Xover super basics

[petemoore]

  I'm sure all these answers are in the equations.
  But I can't work them yet.
  Sometimes reverse equating works.
  Let's start with the HP crossover network, an inductor and a capacitor.
What does the inductor do?
  [This is just type fodder, correct any mistakes].
1. Larger diameter wire core obviously make it able to handle higher currents.
2. An increased number of windings equates to a different Mh rating and frequency attenuation.
  so...When the Winding count goes up [using only 1 type of winding wire] the Mh rating goes up and the frequency attenuation goes down.
3. Smaller capacitors series with diaphram will attenuate lower frequencies, similar to an input cap on a FF, but higher voltages and currents of course.
  So...this nice big crossover [2way]...can I add another inductor [I have a smaller one, much smaller winding wire diameter], or unwrap some of the windings on the inductor...ie increase or decrease Mh rating by seriesing the inductors or de-wrapping windings off the nice big crossover's coil?
  I believe my new compression drivers are over-excursion-ing.
  I can buy the passive treble Xover, or the parts for one and build it, but that's not as much fun.
  I think re-reads on 'Xover's explained' pages would be easier to comprehend if I just knew like...which way to expect the inductor to look, not exact math, but at least a target area.
  Also, I'll try hunting up many application diagrams, see if I can make a head, tails or spider nails on which whats the inductors are howing.
  I am here http://sound.westhost.com/beginners.htm#7_0
  Small inductors are sometimes used in the output of power amplifiers to prevent instability with capacitive loads.
  And these sentences explained about the first half of understanding Xover's in a simple fashion:
  An inductor can be considered the opposite of a capacitor.  It passes DC with little resistance, but becomes more of an obstacle to the signal as frequency increases.
  With a given wire type, the Mh rating will go Up?, or Down? when winding count increases?
  Interesting note here:
  The capacitive voltage divider may come as a surprise, but it is a useful circuit, and is common in RF oscillators and precision attenuators (the latter in conjunction with resistors).  Despite what you may intuitively think, the capacitive divider is not frequency dependent, so long as the source impedance is low, and the load impedance is high compared to the capacitive reactance.
  I think I need a higher Mh inductor to shunt more low frequencies to ground..and the dude who helped me calculate a crossover freq of 3khz said the capacitor value should change at teh same time...so the capacitor would increase or decrease uf value to match the inductor?
  Sorry for the Waltered down versions...I'm trying protect these drivers and get good performance at the same time...dooing a wierd PA type setup.

 

[R.G.]

A simple inductor-capacitor crossover can be considered to be a frequency-sensitive volume pot.

Imagine feeding audio to an inductor in series with a capacitor to ground, and taking audio off the junction in the middle of the inductor/capacitor. What happens? Can we figure this out simply?

We know that an inductor has (ideally) no "impedance" at DC and that its ability to "impede" (imagine resistance, but different) AC goes up linearly with frequency. So at, say 100Hz, the inductor will have X impedance. At 200 Hz, it will have 2*X and at 1000Hz, 1*0X.

Impedance obeys the generalization of ohm's law (I'm simplifying here), so if the inductor lets Y amps through at 100Hz, it will let 2*Y through at 200Hz, if the inductor is the only thing limiting current.

Capacitors are the reverse. A capacitor's impedance is infinite at DC. But it impedes less as frequency goes up. So a cap with an impedance of Z ohms at 100Hz will have 2*Z at 200Hz.

Back at our inductor/capacitor "volume control", at very low frequencies, the capacitor impedance is so big that it can be ignored and the inductor is all that affects the signal. At very high frequencies, the inductor's impedance is so big and the cap's impedance is so small that no signal gets through at all. This is a lowpass filter - only low frequencies get through.

If we put the cap first and the inductor second, we get a highpass filter.

What sets these things apart from resistor-cap and resistor-inductor filters is that both the inductor and capacitor are affecting frequency and they affect it in different ways. So things are not all that simple at the frequency where L and C have the same impedance. First the crossover frequency is F = (1/2*pi)*SQRT(L*C). Next, the AC out at that frequency and at all frequencies is phase shifted from the input phase. Next, both caps and inductors store and release energy, so there can be a resonance at the crossover frequency. Adding resistance to the L/C damps this down. L-C filter design to get the right frequency, right rolloff rate, and proper damping is something of an art.

But back at how you modify what to do what.

The inductance of a coil is proportional to the square of the number of turns. In math-speak,
L = k*N^2 where L is inductance, N is the number of turns, and k is some number which depends on the physical form factor, material inside the coil, phase of the moon, etc. to make the raw numbers come out right. "k" is a parameter, a number you stick in one time for a given circumstance. For instance, 100 turns filled with air has a different inductance than 100 turns filled with iron, so "k" for a 100 turn coil has to have some junk in it for the difference between air and iron.

If you add or remove turns from an inductor, it changes L by the square of the change in the turns. So if you have a 100 turn coil, to double the inductance, you add 41 turns. 141 turns is 1.41 times 100 turns, and 1.41 squared is about 2. If you put on another hundred turns to get to 200, you now have four times as much inductance.

