For a closed 3-way wooden loudspeaker design I am considering two crossovers (see below).
Looking at the quality of the components, which crossover design (A orB) do you recommend me to use?
Does a simpler design mean less distortion and better sound? Or might the heavier design produce a higher fidelity sound separation?
Context: normal private appartment listening room.
Speaker derivers: base woofer for lowest base, paper cone from the seventies to cover warm full midrange, ribbon or cone tweeter for highest notes, driven by a sony ta-n80es amp.
Crossover A:
Crossover frequency points: 500Hz, 3.5KHz
Board Size: 125mm x 95mm
Crossover B:
Crossover frequency points: 250HZ, 3.2KHZ
Board Size:192 x 95mm
Looking at the quality of the components, which crossover design (A orB) do you recommend me to use?
Does a simpler design mean less distortion and better sound? Or might the heavier design produce a higher fidelity sound separation?
Context: normal private appartment listening room.
Speaker derivers: base woofer for lowest base, paper cone from the seventies to cover warm full midrange, ribbon or cone tweeter for highest notes, driven by a sony ta-n80es amp.
Crossover A:
Crossover frequency points: 500Hz, 3.5KHz
Board Size: 125mm x 95mm
Crossover B:
Crossover frequency points: 250HZ, 3.2KHZ
Board Size:192 x 95mm
If they are not designed for the speaker in question, i.e. if they are not designed for those drive units, with their exact frequency, impedance and distortion performances when mounted in that exact cabinet design in exactly the right places: none of them. The results will be so bad the component quality is irrelevant (once you hit a reasonable minimum standard, you're window-dressing anyway).
At the very least, can you tell us what the impedance (in ohms) the drivers are. If you can tell us exactly what they are with model numbers & data, that would of course be the ideal.
At the very least, can you tell us what the impedance (in ohms) the drivers are. If you can tell us exactly what they are with model numbers & data, that would of course be the ideal.
Last edited:
Thanks Scottmoose for your reply! Wow im blown away with your critical reply but I do respect it. So you mean that best is to design your own crossovers? I can solder but where to start? I want a simple 3 way analoge crossover at crossover points of around 300 Hz and 3000 Hz using the best possible componentry (Nichicon capacitors?). I just have no idea how to combine these components to the crossover I want.
Here are all the parameters ive got:
Both crossovers state they are rated 4-8 Ohm.
Drivers:
Woofer: 18" Dayton Audio 8 Ohm woofer
Midrange: Philips 4 Ohm paper cone driver
Tweeter: Fountec Ribbon tweeter 8 Ohms
(basically, im so in love with the warm natural sweet sound of these philips HighFidelity International speakers model 22RH422, I will be re-using the midrange driver adding the base cabinet and tweeter to it)
Philips 22RH422/OIZ | Analog-Audio.info
Here are all the parameters ive got:
Both crossovers state they are rated 4-8 Ohm.
Drivers:
Woofer: 18" Dayton Audio 8 Ohm woofer
Midrange: Philips 4 Ohm paper cone driver
Tweeter: Fountec Ribbon tweeter 8 Ohms
(basically, im so in love with the warm natural sweet sound of these philips HighFidelity International speakers model 22RH422, I will be re-using the midrange driver adding the base cabinet and tweeter to it)
Philips 22RH422/OIZ | Analog-Audio.info
By definition you can't have a crossover that claims to have specific electrical transition frequencies, and in the next breath claims to be rated '4-8ohms'. If it says it has an electrical crossover frequency of x Hz, that can only apply at one impedance, e.g. 8ohms. Because if you present it with a different impedance, that crossover frequency will change dramatically. In other words, if you see a claim like that, unless they've got switchable loads in there (unlikely), it's bollocks.
So from that you can probably see the answer: yes. For quality results, a crossover needs to be designed for your specific drivers, in the specific enclosure. If this sounds negative, it's not meant to, and no criticism is being levelled at you personally. What I'm trying to do is warn you about some pitfalls & point you in a direction where you'll end up with something reasonable. You can learn how to do this properly, or at least do a better job than you'd get from those generic things, and there are plenty of people here who can help, if you wish.
