Introduction to designing crossovers without measurement

Hi
Been trying for some years to adjust my speakers sound to own liking without any luck! Changed x-overs using tables and charts on web pages three times! Till I came upon this thread, really straight forward and step by step guide, thought give it a fourth try! Just one question though. May I come back for help in a new thread hoping to get non-sarcastic, non-insulting answers?!
Thanks for a really fine-tuned thread.
 
The capacitor
Take the value of impedance you wrote down two posts ago (for our example it was 3.75). Multiply it by 12.6 and then multiply it by the crossover frequency. For example, 3.75 x 12.6 x 2000 equals 94500.

Sorry for asking this but I don't really get how you got the figure 12.6 there. I assume you were working backward using the equation for the cutoff frequency:
gif.download

but that works out roughly 6.3, not 12.6. I wonder if I'm missing something here?
 
diyAudio Moderator
Joined 2008
Paid Member
Hi tranhieu, yes there's a little more going on in this case and it's to do with the second order filter.

The formula you've shown gives a capacitance that equals the resistance at the given frequency. It produces a cut of 3dB in the resistance (only 3dB because the capacitor is slightly out of phase with resistance), and is used with first order filters.

With a second order filter, the capacitor and the inductor resonate at some frequency. Doubling one of them and halving the other doesn't change the resonance frequency, which is true in this case. Far above and far below the crossover they'd still behave the same, but at the crossover, the balance of capacitance to inductance controls damping (filter 'Q'). The Q I have used is 0.5, a 'Linkwitz-Riley' filter.
 
diyAudio Moderator
Joined 2008
Paid Member
With the schematic in post 11, the capacitor is used to hold the low frequecies from the tweeter the same as in a first order filter. The inductor is added which returns further lows from ground so they won't be seen by the tweeter.

This capacitor and inductor interact in a unique way, and for this reason I'd be reluctant to call one of them the first or second. In typical use though, you wouldn't choose to run the inductor without the capacitor, although it has been done.
 
diyAudio Moderator
Joined 2008
Paid Member
It can be done (and often is), but try not to cross too close to the tweeter's resonance. This will be necessary to ensure reasonable power handling, and it will reduce the effect of the impedance peak interrupting the crossover...both to which a single capacitor will be more sensitive.
 
Thank you for this excellent series of tutorials. They contain a lot of important points that I only discovered through many years of research and experimentation. Just a couple of questions:

1. Why did you stop at first order filters? I have noticed substantial improvements with second order. My best speakers are fourth order, and I have noted big improvements in clarity with higher order crossovers, and there are excellent reasons to support this.

2. Does anyone have any mathematical models or ray diagrams for the step effect? People often make a big fuss about narrow speaker cabinets for improved imaging, which leads to a multitude of 6.5 inch designs with compromised low frequency performance. This configuration can lead to just as many headaches because of such effects.
 
diyAudio Moderator
Joined 2008
Paid Member
1. Why did you stop at first order filters? I have noticed substantial improvements with second order.
Thanks for your comments, danoz. Using an inductor with a series RC on a woofer is not necessarily a first order filter. In the plot below I've shown a fourth order slope on a woofer that I've used, and it is crossed using only those three components. The acoustic response is the goal, the electrical components are just the tools we use to get there.

You'll notice I decided to follow a second order electrical filter for the tweeter so as not to introduce undue excursion issues, but moreover in this tutorial I wanted to use something that would typically function well in a reasonable cross section of designs to use as a basis. I also needed something that did not turn tweaking into a chore. Furthermore, without phase and delay information available, chasing higher orders can be problematic.

Seasons greetings to all.
 

Attachments

  • Image1.gif
    Image1.gif
    5.9 KB · Views: 4,112
diyAudio Moderator
Joined 2008
Paid Member
In the spirit of the thread I would suggest you use a simple trial and error process. Begin with 10W resistors, such as sand cast (white box) or 5W units if you happen to have them. These units are popular for good reason. Play something through them for a few minutes and feel them, carefully. Any which you can keep your hands on should be OK.

This is a good test for power handling although a resistor can survive beyond temperatures that are comfortable to handle. They may go open circuit if overheated. Keep them from contact with anything flammable until you are sure.

By the way, you should use distortion effects during this test as this should change the tone in a way that puts more pressure on the power handling needs of the resistors.
 
Last edited:
diyAudio Moderator
Joined 2008
Paid Member
Thanks feyrerm,

I was asked in an email about the resistor that is used to attenuate the tweeter. Specifically, whether it is significant that I have put it before the capacitor rather than next to the tweeter.

As I see it this is a compromise based on the fact that this is arguably the most tweaked component. Otherwise I would suggest that a good way of attenuating the tweeter would be to design an L-pad and to include this impedance value within the calculations for the capacitor and inductor.

I didn't do it this way for a couple of reasons, and one of them is the relative difficulty in tweaking compared to the quality of the results obtained.

Focussing on a single resistor in the way described will be reasonably predictable for the smaller amounts of attenuation usually encountered. As the amount of attenuation increases though, the added resistance will cause the tweeter's cutoff point to lower somewhat (although this may not necessarily be a problem, and might be countered by tweaking the other components).

If the resistor were instead next to the tweeter, then it would tend to lower the crossover point somewhat and cause some peaking in the response near it's cutoff as the amount of attenuation was increased.
 
I prefer to always put the L pad right next to the tweeter. The reason is that the resistor across the tweeter provides damping for the tweeter. Given that the tweeter always sees a high impedance at its resonant frequency when looking towards the amplifier, any damping will be a good thing. As such, I often use 50 ohms across the tweeter even when attenuation is not required.

If you embed this with the crossover design, or put the L pad on the other side of the high pass filter, you do not get this benefit.

Dan
 
If you embed this with the crossover design, or put the L pad on the other side of the high pass filter, you do not get this benefit.

Dan
Unless you're using rather thin and flimsy wiring between your crossover board and the tweeter, you're not going to see any practical difference mounting the L-Pad at the tweeter terminals or on the output side of the crossover board...

On your other point - you would never put an L-Pad on the input side of a crossover anyway, it makes no sense to do so and wouldn't work properly. L-Pads always go on the driver/output side of the crossover...