A crossover is not only the electronics. It is the act of passing off to another driver which is typically in a different location to the first. It is always a problem although when done properly, any problems will be close to inaudible. Therefore it is normally only a good idea to have a crossover when there is a good reason.
The first stage in designing a crossover is deciding how the drivers will blend in an acoustic sense. They should be mounted in suitable baffling to achieve this, and treated with electronic filters to finish the job.
You should define the purpose of each of the drivers. Once mounted, measure them to confirm they are doing their job and for designing the filters.
The first stage in designing a crossover is deciding how the drivers will blend in an acoustic sense. They should be mounted in suitable baffling to achieve this, and treated with electronic filters to finish the job.
You should define the purpose of each of the drivers. Once mounted, measure them to confirm they are doing their job and for designing the filters.
Two ways of sending the appropriate energy to the appropriate drivers.
1) "Passive":- a single power stage amplifies full range signal, and divides the output power by frequency band between the different loudspeaker chassis. This separation is done with high power components, inductors and capacitors (which in theory are high current devices, as no power is dissipated in them, but theory doesn't always come up with it) but often resistors to match the efficiency of different drivers and get the impedance somewhere near constant with frequency, or
2) "Active", where each loudspeaker (or group of loudspeakers within a frequency band; in your example, it wouldn't be worth separate amplifiers for the two eight inch if they were aimed in the same direction) has its own power amplifier and the separation of frequency bands is done using active components at line level. The advantages of the former are the simplicity of the wiring and the need for only one power amp per cabinet, which means a standard HiFi set up can drive them. The advantages of the latter are flexibility: frequencies and rolloffs for different bands don't need to be mirror images, phase angles can be tweaked, frequencies changed and sensitivities adjusted at the touch of a switch or turn of a potentiometer, without requiring rewinding inductors or soldering in new value components. You can even introduce specific equalisation for particular drivers. Much more fun for a development set up.
The downside, unless you are building the power amps into the cabinet, is a veritable spider's web of cables, screened and unscreened, a line level input and multiple power amps. And the possibility of running the wrong signal into the wrong driver.
Passive crossovers can be bought ready made for existing driver compliments, and occasionally modified for others but I suspect your "three tweeters – different" is going to require some custom design. Similarly electronic crossovers can be bought, or built, digital or analogue, for quite reasonable sums, but show a tendency to be preprepared for standardised systems; a low to mid frequency selectable, but automatically the same for the two drivers, and automatically the same order of filter, and again, a mid to high.
If all your tweeters have the same recommended low cutoff, this would be fine for you; three different HF power amps, relatively low power, each one its own volume control and perhaps equaliser, but if two of them were 3kHz recommended crossover and the third a 5k hypertweeter you might have difficulties without DIY.
Is that clear? No, it probably isn't. And it's only the first stage of getting those units cooperating.
1) "Passive":- a single power stage amplifies full range signal, and divides the output power by frequency band between the different loudspeaker chassis. This separation is done with high power components, inductors and capacitors (which in theory are high current devices, as no power is dissipated in them, but theory doesn't always come up with it) but often resistors to match the efficiency of different drivers and get the impedance somewhere near constant with frequency, or
2) "Active", where each loudspeaker (or group of loudspeakers within a frequency band; in your example, it wouldn't be worth separate amplifiers for the two eight inch if they were aimed in the same direction) has its own power amplifier and the separation of frequency bands is done using active components at line level. The advantages of the former are the simplicity of the wiring and the need for only one power amp per cabinet, which means a standard HiFi set up can drive them. The advantages of the latter are flexibility: frequencies and rolloffs for different bands don't need to be mirror images, phase angles can be tweaked, frequencies changed and sensitivities adjusted at the touch of a switch or turn of a potentiometer, without requiring rewinding inductors or soldering in new value components. You can even introduce specific equalisation for particular drivers. Much more fun for a development set up.
The downside, unless you are building the power amps into the cabinet, is a veritable spider's web of cables, screened and unscreened, a line level input and multiple power amps. And the possibility of running the wrong signal into the wrong driver.
