Design or use someone else's? This is what I'm looking for.

[icor1031]

If I remember correctly from when I hung out here a decade ago, there were many gentlemen who shared their speaker designs for free: Box dimensions, drivers, and crossovers.

I want your opinion on whether I should use someone else's design, or do one of my own -- to get the results I want, that is; not simply for experience/practice/pride. They will be used for front-left and front-right only. I will not be using a center channel.

• Three-way that will cross well at 60Hz with my subs.
• Very flat response at least to 7k; I'll be using MiniDSP to adjust as needed.
• Uses drivers that perform best at less than 80dB. I never play anything loud.
• From my limited experience and from what I've read, I prefer paper/silk/nomex and similar less-harsh materials. Correct me if my conception is inaccurate.
• Sealed design

Soon I'll be going into a very good career, so cost isn't much of a concern. $300 per driver and possibly even as much as $1,000 is okay.

Suggestions? Any specific builds come to mind?

[ugly woofer]

If you're going to use some one else's design, it's more important to like their voicing than the drivers utilized. Some guys are a bit inconsistent with their voicing and others definitely have a house sound. It's no different than some folks like B&W speakers and some prefer Martin Logan ect.

[icor1031]

If you're going to use some one else's design, it's more important to like their voicing than the drivers utilized. Some guys are a bit inconsistent with their voicing and others definitely have a house sound. It's no different than some folks like B&W speakers and some prefer Martin Logan ect.

Ah... And I suppose the way to figure that out is to attend a meetup?

[djg]

It's a good place to actually hear the diy designs.

[djg]

Dennis Murphy's Philharmonitor, Javad's Enthraals, the late Rick Craig's Integrity, Campbell/Holtz travelers. All upmarket 3 ways with component kits available elsewhere. They meet some of your criteria, maybe not expensive enough.

The high efficiency low watt area is well represented at DIYaudio forum, you might try there. You'll get lots of advice for sure.

AVSforum is into high efficiency drivers, but 1 watt not so much.

[icor1031]

Dennis Murphy's Philharmonitor, Javad's Enthraals, the late Rick Craig's Integrity, Campbell/Holtz travelers. All upmarket 3 ways with component kits available elsewhere. They meet some of your criteria, maybe not expensive enough.

The high efficiency low watt area is well represented at DIYaudio forum, you might try there. You'll get lots of advice for sure.

AVSforum is into high efficiency drivers, but 1 watt not so much.

My description might have been misleading. I used 1W as a proxy for saying that I won't play over ~85dB/1m. And usually I'm at like 50-60dB. I did not mean that they had to be relatively high SPL @ 1W.

For that, is DIYaudio the place to ask?

[djg]

Well, you might just want to graze over there a bit. You can't see pictures unless you are logged in. But I think you'll get a sense of their zeitgeist.

[JRT]

I suspect that notion of nominal playback levels which do not seem loud to you might not be considering the demands of higher crest factor in higher quality highly dynamic recordings. Highly compressed material may not have such difficult requirements, even when played seemingly loud.

I would suggest against sizing the SPL capabilities based on nominal playback levels, but rather to accomodate the worst case maximum crests in the program materials, however brief, unclipped in both the electronics and in the electro-mechanicals.

Recordings with less dynamic compression can have higher crest factor, may sound less loud on average while exhibiting higher crests in the signal. Because those are less loud, the listener will tend to turn up the playback level to hear the recording, and near 0_dB crests are raised similarly.

The monitoring reference level utilized when mastering varies among recordings. Highly compressed loud recordings have lower maximum crests.

The playback level that will sound more correct relative to what was intended would be the same level as was used in monitoring the mastering effort, but that level is rarely if ever disclosed by the publisher. So you are left to guess on playback. However in establishing system design requirements, you can look at worst case (which is arguably the playback of the best engineered highly dynamic recordings).

For high quality recordings, Ioan Allen of Dolby Labs and Bob Katz (author of Mastering Audio, the Art and Science) both have recommended mastering with monitoring reference level of 86_dBC at -20_dBFS signal level summed uncorrelated at the listening position, which means 103_dBC at the listening position at 0_dBFS from each of a pair measured separately. Nominal playback listening levels can be much lower, depending much on what is appropriate for the program material, but a high quality playback system should accomodate 0_dBFS crests, however brief, unclipped.

