Banner
Banner
cookie-banner


Designing

Loudspeakers

 

sfb3ad

 

Open Baffle speakers – the trials


Last month I investigated a classic Open Baffle speaker from the past – the Wharfedale SFB/3 – with regard to its performance in the room. What I found was that the designer, Gilbert Briggs, had mapped out the best position as being alongside a side wall, spaced 1 metre from the nearest corner.

What was interesting about Gilbert Briggs approach was that his aim was for a high efficiency system that I actually measured at 97dB for 2.83v (1 Watt/8 Ohms) input. Yet the fundamental resonance of the bass unit, with its heavy, pulp cone and massive Alnico magnet, is down at a low 25Hz.


This flies in the face of the way most DIY open baffles are configured. Here the concept is again one of high efficiency but achieved by using a lightweight cone and a fairly small magnet to deliver a high ‘Q’ system. This high ‘Q’ driver maximises bass response towards the lowest frequencies, though these are usually limited as much by the fundamental resonance of the driver at around 60Hz as by the limiting size of the baffle.


I was curious. Which is the best route to follow – the Brigg’s way or the modern high ‘Q’ strategy? There was only one way to find out.


graph


‘Q’ AND SYSTEM RESONANCE

First a bit of background as to what we mean by high ‘Q’ and how it affects what we hear. I was provoked in this by a conversation with our esteemed publisher, Noel Keywood, where we ended up discussing the merits of correcting or equalising the impedance ‘humps’ that we see when measuring a bass reflex speaker. Bear with me and I’ll explain why this is relevant.


The ‘Q’ of a resonant system indicates the sharpness of tuning. Originally designated to show the Quality of a tuned filter in radio work it was originally assumed that the higher the ‘Q’ (Quality Factor) the better the tuning quality.

It was quickly found that the value of ‘Q’ could be a useful factor in designs other than electronic tuning circuits, loudspeakers being an obvious candidate. In a drive unit design, for example, there are ‘Q’ values for electrical (Qes) and mechanical (Qms) factors, both of which are summed to give the total system ‘Q’ (Qts).


The value of Qts is one of the Thiele-Small parameters that can be used to calculate the enclosure design for the drive unit. Without knowing this value of Qts you are basically flailing about in the wind when it comes to predicting how a drive unit and associated enclosure are going to behave sonically.


Similarly the drive unit and enclosure combination also have a system Q (Qtc) which adequately describes the way the speaker will behave at low frequencies when the system resonant frequency (fc) is taken into account.


speakerqtc1i

Impedance graph of the small, closed-box bookshelf loudspeaker (Rogers LS3/5a)

showing evidence of the high ‘Q’ design in the tall, narrow impedance peak at the

system resonant frequency.


Now amongst speaker designers there are some well know ‘truths’ when it comes to the desired system Q and a lot of deliberation goes into what level of Qtc should be aimed for to meet commercial considerations.


For example a speaker with a Qtc greater than 1.0 tends to have a robust, thudding bass quality that will make a cheap speaker highly saleable. Where the enclosure is on the small side a high Qtc of 1.2 will have the effect of subjectively extending the bass response below the system resonance. You can see the effect quite clearly in this graph Note how, as the system resonance is approached, the output of the speaker is apparently magnified. This ‘hump’ in the bass response is nowhere near as audibly obnoxious as it looks. Would you believe these traces are from a BBC Ls3/5a design? Well they are, and we all know how listeners revere the Ls3/5a.


speakerqtc1

Bass response of a small, closed-box bookshelf speaker showing

the increase in bass output from a high ‘Q’ design.


Then would you also believe that the Qtc of the Ls3/5a is 1.2! Look at the way the impedance peak at fc is very narrow and shoots off the top of the graph. It just goes to show that you can’t trust measurements to tell you the whole story of how a speaker ‘sounds’.


However the LS3/5a is a closed box loudspeaker. So was my design of the Heybrook HB1 which had a ‘Q’ of 1.1. Both of these speakers belie their size with their apparently extended bass response. Both ‘extend’ the subjective response by using a high Qtc, whilst the back pressure of the air inside the box keeps the drive unit under control.


q-graph

Effect of system Q on the bass response from a closed box design.

The vertical graph axis of 1dB exaggerates the apparent effects which

don’t sound as bad as they look!



‘Q’ AND THE BASS REFLEX LOUDSPEAKER

What happens when you apply this high Q approach to a reflex loudspeaker? Well, usually, disaster. In a bass reflex design the fundamental resonance fc is determined by the mass and compliance (springiness) of the driver, the volume of the enclosure and the tuning of the port (which also has mass and compliance).


bass-reflex-alignments

Low Frequency alignments for bass-reflex systems. SBB4 typically uses

a large box with low tuning frequency and has the best transient performance.

QB3 is commonly used for smaller enclosures because it has the highest efficiency

even though it has the worst transient response.


bassrefleximp

Impedance of a typical bass-reflex loudspeaker. This graph is measured using the bass unit,

so there is a ‘null’ (dip) in the impedance where the port is doing all the work. This null gives the

typical reflex speaker ‘twin hump’ appearance where the dip in the graph is the tuning frequency

of the speaker system.