Passive crossovers can be bought ready made for existing driver compliments, and occasionally modified for others but I suspect your "three tweeters – different" is going to require some custom design. Similarly electronic crossovers can be bought, or built, digital or analogue, for quite reasonable sums, but show a tendency to be preprepared for standardised systems; a low to mid frequency selectable, but automatically the same for the two drivers, and automatically the same order of filter, and again, a mid to high.
If all your tweeters have the same recommended low cutoff, this would be fine for you; three different HF power amps, relatively low power, each one its own volume control and perhaps equaliser, but if two of them were 3kHz recommended crossover and the third a 5k hypertweeter you might have difficulties without DIY.
Is that clear? No, it probably isn't. And it's only the first stage of getting those units cooperating.
I don't want to discourage you from building a speaker but I'd strongly advise you not to build something like this. Designing a crossover will be difficult and using three tweeters is a very bad idea if you want acceptable sound. Have you considered building a nice kit?
Love a challenge
Yes I could do a Kit ........ seems a bit easy though
I do like a good challenge
and I would really like super amazing speakers
I guess I could start with a good 3 way
but that still seems like a easy way - I would end up with a reasonable speaker - but not and Exceptional Speaker
I am happy to take 3 months just to get my head around it all - I am in no hurry
Yes I could do a Kit ........ seems a bit easy though
I do like a good challenge
and I would really like super amazing speakers
I guess I could start with a good 3 way
but that still seems like a easy way - I would end up with a reasonable speaker - but not and Exceptional Speaker
I am happy to take 3 months just to get my head around it all - I am in no hurry
Well that's an excellent starting point
Perhaps you could read up and play around with simulation programs like Boxsim to learn how to design crossovers. You could look for a book about loudspeaker design. You could perhaps even read my personal favourite, Floyd Toole's Sound Reproduction (although it is more about important parameters in faithful reproduction than about speaker design). If you want to design your own speaker, a measurement system is also a must.
Sounds a bit random, especially the 3 different tweeters. One tweeter with 2 mids and 2 woofers could be a winning combination, e.g. in a wmtmw arrangement. Maybe a second tweeter at the back for ambiance.
It's not a seven way; it's maybe a three, maybe (depending on those three tweeters) a four. But the question's right, if you are thinking passive crossover, because impedances and level matching are different with multiple drivers in one frequency range.
I am assuming the tweeters are multiple either for directional reasons or for some scheme to continue into the ultrasonics. I can't see ever needing the extra power in tweeters, and getting phases aligned at those frequencies is a horror.
I am assuming the tweeters are multiple either for directional reasons or for some scheme to continue into the ultrasonics. I can't see ever needing the extra power in tweeters, and getting phases aligned at those frequencies is a horror.
Since you mention super amazing...
The concepts in acoustics are difficult to learn because they are difficult to hear and identify. You may have brains, but it's your ears that need the training.
There is a wealth of knowledge that can be learned and applied via a two way design. There is relatively less further to learn with a more complex speaker but a much greater chance you'll miss even the simple stuff.
Some of the things you'll want to understand (as opposed to simply being told by someone) are: why can two drivers be better than one and how are they used together, why use a baffle and how can one be made better, and you'll need to teach yourself some basic electronics (AC, of course).
You're unlikely to build a super amazing speaker in even three years without a few jump starts, and maybe not even then without being able to demonstrate certain acoustic concepts to yourself....unless your idea of super amazing just means better than all your friends have, then three months should be fine.
You should probably begin with filtering the following things:
The woofer should be low passed within a range that will blend with the mids. This needs to work with the woofer impedance and the woofer response which are both varying with frequency. The mids need to match this in both level and response.
The same should be done with the mid to tweeter transition noting that you'll now have three things going on with the mid (low pass, high pass and level setting), and these filters will interact with each other.
The rear firing tweeter might be better left until later when you can identify what you want this tweeter to be doing.
Numerous proven designs are readily available. A good challenge as beginner is to select one, acquire material, build, and test with measurement system to verify performance.
Most start with monkey coffin sized box instead of human sized box. Don't bury yourself.
The ear training input is good. I started out with 2way, then 3, then 4, then 5 way.
I needed. 5 way as the goal was a 'proper' horn based system. each driver/horn being kept to its best operating freq range.