A brief summary, FWIW:
-A loudspeaker drive unit has an impedance that varies with frequency, and all of them are different. Nominal values (e.g. '8ohms') are just that: nominal, and the driver is usually a long, long way away from that figure across most of its range. It may in fact only cross that value in a couple of places.
-In the same way, a loudspeaker drive unit has a response that varies with frequency. It's not flat by itself, and the second you put it in a box, it will change due to the various loading & diffraction effects involved.
-If, for a moment, we assume that the drivers have a flat impedance that precisely matches an off-the-shelf generic crossover (good luck finding drivers like that
), the crossover frequencies themselves are only electrical transfer functions. Because drivers have a varying frequency response, the actual acoustic behaviour, which is what you listen to, can be & usually is, wildly different. And that's assuming those values are actually appropriate to the real-world impedance load, which they rarely are.
Back onto more specific matters, thanks for the link to the Philips: that's a start. Short of some more detailed data, we're unlikely to be able to help you do a highly refined filter, but we should be able to help you get something reasonable running. Next up: which Dayton woofer, and which Fountek tweeter, specifically? As I recall, there are a number of each. Links are appreciated, as most of us don't have time to go searching. Knowing Dayton & Fountek, they, and the Philips, will all be decent examples of their type. Better still, the Fountek ribbon will have a fairly stable impedance load, which will make filter work a bit easier. I strongly advise you to read AllenB's sticky thread on this forum on how to design crossovers without measurements. Introduction to designing crossovers without measurement This method doesn't pretend to do state-of-the-art / hyper refined, but it's a good start & guide to some of the basics.
So from that you can probably see the answer: yes. For quality results, a crossover needs to be designed for your specific drivers, in the specific enclosure. If this sounds negative, it's not meant to, and no criticism is being levelled at you personally. What I'm trying to do is warn you about some pitfalls & point you in a direction where you'll end up with something reasonable. You can learn how to do this properly, or at least do a better job than you'd get from those generic things, and there are plenty of people here who can help, if you wish.
A brief summary, FWIW:
-A loudspeaker drive unit has an impedance that varies with frequency, and all of them are different. Nominal values (e.g. '8ohms') are just that: nominal, and the driver is usually a long, long way away from that figure across most of its range. It may in fact only cross that value in a couple of places.
-In the same way, a loudspeaker drive unit has a response that varies with frequency. It's not flat by itself, and the second you put it in a box, it will change due to the various loading & diffraction effects involved.
-If, for a moment, we assume that the drivers have a flat impedance that precisely matches an off-the-shelf generic crossover (good luck finding drivers like that
), the crossover frequencies themselves are only electrical transfer functions. Because drivers have a varying frequency response, the actual acoustic behaviour, which is what you listen to, can be & usually is, wildly different. And that's assuming those values are actually appropriate to the real-world impedance load, which they rarely are. Back onto more specific matters, thanks for the link to the Philips: that's a start. Short of some more detailed data, we're unlikely to be able to help you do a highly refined filter, but we should be able to help you get something reasonable running. Next up: which Dayton woofer, and which Fountek tweeter, specifically? As I recall, there are a number of each. Links are appreciated, as most of us don't have time to go searching. Knowing Dayton & Fountek, they, and the Philips, will all be decent examples of their type. Better still, the Fountek ribbon will have a fairly stable impedance load, which will make filter work a bit easier. I strongly advise you to read AllenB's sticky thread on this forum on how to design crossovers without measurements. Introduction to designing crossovers without measurement This method doesn't pretend to do state-of-the-art / hyper refined, but it's a good start & guide to some of the basics.
Last edited:
Not so much critical, but realistic. These will probably work but not well unless your needs are simple. The question of component quality doesn't even come in to it. Nor does the one about simple crossovers having less distortion (which is debatable at best).
By far the best value with a crossover is good workable values. That means blending and trimming well. Your acoustic design needs to be sound to begin with for it to work well.
By far the best value with a crossover is good workable values. That means blending and trimming well. Your acoustic design needs to be sound to begin with for it to work well.