Near the top of the recorded spectrum, crests can exceed 0_dBFS by as much as +3_dB, at peaks between samples at frequencies near half of the sample rate (theoretical max is +4_dB, but that not achievable with practicable anti-aliasing filters which keep aliased content below the noise floor). Averaged power requirements are lower at high frequencies in music, but peak signal voltage can be highest there, though very briefly.

[djg]

This guy has a ton of expensive designs, some (older) designs are free, the newer stuff requires kit purchase for XO specs.

http://www.troelsgravesen.dk/Diy_Loudspeaker_Projects.htm

[Wolf]

I honestly think that my Purveyor qualify here. Sensitivity is 82dB, and that would be 82dB at 2W being a 4 ohm design. Drivers qualify, response qualifies, and it is meant to be a fairly easy build.

https://techtalk.parts-express.com/forum/tech-talk-forum/1486130-purveyor-a-dynamite-small-3-way-with-concussive-tendencies

Wolf

[icor1031]

I'll check them out, thanks Wolf!

[icor1031]

This guy has a ton of expensive designs, some (older) designs are free, the newer stuff requires kit purchase for XO specs.

http://www.troelsgravesen.dk/Diy_Lou...r_Projects.htm

Oh, I'm cool with paying a bit for an already-specced design. I just didn't realize people offer that.

It seems though that he has few, if any, sealed designs?

[djg]

What I meant was: people were cloning his diy designs and selling them commercially, so he stopped publishing the XO parts values to stop intellectual property theft.

[a4eaudio]

I want your opinion on whether I should use someone else's design, or do one of my own -- to get the results I want, that is; not simply for experience/practice/pride. They will be used for front-left and front-right only. I will not be using a center channel.

Doing your own design is pretty intense. especially if you don't have measurement gear to begin with. I would suggest if you want to make a hobby out of it and continue to design your own, you might as well start now. If this is going to be the only pair you plan to design, definitely go with an existing design.

• Three-way that will cross well at 60Hz with my subs.
• Very flat response at least to 7k; I'll be using MiniDSP to adjust as needed.
• Uses drivers that perform best at less than 80dB. I never play anything loud.
• From my limited experience and from what I've read, I prefer paper/silk/nomex and similar less-harsh materials. Correct me if my conception is inaccurate.
• Sealed design

Soon I'll be going into a very good career, so cost isn't much of a concern. $300 per driver and possibly even as much as $1,000 is okay.

Your needs aren't very demanding, so you probably don't need anything TOO expensive. BUT, most DIY speakers are vented/ported and sealed that can go down to 60Hz aren't that common. Many ported designs can be made sealed (for example the Statements) but may not reach down to 60Hz. (It looks like the RS225's used in the Statements have an F3 of about 63Hz sealed. Maybe if you can crossover to the sub at 70Hz or 80Hz, that would open up some additional possibilities.)

Are there any constraints or preferences to narrow down the possibilities? Size constraints? Do you like a tall slim design or something with big 10" or 12" woofers? Desire to try ribbon/AMT/planar tweeters?

Ah... And I suppose the way to figure that out is to attend a meetup?

Pretty much. But if you aren't too picky on what your own preference is for voicing you can read reviews on many of the popular existing designs.

[icor1031]

Doing your own design is pretty intense. especially if you don't have measurement gear to begin with. I would suggest if you want to make a hobby out of it and continue to design your own, you might as well start now. If this is going to be the only pair you plan to design, definitely go with an existing design.



Your needs aren't very demanding, so you probably don't need anything TOO expensive. BUT, most DIY speakers are vented/ported and sealed that can go down to 60Hz aren't that common. Many ported designs can be made sealed (for example the Statements) but may not reach down to 60Hz. (It looks like the RS225's used in the Statements have an F3 of about 63Hz sealed. Maybe if you can crossover to the sub at 70Hz or 80Hz, that would open up some additional possibilities.)