As the system approaches its resonant frequency fc the driver has less work to do, and the port does more. This, you might think, is a good thing because the displacement of the cone is small and therefore the linearity of the cone is good. On this basis it would make sense to tune for a high ‘Q’ for maximum output, and therefore bass assistance, from the port at system resonance.


openbaffle15a

Bass response of open-baffle speaker fitted with Eminence Alpha-15A in free space.


openbaffle15b

Bass response of open-baffle speaker fitted with Eminence Beta-15B in free space.


But what inevitably happens, when you do this, is that the speaker sounds lumpy, boomy and has a prevalent ‘one note bass’ effect. Part of the problem is that you can often hear the ‘honk’ of internal cabinet standing waves exiting through the port. In addition the vast volume of air moving in and out of the port causes ‘chuffing’ and this distortion quite obviously attracts attention to the single frequency of tuning. Also a bass-reflex design with a Qts of 1.0 is a 24dB/octave high pass filter exhibiting its associated ringing and phase shift characteristics. This often compromises the transient performance of the system.


You’ll also observe something much worse than this, however. A typical characteristic of bass reflex boxes is massive cone displacement at low frequencies. How can this be if the port is doing the majority of the work?


lfopenbaffleq

Effect of system Q upon the bass response from an open baffle speaker.

Note this is an anechoic response graph and takes no account of boundary

reinforcement or room gain.


At frequencies below system resonance the bass unit is free to move without restraint from the air in the enclosure. Where low frequencies are present in the source (record warps from turntables/infrasonics from CD) drive unit cone displacement can push the motor system out of its linear region with disastrous effects on midrange distortion.


As a result many designers deliberately ‘de-tune’ a bass reflex system or introduce damping to lower the Qts. Unlike a closed box system one does not aim for a particular ‘Q’ value but, instead, for a bass ‘alignment’. These provide variations from ‘critically damped’ through ‘flat amplitude response’ to ‘maximum efficiency’. Which you choose depends on how you want the speaker to sound.


So how does any of this relate to our open baffle design? The answer lies in a combination of what we have learnt from the closed box and reflex box performance characteristics.


Because the driver is sitting on an open baffle the Qtc is similar to the Qts of the driver modified only by the increased radiation resistance of siting the driver on a baffle. Thus, if we take the Qts of 1.26 for an Eminence Alpha 15A we can expect a high Q system from our open baffle.

Just like the bass reflex system the cone will have no restraining influence below the resonant frequency, there being no back pressure on the cone from a rear enclosure as there would be in a closed box, so where there is considerable bass output from the source the bass unit cone displacement can be considerable.


However the slope of an open baffle below resonance is a theoretical 18dB/octave so it does not suffer so much from transient ringing and phase shift as does the bass reflex. Despite what you might think from an apparently underdamped driver with a high ‘Q’ there is no more transient overhang with this driver than with its more critically damped Alpha 15B brother.


So, given this, let’s look at what happens if we put the Eminence Alpha 15A in a baffle of similar size to the Wharfedale SFB/3 in our listening room.

In the free space position the bass response looks, and sounds, fine and dandy. The high Qts of 1.26 from the driver provides good extension from our open baffle down to the driver resonant frequency at 45Hz. Only the dimensions of the baffle cause the overall power output to fall away below 100Hz.

But what has happened to the specified high efficiency of this driver, according the manufacturer measuring 97dB? Unfortunately this is the upper midrange sensitivity, not that in the bass which, as you can see, is closer to 90dB.


Now compare this to the free space response of the Eminence Beta-15B. The lower Qts of 0.58 means that the bass extension is not handled as well in terms of power into the room, though it doesn’t sound as bad as it looks. Sensitivity, for example, due to the larger magnet system is improved in the upper bass region to 92dB.


openbaffle15c

Bass response of open-baffle speaker fitted with Eminence Beta-15B

abutting side wall of room, 1m from corner.


To my ears, in terms of sound quality, the overall performance with the Beta-15B, though obviously mid-forward, exhibits a more articulate bass quality with better control. This is hardly surprising, considering the bigger magnet system. I can see why the Alpha-15A does well with OB constructors as its bass to midrange balance is obviously superior, but I felt there was further to go to realise the best advantages of open baffle performance.

So, copying Gilbert Briggs advocated position in the room, here’s the response with the Beta-15B with the baffle situated up against the room side wall, 1m from the corner.


Now we can see that the drive unit is coupling better to the room with an associated improvement in bass to midrange balance.

We can see from these graphs that your choice of bass unit for an open baffle design is largely going to depend on whether you want to stick to free standing or close-to-wall configuration. My preference is for the latter because, to my ears, the superior room coupling in the bass, and better control of bass output from the more damped motor system, accords more closely to my loudspeaker design aims.


But is this the best we can do? Find out in next month’s instalment!


nickob4


Typical enthusiasts free space placement of open baffle speakers

(picture courtesy World Designs Forum).

 

Add your comment

Your name:
Subject:
Comment:
  The word for verification. Lowercase letters only with no spaces.
Word verification:

Search

Hi-Fi World, Powered by Joomla!; Hosted by Joomla Wired.