Wow ! I love to see the basics repeated THX Scott... ( Allen already owns a long thread about crossover basics...! )
About that particular speaker ( Philips ), I removed the back of the box and it sounded much better. Probably yours is glued, mine was plastic.
Try to see if you can find an improvement in substituting the only component that fails with age ( if electrolytic ) : the capacitor.
No Nichicon but simply a film type, plastic polypropilene or MKT.
THX Scott... ( Allen already owns a long thread about crossover basics...! )About that particular speaker ( Philips ), I removed the back of the box and it sounded much better. Probably yours is glued, mine was plastic.
Try to see if you can find an improvement in substituting the only component that fails with age ( if electrolytic ) : the capacitor.
No Nichicon but simply a film type, plastic polypropilene or MKT.
Thanks very much ScottMoose and AllenB and PicoWallSpeaker for your input friends! I let the critical view of ScottMoose work in me and read AllenB's introduction to crossovers (and reacted there).
At this moment you convinced me to completely forget about "off-the-shelf" prefabricated 3-way crossovers boards! Your arguments sounds clear and convinced me that IF someones goes the passive way then the only audiophile way would be custom built crossover specific tuned for the used drivers. over here in The Netherlands I researched and found companies that design and build custom crossover filters, costing approx 100~200 $ per channel (labour + high quality parts). 200$ ~ 400$ for a set of crossovers does not sound cheap, but it should be the highfidelity route to go.
Q1. Will the "best" passive crossover boards STILL introduce distortions and phase-problems etc into the sound fidelity?
What if I would skip passive design all together and opt for an active system design?
I have three sony 80ES amplifiers which would be perfect for a 3-way speakerdesign!
Q2. When going the high quality route, is there an ACOUSTIC BENEFIT in using passive over active speaker designs?
Q3. What kinds of active 3-way crossover would you recommend (max 200$ ~ 400$)?
Q4. Is/are there problems if the output of amp goes directly to a 4ohm driver or 8 ohm driver etc? Are there things I should take in to account opting for an active system?
Q5. Would it be possible to just simply rotate the volume knob of Woofer / Mid / Tweeter to achieve the desired (realtime!) loudness level of each of the 3-ways, or is this too simplistic an idea?
Thanks Very Much
At this moment you convinced me to completely forget about "off-the-shelf" prefabricated 3-way crossovers boards! Your arguments sounds clear and convinced me that IF someones goes the passive way then the only audiophile way would be custom built crossover specific tuned for the used drivers. over here in The Netherlands I researched and found companies that design and build custom crossover filters, costing approx 100~200 $ per channel (labour + high quality parts). 200$ ~ 400$ for a set of crossovers does not sound cheap, but it should be the highfidelity route to go.
Q1. Will the "best" passive crossover boards STILL introduce distortions and phase-problems etc into the sound fidelity?
What if I would skip passive design all together and opt for an active system design?
I have three sony 80ES amplifiers which would be perfect for a 3-way speakerdesign!
Q2. When going the high quality route, is there an ACOUSTIC BENEFIT in using passive over active speaker designs?
Q3. What kinds of active 3-way crossover would you recommend (max 200$ ~ 400$)?
Q4. Is/are there problems if the output of amp goes directly to a 4ohm driver or 8 ohm driver etc? Are there things I should take in to account opting for an active system?
Q5. Would it be possible to just simply rotate the volume knob of Woofer / Mid / Tweeter to achieve the desired (realtime!) loudness level of each of the 3-ways, or is this too simplistic an idea?
Thanks Very Much
Last edited:
The questions you have asked will open a can worms.
You will find many people here that will tell you that active crossovers are vastly superior to passive, a few rational folks who say that if the filter function is the same, they sound the same - and one or two kooks like me who prefer passive.
Active is faster and easier to get right, which is why a of people like it. But as has been so well explained in the posts above, the component quality doesn't mean a thing if the crossover is wrong - active or passive.
Since this is a DIY forum, I'd suggest you design and build your own. It may cost you more than the prices you quote above because of the test equipment, but you'll learn something and gain a greater understanding of how speakers work.
