Reasonably priced good quality 6 or 8 channel USB DAC?

[eugenius]

There's no practical problem with the dispersion of an 7-8" midrange with a crossover point as low as 1500hz. And dispersion is important but sacrificing low mid dynamics for it is not a good tradeoff.

Just adding mass isn't enough, in fact the more mass you have the harder it is to damp the vibrations of that mass. Midrange enclosures don't need heroic mass damping, they need broadband attenuation and those steel balls inside the Hawaphon are quite good for that. Also, ideally you would absorb the back-wave by using a tapered midrange enclosure filled with progressively more dense material. You should read what Atohm has to say about their GT3 flagship speaker. That speaker's construction is a source of cheap ways to do good things.
http://www.atohm.com/fileadmin//user_upload/TECHNICAL_BOOK_GT_SERIES_2015.pdf

For nearfield and having that Bliesma tweeter, personally I'd stick to a closed box 2 way. I like the ATC SL midwoofer so I'd buy an used pair of SCM19 mk2 and try and improve on the ATC SCM19 first by activating it, then by replacing the tweeter and dropping the crossover point, then by trying to improve the enclosure damping, then trying to build a better enclosure - and see what each new step really gets me - both objectively and subjectively. Learning by experimentation is the goal.

PS: There's a tendency in DIY to just throw excess at the problem and hope it goes away. That's not engineering.

 

[JustIntonation]

There's no practical problem with the dispersion of an 7-8" midrange with a crossover point as low as 1500hz. And dispersion is important but sacrificing low mid dynamics for it is not a good tradeoff. And the added physical size of an 8"narrows the lobing free angle also for 1500Hz crossover.

Just adding mass isn't enough, in fact the more mass you have the harder it is to damp the vibrations of that mass. Midrange enclosures don't need heroic mass damping, they need broadband attenuation and those steel balls inside the Hawaphon are quite good for that. Also, ideally you would absorb the back-wave by using a tapered midrange enclosure filled with progressively more dense material. You should read what Atohm has to say about their GT3 flagship speaker. That speaker's construction is a source of cheap ways to do good things.
http://www.atohm.com/fileadmin//user_upload/TECHNICAL_BOOK_GT_SERIES_2015.pdf

For nearfield and having that Bliesma tweeter, personally I'd stick to a closed box 2 way. I like the ATC SL midwoofer so I'd buy an used pair of SCM19 mk2 and try and improve on the ATC SCM19 first by activating it, then by replacing the tweeter and dropping the crossover point, then by trying to improve the enclosure damping, then trying to build a better enclosure - and see what each new step really gets me - both objectively and subjectively. Learning by experimentation is the goal.

PS: There's a tendency in DIY to just throw excess at the problem and hope it goes away. That's not engineering.

Well problem is a big word, but there is dropoff there for most 8" woofers. Though the 7" Satori you linked to doesn't have it much. But any extra dB here is a somewhat worse performance off-axis?
But I don't think low mid dynamics are sacrificed in my case. It is for a nearfield monitor, 1 to 1.5 m listening distance I'm not going to go over 100dB 1m ever (perhaps for fun on some bass heavy music but not in the mids). 70 to 90dB RMS is about where I'll spend most of my time there is no compromise anywhere in sight it seems to me regarding low mids? I mean I could take the 6" to 50Hz or so with a port and still be fairly ok as most studio speakers do (though compromised). But for a 120-150Hz crossover, I've looked at the HD plots in that range for the SB17NAC35-4 and taking into acount the -6dB of the crossover it'll do just fine without compromise for a nearfield?

As for the Atohm GT3, that cabinet doesn't look well constructed at all! It'll surely resonate.
Look at actual measurements here (and I've seen others which show the same): http://www.hsi-luidsprekers.nl/index.php/projecten/12-kast-materialen-onderzoek
See the difference for instance between no 6 (18mm MDF) and no 23 (18mm MDF + 11mm massplate). Lead bitumen should do a little bit better damping than the massplate. What it does is dampen the upper resonances and shift the lower resonance lower. But coupled with good bracing (makes things stiffer and brings the resonant frequency up) and 2 different wood plates with different resonant frequencies the result is very dead.
The GT3 cabinet is not dead, only perhaps well damped on the inside but the cabinet itself is very thin it seems. Standing waves are only part of the problem and taken care of enough by good absorption (like fibreglass). The resonances of the wood itself are another thing which are not taken care of in the GT3 it seems.

