New coaxial project

I may have bought two of somethings...



360 eur for both (+25shipping) on ebay.

I'm still debating with myself whether to make a ls50 clone or super tiny sealed box 100hz+, and a sub. We'll see.

 

Oh, no, please no. Unless you like that sound.

 

It's tricky to flush mount these but as coax drivers go, they're very good. Max spl is limited by the tweeter.

My move would be to do a 3-way with a proper bass driver crossed over at 200Hz or higher.

 

Yeah. Saw the thd plot. 200hz it is. Apparently they are really low in distortion up to something like 96db so that low sensitivity shouldn't be that much of a problem. And if I'll use them it'll be nearfield probably.

Why? What's wrong with them?
My main drivers are a full range dipole and a full range IB with complete full range 2pi radiation. This is just for fun.

I heard ls50 in an audio store in the most improper setting possible - with literally everything playing something, so it's basically like I didn't hear them (or any coaxial ever in fact).
I know it's tough to make good coaxials, these kef's are apparently solid, genelec too (but that stuff is super expensive, especially for toying around).
Many people love ls50, some really don't...

That meta absorber is a cool bit of physics and engineering for sure, but I kinda have a feeling it's a solution to a problem that didn't exist.

 

I owned a couple of KEFs with similar coax drivers and felt they were quite bright and not musical as some would say, but like I stated, if you like that type of presentation no problema.

I am more of a old school Tannoy or a bit less bright sounding coax sound. And some coax drivers can cut you like a knife if they are not positioned to alleviate it.

 

Dang, looks like the french site went down. They used to be much cheaper, even the Meta drivers:
https://web.archive.org/web/20221126223419/https://setelec-shop.fr/148-haut-parleur-kef

The KEF coaxials are basically the only ones with a good tweeter frequency response. And the midrange response is quite good given the constraints, too. It can even do some excursion, but it's no woofer.
Making a crossover that's better than the original LS50 crossover should go swimmingly. Maybe I'm not as allergic to the metallic sound, but I think my LS50 Metas are fine even nearfield. I do lower the upper treble in a similar fashion as I did with the 32Sat, though:
https://www.superbestaudiofriends.org/index.php?threads/speaker-listening-position-fr-targets.9821/
(I should update that thread with my current speaker targets as they've changed to something that's less forward in the lower treble)

Even so I gotta admit that the treble timbre of even the Alu sandwich widebanders is less zingy. Is it the 15dB 40kHz peak?
The SB ceramic also seemed smoother,
I think with better XO components it should be fine, however.

Where do you see that? Tweeter distortion seemed quite spectacular to me: 85dB distortion

 

I was thinking about higher SPL's over 100dB. But it's nothing new - most hifi domes have the same limitations. Maybe Morel can go louder due to heroic power handling (at least on paper).

But you are right that KEF coax drivers are some of the best you can get for DIY. At least until you get into large format PA drivers with compression horns.

 

Maybe sustained levels could be a problem, but I still sort of doubt the tweeter will give up before the midrange does.

Then again I'm just guessing with no data to back that up. And I haven't burnt a tweeter voice coil myself, but I know that it happens (though I wonder how much of that is due to ultrasonic noise).

The small tweeters in coaxials have less power handling than standalone tweeters, no?

 

I've replaced a dead LS50 tweeter and there's very little in terms of cooling.

 

Those drivers crinkle up and die. Any excursion makes an audible crunch. Even Genelec greatly limits high frequencies in their metal coax assemblies that sound a bit nicer than KEFs. It's a myth that metal domes can take more level than soft domes ime. The soft domes are physically more likely to survive events that kill the typical metal dome.

Horns and regular non coax soft domes kill the KEFs for power handling and sound: JBL, Radian, Faital, the pre Behringer Tannoys with the super tweeters, almost any monkey coffin two way tbh.

 

The morel domes can die. See all the cooked tweeters on used Questeds that scammers are trying to offload.

 

It's not worth the effort. Buy a cheaper speaker that won't die from over-excursion and has some low end or save money over time and buy the real ls50 meta and deal with the sound issues from the driver

 

Yeah, but studio work is a different ballgame. There are guys who kill compression drivers on mains. You know - the stuff that can do 135dB...

 

My man, I don't give a shit if it explodes on arrival. The effort is what I want it for, but thanks for the advice anyway. Trust me when I say it, I'm not so sure I can find anything in a store, that's better for me than what I already have. For any amount of cash. Speaker-wise.

Also, I'm not sure that meta add-on makes any difference. I'm sure kef would say it does..

