It is easy to understand how an parallel crossover network works.For instance in an first order crossover the high pass path blocks low frequency and the low pass path blocks high frequency...
But what about an series network...how does it function?...suppose we take ar-sxo or lc audio crossover as reference..how does it work? .Can anyone explain it in a simple manner...
But what about an series network...how does it function?...suppose we take ar-sxo or lc audio crossover as reference..how does it work? .Can anyone explain it in a simple manner...
Imagine that bass sound will only go through inductors. Imagine that treble sound will only go through capacitors.
What you mention applies only for first order parallel networks...and absolutely not for series network..which is beyond that...and it is bit complex ....that's why i want an simple explanation about the functioning of series networks
Why 'absolutely' not ?
L and C do their job , and so does R . But ,you'll have to compute also each speaker's model in the circuit , and each speaker has its own L , C , R behavior
which is not straightforward but mixed . So it may seem simpler , a series network , but if you 'open' all the inner parameters you'll find a balance , same as parallel .😱
Basic route is to use tweeters with ferrofluid , so it stops firmly the lower frequencies . I'm talking about 1st order , other orders are...beyond 🙄
Read this link. It was written by one of the most knowledgeable folks I know. He discusses and models the ar-sxo exclusively. It's a very old document and some of the images seems to be missing.
http://web.archive.org/web/20021012193843/http://www.geocities.com/kreskovs/ar.html
this link may be of some help understanding differences between // and series x-over:
Series vs. Parallel Crossover Networks 😉
Series vs. Parallel Crossover Networks 😉
Boys, none of those links explain how a series crossover works, and you know it! 😀
It's really quite simple if you understand a transfer function:
A simple series crossover includes an element of midpass which allows a mathematical chance of achieving matching numerator and denominator to get an allpass response with no phase change or group delay. With conventional crossovers the midpass s term often has a negative sign that equates to group delay. It has such poor rolloff rates that implementing it is not easy with real world drivers. You can create a parallel network including resistors that does the same thing, but what is neat about it is that the resistors can be replaced by impedance corrected drive units.
It's worth mentioning that Bang and Olufsen did something equivalent with a bass and treble unit plus a filler midrange operating over a mere octave or so. I wiull tell you more about higher order series crossovers when I have finished wading through the maths. 🙂
It's really quite simple if you understand a transfer function:
An externally hosted image should be here but it was not working when we last tested it.
A simple series crossover includes an element of midpass which allows a mathematical chance of achieving matching numerator and denominator to get an allpass response with no phase change or group delay. With conventional crossovers the midpass s term often has a negative sign that equates to group delay. It has such poor rolloff rates that implementing it is not easy with real world drivers. You can create a parallel network including resistors that does the same thing, but what is neat about it is that the resistors can be replaced by impedance corrected drive units.
It's worth mentioning that Bang and Olufsen did something equivalent with a bass and treble unit plus a filler midrange operating over a mere octave or so. I wiull tell you more about higher order series crossovers when I have finished wading through the maths. 🙂
system7, what is "s" and what values does it take? When "s" is put in, what is the resulting number representing? Thanks.
In a series crossover, which I have never tried, I would put a Zobel across the woofer, or lower the value of the capacitor in the same area of the crossover to keep the resistance in both legs similar.
In a series crossover, which I have never tried, I would put a Zobel across the woofer, or lower the value of the capacitor in the same area of the crossover to keep the resistance in both legs similar.
Last edited:
Elliot has reached the same conclusion as John K. did in his analysi I referenced earlier regarding series xover having a downside related to the interaction between the woofer's back EMF and the tweeter. So, here we have consensus.
What remains is for the hobbyist to try both approaches and satisfy himself with listening testts.
And, regarding those listening tests oh my, I've had perfect integration, perfectly flat response // filters working on my 2 way monitors that never sounded as musical, as natural as do the AR XOs I'm using right now, and they even measure worst, but sound better.
speakerdoctor, with respect you are missing the point about interaction between the two drive units. For the equivalent parallel topology where one drive unit is replaced with an 8 ohm resistor in the same series circuit and you double up with a voltage amplifier, the drive units are kept entirely separate AND DO NOT INTERACT WHATSOEVER. Don't waste time on the interaction. It's simply the price you pay for the more efficient series circuit. 🙂
@ strawberry, I can't spare you the beauty and power of using complex variables. Essentially s is shorthand for the engineer's jw, or the mathemetician's iw. Engineers don't use i for the root of -1 because i is CURRENT to us! 😉
The impedance of an inductor then becomes sL or jwL. A capacitor has an impedance 1/jwC. A resistor is simply a real number. The transfer function is normalised to resistor value 1 ohm and frequency w=1. Real frequency is then w/2*pi.
The power of this approach is that you are not really very interested whether the speaker is 16 ohms or 1 ohm, or exact capacitor or inductor values. It's the ratios that count. And what it does overall. One of the things a series circuit has is a very nice flattish impedance, though a parallel first order butterworth with single coil and capacitor is not too bad either, and does much the same thing.
@ strawberry, I can't spare you the beauty and power of using complex variables. Essentially s is shorthand for the engineer's jw, or the mathemetician's iw. Engineers don't use i for the root of -1 because i is CURRENT to us! 😉
The impedance of an inductor then becomes sL or jwL. A capacitor has an impedance 1/jwC. A resistor is simply a real number. The transfer function is normalised to resistor value 1 ohm and frequency w=1. Real frequency is then w/2*pi.
The power of this approach is that you are not really very interested whether the speaker is 16 ohms or 1 ohm, or exact capacitor or inductor values. It's the ratios that count. And what it does overall. One of the things a series circuit has is a very nice flattish impedance, though a parallel first order butterworth with single coil and capacitor is not too bad either, and does much the same thing.
Last edited:
A 1.5 way. The cap paralleling/shunting the bottom driver goes to o ohms at high frequencies so that the bottom driver gets no current. ie it is low-passed
A simple 2-way, 1st order series. The cap across the woofer works just like the above, the inductor goes to 0 ohms at low frequencies, so lows are shunted past the mid-tweeter (ie high-passed)
Now let's look at higher orders, this 1st order on the woofer, 2nd order on the tweeter.
The shunt components work as before, the series cap with the tweeter blocks lows, this adds the 2nd pole (ie 2nd order XO now). If we added a series inductor in the woofer leg, it would blook highs, you can extrapolate from here. (the R shunting the tweeter is an attenuation circuit)
dave
Attachments
- Status
- Not open for further replies.