As for closed box 2-way. I've heard a very bad cheap sub in my room before treatment (JBL LSR310) and it showed me my room can do deep bass well. I think its worth it to make a 3-way (more like 2-way + high 120Hz crossover subs) The bass woofers will be in a different cabinet separated by half a meter from the mid-tweeter speakers. I think it will be worth it. I'll be using the new SEAS L26ROY XM001-04 woofers (should be better than the D1001).
As for the ATC. Very good mid driver (edit: I mean the dome SM75, not the woofers) and very good cabinet and very good amps. But not so great tweeters and not so great bass woofers (measure quite a bit more HD than for instance the Seas XM001). Also the crossover points are right where they are most sensitive. In the mids and in the treble. And you can't really cross those tweeters much lower as you'll have strong rising HD and run out of xmax. Also no DSP crossover, and they have a small sweetspot for dispersion. Also don't like their flat baffle design.
Just because it was great for so many years and so expensive doesn't mean its the best.
Take for instance the Grimm LS1. Many think its better than ATC speakers. But it's a W22EX001 + Seas DX tweeter in a simple one layer birch plywood cabinet. And a good DSP crossover. But surely I'm going to build something better here?

Btw, perhaps there's a bit of excess to be on the safe side. But as far as costs, a second hand ATC is still way worse for cost/performance than what I'm building here it seems to me

 

[JustIntonation]

Argh..
You made me doubt again.. haha

Ok the other idea I had which I didn't write about before was this:
ScanSpeak 26w-4867t00 10" aluminium woofer + BlieSma tweeter for a two way.
You can see Hobby Hifi test results of the woofer here:
http://www.audiocomponents.nl/download/scan-speak/26W-4867T00 - Hobby HiFi 2011-2.pdf

It is not complete tight in CSD/amplitude but I think this can most likely be fixed with EQ and since I'm doing DSP (on the computer) I can EQ with great precision without quality loss contrary to a passive crossover. Might as well make full use of this.. (and I'll be able to measure it will the greatest precision as my room will be anechoic in this freq range)
If I were to do it with a 1300Hz crossover to well can just as well pick the T34A-4 tweeter. Then total costs will be considerably lower than my 3-way design. It means 2 less drivers, 2 less amps, 2 less cabinets to construct for the separate subs that are no longer needed and there are many good enough 4 channel DACs available don't need a 6-8 channel DAC anymore then.
The result will not be as good as the 3-way off-axis. But since I'm in an anechoic chamber as far as the mids and highs are concerned this only means a less wide sweetspot for me personally though the resulting sweetspot will still be by far wide enough.
There would be slightly more distortion products falling in the sensitive 2-5kHz range compared to the 3-way, though it would still be very low. And the distortion in the midrange would be lower and no crossover between the mids and bass.
Would probably want to do a 36dB/oct crossover at 1300Hz though, this seems to be getting close to audible for an IR crossover maybe a FIR crossover would indeed be better there..
Another benefit would be easier placement of a single speaker vs two cabinets per speaker one for the mid-tweeter and one for the woofer as in my 3 way design..

But the main reason I chose not to go this path was that I want the speaker to disappear. With the 3 way I can build a small cabinet for the mid-tweeter and put the big woofer cabinet away from this.
What makes a speaker disappear and sound like a point source? It is of course in the first place the time alligned crossover between tweeter and mid at listening position. This will be good for both designs. But one other thing is probably the diffraction of the edges of the speaker enclosure.. these behave as additional point sources. There isn't a lot of info on this on the web that I could find but it makes sense to me. I'll make the speaker enclosures very much rounded at the edges, but will this be enough for the large enclosure of a 10" driver? I don't know. It was this uncertainty of how well a large speaker wil work for nearfield monitoring that made me afraid to go this route when I know a 6" will work well.

So but now I'm doubting again.. A 10" 2-way is really cool in a way
Anybody has any insights into if this will work for a nearfield monitor in the way I described?

 

[andreasmaaan]

@JustIntonation I think your original idea was very good, and although the 10" woofer might produce lower THD when reproducing a sinewave in certain frequency ranges, you also have to consider the IMD products that will be produced by the driver producing multiple bass and midrange frequencies simultaneously with music.

