Crossover! The construction

I have all the necessary components to build my long-awaited box, doubted by many here (and rightly so).
I'm in the construction part now and would like to start building the crossover.
Do you, my friends, happen to have photos or manuals on how I can do this knowing that it will be a 3-way box construction?
I have no doubts about the values of components and such, but about the construction itself and the fear of building it and making some noise or some problem, I don't know...
What would be the best way to build it?

speakers.jpeg
 
I don't see any coils, capacitors or inductors in the picture. Have you already designed the crossover and bought all the parts, or are you starting the crossover design?

To assemble a crossover I like to use a wire wrap approach. I just get some 20 gauge sold copper wire and use it to wrap around the two, three, four or how ever many components need to connect together. I don't solder the connections so they can be easily changed during prototyping. It is easy does not require special tools. Here are pictures of the three way crossover I built a few weeks ago using this technique. I hold the parts to the board with a bit of hot glue. Hot glue can be removed by applying a bit of isopropyl alcohol.

1730419718162.png


1730419906676.png
 
Designing a good passive 3 - way crossover is difficult. It takes me days and lots of parts. If you have not yet bought a bunch of parts, I highly recommend starting out right and buying digital crossovers. You can configure them to do anything in minutes. Depending on your budget there are several to choose from. If you have a couple of amplifiers, the miniDSP is a good bet. If you don't want to have several amplifiers, there are good multi-channel amplifier boards and plate amps with DSP crossovers built in. https://www.parts-express.com/.../audio-amplifier-boards... If you want to buy something for the long term, look at Hypex Fusion DSP plate amplifiers. They sell a plate amp for 3-way speakers. They can be coupled together and you can add a remote and a display to see the volume level etc. The Room EQ Wizard software and a cheap measurement microphone will get you going. I also like Arta software, but the support for that ended. It may still be available to download. There are speaker design books. Look at those. Vance Dickason wrote a good one.

If you really want a passive crossover, you can start with a pre-built crossover and make adjustment. It will never be great, but it will work. https://www.parts-express.com/Dayton-Audio-XO3W-375-3K-3-Way-Crossover-375-3-000-Hz-260-150
 
@Pedroga Can you confirm that the crossover arrangement shown below is still the one you intend to use?

View attachment 1375322

If so, perhaps a kind member who is good with graphics will convert the schematic to a practical layout for you.

Exactly this one!The problem is that I have always made guitar pedals using vero boards, I have never done anything on an MDF board for example

In my head, if you leave the components bare there will be noise, I don't know why I have this in my head, even though I've opened old amplifiers and they barely had any wires connecting the components.:spin:
There are a few 'how to' crossover assembly videos on YouTube, although I can only find videos for two way speakers.

This is quite a good one:


Geoff

I noticed exactly that, dear friend!Can I put the attenuators on the crossover board?As a friend there told me, I'm going to make it as "eclectic" as possible where I can change and add components whenever I want, as if it were a lego.
http://www.troelsgravesen.dk/tips.htm Some general hints / suggestions
http://www.troelsgravesen.dk/coils.htm Advice for coil layout etc.
I read EVERYTHING.
It totally changed most of the things I would use... Glue, connectors, etc.
And, irrespective of Beranek's law, plan for it to change so do not build it permanently until you are satisfied with it. Then, note that there will be human curiosity after listening for a protracted period. Clip leads, wire nuts, wago lever nuts, barrier-strips...all are your friend.
wago lever nuts, This is used a lot in showers in my country, I have some in my "stock"
 
  • Like
Reactions: grindstone
As it is my first project in this field, tips came here on the forum in other topics that I had created about using electrolytic capacitors and such...
I think this time might be a good idea, but I intend to exchange them in the near future.
In my country it's kind of difficult to find resistors of more than 5W, think about the headache this is giving me.
I got some porcelain ones, they are so pretty they make you want to eat them
 