You will find many people here that will tell you that active crossovers are vastly superior to passive, a few rational folks who say that if the filter function is the same, they sound the same - and one or two kooks like me who prefer passive.
Active is faster and easier to get right, which is why a of people like it. But as has been so well explained in the posts above, the component quality doesn't mean a thing if the crossover is wrong - active or passive.
Since this is a DIY forum, I'd suggest you design and build your own. It may cost you more than the prices you quote above because of the test equipment, but you'll learn something and gain a greater understanding of how speakers work.
Hi Eftegari,
You have gotten some excellent replies. Listen, if you would like, you can have some simulation files for a 2-way. You don't have to buy anything, but you can try your hand at designing with a generic crossover calculator and compare your results to the original. It's a great way to get introduced to the issues involved without actually spending money.
A Speaker Maker's Journey: The LM-1 Bookshelf Version
With the simulation and discussion here:
The LM-1 DIY Speaker is Available
Again, the point is not to buy anything, but to use the free information to simulate your own crossover.
Best,
E
You have gotten some excellent replies. Listen, if you would like, you can have some simulation files for a 2-way. You don't have to buy anything, but you can try your hand at designing with a generic crossover calculator and compare your results to the original. It's a great way to get introduced to the issues involved without actually spending money.
A Speaker Maker's Journey: The LM-1 Bookshelf Version
With the simulation and discussion here:
The LM-1 DIY Speaker is Available
Again, the point is not to buy anything, but to use the free information to simulate your own crossover.
Best,
E
It will be better -providing they use the measured responses of your drivers in your enclosure. If not, then they will be no better than generic.
All loudspeakers have distortion. They're the worst part of the entire audio chain. As for phase problems, it depends what is defined as a 'problem'. There is phase and then there is phase, and it depends exactly what is being referred to. Absolute phase? Linear phase? Relative phase? Phase rotation of filters? And at what frequency? All speakers are a compromise & design is about identifying which aspects are important to you & which less. Some tend to obsess over overall phase response -too much in many cases, as its effects are often swamped by other factors, and many supposedly phase-linear systems only actually achieve that at one point in space. Bluntly, I'd say that's a long way down the list of things you should be 'worrying' about at the moment.
Active systems have advantages & disadvantages. They just happen to be different. Theoretically they are more flexible, but that doesn't come without sacrifice, and you still need to design the speaker response; it can't do that for you.
There can be, because not all active speakers are created equal, in the same way not all passive speakers are created equal. With active setups, you're ultimately limited by the quality of the electronics. Whether these are audible relative to the [alternative] compromises of a passive setup will depend on the specific circumstances, so there's no single 'yes' or 'no' answer to that. In many cases the benefits of active will probably outweigh the downsides, but not all.
miniDSP is usually the go-to reasonably priced solution. But: as noted, whether a crossover is active or passive, you still need to design the response. The fact that the filter is done before the amplifier does not remove that requirement.
Not with a half-decent amplifier.
Yes: see above.
It sounds like a good way of blowing your tweeter and midrange units, and getting lousy sound in the process.
Crossovers are not just about adjusting the overall level of the drivers. It's about shaping the response to blend outputs together, including relative phase & delay etc. So unfortunately no, that won't work. I wish it did, life would be a lot simpler!One point: for the love of Mike, don't even think about using on-line crossover calculators, like the one somebody kindly linked to above. The truth is they are the same as generic crossovers: purely mathematical / electrical derivations, and absolutely worthless in practice for most people unless you happen to have drivers with coincident voice coils, which also have a flat impedance and frequency response for several octaves above and below your target filter frequency. Since your drivers do not have coincident voice coils, nor a flat frequency or impedance response for several octaves above and below your unknown crossover frequencies -don't go there.
Last edited:
It's not something I stress about. There are some things a crossover cannot fix (such as the physical aspects of your design) that will eventually take your attention away from this issue, but that for now you can overlook, that set limits on how well a speaker can work when you cross it.
If you want to build speakers with high quality NATURAL sound, you will want to use the best modern drivers with advanced computer designed motors and advanced cone materials, plus attention to controlled directivity which complements your "normal private appartment listening room".