Keep in mind, however, I haven't looked at all these specific drivers. In general though I think if you're going all-out on this system, it makes sense to separate the low-bass driver from the mid-range driver.

And although your near-anechoic environment will reduce the influence of off-axis reflections, you might want to consider potential future uses of the speaker?

OTOH, I think in your original design you may have been overly concerned about midwoofer cone break-up. Generally speaking, if you're using 4th order or higher crossovers, there's no need to keep the cone break-up that far above the crossover point. In your original design I think you were aiming for 2-3 octaves above the crossover point? That's going to be 50-80dB down from the signal, and absolutely inaudible. Assuming 4th order filters, I think a one octave distance from xover point to cone breakup (pushing it 30dB below the signal) is the minimum necessary. A slightly greater distance my be better, but at a certain point there is just not going to be any audible benefit IMHO.

 

[JustIntonation]

@JustIntonation I think your original idea was very good, and although the 10" woofer might produce lower THD when reproducing a sinewave in certain frequency ranges, you also have to consider the IMD products that will be produced by the driver producing multiple bass and midrange frequencies simultaneously with music.

Keep in mind, however, I haven't looked at all these specific drivers. In general though I think if you're going all-out on this system, it makes sense to separate the low-bass driver from the mid-range driver.

And although your near-anechoic environment will reduce the influence of off-axis reflections, you might want to consider potential future uses of the speaker?

OTOH, I think in your original design you may have been overly concerned about midwoofer cone break-up. Generally speaking, if you're using 4th order or higher crossovers, there's no need to keep the cone break-up that far above the crossover point. In your original design I think you were aiming for 2-3 octaves above the crossover point? That's going to be 50-80dB down from the signal, and absolutely inaudible. Assuming 4th order filters, I think a one octave distance from xover point to cone breakup (pushing it 30dB below the signal) is the minimum necessary. A slightly greater distance my be better, but at a certain point there is just not going to be any audible benefit IMHO.

Yep all good points!
And taking them into consideration. Trying to find the right balance..

About the cone breakup, I'm not concerned about the direct breakup itself but I am concerned about the harmonic distortion of the driver itself (and any other distortion occuring after the crossover filter like in the amp). If you look at for instance Seas drivers like the W18EX W22Ex W26FX etc and at their harmonic distortion measurements you'll see that the relatively low harmonic distortion of the driver start to have significant peaks when the harmonic distortion falls on the cone breakup peak. So for instance when the cone breakup has a huge peak at 4kHz then you'll see a huge increase in the 3rd harmonic distortion at 4kHz / 3 = 1.33 kHz and a huge increase in the fifth harmonic at 4kHz / 5 = 800 Hz. These can quickly go above 1% distortion at higher volumes. And coupled with the enhanced sensitivity to our ears of the 2 to 5 Khz range you can add another 6-9 dB or so to the audibility of this distortion. So realistically it can matter a lot. This is why I'm so careful about cone breakup of the mid driver.
But in the case of the ScanSpeak 26W/4867T100 the cone breakup peaks are not nearly as bad as most aluminium drivers and high enough up so that the 3rd and 5th harmonic distortion is generated higher than the 1.3kHz crossover (the really nasty peak is at 7kHz) and the harmonic distortion is so incredibly low there isn't much to be amplified by it falling on the cone breakup to start with. That's why this particular driver will still work very well in this regard I think unlike most other metal woofers.

 

[andreasmaaan]

About the cone breakup, I'm not concerned about the direct breakup itself but I am concerned about the harmonic distortion of the driver itself (and any other distortion occuring after the crossover filter like in the amp).

Ok that clarifies it, and yes it makes sense.

But in the case of the ScanSpeak 26W/4867T100 the cone breakup peaks are not nearly as bad as most aluminium drivers and high enough up so that the 3rd and 5th harmonic distortion is generated higher than the 1.3kHz crossover (the really nasty peak is at 7kHz) and the harmonic distortion is so incredibly low there isn't much to be amplified by it falling on the cone breakup to start with. That's why this particular driver will still work very well in this regard I think unlike most other metal woofers.

Also makes a lot of sense

I think given the level of fidelity you're aiming for, I'd still prefer not to have the same driver handling low bass and midrange for the reasons I mentioned above though.

