Focal Newform Utopia

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
So I am setting off to design and build a three way speaker. The drivers are as follows:

Focal Audiom 6wm
Focal 13w8711

Newform Research R30

The system will use an active crossover. Amplifiers for each speaker.

Ribbon amp: Blue Circle BC2 monoblocks (biased down to 20w so they don't catch on fire again! hopefully!)
Audio quest CV-4 speaker cable, audioquest viper interconnect

Midrange amp: Mark Levinson ML-11 (and I will try an anthem integrated 1 and compare)
Audioquest Type 4 speaker cable, audio quest viper interconnect.

Bass Amp: Mark Levinson ML-9
Audioquest Bedrock or Slate speaker cable, audioquest viper interconnect.

Active Crossover is Behringer dcx2496
REW will be employed with Behringer DEQ2496
Some room Correction with a home made quadratic diffusor. And some absorption will be implented along the way.

I am reading Vance Dickason's Loudspeaker Design Cook Book. I am also running simulations for boxes from WinISD Pro Alpha.

I figure we will start with the woofer. I know this integration has been done before between 13w8711 and audiom 6wm.

Recommended box volume for from Focal is as follows:

80 to 120L box, with Fb 33, and Fb33.6 respectively. It also says superb bass definition from 60 to 70L vented applications. What does this mean? Why are they mentioning it? Is it better transient response from the smaller box? Whats the meaning of superb bass definition?

After running the simulations the response from a box 60L is not desirable.

Looking at the WinISD graphs I am leaning towards the 120 L box because it gives the lowest response, and the dip in the curve is only one dB so I presume can easily be equalized later along with the other dips and peaks due to the room.

In Vance's book he uses the upper cutoff of a speaker to be the -6dB point in the off axis measurement of frequency response. Referring to the graph it appears that the woofer drops about 5db (on and off axis measurements are within 1 dB agreement) between 100Hz and 200Hz, and the spl is constant within a dB from 200hz untill 700Hz. From here there is a dramatic rise and then a sharp rolloff. I imagine the peak above 700HZ can be filtered out in the DCX2496 by doing a response plot with the microphone (Behringer ecm8000) and applying the filter using REW and the DCX.

So the off axis response of the woofer doesn't fall off by 6dB untill we get into the 1050HZ region as far as I can tell. So it appears from this that I can cross over the midrange to the woofer anywhere in the range below 1000Hz with active corrective filtering, or anywhere below 600hz without any active correction.

The lower cutoff for the midrange is determined by the fundamental frequency or the resonant frequency of the midrange. In Vance's book he states that it has to be minimum 2 times the fundamental frequency. The Audiom 6wm has Fs of 116.9Hz. So minimum of 233.8Hz, and its better if its 3 x Fs. Which would give us 350.7Hz. This establishes my lower limit.

So the woofer and Midrange will be crossed over between 350 ,and 600hz, and I imagine the lower I do it The better transient response I will get. At this point I will be varying the xover frequency 20hz at a time, taking notes, and listening to the same music over and over. Something live and acoustic, and it will probably be Eric Clapton Unplugged listening for the ratio of attack and decay (harmonic ringing that lingers after the guitar string is plucked), and also that full body balance, vs too lean sound. I will also graph SPL response. For the xover points of interest and keep them with my subjective notes.

As for the Audiom 6wm enclosure from what I understand the bigger the better beyond a size that will limit the low frequency response at our x over point. I have purchased sonotubes and mounted the drivers to MDF with a flooring adhesive called no more nails. I am considering overlaying another sonotube and filling the space between with either sand, or concrete, but this will be done after the speaker is complete.

I will consider the integration between the ribbon and the midrange in subsequent posts, but for now I wanted to get the general idea laid out for the bottom end.

I have a few questions.

In WinISD Pro Alpha the software determines box volumes based on a driver you select. Does it account for the volume of the speaker itself? Or do I have to subtract this from the given box volume?

What filter slopes have people tried with this driver combination in the past?

Knowing that I am going to use an active setup with the ability to filter the woofer, can I succesfully lower the response of the box a bit and correct for the rolloff? If so by how much?

I will be posting the Graphs I described shortly. I will also be doing the calculations by hand to compare with the software.

Any input appreciated. I will be posting more details , I just wanted to get the general idea out there for some input before I go any further.

Last edited:
Just another noob so cannot comment thoughtfully on some of the issues--actually owned some of the Newform speakers a dozen or so years ago--believe they were mated with a pair of very nice Scan Speak woofers per channel crossed over about 1.8kHz. Liked the sound quite a bit but do remember that they were somewhat less forgiving in terms of listening distance than a conventional cone/dome configuration. Ultimately the aesthetics of the column cabs and big stick led me elsewhere--to an even bigger stick in my case--the B&G 75" magnetoplanar drivers. Now back to doing a studio monitor sized project using Heil tweeters.

What I wanted to say is I would also not hesitate to call or e-mail the owner of Newform for advice if you have not already done so--the man provides world class customer service IME and those of many others. Anyhow best of luck and will be following your project with great interest.
Ok so I have decided to go with the 120L box, and from there I can experiment with different port tuning.

I will be trying out the Equalized Semi Sealed approach with the 2nd order rolloff corrected. I will compare this with the traditional ported tuning with the 3db shelf, and 6db shelf. Both can be accomplished from the same box volume as far as I can tell. The port tuning changes things enough to try both approaches.

The only other question I have is: does WinISD pro alpha give box volumes which account for the volume of the woofer itself? Or do I have to add this to the volume value they give me when calculating my box dimensions?
I am currently trying the configuration with the Focal 6WM, and the scan speak 8935 Carbon fiber woofer.

I have the three way sounding very sweet, and very transparent with the following slopes.

6db/octave at 1.82 kHz high pass on the newform R630.

12/octave butterworth on the 6wm. With some EQ to the individual drivers applied.

The scan speak woofer and the mid are crossed at 345hz, 12db/octave Butterworth on the mid, and linkwitz on the woofer. Sounds very good so far. Minor tweaks are really changing the character now.

Excited. This may turn into a 4 way speaker? .....
Thanks! I assumed as much, I just wasn't sure as Win ISD does know which driver I am using I thought it was a 50 / 50 shot that it would subtract the speaker volume for me.

Thanks for Clarifying!

On a side note. I am finding that the Anthem amp on the midrange driver give a lot of extra sweetness and warmth without changing dynamics much at all.

Quite unexpected to hear a midrange driver in the range of 345hz to 1.21khz make such a difference by changing the amp from transistor to tube.

I think this is magical combination. I may even re wire the anthem to triode mode and see how that works, if I loose any dynamics or not.
Update, new crossover points refined!

I have crossed the scan speak 8535 with a low pass of 384hz 24db/octave, and the focal audiom at 440hz with a 18db/octave filter, sounds a lot better than it did!

I guess with such good drivers even when its not perfect it already sounds really good. Because with these small changes its massive improvements on what already sounds good.

I can tell I'm getting closer to the ideal xover point when each small change causes such a dramatic difference in character.
some updates on the xover process.

In trying to calculate the ideal xover point and considering the recommendations from the founder of the newform research speaker company, I came to a few strange conclusions.

I initially thought that picking the frequency would work best by selecting the same xover point for the high and low pass .

From the definition these are the -xdb down points, so if the drivers are relatively flat in that area they should sum to the reference listening level again.

well the drivers are not relatively flat in the xover region, they are all over the place. So you end up taking a slop which combines with the natural roll off of the driver to produce a new slope. There is no calculating this really, you have to just try all the slopes, and all the points, and move around in frequency and slightly adjust the level of each driver and make notes.

That being said, I have had great success in doing it this way. I now have something which sounds like a high end system. The main criteria I am basing my success on is 3 dimensional soundstage.

