[D OB G]

Some of this has been posted on the full range forum of diyAudio.

Apart from using a high Qt driver (say 0.7 to 1 to 1.2 or even up to 1.6 in some examples), in a conveniently sized baffle, there is another approach, that I use in my designs.

The woofer is wired in series with a large inductor, or, to reduce the impedence, and therefore the size of the inductor required, two 8 ohm woofers are wired in parallel, which obviously doubles the radiating area and reduces the cone excursion. (Using a single 4 ohm woofer helps with the inductor size at least).

Various parameters can be manipulated. The usual 6 dB per octave loss occurs according to the baffle size, which is equalised by the large inductor.

A low fs speaker is chosen, and the size of the inductor is made to accord with this frequency (this is only an approximation, as the impedence characteristics around the resonant frequency are obviously not constant). A typical inductor might be 20- 40 mH !!! (I still use air-cored inductors- quite a size and weight) depending on the woofers. The series resistance of such an inductor with appropriatedly guaged wire might be 2-3 ohms (this can be increased if necessary by using thinner wire, being careful of power constraints, or by adding a resistor).

So by chosing woofers with the appropriate Qt of somewhere between say, 2.8 and say, 3.8, a perfectly damped low frequency response (your choice of 0.56, 0.6, 0.7 e.t.c.- or whatever) can be obtained with a flat response between the inductor high pass frequency (at a 12 dB per octave drop-off below the -3 dB point at the resonant frequency), and what becomes the baffle width determined low pass frequency (at 6 dB per octave- ready to crossover to a full-range or mid range at 6 dB per octave).

This gives great design control, with the only downside being a loss of efficiency. Using something like a MLSSA (I use an old "Imp" from Liberty Instruments) changes can be readily measured. I have been building speaker like this since 1978, and I suspect that the new Open Baffle design by Jamo may use similar principles.

The baffle for the FR or mid-range may need to be slightly smaller, depending on the characteristics of the drivers.

This means there is no such thing as a specifically OB driver.  The main requirements are good linearity with a good Xmax, high efficiency, large total radiating area (multiple drivers ?) with benign cone break-up characteristics- as far as I can see, never achieved with an 18" driver, and unfortunately rare with a 15" driver (remember we're not talking high order crossovers here, or active equalisation, which I also use on some designs).

This the bare bones of the concept.  There are further sophistications, which include: method of magnet/mass retention, baffle shape, baffle materials, edge treatment.

Some advantages of the design are that there is minimum resonating material, and the baffle needs to be narrow by design- the width of the mounting of a 15" driver is ideal.  A disadvantage is that ultimate SPLs
are less than with a larger baffle, but in practice this is not a PA system, and has never been a problem, and sensible amplifier power is needed (although I only use 70 W per channel).

As is very apparent from this forum, there are a multitude of other considerations involved in OB design (some mutually exclusive).  This design approach may have already appeared in this forum, but I can say that it works for me. 

[scorpion]

I think I understand your approach but please give an example from real life. I find your text a little hard to penetrate.
There will be a speaker roll off below Fs, there will be an inductor low-pass 6 dB/octave slope and there will be a dipole roll off at some
frequency depending on baffle size and finally there might be a bass equalization effect depending on raised Qe from the inductor and a possible
resistor. I am probably missing something !

/Erling

[D OB G]

Hi Erling,

Say a baffle has a cut off freq of 150 Hz (at 6dB per octave).

A large inductor that has a turnover frequency of, say, Fs, say 35 Hz, is placed in series with the woofer.

Now the 6dB per octave slope of the inductor exactly balances the 6dB slope of the baffle between 35 and 150 Hz.

Below 35 Hz we have the driver role off, 6dB per ocatve, and the baffle roll off, 6 dB per octave, giving 12 dB per octave below Fs if that is what the inductor is tuned for, or a compound rate below the inductor turnover freq if that is higher.

Above 150 Hz there is no baffle roll off, and the inductor is still rolling off a 6dB per octave, therefore so is the woofer.

By the way, I said that fine tuning of the resistance of the inductor can be had with the addition of a resistor.  In practice I make that another small inductor in series.

I hope that helps.

