My doubt basically about midrange speakers where I have to use a capacitor and an inductor to filter it. I would filter the subwoofer at 400hz, and I was researching on the internet and they said that I should calculate a margin so that there is no cut on top of the other, so I should make the cuts cross a little before, so that there is no peak or with value... so I should do it like this: bass cut at 400hz, the midrange will start at 450hz and end at 4khz, the tweeter will start from 4.2khz to 20khz (this is an example), is this correct or should I make the cuts precisely? 400hz - 2000hz. Also note that I inverted the polarity of the midrange, because as this filter is 6dB per octave it will generate a phase problem between the crossing of the paths (that's what I heard).
The way you've described what you've read, it doesn't seem to add up. There should be no problem crossing at the same frequency, but you can't be sure what is happening without measuring.. do you go higher or lower, not sure.
Also, reverse polarity for 6dB? Not sure about that..
Also, reverse polarity for 6dB? Not sure about that..
You have chosen Fl = 400 Hz and Fh = 2,000 Hz. I read that it is advisable not to have Fl and Fh too close together.
A good 3-way crossover results when Fh = 8Fl. Perhaps the experts can confirm (or deny) that.
So, if your woofer dictates that Fl should be 400 Hz, then a good Fh would be 8 x 400 = 3,200 Hz (3.2 kHz).
A good 3-way crossover results when Fh = 8Fl. Perhaps the experts can confirm (or deny) that.
So, if your woofer dictates that Fl should be 400 Hz, then a good Fh would be 8 x 400 = 3,200 Hz (3.2 kHz).
It is a good idea when you don't have a way of checking their interaction, just like staying 2 octaves from dome tweeter resonance. The response is more predictable that way but essentially there's no (single driver response based) reason you shouldn't do it otherwise.
When it comes to crossover slope (order) and phase relationships, we only care ultimately about the acoustic response. that is:
Acoustic response = Driver (in box) response + Electrical filter.
For example - I can apply a single capacitor on a tweeter. In theory this is a 6dB "first order" electrical filter. However, depending on capacitor value (which frequency this will apply) and what the driver is doing (both frequency response and impedance), this may result in more, or less than 6dB per octave "acoustic" response.
If you are doing your own design (and not planning on doing your own measurements), you should really have a good read of the sticky in this forum Introduction to designing crossovers without measurement
This will show you that starting with a pre-conceived "electrical" filter concept is not how we design crossovers. We select drivers that provide the performance we are looking for in their intended operating range, and that these are compatible with eachother.
Then we select what acoustic slopes we are targeting (e.g. 4th order between tweeter and midrange and 2nd order between woofer and midrange as an arbitrary example). Then we add crossover components, until we see the driver responses changing to align to the desired acoustic targets.
The crossover will often not look anything textbook. And that is ok. We add components to shape the response... all about getting to the acoustic target.
Acoustic response = Driver (in box) response + Electrical filter.
For example - I can apply a single capacitor on a tweeter. In theory this is a 6dB "first order" electrical filter. However, depending on capacitor value (which frequency this will apply) and what the driver is doing (both frequency response and impedance), this may result in more, or less than 6dB per octave "acoustic" response.
If you are doing your own design (and not planning on doing your own measurements), you should really have a good read of the sticky in this forum Introduction to designing crossovers without measurement
This will show you that starting with a pre-conceived "electrical" filter concept is not how we design crossovers. We select drivers that provide the performance we are looking for in their intended operating range, and that these are compatible with eachother.
Then we select what acoustic slopes we are targeting (e.g. 4th order between tweeter and midrange and 2nd order between woofer and midrange as an arbitrary example). Then we add crossover components, until we see the driver responses changing to align to the desired acoustic targets.
The crossover will often not look anything textbook. And that is ok. We add components to shape the response... all about getting to the acoustic target.
If you haven't already done so, please have a good look at Paul Carmody's DIY speaker website as well as the pages referenced by Dave Bullet.
https://sites.google.com/site/undefinition/diy
He has two and three way projects for all budgets, great FAQs and other useful and informative material. He also explains why 3 way speakers are much harder to design than 2 ways and how he 'voices' his speakers.
Geoff
https://sites.google.com/site/undefinition/diy
He has two and three way projects for all budgets, great FAQs and other useful and informative material. He also explains why 3 way speakers are much harder to design than 2 ways and how he 'voices' his speakers.
Geoff
I don't know if it's still advisable to design any XO without measurements. Measuring everything has never been easier and cheapier than it is today. But more important: the result will be much much better.
Using tools like Vituixcad are really just fantastic, and will get you a long way into reaching your goals. Inverting a driver is literally one click. You see what that would change instantly.
You can go even further with things like XSim, which is an automated tool that tries to find the right XO components. It's fun to watch it try out combinations and see the FR of the combination get closer and closer to what you want.
Using tools like Vituixcad are really just fantastic, and will get you a long way into reaching your goals. Inverting a driver is literally one click. You see what that would change instantly.
You can go even further with things like XSim, which is an automated tool that tries to find the right XO components. It's fun to watch it try out combinations and see the FR of the combination get closer and closer to what you want.
I was going to mention that to our student Pedroga as part of his learning curve.
A dome tweeter has a resonant frequency (Fs) which, when using a simple 1st order electrical filter, should be avoided by crossing the tweeter at least two octaves higher than Fs.
Now, an octave is a doubling of the frequency.
So, for a dome tweeter with an Fs of 1,000 Hz, the crossover frequency should not be 1,000 Hz - or even 2,000 Hz - but should be 4,000 Hz.
In your 3-way crossover design above, you could therefore set Fh to 4,000 Hz and work backwards to make Fl equal to 4,000/8 = 500 Hz.
N.B. I supply this as an added lesson in crossover design. It is not intended in any way to negate the excellent advice of the experts above.
ehi, experts...what about ferrofluid? Tweeters with ferrofluid in the gap will behave better with a 1st order filter than a tweeter not equipped with ferrofluid.
So I'm going to make it clear in my head that you should NOT reverse the polarity of the midrange
I used this example value, but my initial idea was 3khz or 4khz
I need to read more about this.
I read and took notes on almost every topic
I'll leave the link to the video I watched about crossovers, the sound from his DIY box was exceptional, I think there's a way to turn on the subtitles
I found the crossovers he made very interesting, but just by looking at the electrical diagrams you can already tell that it is very complex for a beginner like me.
But I will continue studying there and seeing all his projects
However, be aware you would reverse it if it were a 2nd order (12dB/octave) 3-way crossover.
EDIT: Applies to 3-way crossover.
Last edited:
Unfortunately, subtitles are not available.
To remove an unwanted post, look for the Delete button (dustbin symbol) at the bottom of your post.
I'm not so sure, it could go either way. (Either for a reason or accidentally, depending on the details. Expect to experiment.)
Due to the fact that electrical phase shift is less severe in a 1st order crossover than in a 2nd order one I presume?
In practice, Pedroga could determine the best way round by a performing a listening test.