Are there any constraints or preferences to narrow down the possibilities? Size constraints? Do you like a tall slim design or something with big 10" or 12" woofers? Desire to try ribbon/AMT/planar tweeters?




Pretty much. But if you aren't too picky on what your own preference is for voicing you can read reviews on many of the popular existing designs.

I do have a rig for testing impedance and I have a good mic for taking measurements. I also have many acoustic panels (roxul rockwool), but it's a hassle to design my own speakers. I prefer not to. And no, I don't want to make it a hobby and build multiple sets.

I could cross at 70 as long as my subs won't localize. I have two 18" in separate boxes, and I believe they're HO.

The design does *not* need to be compact, so bigger woofers are fine in that regard. As far as ribbon/etc. I'm okay with that if it means higher quality, but for example I have my treble at 42% and I can't remember but I might have a roll off in the XO too, so I probably wouldn't benefit?

[tktran]

what speakers are you used to? What do you have now? What’s your favourite type of music, please describe the room you gonna put it in? Do you have a Music, acoustics, or engineering background?

I concur with other people‘s thoughts you should build an established design, don’t design your own, unless you want wanna play around. If you want something that sounds good then build an established design. My recommendation would to build one of Wolf’s designs, Curt’s designs, or Dennis’s designs
there are probably more but what we designers know if that it takes good science to figure out the stuff that’s known, and good ears to figure out what we are yet to know,

Most people who have designed their own have had a decade or more of experience building speakers.

[icor1031]

The ones I built a decade ago with help here... I think it was paper woofer, silk tweeter. 5.25" and 1.125" as I recall. Peerless? Room would be living room.

[fpitas]

Most people who have designed their own have had a decade or more of experience building speakers.

Even then it's a labor of love. It's a great hobby, but I think designing your own is for those who want the challenge. If you just want great speakers there are farrrr easier ways.

[DeZZar]

They will be used for front-left and front-right only. I will not be using a center channel.

So I assume from your comments that this is really a music focussed system with theatre perhaps a secondary duty?

It seems though that he [Troels] has few, if any, sealed designs?

And for good reason. Is there any specific reason you are limiting your choices to sealed?

If money isn't an issue it would be difficult to do better than one of the designs from Troels.

Troels SBA-741 for instance (http://www.troelsgravesen.dk/SBA-741.htm) is designed with an optional bass module - but can and is intended to be built on its own. It will 'only' perform down into the mid/high 40's in this format and is a good candidate for pairing up with an existing sub.

I work with same drivers a lot and can simply say that you couldn't go wrong with this design.

You could of course up the price point another notch and go for the equivalent scan speak version (http://www.troelsgravesen.dk/Ellipticor-3.htm) - this is end game level hifi for the greater majority of people.

Troels SBA-941 (http://www.troelsgravesen.dk/SBA-941.htm) is another option. I'm currently working on a design for a customer that uses the same drivers in a similar configuration (sort of a nod to the box speaker/ATC classic style) and the sound is impressive. It dips down into the low 30's and arguably doesn't need a sub-woofer but it is an option depending on what sort of style you are going for.

Just note that you don't need a speaker that is limited to playing say 60hz and above in order to use it in conjunction with a subwoofer. Any speaker can be used in conjunction with a subwoofer (or many subwoofers for that matter). A speaker that plays full range gives you the option to also listen without a sub.

[donradick]

If you like to listen at lower volumes, a little known fact is that a high Qms mid or mid-woofer will reveal more micro details. I have a lot of experience with Dayton RSS woofers and some with SB acoustics woofers, but I was shocked when I built a Criton 1TD pair for a friend (from css audio). The amount of detail and clarity blew me away! It's because the LDW7 woofers have high Qms values. I prefer 3 ways, but I didn't feel that I was missing anything in the mids with this design. Sorry, I can't drop a link here.

[ugly woofer]

Icor, would you walk into a shop and pick out a speaker without listening to it? That's the equivalent of building a kit that you have never heard, I don't care who designed it. Just because it's a competent design or others recommend it, doesn't mean that you will like it in the long run. So yes, go to a meet, Iowa would probably be the closest, and have a real listen