 

[JustIntonation]

Ok that clarifies it, and yes it makes sense.



Also makes a lot of sense

I think given the level of fidelity you're aiming for, I'd still prefer not to have the same driver handling low bass and midrange for the reasons I mentioned above though.

Hmm.. IMD should be very low for the ScanSpeak. I've seen IMD measurements for various drivers tested by Mark K ( http://web.archive.org/web/20141015...uristics.org:80/measurements/measurements.htm ) and Linkwitz and Zaph. One can get a good guess at the IMD by looking at the HD.
But of course you're right, splitting bass mid between two drivers doesn't give IMD at all.

And indeed for a normal room the choice would be clearer. But for my room.. well the 3-way is playing it safer.
But the 3-way will cost me about 4000 euro to build. (that is with a finished Okto DAC8 and T34B-4 tweeter)
But cutting costs at all reasonable places the two way will cost me about 2500 (T34A-4 tweeter)..
That's quite a difference. Hence my continued doubt

 

[andreasmaaan]

Hmm.. IMD should be very low for the ScanSpeak. I've seen IMD measurements for various drivers tested by Mark K ( http://web.archive.org/web/20141015...uristics.org:80/measurements/measurements.htm ) and Linkwitz and Zaph. One can get a good guess at the IMD by looking at the HD.
But of course you're right, splitting bass mid between two drivers doesn't give IMD at all.

And indeed for a normal room the choice would be clearer. But for my room.. well the 3-way is playing it safer.
But the 3-way will cost me about 4000 euro to build. (that is with a finished Okto DAC8 and T34B-4 tweeter)
But cutting costs at all reasonable places the two way will cost me about 2500 (T34A-4 tweeter)..
That's quite a difference. Hence my continued doubt

Yeh, I'm not suggesting this IMD will be catastrophic, and maybe the cost savings will be worthwhile

But it's hard to conclude too much looking at the tests at that site. Firstly, the multitones tested are all clustered close to each other so it's difficult to gauge the effects of a low frequency tone on the midrange. Secondly, the 5-tone test is at a much higher SPL than the single-tone and 2-tone tests.

Have a look at these measurements from driver vault of two high-end PA woofers (not that these graphs are ideal for our purposes either - the "low" tone is 300Hz so still in the midrange). Firstly, the excellent 18Sound 12NDA520:

Here we see the worst IM products higher in level than the worst harmonics.

And then the also quite good AE speakers TD15X:

I didn't bother labelling this one but you can see that again, the IM products are higher in level than the harmonics.

Now of course you won't eliminate this problem by splitting the bass and mids between two separate drivers, but you will mitigate it somewhat.

Maybe not enough to justify almost doubling the costs though!

And you can always add subs later if you decide to build the 2-way now...

EDIT: on reflection, I think I may have mislabelled that 12NFA520 graph. Both of those higher distortion components above 1KHz may actually be IMD products, with HD2 of F2 actually about 15dB lower in frequency/level. Hard to say for sure but I think that's what's happening there...

 

[JustIntonation]

Yeh, I'm not suggesting this IMD will be catastrophic, and maybe the cost savings will be worthwhile

But it's hard to conclude too much looking at the tests at that site. Firstly, the multitones tested are all clustered close to each other so it's difficult to gauge the effects of a low frequency tone on the midrange. Secondly, the 5-tone test is at a much higher SPL than the single-tone and 2-tone tests.

Have a look at these measurements from driver vault of two high-end PA woofers (not that these graphs are ideal for our purposes either - the "low" tone is 300Hz so still in the midrange). Firstly, the excellent 18Sound 12NDA520:

View attachment 17992

Here we see the worst IM products higher in level than the worst harmonics.

And then the also quite good AE speakers TD15X:

View attachment 17993

I didn't bother labelling this one but you can see that again, the IM products are higher in level than the harmonics.

Now of course you won't eliminate this problem by splitting the bass and mids between two separate drivers, but you will mitigate it somewhat.

Maybe not enough to justify almost doubling the costs though!

And you can always add subs later if you decide to build the 2-way now...

EDIT: on reflection, I think I may have mislabelled that 12NFA520 graph. Both of those higher distortion components above 1KHz may actually be IMD products, with HD2 of F2 actually about 15dB lower in frequency/level. Hard to say for sure but I think that's what's happening there...