You can get an even frequency response from many xover arrangements regardless of phase relationships between drivers. The gross misalignment in this aspect is very apparent. You can set the slopes, and points and levels by trial and error,

But you cannot get 3d imaging, by playing with these alone. Not really at all, because in the drivers are out of phase with eachother, there will be peaks and dips. But they wont affect the tone balance in a gross way. They will affect it in a subtle way. It wont sound like a tone imbalance, it will sound like fuzziness, lack of snap, and no 3d imaging.

So having said all this, here is the procedure I have followed to get very 3d imaging, and very refined mating of the drivers.

No EQ to any of the individual LMH bands. Tried this with real close up measurements and it didn’t sound good.

First measure the entire output on REW and try to center the average of the responses between low mid and high.

Set the low and high pass points of the drivers to be at such a point that you get a desired slope near their roll off. For example on the ribbon I set the lowe pass to 6db/8ve and set it at 1.82 so that by 900hz it is down 6db, but it already has substantial rolloff so that I am down 12 by about 1.2

The midrange I set around 1.2khz with 12db/8ve, then I move the points on one driver up and down and note if it sounds thicker, or thinner, and when its too thick or too thin I go back the other way. I take my time doing this, maybe a few days or a week at this stage, starting all over again a few times in a very different region and always listening fresh and not for extended periods of time.

Now the transient response of each driver successively gets worse as we go from ribbon , to mid, to woofer, so it would make sense that the upper frequency components be squeezed out drop by drop from the faster driver in the xover region. So we want the overall rolloff of the faster driver to be a bit less than the one below it. So we hear a bit more of that driver than the precedeing driver.

The same principle for the next coupling of mid and woofer. And I came up with the following.

So far it sounds great, and I have not even refined it as far as it can go yet.

|Heres the process

Make small 0.2db increment changes up and down , don’t rush, listen to three songs with each change, on each driver, live with it for a day maybe, record settings.

Run the long and short delay correction on the DCX 2496.

Turn the short delays on in the xover setting for each band.

Adjust the time in ms back and for individually by a few turns of the knob, three or four turns max in either direction or +-0.03 ms.

Listen for quick precusion hits, and listen for transient, and steady state separation on the guitar chords.

Now adjust the polarities for fine adjustment.
First try inverting, non inverting, then try 90deg in both, then 45 in both.
Listen to the voice, and listen to high hats, when the voice sounds full, with no grain, the xover position is close. When the high hats have sharp loud attack the phase is correct.

Then the following arrangement of amplifiers and speakers.

Newform R30 ribbon + Blue circle BC2’s

284mm or 0.83 ms
phase inverted 90 deg
DCX Gain +15db
High Pass none
Low Pass butterworth 6 db/8ve1.82 khz

Focal Audiom 6wm + Anthem int 1 (volume knob set to 12 o clock)

Short delay112 mm or 0.33ms
Inverted 145deg
DCX Gain -5db
High Pass 345 butterworth 12
Low Pass Bessel 12db/8ve at 1.21khz

Scan speak 8935 + Yamaha (+16db, not sure if this even has effect with mc input)

Short delay 124mm or 0.36ms
Phase normal 90 deg
DCX Gain -7db
High Pass 20hz butterworth 24
Low Pass butterworth 18db/8ve 243hz

I have tweaked the phase settings a few degrees here and there, and I am absolutely blown away by this setup now. Unbelievable, percussion is unreal. Strings are unreal, imaging is unreal. I never heard a system like this before anywhere.

The snap in the drums, and the fundamental on a drum hit is just so real, I am so happy I decided to put the focal driver in the mix. I can't wait to see what it does when I add in the big woofers!
Last edited:
Ok so after two years of listening to different driver configurations for the main speaker I finally got the urge to do the bass section.

I am using

Focal Audiom 15wx pair
Focal 13w8711 Pair
focal 10w6411 pair

Basically thinking single cabinets for each woofer

Focal Audiom 15wx pair

20hz tuned 9 Cu Ft

Focal 13w8711 Pair

33.7 Hz Tuned 60 L^3

focal 10w6411 pair

Sealed enclosure of 2 L

The only thing I am not sure of is if I should make the 13w woofers in a larger vented enclosed to get more output from them (at the cost of a much greater group delay)

How do I post the graphs? I can only post links ? not jpegs?