  • Like
Reactions: olsond3
The rounding of values worries me too... although it is not the focus of this topic.
For example, I need the resistors: 1.53 Ω, 6.43 Ω, 0.43 Ω, 32.78 Ω.
So I would do it the following way, the same way I did in high school. 1.53 Ω

Without the 4.7 Ω resistor, a combination that still comes close would be:
2.2 Ω in parallel with 10 Ω: 1.64Ω
6.43 Ω

Combining 10 Ω in parallel with 15 Ω gives a very close approximation:6.43Ω
0.43 Ω

Without the 0.47 Ω resistor, you can combine 0.68 Ω in parallel with another 0.68 Ω: 0.34Ω
32.78 Ω

For this value, combining 22 Ω + 10 Ω resistors in series results in: 32Ω
I don't know if these roundings are very serious in this type of project
 
Don't worry about the values being perfect. The schematic is a bare bones start and looks too simple. I expect it will
need a lot of tuning and more parts to actually work properly. Without some filters to control the impedance of the
drivers the high pass filters will tend to be at the low end resonance values of the drivers. The high end of the woofer
will likely not roll off without a resistor and capacitor network to flatten the impedance at midrange frequencies.

Did someone create that schematic using crossover design software that shows the expected frequency response? Xsim4, VituixCad or boxsim could be used to see how this will work.

You will be changing many of those values and adding more parts as soon as you make your first measurement of the speaker.

Do you have any type of microphone you can plug into a computer to make a measurement? That will help you a lot.
 
  • Like
Reactions: Pedroga
Don't worry about the values being perfect. The schematic is a bare bones start and looks too simple. I expect it will
need a lot of tuning and more parts to actually work properly. Without some filters to control the impedance of the
drivers the high pass filters will tend to be at the low end resonance values of the drivers. The high end of the woofer
will likely not roll off without a resistor and capacitor network to flatten the impedance at midrange frequencies.

Did someone create that schematic using crossover design software that shows the expected frequency response? Xsim4, VituixCad or boxsim could be used to see how this will work.

You will be changing many of those values and adding more parts as soon as you make your first measurement of the speaker.

Do you have any type of microphone you can plug into a computer to make a measurement? That will help you a lot.
I used simulators
 
  • Like
Reactions: olsond3
Simply use the standard resistance value closest to your calculated value.

The standard values you need to match the above calculated values are:

1.5 Ω, 6.8 Ω, 0.47 Ω (or 2 x 1.0Ω in parallel) and 33 Ω.
This is the way I can get closest to the value of the resistors with at least 5w...
I don't have many resistor options.
330R
360R
390R 470R 47K 47R 68R 6K8 0R05 0R08 0R47 100R 18K 27K 2K2 33R 3K3 2K2 51K
100R 2R2 0R18 0R27 0R10 0R15 0R22 0R33 0R47 0R68 100K 100R 10R 12K 150R 15R 180R
18K 1K 1K5 1R 220R 22R 270R 27K 27R 2R2 300R

So I did some rounding of values but with these roundings I'm afraid of causing a problem as this directly affects the speaker's impedance from what I understand.
Without the 4.7 Ω resistor, a combination that still comes close would be:
2.2 Ω in parallel with 10 Ω: 1.64Ω
6.43 Ω

Combining 10 Ω in parallel with 15 Ω gives a very close approximation:6.43Ω
0.43 Ω

Without the 0.47 Ω resistor, you can combine 0.68 Ω in parallel with another 0.68 Ω: 0.34Ω
32.78 Ω

For this value, combining 22 Ω + 10 Ω resistors in series results in: 32Ω
 
Did someone create that schematic using crossover design software that shows the expected frequency response?

Pedroga has simply been tutored on the basics of first order crossover design including the use of simple crossover calculators.

What he has now is a an improvement on what he had to start with, but it's a long way removed from a ideal crossover designed with the aid of computer design software.

It's been a long journey to get this far - check out his previous posts.
 
Last edited:
  • Like
  • Thank You
Reactions: olsond3 and Pedroga