I included external dimensions because it's the form factor that makes people (not me) interested in this type of loudspeaker. All internal volumes eg. for woofer, for midrange can be derived from the external dimension. The dimensions suggested was equal to the L100 Classic (20218) released. Feel free to modify to whatever dimensions.
I just wanted to start somewhere
For greatest compatibility, yes 8 ohms nominal.
I noted that the L100 Classic is rated at 4 ohms nominal, with a 90dB/2.83V sensitivity. To compare like with like (build it yourself vs buy it yourself, I wrote "shall" have 4 ohms impedance. Perhaps we should write that as "may" have.
If were were to targeting 4 ohms impedance, then perhaps we would write "will have"
I would have chosen twin 10" 830668 which is both higher in sensitivity, more output and needs a lower box than a single 830669. But that it an different (slimmer floor-stander speaker, with higher SPL)
Reference:
Reference: placement of (full range) speakers
That tweeter is cheap and so, but not usable low enough. I would not cross that below 3kHz because distortion rises fast below that. The SB26ADC is also relative cheap and a lot better on that. And if you insist on a build-in waveguide the SB26STWGC-4 is a way better choice and the price is still reasonable for a tweeter.
...and how the transition of a big Sd to a too much lower one. Does it make sense to have a 12" then a 4"?!
In my humble opinion, rather than debating back and forth and trying to find that one driver available everywhere in the world, this kind of project would be more successful, practical, and easy for others to execute if someone buys a few of the shortlisted midrange drivers (of the same class/diameter/based on other requirements), take measurements, and create crossovers for the different variants.
In my view, crossover design may be the easiest part of this whole attempt as it just takes a couple/few hours with VituixCAD to create one/some versions that generate good-looking plots and sound good if drivers are sensibly selected based on power handling/max SPL/cabinet size-shape/directivity requirements.
In my view, crossover design may be the easiest part of this whole attempt as it just takes a couple/few hours with VituixCAD to create one/some versions that generate good-looking plots and sound good if drivers are sensibly selected based on power handling/max SPL/cabinet size-shape/directivity requirements.
That Genelec reference is simply saying that their loudspeakers are sissies and don't work well in a free-field environment 🙄
Let's assume that we want a smooth Directivity Index through the M-T crossover region, basically from 1k - 6k. By "smooth" I mean a DI which is either flat or consistently rising, and which has minor peaks/dips which are less than 1/3 octave wide.
Let's also assume that we will be using a non-waveguide flat-faced tweeter.
A 4" midrange (Sd ~ 50 cm^2) is easier to cross to a flat-faced tweeter. The crossover frequency can be as high as 3k with good DI performance, and there is more flexibility in the crossover slopes.
A 5" midrange (Sd ~ 80 cm^2) can work, but my past projects using 5" mids all forced me to have a crossover of about 2k. At 2k, a 5" cone has a very wide dispersion which matches the tweeter. A higher Fc resulted in an unacceptable peak/dip in the DI curve.
There is nothing wrong with a 5" mid and a 2k crossover, but we must accept the fact that this will affect our future tweeter choices. A 4" mid opens the possibility of using a wider variety of tweeters.
At the lower crossover from woofer to mid, there will be no measurable directivity difference between a 4" and a 5" driver. Both are effectively omnidirectional in the 300 - 600 Hz region. I also feel that differences in 100-500 Hz directivity between monopole speakers are meaningless in normal rooms... Of course dipole and cardiod patterns can make a big difference.
A small waveguide on the tweeter changes the situation, and something like a Seas DXT tweeter or the SB26STWGC allows a 5" mid to cross in the 2.5k - 3k region with a very nice DI curve. But again, this path starts to really limit our tweeter options.
Something to think about...
j.
I agree that both 4" and 5" should work in terms of directivity and xover to the tweeter.
Not sure a 4" will be sensitive enough and provide sufficient low-frequency dynamics/SPL if the xover to the woofer happens at 400 Hz or lower.
I personally prefer a 5" not only for aesthetics, but also because the cone will work at lower excursion.
Not sure a 4" will be sensitive enough and provide sufficient low-frequency dynamics/SPL if the xover to the woofer happens at 400 Hz or lower.
I personally prefer a 5" not only for aesthetics, but also because the cone will work at lower excursion.
@tktran303 , hi,
I found the sim 2 to 4 pi quite interresting, but what I don't understand is why you loaded the 12" with 60 L and the 10" with 72 L. Is not the opposite that make them working better ?
I found the sim 2 to 4 pi quite interresting, but what I don't understand is why you loaded the 12" with 60 L and the 10" with 72 L. Is not the opposite that make them working better ?
My fear, indeed...
It is not hard to find cheap tweeter that cross at 2K. The 26ADC, has very low THD, easy to work with, a 12dB is certainly feasible at 2K hz if one accepts the not so good waterfall below 2k5 Hz. But perhaps a 26WG makes a 12dB filter easier due to the recess...
I'm not to worry about the tweeter choice. We stay a little too much stickyy on this midwoof choice... That NE149W-08 availability about the MOQ at PE needs to be known. And we need to reduce the midwoof list now with objectiv criterias.
-crossover easyness
- price/sourcinhg
- sound (I assume a fiber glass, polypro or paper don't sound the same and that's matter imo if you cross to high the midwoof with the SLS woofer
- crossover easyness (less parts : slope, notches, etc.
For I, the 5" size reduces the crossover slope constraints if you choose a tweeter quiet enough like a 26ADC for instance.
I 'd reduce the list of the woofmid with that in mind... simply remove the units with breakups that can be handled with the low pass inductor that rules the BSC too. Trying to do 12dB for this unit for High/low pass is the good path due to the constraints of cost imho. So 5" again !
YMMV...
Last edited:
... and on the other hand, maybe we don't care about DI curve performance? The designers of the original 1960 - 1980 "monkey coffins" obviously did not concern themselves with this.
The problem of one's own motivation to do all of the aforementioned has not been solved yet.
If PE would believe this loudspeaker kit might actually sell, they may be willing to donate prototype parts,
and perhaps a cut for the diyAudio from kit sales?
I can think of 3-5 good tweeters that can cross at 2kHz and are priced reasonably.
Is there a consensus on any of the current midranges under consideration that they can be eliminated?
Is there a consensus on any of the current midranges under consideration that they can be eliminated?
- I'm thinking we drop the Peerless NE series from our list, at least for now. They simply aren't available to Europe at a reasonable price. If that changes before we have made significant progress, we can reconsider them.
- SB15NBAC is the most expensive option and would require a notch filter - driving up the cost even further. In contrast the Scanspeak 15W is almost the same cost but drastically smoother, without the aluminum-cone breakup to deal with.
Speaking only for myself, I care about directivity performance and how we can IMPROVE those classic "monkey coffins'. Which is one reason I have said that I will measure a version with/and without a waveguide.
With your past projects with the 5" mid...at 2kHz did you feel like the end result with respect to directivity was GOOD, or still just ACCEPTABLE given the tradeoff of sticking with the 5" driver?
I vote to remove the SB15NBAC for the reasons you mentioned.
I have a pair of SB12MNRX2-25-4 drivers sitting around waiting for some future project. When we have decided on a baffle size/shape, I would be willing to put together an XPS foam board prototype and make polar response measurements and impedance measurements.
I have a pair of SB12MNRX2-25-4 drivers sitting around waiting for some future project. When we have decided on a baffle size/shape, I would be willing to put together an XPS foam board prototype and make polar response measurements and impedance measurements.
Dispersion turned out to be the control of how the loudspeaker sounds in a room. I don't see why anyone would want to ignore that in a new design.
I think I was able to achieve quite good directivity performance with 5" drivers. But in both cases I was willing to (1) shape the baffle and optimally space the drivers and (2) profile the cabinet edges. This enabled me to maximize the DI potential of the drivers I was using.
First example: SB15CAC30 in a 3-way active speaker, two 8" woofers, and SB26CDC tweeter.
The DI performance is very nice. I experimented with voicing in the 3k - 6k region, and in my room, this the voicing that sounded best... despite the PIR curve which rises a bit in this region.
Second example: SB15NBAC30 in a semi-active speaker with passive crossover from mid to tweeter (SB26STAC) and active crossover to Dayton RS270 woofer.
The DI curve is again very nice. This speaker is used in a different room, and I voiced it to have a little more bass below 200 Hz. Here again I experimented with making the PIR curve flat from 2k-7k, but it sounded, well, a little flat. This is the final voicing.
A 5" driver can be crossed to a 1" dome at 2k with good directivity performance... BUT, the baffle area around the tweeter needs to be shaped to minimize baffle gain, and the CtC spacing between the mid and tweeter needs to be optimized, even if it makes the speaker look strange.
I am not sure how much of this is applicable to a wide baffle speaker with hard square edges...
j.
Attachments
And all of this directivity focus may not matter that much in real application... I have another speaker that I use in my main system which has a 4 dB dip/peak in the directivity index. With careful voicing, and proper placement, it sounds excellent, better in fact than the two systems I showed above.
Yes, and this is my concern...we are locked in at design goal #1 and not able to do those things.
But this is why I'm not THAT concerned.
If you sim the optimal LW, PIR, DI, etc. taking as given the drivers and baffle, then by definition any change you make in the voicing process to actually make the speaker SOUND better will make the squiggly lines LOOK worse. I'll take a better sounding speaker over better looking squiggly lines any day.
https://www.erinsaudiocorner.com/loudspeakers/wharfedale_linton_85/
So, don't be osessed with directivity slope alone ...
- This is truthfully one of my favorite speakers so far. On sight, I half expected this speaker to sound quite colored and have resonances; an experience I had with the Klipsch Heresy IV. However, my subjective listening experience was the complete opposite. For the first time in a while I found myself going through an extensive number of songs and just enjoying the music. There were no “standouts” that bothered me and pulled me “out of my element”. Even upon additional listening sessions I continued to find myself more and more happy with the sound from the Lintons.
So, don't be osessed with directivity slope alone ...