be appreciated

thanks
all the best, and guess what CHRISTMAS SOON! YAY!!!

simon
 

crutschow

Joined Mar 14, 2008
34,285
It sounds like your biggest issue may be the meaning of an octave. An octave is simply a doubling of the frequency. Thus, for example, the frequency an octave above 7kHz is 14kHz.

Conventionally a filter rolloff or corner frequency is stated at the -3dB (half power) point. Above or below that point (depending upon whether it's a low-pass or high-pass filter) the response rolls off at some dB amount per octave. This rolloff amount depends upon the filter order and generally equals 6db additional rolloff per filter order. Thus a 1st order filter rolls off at -6dB/octave, a 2nd order filter rolls off at -12dB/octave, a third-order filter rolls off at -18dB/octave, etc.

Sometimes the rolloff is expressed as dB per decade (factor of 10 or about 2 1/4 octaves) of frequency. In that case a 1st order filter rolls off at -20dB/decade, a 2nd order filter rolls off at -40dB/decade, etc.

-dB is a reduction in output level. +dB is a increase in output level. Have you read this.
 

Thread Starter

ninjaman

Joined May 18, 2013
341
thankyou, I understand a little more about the octave.

what about the inductor and capacitors themselves. my lecturer mentioned XL = XC, where does this come into filters. I know that if XL = XC there will be maximum power transfer. but im not sure how this fits in. I had to use an online calculator to get the values of inductor and capacitor for the type of filter that I was using. I don't have any idea about how these work together. I know the formula for XC and XL, also for resonant frequency. just to check that:

Fo = 1/2(pie)(sqrt)LC, this is correct?

im guessing that at this frequency the cross over will happen. the 0.707 point?
is this correct?

I am planning on making a video on you tube to explain all this.
so any help would really be appreciated.
many thanks
all the best
simon
 

LvW

Joined Jun 13, 2013
1,752
..........
Is there a math formula that i could use to help me understand. i have looked. i went to the library and found that the only books that would describe anything like this was car audio books and newns audio hifi handbook. though only limited information is available and no answer to my question.
........
I am planning on making a video on you tube to explain all this.
........
ninjaman, I only can say: Good luck for your project. And reserve a lot of time.
Filter design is one of the most challenging task in electronics. There is not only one single "math formula that i could use to help me understand".
It´s a bit more.
But it is a good start to learn about "octaves" and "dB`s".
 

Thread Starter

ninjaman

Joined May 18, 2013
341
hello

octaves are doubling or halving of frequency, so one octave is 5, the second octave is 10, the third is 15.
dBs are a measurement of something. here I think it is signals. attenuation is "absorbing" the signal. like a sponge absorbing water.
example, an inductor allows low frequencies until the roll off(I believe this to be 0.707, the half power bit?), then the signals fight against the inductive reactance to get through the inductor.
the higher the order of filter, the more drop off the signal. this is due to more inductive components in the circuit. and the values of these inductors gets (less/more?) to offer more inductive reactance to the signal?

if frequency gets higher the inductive reactance gets higher?
the signals get higher past 0.707?
so for a steeper curve you would require more inductors getting higher in value each time for an almost vertical drop?
is this right?
 

Thread Starter

ninjaman

Joined May 18, 2013
341
hello again

ok, I have gone through the formulas for XL and XC. I have two values for inductors.
the network is a T network(I think). an inductor in series, then capacitor in parallel and another inductor in series. I look at the results with as the frequency increases. the current goes down with the inductors and capacitors. but I did something wrong. I forgot they were connected in a network. so the first inductor and capacitor make a voltage divider. there would be a voltage drop across the inductor and it would be that voltage used to calculate the current through the second inductor.

could anyone comment on whether this is correct or not, PLEASE.

I feel im getting closer. I just have to work out the power and then try to attach it to dB/octaves?

cheers to everyone who helped so far!

simon
 

shteii01

Joined Feb 19, 2010
4,644
hello again

ok, I have gone through the formulas for XL and XC. I have two values for inductors.
the network is a T network(I think). an inductor in series, then capacitor in parallel and another inductor in series. I look at the results with as the frequency increases. the current goes down with the inductors and capacitors. but I did something wrong. I forgot they were connected in a network. so the first inductor and capacitor make a voltage divider. there would be a voltage drop across the inductor and it would be that voltage used to calculate the current through the second inductor.
I would like to see this circuit.
 

#12

Joined Nov 30, 2010
18,224
That is not a crossover circuit for a speaker. C1 is a capacitor, not an inductor. L1 is an inductor, not a capacitor.

If you're going to mix radio signals with speaker frequencies, it's no surprise that it isn't making sense.

Maximum power transfer and resonance are relevant in radio circuits. They are (almost) not relevant in a speaker crossover. The radio circuit is trying to enhance and pass the best quality possible. The speaker crossover is trying to split up and redirect the energy to the various speakers.

Try working with this for a while.
 

Attachments

Last edited:

Thread Starter

ninjaman

Joined May 18, 2013
341
http://www.apicsllc.com/apics/Misc/filter2.html

on here, the third order is what I mean.

I have currents through the inductors and capacitors. I have an input voltage. from this im guessing I can find the power. I would put this in
10 log10 (powerout/powerin) and find dB.

then for octave, I would total all the frequencies after the 0.707 point to zero. I would divide the dB by this number and get dB/octave?

is this correct?

im guessing that this is signal attenuation?

any help,

cheers
simon
 

#12

Joined Nov 30, 2010
18,224
When you got to, "octave" you went off the track. I don't know how or why you would "total all the frequencies".

I think you need Xc = 1/(2PiFC) and Xl = 2PiFL
That means that at any given frequency the inductor or the capacitor will have an impedance that is related to the frequency you chose. For a rough estimation, you can pretend that the impedance of each part acts like a resistor. At any given frequency you can imagine a voltage divider made of resistors and find the amount of current that will pass through a particular filter. Do not forget that each speaker has resistance and inductance, and they are of size that is relevant.
 
Last edited:

Thread Starter

ninjaman

Joined May 18, 2013
341
how do you figure out 18dB/octave then?

that part im lost with. I understand the part about the XC and XL being single resistors. im using 3000Hz as cross over frequency. so I would use the XL1, XC as a voltage divider? that gives a voltage to go through the XL2 resistor? then through the L pad/Zobel to the speaker? is it then that I have the 18dB/octave?

anygood? cause Im getting more confused

cheers though
simon
 

#12

Joined Nov 30, 2010
18,224
Yes. Do the math as if the inductors and capacitors are resistors, for which you calculated their impedance at 3000 Hz.

I have to go to work now.
 
Top