Crossover Design Feedback please

Hello, Not sure if this is the right place to post this but I am designing a 3 way crossover for my first diy audio project.
I am completely new to this so I don't really know what I'm doing.

The drivers I'm planning on using are:

Sub: Dayton Audio SD270A-88
Mid: Dayton Audio SIG180-4
Tweeter: Dayton Audio ND16FA-4

I have come up with this design and the system curve looks reasonable to me but again I have no Idea what I'm doing so I would love some feedback!

Screenshot (76).png


Btw the way I came up with this is by using a crossover calculator online and then tweaking the values to get a better response curve
 
A few initial thoughts:

crossover design is hard, I don't pretend to understand it, but three way crossovers are much harder to design than two ways, and either type of crossover needs experience, measuring equipment and expertise in how to do it all. Ideally, the drivers should be measured in the cabinet you intend to use.

your design looks like it doesn't have "Baffle Step Compensation" so would sound bright with less bass than you would like

the ND16 tweeter has a stated 'resonant frequency' of 2125Khz, so you don't really want to cross it below twice that, i.e. about 4200. The SIG180 - which looks like a nice driver - starts to 'roll off after 3kHz, so it might be hard to get the woofer and tweeter to work together. A 1'' tweeter might be preferable, it can cross lower.

as all your drivers are 4 ohms nominal, your impedance plot looks like it would be a '2 ohm' nominal speaker, which will kill your amplifier; and

lastly: you're much, much better off using a proved project by a noted designer which you know will work well.

I know this sounds negative, but you'll be spending a lot of Euros and a lot of your time and you want to be happy with your project. If you let us know things like your budget, room size, music taste and sources and desired cabinet size etc people will be pleased to help you with some project options.

Geoff
 
Last edited:
Part of what is driving impedance down is the excessive overlap between drivers. For a 300 Hz crossover at 12 dB/octave, target curves like this are more typical:

1694997274169.png


I agree with several of the points above. But if you really want to learn, it's not an insurmountable problem. As long as you expect a lot of intricacies and want to put in the effort, it's doable. It often comes down to whether you enjoy the hobby or are just looking to build one nice pair of speakers. Your performance expectations are also important. Some people enjoy the building process and aren't looking for perfection in their projects.
 
Last edited:
I don't know very much about cabinet design, but my very basic understanding is the while the volume needs to be about the same, cabinet dimensions, especially the baffle width, can be varied by up to 10% without too much affect on the sound. But I'm very happy to be corrected on that.

Driver placement on the baffle (that is, their distance from each other and offset from centre line) needs to be the same, because the crossover has taken those dimensions into account.

To give you some inspiration, here are some links to interesting designs which are a bit different to normal:

https://projectgallery.parts-express.com/speaker-projects/the-knuckleheads/

https://sites.google.com/site/undefinition/floorstanding-speakers/tarkus


The video has some really interesting projects

Geoff

 
your design looks like it doesn't have "Baffle Step Compensation" so would sound bright with less bass than you would like
Experts on this forum are obsessed with building speakers to perform on an open floor in the middle of a room at a trade show or diy meet.
90% of home built speakers are located against a wall, which boosts the bass about 3 db. Corner location boosts about 6 db. Harder walls are better. Mine is 1" of plaster over wood lathe. My Peavey SP2(2004) have no BSC and the bass is fine in my installation against that hard wall.
Agree with the comment about building a 2 ohm speaker for a home amplifier, ie single output pair transistors. A triple CS800x could handle it.
 
Last edited:
Are you dead set on using those drivers? 16 mm tweeter is too small and fragile to be crossed reasonably low, basically won't work with a 180 mm mid. 3" or 4" mid is far more suitable to be placed between 10" woofer and 5/8" tweeter. Crossed at 400 Hz and 3 kHz that 10-3-5/8 combo will have not only linear on-axis curve, but smooth directivity and power response as well.
 