Hmm I do think you labeled it correctly first time though. The level matches the HD level for that SPL it seems.

But I just got the idea.. I can build the 2 way with the ScanSpeak 10" and if not great enough I can turn it into a 3 way later by plugging the tweeter hole. Not by adding a sub as the ScanSpeak 10" is a great sub already I don't really need more volume than it gives (it has +-9mm linear travel +-14mm max and it does it with the very lowest distortion), but by removing the tweeter from the enclosure and putting it in a new one with a SB17NAC35-4 Can simply keep the 26W/4867T00 in the old enclosure as a top quality subwoofer that can be crossed at any frequency of choice to the mid-tweeter box, then nothing is wasted.
Thanks for the suggestions both Andreas and Eugenius! I think this is the most reasonable way to do it. Complete package for lowest price, if not perfect enough then can change it later without having wasted money or effort.

 

[andreasmaaan]

Hmm I do think you labeled it correctly first time though. The level matches the HD level for that SPL it seems.

Possibly yeh That would make the IMD product higher in level than the harmonic for the other driver in that case. But yeh, I’m really not sure which is which...

I can take some measurements of some woofers with some more analogous frequencies (eg 50hz and 900hz) for you when I’m back in the workshop in January if you like? Or perhaps you could start with the scansoeak woofer and measure it yourself before deciding how to proceed with the rest.

But yeh, I don’t think you’d be making a terrible decision with the two-way. Just not what I’d choose myself TBH.

 

[JustIntonation]

Possibly yeh That would make the IMD product higher in level than the harmonic for the other driver in that case. But yeh, I’m really not sure which is which...

I can take some measurements of some woofers with some more analogous frequencies (eg 50hz and 900hz) for you when I’m back in the workshop in January if you like? Or perhaps you could start with the scansoeak woofer and measure it yourself before deciding how to proceed with the rest.

But yeh, I don’t think you’d be making a terrible decision with the two-way. Just not what I’d choose myself TBH.

Oh yes I'll be measuring the hell out of my system and post it here Will also likely be January seeing my slow pace.. Maybe we can come up with a good test for what is most audible in speaker design regarding harmonic distortion, IMD etc.
What I also like about the cheaper 2-way btw is that if it turns out as good as I hope then it seems fairly easy to turn it into a commercial speaker? I mean why not.. there isn't really anything like it on the market and Hypex is about to come out with a very good yet cheap 4 channel OEM DSP module which would make it useable for a wide range of people. If I were to sell it direct with a 1000 euro profit for 3500 a pair this could still be about the best deal for people. Don't know if I'd want to get into this though.. (edit: as a comparison the Grimm LS1 has an 8" Seas woofer and DXT tweeter and goes for about 18,000 euro a pair, the LS1Be for double that, and my two way drivers are better)

BTW will be ordering the BlieSMa tweeters tomorrow already, they're having a black friday sale till tomorrow. www.bliesma.de

 

[eugenius]

The Atohm GT3 is for me an intelligent 80/20 Pareto type solution which is great for a manufacturer who needs to make a profit. I pointed it out to you because of the internal ideas, not the profit-enabling box construction. Heroic Rockport Lyra type enclosures don’t sell in significant numbers.

Let's agree to disagree on the the ATC cone/dome hybrid super linear midwoofer. For me it's still the best small midwoofer for a closed box, well damped, exceptionally low harmonic distortion between 100-150Hz at a normal listening level, and the dome shape gives it good dispersion too. And it's also a monster of an under hung driver, with generous linear xmax to give good enough bass for nearfield or small rooms. But it's very inefficient so the tweeter is a bit wasted.

-----

First of all, buying Bliesma tweeters says that price to performance talk and saving money is not in the cards for you.

On the Scanspeak aluminium 10" two way - for me it's a fuck no! If you're going to use a 10" for a two way, it needs to be something very well damped - like the TD10M - I picture a two way like the Vapor Arcus with the Beyma AMT replaced with the cheaper Al-Mg Bliesma tweeter.