I used passive x/o as things grew. I had issues with freq rising impedance on my 15'' drivers - needed Nobel correction... then there are Lpads to level the driver efficiency - spawn of Satan these, sapping dynamics and clarity. Very detrimental to high efficiency super dynamic compression drivers.
Inductors and caps for my low 90Hz crossovers are not cheap, so cost carefully. The whole lot runs to $$$.
You can easily get a really good fit dsp x/o, I use WAF Najda (search / look under line level forum), works out cheaper on the x/o front, power amps cost of course but you don't need dacs or pre amps, and yes Najda really is good enough SQ wise where other dsp x/o's were not.
You need to measure to tune, look into Holimpulse, and calibrated behringer Mic and phantom pre.
I needed. 5 way as the goal was a 'proper' horn based system. each driver/horn being kept to its best operating freq range.
I used passive x/o as things grew. I had issues with freq rising impedance on my 15'' drivers - needed Nobel correction... then there are Lpads to level the driver efficiency - spawn of Satan these, sapping dynamics and clarity. Very detrimental to high efficiency super dynamic compression drivers.
Inductors and caps for my low 90Hz crossovers are not cheap, so cost carefully. The whole lot runs to $$$.
You can easily get a really good fit dsp x/o, I use WAF Najda (search / look under line level forum), works out cheaper on the x/o front, power amps cost of course but you don't need dacs or pre amps, and yes Najda really is good enough SQ wise where other dsp x/o's were not.
You need to measure to tune, look into Holimpulse, and calibrated behringer Mic and phantom pre.
Last edited:
The basic cabinet size (internal volume) depends on the type (bass reflex, acoustic suspension, various other types), as well as the Thiele-Small parameters of the woofer. Offhand that seems like a huge cabinet, maybe even for TWO 15" woofers. If you really want a cabinet that big, maybe a horn design could work.
Yes, the crossovers are filters, but the crossover parameters are interrelated with the cabinet/baffles. For one thing, a "great" crossover compensates for the baffle step, and that's directly determined by the size of the baffle.
Here's a good book to read:
Loudspeaker Design Cookbook: Vance Dickason: 9781882580477: Amazon.com: Books
Also available here:
Loudspeaker Design Cookbook 7th Edition Book 500-035
Yes, designing speakers is definitely a "good challenge." It's sort of like the Chinese saying "may you live in interesting times."
The short answer to this is yes, but having different models of tweeters makes things a lot more complicated, both electrically (the design of the crossover) and acoustically (where the tweeters are mounted on the cabinet). I'd pick one tweeter (perhaps the one with the most/best frequency response overlap with the midrange), design it for that, and then maybe add another tweeter, perhaps back-facing for ambiance. Even there I can only imagine it would be preferable to use the same model as the forward-facing tweeter, to get the same "tone" or coloration out of it.
The problem is there's enough to know about designing speakers that when you're starting with little knowledge, you don't know how much there is that you don't know. If the Dickason book might be overwhelming (it DOES cover a lot of info), maybe you should start with Weems' book:
Designing, Building, and Testing Your Own Speaker System with Projects: David Weems: 9780070694293: Amazon.com: Books
Get some drivers and cabinet for a two-way speaker (because it's a lot easier than 3-way), and toss them together in the cabinet, connect a crossover and hear how it sounds. It can be humblingly disappointing compared to a decent mid-fi speaker (lower end JBL or Infinity). Don't ask how I know...
Indeed, a great crossover specifies the parameters of the baffle before it's built. This is to create the desired directivity of the system, and to have the sections operating toward this end within their expected bandwidth.
OK Thanks all for shearing your experience - what I will do first is build a 2 way pair
and then I will do a 3 way
Now I want the 2 way to be made with the best components So ....... what would you all suggest for the Drivers ? I am happy to spend a few good dollars
as I have learned that many 2 way systems are far better them a 3 or 4 way system - it all depends in the quality of the components used -
and later I will get back to the Big box idea !
and then I will do a 3 way
Now I want the 2 way to be made with the best components So ....... what would you all suggest for the Drivers ? I am happy to spend a few good dollars
as I have learned that many 2 way systems are far better them a 3 or 4 way system - it all depends in the quality of the components used -
and later I will get back to the Big box idea !
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.