Caps are similar about area and spacing, but you can't vary that, so we won't go into it.

Inductors in series add like resistors. L = L1 +L2. Inductors in parallel act like resistors: L = (L1*L2)/(L1+L2) in parallel.

Adding inductance moves the filter frequency down; adding capacitance moves the frequency down, too. Small inductors, small capacitors are high frequency.  It also changes the resonance and may introduce unwanted phase shifts at other frequencies.

My recommendation?

Don't do this.

Go to sound.westhost.com and get the canned crossovers for a Linkwitz-Riley crossover and do your crossover at signal level, not speaker level. Then put another power amplifier chip on the output of the crossover to drive the second speaker. Active crossovers are vastly superior to passive crossovers and suffer only from needing another power amp channel. With an LM3886, the power amp is so cheap and easy that it overwhelms the difficulties of a passive crossover.

[Meanderthal]

 Arrgh... well that just blew my fantasy of a "Hey, it's easy, just  **** ** ******"  kinda answer. Been wanting to upgrade some floor monitors for a while now... rat shak no longer sells the wimpy but effective 100 watt 2 way 2500hz crossovers, and those USED to work particularly well... for vocal monitors anyway... 

I guess I could pull one and reverse engineer it, but the inductors... would no doubt be a PITA!

[petemoore]

If you add or remove turns from an inductor, it changes L by the square of the change in the turns. So if you have a 100 turn coil, to double the inductance, you add 41 turns. 141 turns is 1.41 times 100 turns, and 1.41 squared is about 2. If you put on another hundred turns to get to 200, you now have four times as much inductance.
  This helps!
  My recommendation?

Don't do this.
  But this just overpowered that.
Go to sound.westhost.com and get the canned crossovers for a Linkwitz-Riley crossover and do your crossover at signal level, not speaker level. Then put another power amplifier chip on the output of the crossover to drive the second speaker. Active crossovers are vastly superior to passive crossovers and suffer only from needing another power amp channel. With an LM3886, the power amp is so cheap and easy that it overwhelms the difficulties of a passive crossover.
  Ok..I've had enough passive discussions.
 

[RDV]

The little simple active x-over at ESP is as easy as building a pedal except for the bipolar power and that's not too hard either once you get used to it.

RDV

[Ben N]

Impedance obeys the generalization of ohm's law (I'm simplifying here), so if the inductor lets Y amps through at 100Hz, it will let 2*Y through at 200Hz, if the inductor is the only thing limiting current.

Capacitors are the reverse. A capacitor's impedance is infinite at DC. But it impedes less as frequency goes up. So a cap with an impedance of Z ohms at 100Hz will have 2*Z at 200Hz.

I think that was supposed to be "Y/2" rather than 2*Y, and Z/2 instead of 2*Z, yes?

Ben

[R.G.]

Yeah, yeah, confuse me with what I wrote, not what I meant... 

It's that old "can't type as fast as I think" stuff. My mind goes on and leaves my fingers to cope with whatever they can remember of what was supposed to go out.

You're absolutely correct - Y/2 and Z/2.

I'll go back and edit the post for people who pick this up later in the archives and don't get down this far.

[petemoore]

   Thanks for the helps guys, I'll see if I can get this sorted out better in my mind here..

[petemoore]

  I asked for the basics, super basics, and thank you for letting me have it!
  Instead of hunting down all the right capacitors which I don't have yet, I pushed on to the board build, using 3/4 of 2x20 pin IC sockets, OA's in the top halves, capacitors assigned seriesed up and down on each sockets remaining bottom lugs.
  I did have a slough of caps marked .0018uf, so I stuck those in to test the circuit.
  First through the woofers I heard nothing, except faint clicking as I touched the metal to circuit points, output, then working east, pins 5 of course had boosted sound.
  Plugging it into the HF drivers, I got faint extreme high freqs !!!
  Working with these amps and Opamps for a Bi-amp system has advantages...I can build mostly from what's here, inexpensively, and have a more effecient, much more flexible system in terms of application...ie all the crossing over, done 'early', means there is no further problem as long as the amp/speaker can do it's job, ohmages become much easier to work with, more of 'this' or 'that' becomes a matter of simple 'adding' [another amp to say bass, but this might take some de-coupling]. some further tuning can be done with volume controls and EQ, however I suspect I'll build at least another LP or bandpass output.
  I can use the passive/inductor board maybe on some other system, one with 8ohms high and low, and a lower crossover freq.
  I'll shop 1% .0033uf =3n3 capacitors tomorrow !
  The 11k resistors turned out to be 11.73k resistors, I suppose this ups the gain a notch, I had many 5k6 and 6k8 resistors to build matching 11.73's with...that's what went in there !! Bias voltages are as perfect as my batteries are matched.
  Yes, although it involves another amp, another amp has advantages, as does the active system...I am already well on my way to sweetness in the PA!...you can bet I'll be trying this system in various configurations. 
  Also [it seems] re-tuning the Xover freqs is a matter of obtaining enough 1% caps of slightly different value.
  I am already liking this much better than fretting around with inductors.
  Thanks for the link, and I'll watch myself RDV, Nice amp BTW !