Lets assume a crossover frequency of 1000Hz and 2 drivers each having an 8 ohm impedance. We feed that into an online calculator and it spits out the values for a 2nd order L-R of 1.3mH and 20Uf. For the LPF; its a series inductor and a parallel capacitor. The HPF, just reverse the components. I understand how the impedance increases with rising frequency and how the capacitor bleeds off more signal. Different values different cut-off. Cool! we can build that.
What happens if we change just one of those values? Coils are expensive but its relatively cheap to add or subtract capacitance. What effect does changing only one component have? Change the frequency? Change the linearity? Multi phase shifts. Space time distortion? I cant find anything on the web.
What happens if we change just one of those values? Coils are expensive but its relatively cheap to add or subtract capacitance. What effect does changing only one component have? Change the frequency? Change the linearity? Multi phase shifts. Space time distortion? I cant find anything on the web.
Changing only a series or shunt capacitor value in a basic crossover will alter the crossover
frequency, but may also change the frequency/phase response due to the driver characteristics.
For a tweeter, smaller series C -> higher tweeter crossover frequency.
For a woofer, smaller shunt C -> higher woofer crossover frequency.
The crossover will remain linear if operated within the limits of the (metal core) inductor.
Air core inductors will remain linear, like capacitors.
frequency, but may also change the frequency/phase response due to the driver characteristics.
For a tweeter, smaller series C -> higher tweeter crossover frequency.
For a woofer, smaller shunt C -> higher woofer crossover frequency.
The crossover will remain linear if operated within the limits of the (metal core) inductor.
Air core inductors will remain linear, like capacitors.
Last edited:
No free lunch
You seem to be looking for a cheap simple way out. NO FREE LUNCH!!
If you are serious about doing it right, install simple zobels first. If you have access to a measuring system (Parts Express DATS) for example, things get a LOT easier. Worth the $100 many times over. You can monitor the crossover design from initial component set up to when you finish. Validating the actual performance of the zobel(s) and or the true values of any components used for example. NEVER trust anything! There are many issues to solve that have big effects on how a crossover performs. Horizontal driver offset, Baffle step, individual driver efficiencies, Driver separation (vertical on baffle), driver native frequency response both on and off axis (break ups and deviations from flat also). Parasitic effects of the crossover components used. And this is just the start. BTW, you do not have to use zobels, it just makes the initial design a lot easier.
Formulas and simulators are starting points at best and can often lead one in to the wrong direction. Not trying to discourage you, but the above info needs to be kept in mind throughout the design process. J
You seem to be looking for a cheap simple way out. NO FREE LUNCH!!
If you are serious about doing it right, install simple zobels first. If you have access to a measuring system (Parts Express DATS) for example, things get a LOT easier. Worth the $100 many times over. You can monitor the crossover design from initial component set up to when you finish. Validating the actual performance of the zobel(s) and or the true values of any components used for example. NEVER trust anything! There are many issues to solve that have big effects on how a crossover performs. Horizontal driver offset, Baffle step, individual driver efficiencies, Driver separation (vertical on baffle), driver native frequency response both on and off axis (break ups and deviations from flat also). Parasitic effects of the crossover components used. And this is just the start. BTW, you do not have to use zobels, it just makes the initial design a lot easier.
Formulas and simulators are starting points at best and can often lead one in to the wrong direction. Not trying to discourage you, but the above info needs to be kept in mind throughout the design process. J
He's a curious beginner, with an entirely reasonable question. A straw, not a fire hose.
Last edited:
If I remember correctly in your case you have a 4 ohm Kappa 15 which you want acoustically match with using a custom LPF to compliment PSD2002 already fitted with 18dB/oct 1,6kHz stock HPF. Due to the response and cone break up of the Kappa you will need to cut low or steep. Glad you made a thread about it to get more opinions but you should tell the whole story if you want good advice 😉
I’ve suggested a 12dB at 800Hz in your other thread and also a -10dB Lpad for the highs. Would be great if someone could do a sim.
I’ve suggested a 12dB at 800Hz in your other thread and also a -10dB Lpad for the highs. Would be great if someone could do a sim.
I'm not trying to do anything specific. This is not the first time I've made a filter. It just occurred to me that there are no calculators for anything outside the accepted formulae. Its never discussed. It doesn't look like anyone has experimented. Perhaps we all just like listening to music and there's nothing wrong with that. Only Rayma attempted to answer the question and only as far as I could have guessed myself. A bigger cap bleeds more LF signal.
The formula for the Q of a filter is the square root of the capacitance multiplied by the square of the load, and divided by the inductance. So you can devise that increasing capacitance in either the high or low part of the circuit will move toward an underdamped response.
The field of filters is large, and requires lots of math. There are both analog and digital filters.
For analog filters, there are both active and passive. Also there are many filter types, such as Butterworth,
Bessel, Chebyshev, Gaussian, and elliptic, among others. All have their own behaviors and uses.
Passive speaker crossovers are very limited, but are low cost (important in a consumer product),
and can work reasonably well if carefully desiigned. Once you need more than a simple calculator,
next would be simulation. The equations for many filters quickly become very complex and unwieldly,
not to mention the behavior of the drivers.
Here is a good public domain book on filters.
Electronics: Analog Electronic Filters Theory, Design and Synthesis H. Dimopoulos (Springer, 2012) BBS : Free Download, Borrow, and Streaming : Internet Archive
Last edited:
A good way to start learning is to download Xsim software and begin experimenting with simulations. I found that making a variety of 1st order and 2nd order simulations helped develop a feel for how components interact (inductors, capacitors, resistors). You can download some driver responses and impedance data from parts express: the Dayton drivers have the frd and zma files available for download.
You could also start with VituixCad2. It is a much more fully capable software design and simulation software suite. It is a professional level software application. But with that capability comes complexity and a more lengthy learning curve.
You could also start with VituixCad2. It is a much more fully capable software design and simulation software suite. It is a professional level software application. But with that capability comes complexity and a more lengthy learning curve.
Rayma... Not trying to "hose" anyone, sad that some may take it that way. Basically cutting to the chase. Wish someone when I started out years ago would have let me know the issues I mentioned in my post. Understand the initial basic question. But to answer it you must factor in the major design issues one faces building a credible crossover. Yes it is fun to play and learn. What I did for many years. Still do on occation. What happens to many is they implement the values provided by a simple caculator and end up with nothing very usable. And not have a clue why. Not a confidence builder to be sure. J
Sure, and even the most experienced have real limits to what they can do with passive crossovers.
As far back as 50 years ago, active crossovers and multiple amps were thought to be better but
too expensive for most (and they were). Audio Research sold their EC-1 crossover and 3 stereo amps
to more people than you'd think, to use with Magneplanars. And no DSP.
More good background for beginners:
Crossovers
Active Filters
As far back as 50 years ago, active crossovers and multiple amps were thought to be better but
too expensive for most (and they were). Audio Research sold their EC-1 crossover and 3 stereo amps
to more people than you'd think, to use with Magneplanars. And no DSP.
More good background for beginners:
Crossovers
Active Filters
Last edited: