4X8" DIY Mid-High

Robert Kasleder

New member
Feb 24, 2018
8
0
1
Romania
Hi.

For years I have been thinking about a high quality mid high cabinet that would have the "sound quality of a studio monitor", high SPL, relatively low wight and acceptable size and a good dispersion trough out it's frequency range (preferably 90x40-60). It should also go down to 80hz.

I bought 2 BMS 4594 drivers about 4-5 years ago, but newer decided on the final plan.

Some types of Synergy horns where serious candidates but despite the good horizontal dispersion of a 90x50 synergy horn, the vertical dispersion is not that good due to it's size. And a lot of times it's difficult to put the wide cabinet in a narrow space. Another issue is if you have a low sealing than all of the low-mid frequencies will reflect back due to the wide vertical dispersion at those frequencies.

Peter Morris's and other members dipole designs on this forum inspired me a lot.
So I would like to thank you guys for that.

Counter to the synergy horn a dipole design has a very good vertical dispersion even at lower frequencies. And getting the crossover frequency low and the drivers close enough it can act as a good point source to. It can fit in narrow spaces, you can use it even at home. And when used in a low venue, you have less reflections of the sealing.

So here is my approach on a dipole mid-high design based on my priorities:

It would be very similar to Renkus-Heinz's STX4.
https://www.renkus-heinz.com/upload/st-stx4-datasheet-1.pdf
But less heavy (the STX4 weighs 50 Kg's) , it won't go as low (50Hz for the STX4, 80HZ for my design), and maybe a little bit smaller.
So I would use 4x8" drivers, 2 on the top and 2 on the bottom, with each 2 facing each other in 90 deg. In the middle the RCF HF950 with the BMS 4594 driver. But the cabinet should be wide enough to fit the 18 Sound XT1464 in case I would like to change the horn.
A 8" driver has 220 cm diameter. So the distance between the horn's center and the acoustical center of the drivers is 26cm.
650HZ 1/2WL = 26cm. So I guess it should work fine for the vertical dispersion.

I haven't decided what 8" driver should I use.
The B&C 8NDL51 is probably the best candidate. It has the best price/performance ratio.
The Beyma 8LW30 has a 2.5" VC and 250W, but it's a little bit heavier, more expensive, and it doesn't have a treated cone.
The 18Sound 8NMB420 has the highest SPL acording to sims (+3db in WinISD, +2db in Hornresp compared to the B&C), but it's the most expensive.

Attached are some comparison between the B&C and the 18Sound in WinISD and Hornresp.
The WinISD is 4x drivers at 1000w total (green is the B&C).
The Hornresp is 2x drivers facing each other in 90 deg with a vented back chamber at 500w total in 4Pi (black is the B&C).
 

Attachments

  • B&C vs 18Sound WI.png
    B&C vs 18Sound WI.png
    28.1 KB · Views: 82
  • B&C vs 18Sound HR.png
    B&C vs 18Sound HR.png
    44.6 KB · Views: 87
Thanks for your suggestions, but the B&C 8NDL64 is to heavy(2.8Kg each) and although the 18sound 8NW650 can handle more power, it's less sensitive according to simulation.

Below is a simulation in WinISD: Green- 4xB&C 8NDL64 in 48L/70Hz @1000w
Yellow- 4x18Sound 8NMB420 in 85L/75Hz @1000w
Red- 4x18Sound 8NW650 in 50L/75Hz @1600w

Note how the 8NW650 requires 1600w to be at the same SPL as the 8NMB420 at 1000w.
 

Attachments

  • 8NMB420 8NDL51 8NW650.png
    8NMB420 8NDL51 8NW650.png
    106.4 KB · Views: 37
  • 8NMB420 8NDL51 8NW650.png
    8NMB420 8NDL51 8NW650.png
    106.4 KB · Views: 37
Last edited:
I assume you have checked Xmax, for LF output this is normally the limiting factor. Some times a less efficient speaker may give you more SPL because its still operating within its Xmax range. All of the ones you mentioned have a similar Xmax ... and the MB420 does look good on paper.

I used the NW650 in my line array but for different reasons - it operates in its piston range over a wider frequency range.
 
Yes that's exactly what I was looking for, since I will only operate the speaker up to 800Hz maximum. But the low end is important, and Xmax plays a big role in that.

I don't know if the Xmax values of the 18Sound speakers are truly matemathical or no since they don't give you the magnetic gap height and the winding depth.
And I don't know if in reality i get that SPL advantage from the NMB420 over the 8NDL51

Here is an Xmax sym: Yellow- 18Sound NMB420 Xmax 5.8 - excursion: 6.3 at 100hz at 1000w
Green- B&C 8NDL51 Xmax 4.5 -excursion 5.3 at 100hz at 1000w
Red- 18Sound NW650 Xmax 5.5 - excursion 6 at 100hz at 1600w

So at the and the my question is if the 18Sound 8NMB420 is worth the extra 40Euro/ piece over the B&C.
 

Attachments

  • 8NMB420 8NDL51 8NW650 Xmax.png
    8NMB420 8NDL51 8NW650 Xmax.png
    105.8 KB · Views: 24
although the 18sound 8NW650 can handle more power, it's less sensitive according to simulation.

Below is a simulation in WinISD: Green- 4xB&C 8NDL64 in 48L/70Hz @1000w
Yellow- 4x18Sound 8NMB420 in 85L/75Hz @1000w
Red- 4x18Sound 8NMB420 in 50L/75Hz @1600w

Note how the 8NW650 requires 1600w to be at the same SPL as the 8NMB420 at 1000w.

Note it's also using a box more than 40% smaller than that required for the more efficient driver.
That's Hoffman's Iron Law in full force right there.
 

What you say it's correct, but those values are not for the mathematical Xmax (voice coil height-magnetic gap)/2
Personally I don't like to look at those values I prefer to calculate the mathematical Xmax. Especially when I compare speakers from different manufacturers which can have different formula for calculating Xmax.

8NDL51 (17-8)/2=4.5
8MND51 (16-8)/2=4
8NW51 (19-10)/2=4.5
8MBX51 (15-7)/2=4

18Sound states in their specs that the 8NMB420 has a "Linear Mathematical Xmax± 5,75 mm (±0,23 in)"
But they don't give you the voice coil height and the magnetic gap height so you could double check.

So for me It's really down tho this to justify that extra 40 EUR/driver for the 8NMB420 over the B&C8NDL51:
Thus the 8NMB420 really has a mathematical Xmax of 5.75?
And if it's really 2-3 dB louder than the B&C 8NDL51

There are other differences between the two drivers:
The 8NMB42 has a lighter cone but also a weaker motor and requires a bigger cabinet than the 8NDL51.
Also the 8NMB420 has an aluminium VC, while the 8NDL51 has a copper VC.
 
Last edited:
What you say it's correct, but those values are not for the mathematical Xmax (voice coil height-magnetic gap)/2
Personally I don't like to look at those values I prefer to calculate the mathematical Xmax. Especially when I compare speakers from different manufacturers which can have different formula for calculating Xmax.

8NDL51 (17-8)/2=4.5
8MND51 (16-8)/2=4
8NW51 (19-10)/2=4.5
8MBX51 (15-7)/2=4

18Sound states in their specs that the 8NMB420 has a "Linear Mathematical Xmax± 5,75 mm (±0,23 in)"
But they don't give you the voice coil height and the magnetic gap height so you could double check.

So for me It's really down tho this to justify that extra 40 EUR/driver for the 8NMB420 over the B&C8NDL51:
Thus the 8NMB420 really has a mathematical Xmax of 5.75?
And if it's really 2-3 dB louder than the B&C 8NDL51

There are other differences between the two drivers:
The 8NMB42 has a lighter cone but also a weaker motor and requires a bigger cabinet than the 8NDL51.
Also the 8NMB420 has an aluminium VC, while the 8NDL51 has a copper VC.

18 Sound use - "Linear Math. Xmax is calculated as(Hvc-Hg)/2 + Hg/4 where Hvc is the coil depth and Hg is the gap depth." Which is probably the point at which distortion and non - linear behaviour starts to rise, but it can depend a little on the drivers design.

With the 8NMB420 it has an RMS power of 280 watts, but the continuous program power is only 400 watts not 560 watts. If you use the program power for your calculations what you find is that the is the 420, 650 and the 8NDL64 all make about the same SPL (based on the manufactures Xmax) the 8NDL51 makes about 3 dB less but goes lower.

The reason I chose the 650 was that it had a stiffer cone and as you can see from the frequency response it operated in the piston range up to about 1200 - 1500 Hz. You can get a rough idea where this happens by looking at the off axis response compared to the on axis response. For a line - array and coupling between drivers this is important and if you read the spec's 18 sound say that its recommended for line array applications.

B&C also specify Xvar which is a good indication of the useful Xmax

"Evolution is a process that affects not only products, but also their technical specifications. Constant advances in research provide more and more precise methods to measure the performance of loudspeakers, and describe their features. Thiele – Small parameters have become the universal language for describing loudspeaker behaviour in the small signal domain. Nevertheless, they comment little on the working limits of loudspeakers in the large signal regime.

These limits are customarily indicated by Xmax, the maximum linear excursion. This value is typically measured according to the AES2-1984 standard, corresponding to a maximum of 10% total harmonic distortion (THD) with a sinusoidal signal (though most manufacturers, including B&C, now typically provide data for Linear Mathematical Xmax, not measured Xmax). Recent research shows that this method can yield ambiguous results, and even different numerical values for the same loudspeaker. The main limit of this measurement is that it looks at the output signal instead of the physical features of the driver itself. On the contrary, the most up-to-date instruments for distortion analysis can measure the variations in loudspeaker parameters when they are fed with high-level signals. In this way, an excursion limit can be fixed, beyond which the parameter’s variation becomes excessive.