Are you dead set on using those drivers?
Definitely not, I'm considering making a two way design now because it will be easier designing a crossover. I'm making this project just so I have a little practice building speaker cabinets before starting to build a big Reggae soundsystem. I'm planning on handling crossover with a dsp when I build the reggae soundsystem so my focus is not really on learning to make crossovers.
 
Experts on this forum are obsessed with building speakers to perform on an open floor in the middle of a room at a trade show or diy meet.
90% of home built speakers are located against a wall, which boosts the bass about 3 db. Corner location boosts about 6 db.
Point taken, but if some BSC is built into a design it could give more flexibility in placement. I'm not in any way an expert and just make suggestions from my experience

Geoff
 
If you're going to make a two way and design the crossover yourself, I suggest buying drivers which have smooth responses and can work together well without too much in the way of crossover parts. Some drivers need quite a few parts to tame cone break up or deal with impedance issues, etc.

There are many discussion threads around, on this and other Forums, just Google 'designing crossovers without measurements'. Here's some I found particularly useful.

https://www.diyaudio.com/community/...igning-crossovers-without-measurement.189847/

https://sites.google.com/site/undefinition/diy-speaker-faqs/simulated-measurements

https://speakerdesignworks.com/speakerbuilder.html

For software, I use "Xsim", it's easier for me than Vituix or Passive Crossover Designer.

What brands are available for you apart from Dayton? Nothing against Dayton, I've used many of their drivers but SB Acoustics, Peerless and SEAS, for example also make very good drivers at various price points. What's your budget, please?

Some websites have actual measured data for some drivers, of course those measurements are taken from their cabinets. Some makers' data and specifications is regarded as accurate, others less so.

I use "fptrace" (free) to translate makers' graphs into frequency response and impedance curves into "frd" and "zma" files which can be used in crossover software. Of course, they're not measured in your cabinet and may not be accurate anyway but it's a start.

Have fun!

Geoff
 
Last edited:
Definitely not, I'm considering making a two way design now because it will be easier designing a crossover. I'm making this project just so I have a little practice building speaker cabinets before starting to build a big Reggae soundsystem. I'm planning on handling crossover with a dsp when I build the reggae soundsystem so my focus is not really on learning to make crossovers.
If it helps, the need for quality frequency response data is still the same whether you design an active or passive crossover.

We can make your 3 way work in your first post, but it requires knowing the baffle design and driver placement on the baffle. The reason for this is we can apply the baffle step and diffraction response to the manufacturer curves which gives a closer approximation to what the drivers will actually do in your enclosure.

We'll want to aim for a crossover that presents no less than 3 ohm load at any point to the drivers, this is especially important in the bass frequencies (greatest current demands) and very high frequencies (no "shorts" leading to ultrasonic oscillation).

Measurements are always preferred however, and buying a USB mic is the simplest, cheapest option that gets the job done. You'll need this capability for more complex builds and to optimise whatever you build (active or passive).
 
diyAudio Moderator
Joined 2008
Paid Member
GeoffMillar, the Behringer ECM8000 non-USB mic seems to be widely available in Australia for $72.

It requires phantom power, which can be DIY into a cable for the cost of a few 9V batteries and a few simple components. Otherwise it needs no special interface, just the regular line in on a computer. There are those who buy an additional interface just for the simplicity of it.
 
The only problem with the ECM8000 is no calibration file. You'll need to perform a version of self calibration. That along with buying / making an XLR -> single ended cable plus phantom power supply (+48V), I'd recommend the UMIK-1 as something that would possibly be cheaper overall.
 
I've said this before. Image stolen from the net. My own personal ECM800 follows a similar pattern, but not the drastic +/- 5dB of some of these mics:
1695181844356.png


Whilst agree the ECM8000 is likely accurate around your typical crossover points for a 2 or 3 way, it could create mis-leading data when wanting to apply DSP / LT to adjust bass response or "false" attenuation of high frequencies / misunderstanding of tweeter or diffraction ripple.

Having a mic with some sort of linear +/- 1dB response is preferable (to me anyway) than one without any calibration at all.