There's something wrong with this set-up. What is the main goal? one speaker driver in each box. Is this to work with what driver for the mid section? Are they going to work all together in the same speaker. A few lines for a starting point:

FOCAL AUDIOM 15WX, a starting point would be:
VB = 198.4 L, FB = 21.5 Hz, (198.4L= 7.00ft³)
3x ports with Sd surface area 129.03 cm2 - 7.4 cm x 34.27 cm (D/L).
FOCAL 13W8711, VB = 43.6 L, FB = 28.0 Hz (43.6L= 1.54ft³)
FOCAL 10W6411 - Then the 10" is not a mid and as a woofer should not be in a sealed box, but if working as a mid section sealed then, ~20L. (20L= 0.706ft³)

Note if drivers are used in some kind of 3-way (other) then the all process should be revised (baffles/crossover/BSC) to achieve best results with the right alignments and volume of enclosures. :D
The goal is a speaker that is to be something like a grand utopia, but with a few extra drivers. Crossover complexity is not as much of an issue because the system is completely active. Behringer DEQ , and Behringer DCX employed throughout.

Grand utopia has traditionaly had a 10 inch or a 12 inch crossed over to the 6 inch midrange pair.

I am going to try to really limit the working range of each driver. Also I just happen to have a certain compliment of drivers to use. So My setup will be a bit different.

These will be single drivers per speaker, I do not have pairs for each speaker.
Focal berrliyum tweeter
Focal 6wm mid range driver, (I have two different models to try)
focal 8 inch w cone
focal 10w
focal 12w

I am not really concerned with baffle width influence on frequency response in the bottom three drivers.

The entire system is active with behringer crossovers, and Behringer DEQ. Drivers will be equalized individually and the main speaker as a whole with respect to the room.

I am treating the 15, 12, and 10, as subwoofer component working up to the range of 100hz.

I may however put the 10" speaker in the same cabinet as the main speaker but I am not decided yet.

I want the main speaker baffle to be the most narrow as possible. It just sounds better this way.

So most likely the main speakers will have the tweeter, mid, and 8 inch in a bookshelf type cabinet. Every driver will be a seperate box. for this entire project.

SPace considerations are forcing me to put the three bass unites behind a projector screen, and the bookshelf speaker in front of the screen.

How did you come to the conclusion of FOCAL 13W8711, VB = 43.6 L, FB = 28.0 Hz (43.6L= 1.54ft³) ?

I am going by the focal PDF of the driver which recommends 60 to 70L

I have plotted the response and group delay in win ISD but I cant seem to post it here for you to consider. Any thoughts on how to do that?

Thanks for you interest in the project.
Go advanced, manage attachments and see
that your files are in accordance with this.


  • File type and size.jpg
    File type and size.jpg
    66.2 KB · Views: 303
Not sure I know what you mean by a woofer should not be in a sealed box. Original grand utopias have 10 inch woofers in sealed boxes.

Sealed boxes have better group delay qualities. Because there is a 12 inch woofer crossed over to the 10 inch woofer, the range of 10 inch really doesnt need to be low at all.
I think its rather odd to hear this response "10W6411 - Focal is a good driver for a Transmission Line or Quarter Wave design. (not a sealed woofer)"

I think you guys get very caught up in the words "woofer" and "midrange"

seems kind of funny to me.

The driver only has to play sound from about 60 or 70 hz , to about 90 or 100 hz in this configuration.

It is doing almost nothing. That is the point so many drivers. The main reason no one divides up the frequency band this much is because they use passive x overs. And the drivers sound different. Neither of which are the case here. W cone 12 inches, crossoved over to the same material woofer a bit smaller all the way up the frequency band. The slopes will have to be steep so that maybe a limiting factor.
Last edited:
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.