The “X var” value reported in our data (generally after the traditional “Xmax” value) is measured this way. Beyond this excursion limit, the magnetic field seen by the voice coil, or the total suspension compliance, or both, drops to less than 50% of their small signal value, producing high distortion levels, strong variations from small signal behaviour and power compression. The new technique yields different results from the standard measurement based on THD. B&C Speakers believes that this added information gives a more accurate and reliable description on loudspeakers behaviour in actual operating conditions."


See ... http://www.bcspeakers.com/resources/faq-frequently-asked-questions/
 
Last edited:
Thanks for clarifying how 18Sound calculate their Xmax.
If I use that formula to calculate the B&C 8NDL51 Xmax the I get 6.5mm, B&C specs say 7mm.
So that means B&C is using a different method.

But my conclusion from this is that whit the 18Sound's formula the 8NMB420 having 5.8mm, and the 8NDL51 having 6.5mm, probably the 8NDL51 is beter in this case. The 8NMB420 will have 3dB more, but with larger displacement and less Xmax, which leads to higher distorsions.
Below is a simulation of displacement with the 8NMB420 and 8NDL51 at program power(1600w for 4 speakers).
The 8NMB420 has 8mm at 400w. 2x8=19 which is it's max peak to peak excursion.

The reason why I'm not looking at the bigger VC drivers is weight. All of them exceed 2 Kg's. And I have to put 4 drivers in one cabinet.
Of course if weight is not an issue you could get the most powerful 8" driver and have more SPL.
The only other driver that I could find under 2 Kg's that is louder than the 8NDL51 without exceeding it's Xmax is the BMS 8N519.
But that cost double compared to the B&C.
 

Attachments

  • 8NMB420 8NDL51 Xmax p.power.png
    8NMB420 8NDL51 Xmax p.power.png
    104.8 KB · Views: 22
In addition to Marjan's point, the box is the major parameter in optimizing the weight, in my experience. Consequently a more heavy driver with higher motorforce, may give you more total output for a given box volume (assuming you have DSP at your disposal to correct the non-flat frequency response), thus decreasing box volume and total weigh, compared to the less powerful driver in a larger box.

Best regards,
Fred
 
The cabinet has to be wide enough to fit 2x8" drivers(this can be reduced by angling the 2 drivers in 90deg) , and also deep enough for the HF950 horn + the BMS4594 driver, and tall enough for 2x8" drivers + the HF950.
So I have physical limitations on how small the box can be. Like I said I don't have a finished plan yet but based on rough calculation the box should have around 60-70L.
Being limited by these factors the only way to keep the weight down and to have the highest SPL is to use the lightest driver which can produce the most SPL for this size of a box.

And like Peter Morris pointed out in POST #4 linear excursion is also a limiting factor.

For example here is a comparison of 4 speakers in their optimum sized box/tuning freq, and at their program power:
4x BMS 8N519(RED) -45L/75Hz at 2000w
4x B&C 8NDL64(YELLOW) -30L/75Hz at 2800w
4x Beyma 8LW30(GREEN) -60L/73Hz at 2000w
4x B&C 8NDL51(BLUE) -50L/70Hz at 1600w

The 8NDL64 has the smallest box(30L) but it needs 2800w to have the same output as the BMS.
Plus at this power the BMS exceeds it's mathematical Xmax by only 0.5mm, while the B&C exceeds it by 3.5.
The Beyma is about 1.5-2 dB louder than the BMS and the 8NDL64, but exceeding it's mathematical Xmax by 3.8mm.
The 8NDL51 is 1dB below the BMS and the 8NDL64, but exceeding it's mathematical Xmax by 2.3mm.

So my top pick would be the BMS 8N519, but it costs 200EUR/piece versus the 8NDL51 at 100EUR.
That means 400 EUR more/cabinet and 800 EUR more for one pair of cabinets, to gain 1dB and have less distorsion.

The Beyma 8LW30 is the loudest but when driven with the same power as the 8NDL51(1600w) it's only 1.6 dB louder, and exceeding mathematical Xmax by 3mm. It should also have less power compression at this power compared to the 8NDL51 due to it's larger (2.5") VC and 50w more power handling. It's 20 Euro more expencive and 0.55Kg heavier thatn the 8NDL51.
So this looks like a good trade off: +2dB for 20Eur, 10L and 2.1Kg/cabinet, and 0.7mm more excursion.

The B&C 8MBX51 is almost identical to the 8LW30 in the simulations, but it has smaller VC, 50w less power handling, and 0.5mm less mathematical Xmax.
 

Attachments

  • 8N519 8NDL64 8LW30 8NDL51 Xmax.png
    8N519 8NDL64 8LW30 8NDL51 Xmax.png
    113.3 KB · Views: 32
  • 8N519 8NDL64 8LW30 8NDL51.png
    8N519 8NDL64 8LW30 8NDL51.png
    94.8 KB · Views: 30
Last edited: