Introduction to designing crossovers without measurement

Your drivers are very different from my B&W P6 speakers. So don't know if this will help you but I built it by ear and the tweeter is reversed polarity. I've been told it is a 2 1/2 way XO. The P6 has a 6.5" woofer, 6.5" midrange and a 1" tweeter.
P6 final XO.jpg
 
  • Like
Reactions: 1 user
I'm working on it now, but it will take some time.

How about an L-pad next to the driver to start. This will bring it down 6dB and smooth the impedance. It will also mean you don't need the same resistor and capacitor as the woofer circuit. (By the way, you can find L-pad calculators online just for now)

Next you can use the capacitor and inductor from the tweeter circuit, or just the capacitor.

Then add the woofer inductor. Total of 4 or 5 components. (I'll have to review this after making the demonstration graphics.)

Not sure if I understood correctly and based on what I understood, I drew the crossover for 3 ways as attached. I added L-Pad for both Tweeter and Midrange and added Zobel circuit on Woofer. Not sure if I need to add an inductor and caps on midrange though. Hope I am not totally misunderstood.
 

Attachments

  • IMG_4444.jpg
    IMG_4444.jpg
    400.5 KB · Views: 121
diyAudio Moderator
Joined 2008
Paid Member
My bad, it's the way I said it.

The L-pad looks OK. Now you can alternately pretend the mid is the tweeter being crossed to the woofer as a two way, then pretend it is the woofer being crossed to the tweeter. This way those extra components get added to the mid to the left of the 4 ohm resistor... except with different values that suit the mid itself.

Tell me if I'm making sense yet. Here's what I wrote earlier about this...

Next you can use the capacitor and inductor from the tweeter circuit, or just the capacitor.

Then add the woofer inductor.
but instead of 'tweeter circuit' I meant to say from the 'tweeter circuit post' using the midrange in its place.
 
My bad, it's the way I said it.

The L-pad looks OK. Now you can alternately pretend the mid is the tweeter being crossed to the woofer as a two way, then pretend it is the woofer being crossed to the tweeter. This way those extra components get added to the mid to the left of the 4 ohm resistor... except with different values that suit the mid itself.

Tell me if I'm making sense yet. Here's what I wrote earlier about this...


but instead of 'tweeter circuit' I meant to say from the 'tweeter circuit post' using the midrange in its place.

Thank you, Allen. To be honest, I am not 100% sure if I followed you correctly and treat the mid-driver as tweeter as well as woofer. Then, came up with this modified crossover schematic.

In short, 2 additional inductors and 2 additional capacitors were added in the Midrange and also replaced L-Pad resistor values to 6 ohm & 7.5 ohm respectively.
 

Attachments

  • IMG_4445.jpg
    IMG_4445.jpg
    625.4 KB · Views: 106
diyAudio Moderator
Joined 2008
Paid Member
That's the way, and that circuit should work.

One of the things to be aware of is that if you bring the two crossover points close together then they can interact.. not that it has to be a problem. If you're ever looking for an excuse to do some simulating.

The question of first or second order (for both the low and high pass sections, and whether you should keep one or two of those components) is more a matter for experimentation here. There was some mention of it in the tutorial but the factors involved (the driver depths, responses and frequencies involved) may be different with a 3 way.
 
  • Like
Reactions: 1 users
Hi, would someone be interested to suggest a cross over for the following MTM:

https://www.parts-express.com/TC-8034-8-Treated-Paper-Cone-Woofer-with-Foam-Surround-4-Oh-299-2200

https://www.newark.com/mcm-audio-select/53-805/3-x-7-piezo-horn-tweeter/dp/96K1143?st=tweeter

I'm building a set of high efficiency speakers for outdoor parties. I like the MTM arrangement to get better power handling and loudness.
There are no measurements for these drivers I'm aware of. So I'd like to keep the crossovers as simple as possible. From initial listening crossoverless (I've built the boxes already), The tweeters wont need any attenuation. And I have a good feeling about crossing 2.5k and higher. They will be used with subs. I've had good success in the past with the online calculators, at least for party speakers, but I don't know how to design for piezos. The woofers are 4ohms so would I series the woofers and then parallel with the tweeter etc?
 
diyAudio Moderator
Joined 2008
Paid Member
You can treat the two woofers as one 8 ohm unit and follow the tutorial as normal... However, MTM is a good opportunity for a 2.5 way. In a 2.5 way both woofers play bass, and one rolls off at the baffle step frequency (say 600Hz) and the other plays up to the crossover. This usually gives a fairly easy baffle step compensation, which helps if you do this without measurement. Another is that it avoids the narrowing vertical dispersion that MTM unfortunately has below the crossover if you can't cross them really low.

Since you have them in series, you'll need a couple of extra components. C1 causes the woofer labelled S2 to roll off at 600Hz. R1 and L2 make sure woofer S1 stays steady while that happens. L1 is the actual crossover at 2,500Hz. If you have any problems at the crossover you can add the RC zobel circuit across S1 as described in the section about flattening the impedance of the woofer.

Untitled.png


Crossing piezos is different because they are a capacitive load for the amplifier. Here is a small writeup on the basics - http://www.frugal-phile.com/piezo-XO.html
 
  • Like
Reactions: 1 user
You can treat the two woofers as one 8 ohm unit and follow the tutorial as normal... However, MTM is a good opportunity for a 2.5 way. In a 2.5 way both woofers play bass, and one rolls off at the baffle step frequency (say 600Hz) and the other plays up to the crossover. This usually gives a fairly easy baffle step compensation, which helps if you do this without measurement. Another is that it avoids the narrowing vertical dispersion that MTM unfortunately has below the crossover if you can't cross them really low.

Since you have them in series, you'll need a couple of extra components. C1 causes the woofer labelled S2 to roll off at 600Hz. R1 and L2 make sure woofer S1 stays steady while that happens. L1 is the actual crossover at 2,500Hz. If you have any problems at the crossover you can add the RC zobel circuit across S1 as described in the section about flattening the impedance of the woofer.

View attachment 1095128

Crossing piezos is different because they are a capacitive load for the amplifier. Here is a small writeup on the basics - http://www.frugal-phile.com/piezo-XO.html
Ok! So from the piezo article. It says I should use a 22ohm resistor parallel to the tweeter, and then a 2.2uF cap inline, this would effectively be a 3khz highpass. At what slope I have no idea. probably doesn't matter.

And then if I follow your crossover for the woofers, I'm low passing one woofer at 600Hz, and the other at 2.5kHz.

And then I should wire the woofers and tweeter in parallel? Then what impedance range will the amp see? Will this be more of a 4ohm or 8ohm speaker or somewhere in between?

What kind of problems might I experience if I don't flatten the impedance of the woofer?

Is impedance flattening and/or baffle step compensation always worth it?
 
diyAudio Moderator
Joined 2008
Paid Member
At what slope
The piezo would be first order. They can handle being crossed like this and the only reason to change it is if you think it sounds wrong.
I'm low passing one woofer at 600Hz, and the other at 2.5kHz.
Yes.. Wait, I used generic 8 ohm drivers in that sim by mistake. No problem... for the starter values for 4 ohms, halve the inductor and resistor values and double the capacitor.
Will this be more of a 4ohm or 8ohm speaker
Closer to 8 ohms.
What kind of problems might I experience if I don't flatten the impedance of the woofer?
The woofer might have more high frequency energy than you want. If this happens you could even try a larger inductor before instead trying the zobel.
Is impedance flattening and/or baffle step compensation always worth it?
You'll probably miss the baffle step compensation, especially if you go taking them outside.
 
diyAudio Moderator
Joined 2008
Paid Member
You seem to be referring to the side benefit of using second order causing a little extra delay because the tweeter is closer to the surface of the baffle. Without measurement it's still only a guess but if you are not sitting at exactly tweeter level, switching the upper and lower woofers places could give you the variation you need.
 
In a 2.5 way both woofers play bass, and one rolls off at the baffle step frequency (say 600Hz) and the other plays up to the crossover.
If my boxes are 12" wide, should I roll off S2 at 380Hz or so instead of 600? If so how should the values change for these 4ohm drivers?

Also how/why would the woofers being in series increase the number of components? How would it look like in parallel? Not that I want my speaker to be 2ohms, I just want to understand the circuit more. (though I do have both car amps and PA amps that can drive 2 ohm loads for extra power.)
 
diyAudio Moderator
Joined 2008
Paid Member
First with this post I'll describe the series circuit.
how should the values change for these 4ohm drivers?
Everything is scalable. There are 3 things to note...

1. I've put 8 ohm resistors in parallel with the drivers for the simulation to make them act like 4 ohm drivers. Do not add them to your crossover.

2. I made this change...
for 4 ohms, halve the inductor and resistor values and double the capacitor.

3. Then I adjusted the capacitor and the middle inductor together to make the baffle step frequency change.

Tweaking: While everything is open to tweaking, notice that the top end is down 3dB at 2.5kHz. Some crossovers are down at 6dB at the cross. These are not the frequencies you want too much energy. It might turn out necessary to try larger values for the top inductor, even twice as much.

Untitled.png
 
Last edited:
  • Like
Reactions: 1 user
diyAudio Moderator
Joined 2008
Paid Member
How would it look like in parallel?
An easy circuit when in parallel. I normally don't go for the series option when I have a choice when doing a 2.5 way.. in this case it's just that the MTM is also a tricky option and it's hard to say which is better.

The simplest parallel 2.5 way...

Untitled.png


In addition, the Zobel (RC) circuit may be added to the woofer that plays up to the crossover frequency.

Also how/why would the woofers being in series increase the number of components?
They share the amplifier drive Voltage. As I remove the higher frequency energy from the baffle step woofer, the other one tries to compensate. (Recreating a finite, non-infinite, non-zero impedance from within the middle of a circuit requires both series and parallel components.)
 
  • Like
Reactions: 1 user
diyAudio Moderator
Joined 2008
Paid Member
Ok, you can treat the two as a single 8 ohm driver when working out the inductor. Work out the resistor and capacitor as for one woofer, then double the resistance and halve the capacitance.

The other thing to note when using two in series is that the total sensitivity stays the same as for one woofer.

Edit: Added.
 
Last edited:
  • Like
Reactions: 1 user
diyAudio Moderator
Joined 2008
Paid Member
The capacitor is fine, it's polypropylene film with a Voltage rating above 50V.

I wouldn't recommend the resistor unless you are certain otherwise, as it is low powered and metal film. The standard resistor for speakers is wirewound and 5W. The power requirements do vary greatly but these cover most situations. As previously discussed, resistors should be checked in use to make sure the power rating is satisfactory.