 

I prefer the Sica 6.5 C 1.5 CP coaxial drivers over the KEF drivers. While the Sica speakers may not be as smooth or measure as well, they excel in power handling and offer better bass response. Additionally, they feature a substantial softdome tweeter. Their sound is also less bright compared to the KEF LS50s, making them preferable to me even in nearfield listening situations. Nonetheless, I'm interested in seeing what capabilities you can extract from those KEF drivers.

 

Man, we're all talking about entirely different things.

@Ardacer wants something that measures impeccably. In this sense the LS50 is probably the single best choice for coaxials. The Genelec coaxials have third order tweeter distortion, 8351 and 8361 have a less smooth tweeter FR (than the Meta) and less uniform dispersion. And the drivers aren't available for DIY, anyway.
However the normal LS50 does seem to have a less regular tweeter response. But with the Meta drivers not available anymore it does seem a logical choice for the project.*

Though I don't understand why you wouldn't do a cardioid three way, Much more even dispersion. It could work in a small enclosure.
LS50 Meta widens considerably for frequencies lower than 1kHz. You could easily do something that measures like Hathor for low frequencies and like the LS50 for high frequencies. In fact that's what I had planned with a full range cardioid with some large woofers and an eclosure like what Geithain does.


@Psalmanazar and @Hrodulf want incredible dynamics at any SPL, as far as I can tell. Psalm seems to want monitors he can mix on that won't fail him. He has tons of experience with all the brands and knows what he wants.
Hrodulf is a madman as far as I can tell and I mean that in the best possible sense. I think he runs some custom horns with a compression driver and high end midwoofers in a pseduo coaxial configuration, if things haven't changed.


I'm a subjectivist first and foremost. I want all the resolution I can get with laser precise imaging. And the most resolving systems I've heard have been based on widebanders so far. In my post about the speaker in room targets the three speakers are in order of how resolving they are to me. So Amon > Hathor > Canton Reference 3K. Though IMO the difference between Amon and Hathor isn't as significant.

In terms of measurements max SPL and AM from IMD don't interest me as much as excess phase and dispersion. I do care a lot about low THD sub 1kHz and don't care as much about THD products past 4kHz. While it seems to be the other way around for most hifi companies.

*FWIW the LS50 always sounded too sharp and uncontrolled for me. The Meta doesn't. Again, not as realistic a timbre as the metal widebanders I use, but really not bad.

 

Hey, we all started somewhere...





That's a 60-measurement smart average at the listening spot. Feels like yesterday even if it was 10 years ago.

Pretty sure they won't. 2xSB34SWNRX-S75-6+4x15MU+1xHF1440/ch should be enough for my tastes.

 

Actually, I am considering making a 3-way with a sb acoustics 8 inch sub drivers that I used for my full range dipole build. These things are awesome, I can only imagine that Hrodulf's SB34SWNRX-S75 works like a charm.

It's a good idea, but mainly for better bass response, as far as I can tell. It's also probably possible to fix the dispersion, but not all the way down to the rock bottom. But that's ok, as I already have multille speakers that have no problems doing just that. This project was originally my idea of building a "normal" and small speaker for a change. To make a proper cardioid (full range).. it's possible, but it would need at least 2 big woofers per unit, ideally 3. (Or one, and a really big horn). I'm not sure I'm willing to throw that amount of cash on a side thing.

 

I've done some real quick unibox back-of-the-envelope modeling. It's an older and less powerful excel spreadsheet, but is quick and I find it useful. Ignore the focal 10K, I just overwrote the KEF ls50 parameters over it.



This is the result for 42Hz port tune, with 5,5cm diameter and 18cm length which is kinda similar to what you see on the photos.



This is at 10 watts. And...



This is at 100W (that ls50 is supposed to be able to handle). 50w is not much better.
Could the lack of low-end protection be the cause of the death of so many ls50/q150/q100 woofers? Lots of pics on the internet of broken drivers - usually ripped cones of woofers.


https://www.avsforum.com/attachments/speakers-jpg.3057696/

They may know how to make a proper coaxial driver, but this would hardly be the "speaker for every man" or whatever they said in promos in that case.

 

Not all of the 100W will go to the midrange.

For example my little widebander is rated at 30W, yet I've pumped 100W for a bit into the speaker without it even sounding close to breaking. You have to consider the spectrum of the music and crest factor, too.

But this seems more about hitting Xdamage than actual power limits. Maybe the magnet gap height isn't too far from Xmax.