Part 15 – Open Baffles and Bass
When you think of a speaker, you think of a speaker in a box. Peter Comeau explores the alternatives...
From Hi-Fi World - February 2008 issue
For as long as many of us can remember, loudspeakers have come in boxes, but that wasn’t always the case. With the exception of panel speakers like electrostatics, an enclosure and a drive unit seem inseparable. If we look at the early history of loudspeakers, however, it is littered with, what we now call, open baffle speakers.
Having been brought up in the hi-fi tradition of speakers in boxes I remember being horrified when I first discovered an old radio with a perforated back plate. How could an ‘open box’ possibly produce any bass? Surely the front and back waves would cancel each other out?
Table radios of the 1930 – 1950 era, such as this EKCO 75, used the deep,
Bakelite cabinet housing to act as an extended baffle, normally fitted with a
perforated back panel for ventilation.
Yet bass did come out of this radio, perhaps not very deep bass but bass was present nevertheless! The answer, as I came to find out when I studied speakers rather more intently, is that frequencies are only cancelled when their half-wavelength exceeds the distance between the front and back of the drive unit.
When the drive unit is mounted on a baffle this distance is markedly increased. For example a circular baffle of diameter 215cm will allow strong radiation down to 80Hz and below. This sounds frighteningly large to the home constructor dreaming of attractive wood enclosures in the living room, but all is not always as it seems in acoustics.
For a start only an idiot would make a circular baffle. Who wants total cancellation at the baffle cut-off frequency? Change the baffle to a rectangular one and you can see that there are three dimensions that determine the half-wavelength cancellation, namely height, width and corner to corner. Obviously you will start to get cancellation at the smallest dimension, say the width, but this will be blurred by the reinforcement of radiation occurring at the larger dimensions.
As a result of the differing dimensions, complete cancellation will not occur until a much lower frequency. This might be as much as half that calculated as, at frequencies below 100Hz, room gain becomes considerable.
For example, if we take room gain at 40Hz to be 6dB above that at 80Hz – a not unreasonable premise for speakers placed on the floor and near a side wall – then our 215cm round baffle can generate substantial output down to 40Hz (where it would normally be -6dB measured under anechoic conditions).
This table shows the relationship between baffle diameter (in metres) and the
cut off frequency Fc. Below this frequency output will fall by 6dB per octave
until the driver resonance is reached when the slope changes to 18dB per
octave. For a circular baffle there will also be a peak where front and rear
radiation become coincident, shown here as Fp. There will also be further
peaks at multiples of Fp which is why, in general, circular baffles
should be avoided!
Now we are getting somewhere, and it starts to all make sense when we look at panel loudspeakers like the Quad Electrostatic where the baffle width clearly isn’t over 2 metres! By increasing just one dimension, say height, and allowing the speaker to couple to the floor and, possibly, the side wall, strong bass output can be achieved to satisfy even the organ enthusiast!
I’m not going to claim that you’ll achieve the same bass power in the room from an open baffle speaker as you can from, say, a transmission line speaker of similar overall dimensions. As my wife pointed out when hearing the reproduction of a Bach organ work on the radio “organ music sounds better in a church where you can feel the power of the low notes”.
I know what she means. That ability to really move the air, so that it has visceral as well as audible impact, is something lacking in most hi-fi systems, box speakers or not. But you do need large speakers to really make it happen.
THE BOX IS MISSING!
So, considering that most people like small speakers in their living rooms, what is the point of pursuing the open baffle? I’ll tell you – it is that the box is missing!
Despite the best intentions of the loudspeaker designer in providing bracing, damping, internal absorption, adding a port, a horn or a quarter wave pipe, there is no getting away from the fact that putting a box behind a speaker just encourages resonance.
Now most of us grow up hearing these box resonances from every loudspeaker we listen to. So we are used to it. In fact I formulated a theory many years ago that we are so used to hearing box resonances that the sound seems ‘wrong’ when you take them away.
That might partially explain why panel speakers are often described as ‘thin’ or ‘lightweight’ by some listeners on first hearing an electrostatic. But when you talk to panel loudspeaker adherents they will be the first to describe box speakers as sounding plummy and coloured.
Theoretical plot of driver response on a circular baffle. Note that peaks occur at Fp and its multiples.
Is there somewhere in between? Would it be possible to have one’s cake and eat it? At one point in my youth I was exposed to the sound from Gilbert Briggs Wharfedale SFB3. This Sand Filled Baffle loudspeaker with its three drive units gave the same impression of bass power as its ported competitors but without their lumpy, ill defined bass and coloured midrange. Clearly the open baffle system can be made to work.
Our first problem when looking at how to design an open baffle speaker is that of drive units. Nearly every single drive unit manufactured commercially today is destined to end up in a box of one sort or another, so it is hardly surprising that drive units from all the major manufacturers are developed to make the most of an enclosure, usually a ported one.
In Briggs day this wasn’t so. With little conception of the exact science behind the relationship between compliance, mass, damping and Helmholtz resonance, drive unit designers of the ‘40s to ‘50s made sure that their units would produce good power at low frequencies without falling to bits on an open baffle. Hardly surprising, then, that they were suited to unreasonably large enclosures, whether ported or closed box types, if you wanted to maintain their bass characteristics.
Plot of 100mm driver on a 300mm circular baffle showing the relationship between Fc,
Fp and the midrange power level. Circular baffles are to be avoided!
Effect of driver Qts on LF response. As Qts is increased, so is bass output towards
the driver free air resonance. A Qts of around 1 is acceptable for an open baffle bass unit.
But there is one area of the industry where bass drive units still have to be stable in open, or semi-open, cabinets just like the table radios of old – the Guitar industry! When it comes to guitar combo cabs you can still see ten or twelve inch monster drive units in virtually open back cabinets pounded to near destruction on stage by rivetingly hard transients.
Of course, in the main, these units are not suitable for hi-fi use. The ideals of low distortion and vanishingly small levels of colouration are not foremost in the minds of guitar speaker designers! Occasionally, however, one finds a gem amongst drivers made for the Pro-Audio market and, when you do, it is usually no problem adapting it to an open baffle speaker.
CHOOSING A BASS DRIVER
So what, exactly, are we looking for from a driver that makes it suitable for our Open Baffle? The first of our interests is excursion, particularly excursion through a linear operating region. The reason this is so important is that there is no restoring force on the diaphragm as there would be in an enclosure. If you are considering adding bass equalisation to extend the response of your driver, then its excursion limits become more critical as you ask the driver to do more work as the frequency is lowered.
Goodmans Axiom 150 dating from the late 1950s is typical of large bass drivers
of its day (dust cap and whizzer cone removed in this example). These were ideally
suited to large, aperiodic, cabinets and Goodmans marketed one of the first
commercial aperiodic vents!
Compare to a modern Eminence ProAudio driver – not too much difference?
Next we need to look at the Qts of the driver which is defined by its moving mass, the springiness of the suspension and the electrical damping from its magnetic system. Again, for a speaker in an enclosure, we would normally select a Q > 0.3 for a closed box speaker and a Q < 0.5 for a reflex design.
For an Open Baffle speaker using a passive crossover we actually require a driver with a high Qts. A Qts > 0.77 will mean that the amplifier will find it easier to drive the speaker as the frequency becomes closer to the system resonance. By careful choice of resonant frequency we can use this to extend the response below the baffle cut off frequency.
Finally the search for low Fs, the frequency of fundamental resonance of the driver, goes out of the window. The only reason we look for a low Fs in a driver destined for use in an enclosure is because we know full well that the resonant frequency will be raised considerably, in many cases doubled, by the reactance of the air in the enclosure. Thus, for a bass reflex or closed box speaker we normally look for an Fs of 30Hz or lower to achieve a system resonance, once the driver is in the box, of below 60Hz.
Plots of driver mounted low on baffle, in proximity to floor, showing
cancellation and reinforcement effects between direct and reflected sound.
Our Open Baffle, however, is not going to influence the driver resonance in this way. So we can choose a driver Fs knowing full well that this is going to be close to our final system Fs. Figures in the 40-60Hz region will do just fine.
So what we have outlined is a drive unit that doesn’t normally exist – good linearity with controlled excursion, high Qts, and a relatively high Fs, oh, and high sensitivity too!
That’s why drivers designed for Pro-Audio use suddenly start to look attractive. We can start investigating the 12in (30cm) and 15in (38cm) drivers that would normally need a whopping big enclosure in order to achieve any bass extension. Initially these drivers look so wrong for hi-fi – lightweight paper cones, folded ‘accordion’ surrounds, limited Xmax – but remember, we don’t need these drivers to do the same job as if they were in an enclosure.
