How should I go to work making my new speakers?

Hello,

Here I am again. I'm willing to rebuild my system because my current doesn't sound that plesent.
I've got a better job so even though I've already got amazing midranges and tweeters I'm going to upgrade. To the absolute top. For the midrange, according to a lot of looking at measurments I've come to the conclusion that the SB-Acoustics MW16P-8 is the best midrange out there. It has distortion as low as it can get and it has the best decay on the market. Currently I have the MR16P-4 which is it's brother though the decay and distortion of the MW is a lot better.

Then for the tweeter, and this one is more of a quistion because it also has a lot to do with the resulting character. I currently use beryllium tweeters and according to the measurments beryllium is the absolute best. If I had to chose based on measurments the tweeter I would be taking is the Scan-Speak Illuminator D3004/664000. Though, in terms of sound I'm aiming at. I want my speakers to have a character like that of Wilson Audio and Dali speakers. Very natural and with great texture. Though, these speakers use soft dome tweeters. If it's possible to obtain the same kind of character with beryllium I'll stay with D3004/664000 because it's technically the best. If not I'm going to soft dome tweeters. As far as I could see the Scan-Speak Illuminator D3004/662000 is the best out there but I know the Seas Excel F29C002 should also be phenomenal though I couldn't find any measurments of it. I could find measurments of it's smaller brother the F25C002 which had outstanding decay. Even better than that of the beryllium Scan-Speak. Though I don't know about the F29C002. Anyone know about the performance of the F29C002?

I'm sticking with my SB-Acoustics WO24P-8 because it's doesn't really get any better than those. Rather different.

What I would like to do is use two MW16P-8's to create a 2-1/2 way. To save the midrange performance and dispersion of a single driver and reinforce the mid-bass and reduce distortion in the mid-bass, reduce stress on the midrange driver and overall get some more cone area/power in the mid bass. A single MW16P-8 is only a small mid-woofer and I want some more speaker than that to do the mid bass. The woofers will be actively operated. The mid-high module I want to make passive using (nearly) top of the line components so I can drive it with a much better amplifier than the Hypex FA123. Which is also one of the main problems in the current system, the amplifiers. Even though the Hypex are fun and cheap and all. They don't offer the sound quality I desire. They don't offer a black background and are noisy. And you can't hook them up analoguely. The ADC's inside are absolutely terrible.

Though than I have to make a passive crossover. And I don't know a single thing about how one does that, let alone properly. I can't even set a digital crossover to sound good. So tel me. What are the steps one should take one by one when designing a crossover. Again, I want them to have a character like that of Wilson Audio and Dali speakers. What kind of frequency response should I be aiming for and all? And what is the best way to measure my speakers when te most anechoic space I have is outside in the open air?

And can anyone tell me how to properly measure with REW and the MiniDSP UMIK-1. Because it's not working properly for me. It's extremely inconsistent and partly doesn't work at all. Everytime I restart REW and plug in the mic I get a different measurment of the same speaker, it doesn't work under about 300 Hz, no matter what changes I make to the frequency response of the woofer, the measurments don't show a difference and in the top register the measurments only change very little when changes are made.
The measurments also differ from laptop to laptop. I get different results on my Macbook than on my HP. So basically, I can't make measurments. Anyone know how to fix this?

And can you help and tell me exactly what I have to do when I get to designing the crossovers? (will be a few months). I'll be using mundorf foil coils because they are the best though for capacitors I don't know. What do you think are the best crossover capacitors? The higher-end Mundorf, Jantzen or Audyn? (The Audyn True Copper Max are too expensive though, even for my budget)
 
From Sound System Engineering;

“When the number of Variables approach an order of magnitude, I turn in despair to my measuring apparatus.”

I believe you are right to first get your measurement platform working correctly and reliably. There are so many variables that are very difficult if not impossible to get right by ear.

Some project management guidelines that I live by:

Rules of Data Collection:
1 What am I trying to measure?
2 Why am I trying to measure it?
3 is it relevant?
4 is it audible?

Rules of Interpreting Data:
1 Is it believable?
2 Is it repeatable?

I think objectively identifying via measurement if possible what you don’t like with your current system is a must to minimize the probability of ending up in the same position.

Barry.
 
3wayaddict

You've apparently already tried to design your own speaker and don't like the result. You're asking lots of questions for which there are no simple answers. And yet you expect people to somehow magically give you the information needed to successfully replicate some of the very best and most expensive high end speakers in the world.

Since you have already invested in some very good SBA Satori drivers, there is a much more practical approach. Look at several of Troels SBA 3-way designs. Specifically the SBA 741, 761, 941, or 10. It will take you years of building and experimenting to even come close to the performance of one of those speakers on your own.
 
Ditto on kits if you want decent results in the near term. Good for learning on as well.

1. Build it.

2. Measure outdoors from all conceivable angles. Preferably with the mic on a stand so you can see the changes in real time on screen. Move the stand with one arm while looking at the screen looking for general overall patterns. Try ARTA measurement software.

3. Once you have a good idea of what you're working with, move indoors and do the same thing. Look for changes in response. The room will most likely make a mess of the measurements with increasing distance from the speaker. Worse response changes will be below 1khz or so.