If you actually go through with the Alu 10" Scanspeak, despite steep active filters, be prepared to add something like this Eton driver down the line and turn it into a 3way. And if you do, that big tweeter is also kind of wasted - unless you become a masochist and you try for some kind of Jim Thiel approach with first order crossovers and metal drivers.
https://www.eton-gmbh.com/fileadmin/user_upload/Produkte/Home-Hifi/Symphony/3_400_A8_25MG.pdf

Legend Acoustics used this Eton mid to turn their 3way Tikandi DEQX speaker into a 4way with the 22 cm Seas Magnesium midwoofer (the same driver Grimm uses) and the Scanspeak Beryllium dome tweeter - presumably for good reasons. And it's not like they didn't have driver correction or steep linear phase filters available on the Deqx. I presume Grimm also uses some amount amount of driver correction to make the LS1 work - Putzeys says it's one of his areas of research. Maybe he knows how to do driver correction better. Maybe his Seas magnesium midwoofer is customized and better damped than stock.

If you want to go for the more expensive beryllium Bliesma and you intend to use DSP crossovers, I say just follow the "get what looks the very best" logic to it's end and go all in with hard material drivers like this new accuton mid and also a suitable dedicated midbass/multiple subs 4 way setup. End of game type stuff - if you can afford it.
https://accuton.com/en-home/produkte/lautsprecher/keramik/Keramik-C168-6-990-CELL

Just don't get ideas about turning a profit from it. If you want to make a profit selling high end speakers, making good speakers is not enough. You need to be a shrewd marketing bastard.


PS: What are the prices on the bliesma tweeters, after the 25% Black Friday promo? They didn't list them on the website.

 

[JustIntonation]

First of all, buying Bliesma tweeters says that price to performance talk and saving money is not in the cards for you.

Hehe ok maybe a point there.. but this is something I do not want to save on because of my room. You don't want to know what an anechoic room does to tweeters.. Anything less than totally transparent in I'm now guessing the 2-5kHz range is incredibly audible in the most harsh way you can imagine. (have experience listening in such a room for a long time but my knowledge on speaker design wasn't there then)

On the Scanspeak aluminium 10" two way - for me it's a fuck no! If you're going to use a 10" for a two way, it needs to be something very well damped - like the TD10M - I picture a two way like the Vapor Arcus with the Beyma AMT replaced with the cheaper Al-Mg Bliesma tweeter.

If you actually go through with the Alu 10" Scanspeak, despite steep active filters, be prepared to add something like this Eton driver down the line and turn it into a 3way. And if you do, that big tweeter is also kind of wasted - unless you become a masochist and you try for some kind of Jim Thiel approach with first order crossovers and metal drivers.
https://www.eton-gmbh.com/fileadmin/user_upload/Produkte/Home-Hifi/Symphony/3_400_A8_25MG.pdf

Can you explain why you think the Scanspeak 10" alu won't work?
I'm pretty sure I can make it ruler flat in frequency untill about 2.3kHz and this should give a cleaner CSD as well. Above 2.3kHz I can make it fairly flat on-axis but it will drop off quickly off-axis.
It is also very well damped for an aluminium 10". Probably the best damped 10" aluminium driver there is. The ringing is very short it start at about 4kHz and is the worst at about 7-8kHz but it's nowhere near Seas excell drivers ringing for instance.
I don't think there is a 10" paper woofer which can be taken that high (1300Hz crossover) that clean, they all seem to have minor cone breakup starting much lower. May not sound harsh but it's not clean either.
I can't really go with any Acoustic Elegance drivers as first of all I've come to think after seeing so many driver measurments that with DSP filters and EQ that good metal drivers behave better in the passband and the cone breakup above this can be made inaudible so one only has to watch out for harmonic distortion products of the driver falling on the cone breakup. And secondly because I have not seen any good measurements of Acoustic Elegance drivers.
Here are the measurements of the Scanspeak 10" http://www.audiocomponents.nl/download/scan-speak/26W-4867T00 - Hobby HiFi 2011-2.pdf

edit: and btw thanks for the Eton suggestion, but the SB17NAC35-4 which I picked for my 3-way actually has less distortion than the Eton driver you linked to. From 600Hz upwards the 3rd HD is significantly lower and since the SB17 is a much larger driver it can be taken much lower, it's distortion is much lower below 400Hz or so. And since I don't want a crossover in the midrange 160-1200Hz or so the Eton won't work for me in a 3-way.