Roll surrounds allow big excursions, necessary in a reflex loudspeaker to handle the considerable cone movement. Similarly the huge cone excursions, defined as Xmax, necessary for smaller diaphragm speakers just are not required with a 30cm cone or above. For the small movements of the cone the folded surround, which provides a good restoring force, starts to look like a good idea.
Of course these driver features are exactly what you will see in speakers dating back to the 1940s and ‘50s. We’re not reinventing the wheel here! That doesn’t make them any less relevant in a modern loudspeaker system. We just have to choose the engineering of the rest of the speaker carefully.
WHAT ABOUT FULL RANGE
At this stage, up crops the usual interested parties from the ‘full range driver’ fraternity. Yes, many of the features I have described do accord with the smaller full range drivers too. But, as many DIY speaker experimenters have found out, the small diameter Fostex just don’t produce ‘real’ bass when the music asks them to. Yes, you can stuff them into quarter wave resonators to ‘enhance’ what little bass output they have, but on an Open Baffle they just cannot move enough air on their own.
Augment a full range unit with a big bass driver, however, and you can have your cake and eat it! There’s another reason for running the two units together and this has to do with the ideal position for the drivers on our Open Baffle.
Wharfedale SFB3 rear view. Note cutouts in side ‘wings’ to reduce resonance
and bracing to improve stability. As expected the bass unit is positioned closer
to the floor boundary and the midrange unit higher up. A cone treble unit, on its
own baffle, points upwards at the rear!
Go back to the beginning of this piece and you’ll see that I mentioned how the floor, and possibly side wall, comes into our baffle design when we put the speakers in a real room. It doesn’t take a nuclear scientist to see that, if we put the bass driver close to the floor, the baffle is naturally extended by the floor. This leaves a three sided baffle and, by putting the driver close to the floor, the vertical dimension of the baffle is increased.
Now that doesn’t mean that you can put the driver bang against the bottom edge. You do have to consider the floor reflection which, if it occurs too high in frequency, will notch out the upper bass response of the driver. As a compromise, however, this position for our drive unit has considerable gains and few detractions. Except one.
If you want the driver to be full range, the last place you should consider is close to the floor. Various reasons crop up as we start to consider the pros and cons, the most obvious of which is that the driver is not at ear level. Full range drivers have relatively poor dispersion at high frequencies and really do demand that they are positioned somewhere around ear level for a seated listener.
Next, remember that the floor reflection is going to occur somewhere in the midrange – there’s no avoiding it – so the further away from the floor the better. If you have carpet on the floor under your speaker, and who doesn’t these days, then this will have a considerable damping effect on the mid and high frequency output of any drive unit near it.
All of which points towards a two-way or three-way speaker. We can position the bass unit where it gains most benefit from the baffle and floor extension, leaving the mid/treble to be housed where it/they ought to be.
Celestion's SL6000 subwoofer, seen here beneath an SL600, had no
cabinet and reached low, but needed an electrical equalisation network.
THE PERILS OF HAVING WINGS
There’s only a couple more aspects that we have to consider before embarking on the final design of our Open Baffle. One is should we have ‘wings’ folding back from the sides of the baffle? The simple answer to this one is ‘yes’ with a proviso.
Folding back the sides of the baffle obviously extends the baffle and helps lower the cut off frequency. It is especially useful for increasing the lateral path length and making it considerably different to the vertical baffle dimension, all of which helps to ‘blur’ and smooth out the baffle step in the frequency response.
In addition these side ‘wings’ increase the baffle’s stability. How else are you going to keep it standing up? Watch out, however, for the reflective standing waves that are set up by these folded sides. Think of it in terms of a box with the rear panel missing. What you will hear, with deep ‘wings’ is a hollow, resonant honky nature to the midrange that is really very unpleasant, especially on vocals and piano. You would be better off putting the drivers in a box rather than have deep side wings – at least then you could include some interior absorbent to reduce the effects of the resonance.
A modern take on the open baffle loudspeaker - Jamo's flagship R909.
So keep the side folds shallow and taper them so that any standing waves don’t occur at just one or two audibly obvious frequencies. The attraction of tapering the ‘wings’ in a two-way system is that the bass unit receives the maximum benefit from the extended baffle while the midrange unit, which ‘sees’ the narrower section of the taper, is more ‘open’ and so suffers less reflection.
I’m off now to choose some suitable drive units and, as this coincides with the Christmas break, I’ll be back in the first quarter of 2008 with our first Open Baffle design kit.
Happy New Year!