David

[scorpion]

Now the 6dB per octave slope of the inductor exactly balances the 6dB slope of the baffle between 35 and 150 Hz.

This cannot possibly garantee flat bass response between 35 and 150 Hz ? 
To me the inductor is a low-pass device taking down bass response with 6 dB/octave from 35 Hz at the same time the Baffle is diminishing Bass output 6 dB/octave from 150 Hz downwards in the other end.

/Erling

[D OB G]

Hi Erling,

Perhaps try drawing a graph.

With the inductor, start at 0 Hz and draw a flat line to 35 Hz, at 0 dB, then slope it down to the right at 6 dB per octave indefinitely.

With the baffle start at 150 Hz and draw a flat line indefinitely to the right, at 0 dB, but to the left slope it down at 6 dB per octave.

If you superimpose these lines you get a flat response between 35 and 150 Hz, albeit the bass is x dB down (~6 dB in this case), which is the efficiency penalty.

David

[scorpion]

D OB G,

But you cannot add these graphs together in that way. Let's go out from the bass unit, Level at (or may be - 3 dB) 35 Hz, - 6 dB at 70 Hz, - 12 dB at 140 Hz by the inductor low-pass. From the other side - 6dB at 75 Hz and - 12 dB at 37.5 Hz from the Bass roll off.

/Erling

[D OB G]

Hi Erling,

Perhaps a better way to draw a graph would be a "straight line" from 0 Hz to, say, 300 Hz.

Draw the "inductor line" from 0 Hz,  6 dB above the "straight line" till 35 Hz, then slope it down to the right at 6 dB per octave through the 70, 75 Hz point, and continue sloping.

Draw the "baffle line" from 150 to, say, 300 Hz 6 dB above the "straight line", then slope it down to the left at 6 dB per octave through the 70, 75 Hz point, and continue sloping.

At every point between 35 and 150 Hz, the amount above the "straight line" is equal to the amount below, with the straight line being ~6 dB down from the "inductor line".

David

[scorpion]

D OB G,

Now you are beginning to talk. I was triggered by your reference to the JAMO OB.
Why draw the lines as you suggest ?

/Erling

[MJK]

If you have a driver with a higher Qts, the impedance curve around the resonance is going to be peaking significantly. How are you going to size the inductor for a nonconstant impedance and still produce the constant 6 dB roll-off? The only way I see to accomplish this is by going active where the drivers impedance does not impact the "crossover" filter component sizing. Doing this would also require a highly efficient woofer to be able to give away significant SPL and still be able to transition to a midrange or full range driver.

[scorpion]

MJK,

That's why I think the Jamo is a very remarkable design ! 

/Erling

[D OB G]

Hi MJK,

Obviously Linkwitz goes active, and uses his Peerless XXLS drivers to compensate for the SPL (even though they are quite inefficient).

To answer your question, that's why I go for middle Qt drivers, to avoid the resonant peak, and to give room to fine tune the inductor impedance and dc resistance.

I still need efficient, long stroke drivers.

With a Qt of 0.3 and an inductor of 32 mH (higher than would be calculated using a constant impedance), Fs of 29 Hz, I am -3 dB at 35 Hz, (as against the theoretical 29 Hz) and flat (really flat) to the baffle cut off.

David

David

[MJK]

To answer your question, that's why I go for middle Qt drivers, to avoid the resonant peak, and to give room to fine tune the inductor impedance and dc resistance.

I was referring to the electrical impedance peak around driver resonance, this will be quite a tall peak. How do you size the inductor to account for this very non constant electrical impedance right in the middle of the intended range of operation?

[scorpion]

I do not think I can contribute anythingmore, but to think that this Jamo design actually is quite something.

/Erling

[D OB G]

Hi MJK,


I agree, I do have a theoretical problem.

In this design I am using 2 X Peerless extra long stroke woofers 830491 (no longer available). I don't have the impedence plot.

I have the impedance plot for the 830869, which may be similar, may be not.

At the -3 dB point, at 35 Hz, imedance is 35 ohms.

Using 2 drivers, at the -9dB point the impedance is about 18 ohms.

Using my formulae, the inductance should be about 82 mH.

I am using 32 mH, which is obviously lower than the impedance calculation, and higher than the DC resistance calculation (27 mH).