Legend Acoustics used this Eton mid to turn their 3way Tikandi DEQX speaker into a 4way with the 22 cm Seas Magnesium midwoofer (the same driver Grimm uses) and the Scanspeak Beryllium dome tweeter - presumably for good reasons. And it's not like they didn't have speaker correction or steep filters available on the Deqx. I presume Grimm also uses some amount amount of driver correction to make the LS1 work - Putzeys says it's one of his areas of research. Maybe he knows how to do driver correction better. Maybe his Seas magnesium midwoofer is customized and better damped than stock.

Ah yes that could be possible. Though I've come to think that the trick for a great metal driver is in the shape of the cone. Seas makes fairly flat metal cones and these are sure to breakup hard relatively low. And for instance the scanspeak metal cone is much deeper giving it more strength, or the BlieSMA alu dome which is very round or the SB Acoustics metal driver which has reinforcing ribs etc.
In any case the stock W22EX001 I'd presonally never would dare to use for my 2way with 1300Hz crossover, and Grimm is using it at 1550Hz crossover. It will surely have quite a high 3rd and fifth harmonic distortion right below the crossover which is falling on the cone breakup.

If you want to go for the more expensive beryllium Bliesma and you intend to use DSP crossovers, I say just follow the "get what looks the very best" logic to it's end and go all in with hard material drivers like this new accuton mid and also a suitable dedicated midbass/multiple subs 4 way setup. End of game type stuff - if you can afford it.
https://accuton.com/en-home/produkte/lautsprecher/keramik/Keramik-C168-6-990-CELL

Have you seen the measurements of the Accuton ceramic drivers? They're not the greatest and certainly not worth the money it seems to me. There's nothing revolutionary about cermaic drivers. You can get the SB17NAC35-4 in a ceramic version as well as the SB17CAC35-4, it measures almost the same as the alu version only very slightly worse.

PS: What are the prices on the bliesma tweeters, after the 25% Black Friday promo? They didn't list them on the website.

PS2: If you want to make a profit selling high end speakers, being a good speaker engineer is not enough. You need to be a shrewd marketing bastard.

The price for one pair of T34A is */ Euro and for T34B is */ Euro + 19% German VAT, if you are a private person. (in the EU)
(I'll delete this pricing info after a few days as I'm not sure the designer likes it if this info remains lingering on the web)

As for selling speakers. I'm not sure if I want to do it, probably not
But if I do then the forums for studio people is a good place to make people aware. And they kinda need something really good that's not expensive I'm sure it'll sell if explained well. Not looking at all to get rich here. I've gone for that in a distant past and I prefer to do good in the world now and not for instance prey on uninformed music lovers like many "audiophile" brands do with their crazy expensive and not so great speakers.

 

[andreasmaaan]

Maybe we can come up with a good test for what is most audible in speaker design regarding harmonic distortion, IMD etc.

It depends on how the driver is going to be used IMHO. This is why I suggested that for your intended use in the two-way, you need to test IMD using a very low frequency combined with a midrange frequency.

What I also like about the cheaper 2-way btw is that if it turns out as good as I hope then it seems fairly easy to turn it into a commercial speaker? I mean why not.. there isn't really anything like it on the market and Hypex is about to come out with a very good yet cheap 4 channel OEM DSP module which would make it useable for a wide range of people. If I were to sell it direct with a 1000 euro profit for 3500 a pair this could still be about the best deal for people. Don't know if I'd want to get into this though.. (edit: as a comparison the Grimm LS1 has an 8" Seas woofer and DXT tweeter and goes for about 18,000 euro a pair, the LS1Be for double that, and my two way drivers are better)

Why not?

If you're planning on making it usable for a wide range of people, maybe you want to think more about polar response though?

 

[JustIntonation]

I asked BlieSMa about advice on which tweeter to use for my project. And the answer was that for a studio monitor the T34B is adviced.
It has a more detailed, more transparent and more natural sound than the T34A. I was also told that several studio monitor companies ordered both the T34A and T34B to test and all chose the T34B.
So I've just ordered a pair of T34B

It depends on how the driver is going to be used IMHO. This is why I suggested that for your intended use in the two-way, you need to test IMD using a very low frequency combined with a midrange frequency.



Why not?

If you're planning on making it usable for a wide range of people, maybe you want to think more about polar response though?

True.