--> Find a well reviewed modern 3-way design which meets your musical tastes and room demands.
--> iPod Bedroom romance from philips HighFidelity International Speakers
-Have you listened to ribbon tweeters and compared their sound with dome tweeters? Ribbon tweeters require a higher crossover frequency, which requires a smaller diameter midrange to generate a good polar pattern(directivity). Not a good crossover partner to your 18" woofer.
-A modern dome tweeter allows a lower frequency crossover to a larger diameter midrange - a midrange capable of covering the human vocal range. A good partner to a larger woofer.
-Very few "normal private appartment listening room" require an 18" woofer in a BIG cabinet. A 12" woofer in a 4 cuft cabinet can provide strong bass. The woofer can be on the front baffle or side baffle.
-Two separate cabinets provide some usage flexibility(TM for desk, bedroom ), but two separate boxes are much larger than one combined volume for the woofer.
The woodshop capability often sets the design direction. Basic box, vs. a few bevel cuts for style.
Are you interested in getting recommendations for a new tech 3-way from this forum?
--> Find a well reviewed modern 3-way design which meets your musical tastes and room demands.
--> iPod Bedroom romance from philips HighFidelity International Speakers
-Have you listened to ribbon tweeters and compared their sound with dome tweeters? Ribbon tweeters require a higher crossover frequency, which requires a smaller diameter midrange to generate a good polar pattern(directivity). Not a good crossover partner to your 18" woofer.
-A modern dome tweeter allows a lower frequency crossover to a larger diameter midrange - a midrange capable of covering the human vocal range. A good partner to a larger woofer.
-Very few "normal private appartment listening room" require an 18" woofer in a BIG cabinet. A 12" woofer in a 4 cuft cabinet can provide strong bass. The woofer can be on the front baffle or side baffle.
-Two separate cabinets provide some usage flexibility(TM for desk, bedroom ), but two separate boxes are much larger than one combined volume for the woofer.
The woodshop capability often sets the design direction. Basic box, vs. a few bevel cuts for style.
Are you interested in getting recommendations for a new tech 3-way from this forum?
for the two boards, besides drivers not having infinite/flat acoustic response and impedance, those woofer inductors appear to be air core and ~18-19 (?) gauge, so will add some resistance to the woofer. They may not even have inductance associated with lower F crossover work. Once you have a design, ERSE steel bar core are good for woofer - Jantzen air core for tweeter work.
(a lot of the Klipsch "heritage" stuff have very simple crossovers)
(a lot of the Klipsch "heritage" stuff have very simple crossovers)
I am not going to answer the rest of the questions because others have already done so.
Active: direct coupling between driver and amp; some say less energy loss in the crossover since the crossover is before the amp; opportunity to Time align via DSP.
Passive: Needs only 1 amp instead of 2 or more hence cheaper; do not need a power source near the speakers; preamp is close to power amp so preamp can drive power amp better (some preamps cant drive long cables).
To add to what LineSource, ScottMoose and others have said.
Building your own 6" 2 way from a kit will not save you a lot of money over store bought brands. For example if you buy some speaker like a Elac Debut B6 for say $250/pr you would be hard pressed to build anything much better for the same price.
When you move up the value chain you start to find that you can build for less than a branded speaker for example if you consider a 2.5 way like the Sonus Faber Venere 2.5 at $2500/pr you can build a similar SB Acoustics Tower kit from Madisound for $1000/pr ($500 for the drivers and crossover and $500 for everything else - cabinet, etc).
As you move further up the value chain you can save more for example the Joesph Audio Pearl 2 retailed for about $25k/pr in 2011. I built a speaker using the same Magnesium SEAS Excel drivers for a fraction of that cost (maybe around $4000) at about the same time. You can see a pic of the system here (different topic) Corrosion SEAS Magnesium Drivers
BTW the same SEAS Excel drivers are used in speakers costing a lot less than the Pearls, like the Salk Sonud HT2TL and Tyler Acoustics Decade SE.
Salk Sound
Decade X SE
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.