I can't explain the difference, except to say that it works (I've been designing OB speakers this way since 1978).  There is also the effect of the DC insertion of the inductors to explain!

I Invite readers to try some simple tests on small baffles themselves, if they have some large inductors floating around.

To calibrate this process, I made a variable tapped inductor, which gives the empirical results I've come to expect.

I admit that I can never quite make the theoretical -3 dB point at Fs, but I come mighty close.

Does anyone have a theoretical basis for what I find empirically?

David

[JohninCR]

David,

I don't understand why anyone is disagreeing with your approach.  I use active filters, so I don't have to worry about the electrical problem of large inductors, but a 6th order low pass filter that starts very low flattens response nicely for a woofer below the baffle cutoff.  The reason you end up using much smaller inductors is that their effect isn't linear and starts much lower.  Plus it's unlikely that the driver's natural response (disregarding dipole cancellation) is flat down that low anyway.  The bottom line is that a 6db/oct low pass filter offsets the 6db/oct dipole cancellation, and it is especially useful with smaller baffles because the flattening effect extends to a higher frequency resulting in more flexibility in where and how you cross to the next driver.  As long as you are bi-amping, it makes things pretty easy.

[scorpion]

David,

JohninCR is right, some erroneous thinking of mine yesterday. But you will of course have a level and a dynamic problem if you are starting from
- 12-15 dB over the 40 - 140 Hz range and try to mate the bass with a potent fullranger or midrange. I am still a bit obsessed by the Jamo solution.
You are right, I think, in your assumption about their crossover, so I am sitting with paper, pencil and ruler trying to catch some of the spirit in the design. 
The crossover figure in their R909 White Paper is not very accurate.

/Erling

[MJK]

The crossover figure in their R909 White Paper is not very accurate.

Erling,

I had not really paid much attention to this speaker. This morning I looked a little closer at the information available on the Jamo website and found it very interesting. It would not be too difficult to reverse engineer the design and come up with something similar using readily available drivers, no magic in what they are doing. A DIY design would definitely be a lot cheaper. Simulating the design and coming up with a passive crossover concept that does the same thing should not be a problem. I am already working on a design for a two way, that uses a similar approach to reduce baffle size and still provide decent bass, so a bigger three way with multiple bass drivers would only be an extrapolation to something bigger.

Martin

[scorpion]

MJK,

Yes, interesting speaker, isn't it. And they have managed to go home with 89 dB sensitivity. I have heard the Jamo. The Bass is very good althoug the dipole cut off is something like 160 Hz.  I am not really a fan of the Seas magnesium unit. I think my B200 is way better, for instance, so there is scope for improvement. But it is the sensitivity that impress, which means they have worked a lot with the bass-units. I like the concept of coming up with a good 2- or 3-way all passive DIY OB design.

/Erling

[D OB G]

Hi Martin,

Re the Jamo:

I've had a set of OBs going for a few years now that use 4 X Kef B139s per side, re-voicecoiled to 16 ohms each. (Same radiating area as the Jamos).

The midrange is the Seas Excel M15CH001 (the immediate precursor to the driver Jamo uses- mine is filled poly v's magnesium? -flatter response anyway with the polypropylene.)  One virtue of this (these) drivers is the neodymium magnet assembly which would have to have the least rear wave obstruction of any driver I have seen.

I use a Dynaudio Esotar v's Jamo's Revelator (you takes your pick).

The whole system is active, with active equalisation, crossing over at 125 Hz v's 250 for the Jamo I believe.  The dynamic range is stupendous.

The baffle is a bit wider than the Jamos to accommodate the B139s (-3 dB at 25 Hz, same as the Jamo), and was built when I believed in massive baffles (can be taken apart, otherwise takes three people to move it).  At least the woofers were magnet mounted (as well as baffle mounted).

While a lot of people have talked about the Jamos, I'm one of the many who have yet to hear them!

I still find the task of passive inductor equalisation a more interesting challenge.

David

[MJK]

But it is the sensitivity that impress, which means they have worked a lot with the bass-units.

Erling,

My passive two way design is calculating to be at 91 dB with bass down to 50 Hz before rolling off and less than $300/pair.

Martin