And after a good nights sleep I've come back from my crazy thinking of yesterday
Its cutting it a bit too close and too risky with the Scanspeak 10".. And when I've built it sure it will sound great but I'll allways wonder how it will sound with the SB17NAC.
I'll build it first with the SB17NAC35-4 and BlieSMa T34B-4, this is sure to be most transparent and clean, and then later decide which bass/sub-woofer to add after measurements and listening. Maybe I can add the Scanspeak 10" and make it so I can switch, maybe the Seas XM001-04 for loudest volume, maybe something else.
Sorry for being like a pingpong ball sometimes with my doubts I can get those moments

 

[eugenius]

The 10" Alu Scanspeak starts to ring at 1kHz. Yes, it's quite low level for an aluminum driver but it's high Q and will be audible and unpleasant no matter what you do in the crossover. 10" is also too big for a two way IMO.

Grimm certainly does something right to that midwoofer in the LS1:

For comparison, here's the SCM19 mk2:

And the Elac 330 CE (another speaker with a high quality underhung midwoofer):

Personally I'd rather not go larger than 8" for a two way with the Bliesma, to keep good dispersion.

Left field idea: try the JBL 728G (the midwoofer from the JBL 7 Master Reference monitors - where they cross it at 1.7 kHz to a trick compression driver with a M2-like waveguide). The JBL guys are puffing their chests about it, it's got great bass extension and it's available for 325$ here:
https://reconingspeakers.com/product/jbl-5047951x-model-728g-8-woofer-for-lsr708i/

 

[yummy]

what you mean is AV product, good AV SUPER expensive, those guys are rich

 

[JustIntonation]

The 10" Alu Scanspeak starts to ring at 1kHz. Yes, it's quite low level for an aluminum driver but it's high Q and will be audible and unpleasant no matter what you do in the crossover. 10" is also too big for a two way IMO.

Grimm certainly does something right to that midwoofer in the LS1:
View attachment 18066

What we're seeing here is the result of a DSP linear phase FIR crossover.
The ringing of the W22EX001 starts at about 4kHz and is filtered out by the crossover at 1550Hz.
But what is not filtered out are the harmonic distortion products of frequencies occuring up untill the crossover of 1550Hz which are falling on the cone breakup peak. This would be visible in the harmonic disortion measurement of the LS1. I don't know if anybody published such a measurement of the LS1? But here is the harmomic distortion measurement of the W22EX001 for about 88-90dB SPL:

See how the 3rd and fifth (and higher) harmonic distortion will not be filtered out by the crossover? Even when taking into account the crossover I'm guessing the LS1 has about 1% distortion at 96dB SPL for the 3rd harmonic and it will fall on 4kHz which is the most audible frequency.
I'm guessing this must be audible probably as a forwardness in the sound.

I don't get the Seas DXT tweeter used either.. That's a €70 tweeter on an €18.000 speaker. Doesn't look particularly clean on the CSD you posted either.

 

[JustIntonation]

Just a little update.
Have decided to give the Hypex Fusion amp a go. Just ordered two FA123.
Dove in a bit more into the way it handles the signal an convinced myself it should have perfect imaging and any artifacts from 2 clocks for left and right should be inaudible. I'll test this once I receive them with just the T34B-4 tweeters (already have them, they're beautiful) and compare it to the AMS0100 amp I have here, have two weeks to send them back if there's degraded imaging with the FA123.
Also did a few more tests and reading up on the AMS amps vs Hypex NCore and I'm beginning to suspect the AMS amps perform quite a bit less good than published specs in real world circumstances due to the lack of input stage probably and the Hypex NCore amps are probably better.
So no longer in need for a seperate 6-8 channel DAC as a good enough multichannel AKM DAC is in the Fusion amps. Other benefits are lower price total and built in DSP in the Fusion amps.
Also decided on the woofer for my speakers, it will be the Satori WO24P-4.
And for USB to SPDIF I'll use the Armature Hecate.
So it'll be Armature Hecate - Hypex FA123 - Satori WO24P-4 + SBAcoustics SB17NBAC35-4 + BlieSMa T34B-4 and I'll put the drivers in one single 3-way box. Now finalizing the design on the box and then ordering the wood and tools.
Things have been snails pace but slowly getting there

 

[maverickronin]

Things have been snails pace but slowly getting there

Faster progress than I usually make on my projects.

Maybe in the spring I'll be able to actually make the SB12PAC25 + SB26ADC minimonitors I've been planning...