4. Learn what can be corrected and what cannot.
 
Last edited:
I don't use my own speakers just because I can make one. I don't use my own amp just because I can make one... I built other people design as a benchmark/reference and tried (for years) to better the reference using the same set of drivers. Or, the drivers don't have to be the same. If you have a reference speakers design for $1000, and your DIY costs more, you will have to make speakers that is better than that reference. This will create big frustration, especially if you think that Beryllium is the absolute best 😀
 
Buy a DEQX unit (you stated you had plenty of money) and using the measuring power of it you may fall in love with the speakers you already own plus you can get expert help both online and have a home visit . You can of course use it in many projects and properly used it can produce phenomenal results
 
Though than I have to make a passive crossover. And I don't know a single thing about how one does that, let alone properly. I can't even set a digital crossover to sound good. So tel me. What are the steps one should take one by one when designing a crossover. Again, I want them to have a character like that of Wilson Audio and Dali speakers. What kind of frequency response should I be aiming for and all? And what is the best way to measure my speakers when te most anechoic space I have is outside in the open air?

And can anyone tell me how to properly measure with REW and the MiniDSP UMIK-1. Because it's not working properly for me. It's extremely inconsistent and partly doesn't work at all. Everytime I restart REW and plug in the mic I get a different measurment of the same speaker, it doesn't work under about 300 Hz, no matter what changes I make to the frequency response of the woofer, the measurments don't show a difference and in the top register the measurments only change very little when changes are made.
The measurments also differ from laptop to laptop. I get different results on my Macbook than on my HP. So basically, I can't make measurments. Anyone know how to fix this?

And can you help and tell me exactly what I have to do when I get to designing the crossovers? (will be a few months). I'll be using mundorf foil coils because they are the best though for capacitors I don't know. What do you think are the best crossover capacitors? The higher-end Mundorf, Jantzen or Audyn? (The Audyn True Copper Max are too expensive though, even for my budget)


It is the crossover which makes the loudspeaker. If it is not properly set, changing drivers won't help. When choosing drivers, their properties shall be choosen so that they blend each other properly. Crossover design can help in that area.

Directivity (dispersion) control through the whole frequency range is one difficult design issue, which affects a lot how the loudspeaker will sound.

To design a crossover, this sw is very good and professional: Software
There you can find a lot of links to helpful articles and guides. Some learning curve can be expected though.

For consistent measurements anechoing space would be best, but is usually unobtainable.
In-room measurements are a pain, I can tell. Gated measurements can help up to some degree, but still when I change mic position by 10cm, the result is different. Some people are using averaged measurements, making 5..9 measurements while changing mic position 20..50cm between every measurement. Acc. to my experience, this gives most consistent results I can achieve in-room.
For frequencies below ~300Hz the room starts to affec a lot. Below ~100Hz response is mostly defined by the room and speaker placement. See Geddes on multiple subs.

And last, designing and buildind a speaker is largely question of compromises, what to compromise and how much to compromise. Avoiding all compromises...just not possible.
 
About the passive crossover, as you already have digital available, I would stongly suggest to adjust all parameters with the digital one. Then you can transfer the design to passive, without need to purchase all components three times with different values... due to obvious need to test different filter parameters. After 2 years of build, I'm still adjusting the XO occasionally.

And the high-end high-price components are not needed, normal good quality parts will do exactly as well. But this is engineers opinion, gold-ear audiophiles may have different opinion.
 
3wayaddict,


perhaps it would make sense to invest in a solid quality piece of a measurement gear? I think the measurement consistency would be there by default.

It does seem Troels had a good reason to buy a factory made measurement rig, at least to make a better impression on future customers if not for practical reasons also.
 
I think you're being a little humble. The selection of drivers is good. The concept of a dsp based active amplfication is good. You also wrote that your loudspeakers sound better than many expensive off-the-shelf loudspeakers. Where I see potential for improvement is therefore not the driver selection or the concept. I see optimization possibilities in the cabinet. Especially the avoidance of unwanted resonances. In addition, it can make sense to invest in room acoustics. In addition, you can possibly optimize the crossover: other filter types, slopes, etc..
 
download xsim

get audiomatica clio pocket for measurements (first make sure you can import clio files to xsim)

then get a pair of seas 18cm coaxials

then play with measurements and simulations and build and test loads of crossovers for the seas
 
Since you have identified the digital crossover and amps (Hypex) as a weak point why not try analogue active (op-amps). You could get a cheap variable PA style crossover to experiment with. One where you can vary the crossover frequency and output levels as much as you want (the slopes are usually fixed though). Then when you've decided on the best frequencies you can build an active xover, there are plenty of pcbs available on ebay to make it easier then you can experiment with amps. You would still have control over the volume levels of the individual drive units. You could have different amps for different purposes e.g. a small class A for the treble and maybe A/B or class D for the mids and lows. Just a thought!
 
I've taken on the room acoustics for a bit. I put a absorption panel in-between the speakers which really enlarged the soundstage and got rid of quite some excess treble. I've also but small absorption blocks in every corner of the room which really tightened up the bass but didn't weaken it.

I already knew perfecting a crossover to get the best out of it, especially with what these drivers are capable of, takes loads of time. I don't have that time right now. So I think I'll save this hobby when I've got some more time, knowledge, space and perhaps even money to work with. And I'll be better of buying brand speakers right now, and I'm not at all sad about that. My main interest are the DALI Rubicon 5 or two (depending on what's better for my room. I know bookshelf is obviously more logical but if the floorstander is manageable I'd like the dedicated midrange and hybrid tweeter of the floorstander.) Followed by the Dynaudio Special Forty, Monitor Audio Good 200 or 100, Sonus Faber Sonetto III or II and Focal Aria 926 or 906. I heard the DALI Epicon 6 last week and the sound amazing. Of course those are a level higher than the Rubicon but I REALLY liked the DALI-character. The other brands I haven't really heard before but I'm gonna try those as well.

I believe in terms of amplifier I'd like something fuller and meatier so I've got my eyes on the Schiit Aegir.
 
If you recorded the whole design process in the exact way you did it, someone could probably point out the mistakes that led to a suboptimal performance. Normally these should sound like any high quality factory made commercial speakers.