Home Built Subwoofers for the
LS3/5a
 |
This project would have been
impossible without the help and support of Jo Ki, Andy Whittle and Ken
Kessler. As well as inspiring me to add some subs to my LS3/5as, Jo went
way beyond the call of duty and opened up his AB1s in order to check the
dimensions I had calculated. Andy is the designer of the original AB1
and following an approach from Ken, very generously agreed to supply details
of the AB1's crossover for inclusion here.
Although I have owned LS3/5as
for perhaps 25 years I have to confess that it was only this year (2000)
that I first heard them partnered with the Rogers AB1s. I was very fortunate
to receive an invitation to visit Joseph Ki when I was in Malaysia during
August. Jo's system really knocked me out, it sounded so good that I was
very tempted to buy a pair of AB1s and bring them back to the UK. The
price of AB1s there was very good compared to the UK and finally it was
simply lack of sufficient luggage allowance and the risk of a nasty excess
baggage charge that stopped me doing that
|
|
On my return to the UK, without
the AB1s, it dawned on me that I might try building some subwoofers based
on the concept of the excellent AB1s. I already had a spare pair of the
B110 drive units and so I decided to find out as much as I could about
the Rogers' design.
Quite a lot of information
about the AB1 is available from reviews and from Rogers' own literature.
The design is a fourth order bandpass subwoofer with -3dB point at 55Hz.
The LS3/5a is fed from the subwoofer through a second order high pass
filter which takes a lot of the strain off from the LS3/5a and as well
as extending the frequency response downwards also allows a couple of
dB extra spl from the combo.
|
 |
An email to
KEF resulted in the data sheet for the B110 arriving next morning in the
post (thank you Ron). It's great to know that KEF still provide excellent service
for amateur constructors. The B110 shown here is the SP1003 unit which was used
in the 15 ohm LS3/5a, but any of the B110 series will work fine in this application
and the SP1228 version is usually used.
|
Design information for sub
woofers including Fourth Order Bandpass systems is available from
the DIY Subwoofers pages which provide an excellent resource for
sub woofer builders. From this site a design spreadsheet written in Excel
called "Bandpass" can be downloaded.
|
 |
"Plugging in" the values for Vas,
Fs and Qts from the B110 data sheet and using a value of -3dB at 55Hz into "Bandpass"
gave me values for Vf, (the net volume of the vented section of the enclosure)
and for Vr (the sealed section).
The external dimensions are 570mm
high (the original AB1 has a 25mm plinth bringing the height up to 595mm), 190mm
wide and 160mm deep.The cabinet walls are made from 18mm MDF (also own as Medite)
and the internal baffle for the B110 drive unit is made from 12mm MDF.
| Note that the drive
unit is a little too big to fit in the cabinet. To overcome this the original
AB1 has a short slot cut in both the front and back panels. The slot is
shown in Jo Ki's drawing. The rear panel is slotted to accept the 12mm baffle.
In order to ensure that the seal between the two compartments is airtight
when the back panel is screwed on the rear edge of the 12mm baffle has a
rubber strip applied. I used self adhesive foam draft excluder strip. |
|
|
A local timber merchant cut the panels
from a single sheet of 4ft x 4ft MDF and the panels (with the exception of the
back) were glued and screwed together. All joints were sealed using frame sealant,
it is vital they are airtight.
 |
The AB1's port is 4.2cm in
diameter and 8.5cm long, I hunted round for some plastic tube 4.2cm in
diameter before realising that the plastic tube that the frame sealant
had come in was 4.6cm inside diameter. Re-calculating the port gave a
new length of 10.5cm for this diameter, an excellent result given that
the frame sealant tube was exactly 21cm in length!
|
| Car body damping pads were applied
to all internal cabinet walls and on top of these (12 - 25mm) high density
foam was glued. The drive units were mounted onto the baffles using grommet
strip to make a good air seal. A loose roll of foam is then placed in the
bottom section of each cabinet. It is important that the rear panel makes
a good air tight seal and so soft rubber draught excluder strip was used
on all the edges. Foam draught excluder strip works well too. |
 |
Andy Whittle, the designer
of the AB1 kindly supplied the circuit of the AB1 crossover for publication
here. Thanks both to Andy and to Ken Kessler for their help.
I was unable to obtain 16mH
inductors off the shelf so I used 15mm Ferrite Cored inductors obtained
from Wilmslow Audio and
increased the value of the capacitors to 240uF to keep the crossover turnover
frequency the same. Wilmslow Audio also sell B110 drive units, damping
pads and suitable non-polarised electrolytic capacitors.
|
Stirling
Broadcast can supply high quality "E core" 16mH crossover inductors and
may also be able to supply other AB1 spare parts including the vent tubes and
drive units. The Stirling crossover inductors are available as an upgrade for
owners of real AB1s. I intend to test a pair in my home made crossovers soon.
The
crossover components were mounted on a simple PCB. Unwanted areas of copper
can be removed with a craft knife. Components are mounted on the copper side
of the board and the ferrite inductor is secured to the PCB using cable ties
through 4 small holes drilled in the board. I mounted my crossovers in small
ABS boxes on the rear of the cabinets because in the future I want to experiment
with bi-amping and active crossovers. In the real AB1 the crossover PCBs are
mounted in the sealed section and a second set of terminals feeds the LS3/5as.
I ran the AB1s with the standard
crossover circuit for about a year before deciding that i wanted to see if I
could improve the performance. The AB1s polarise LS3/5a owners, some of whom
cannot live without them and some of whom cannot live with them! My conclusion
is that, if you have what I will call a high-resolution system, one that throws
up a big stable 3D soundstage, the standard AB1 may not be for you.
Examination of the AB1 crossover
schematic shows that the LS3/5a signal is fed through the reversible electrolytic
capacitors in the AB1 crossover and I wondered if these were the problem? There
are two solutions to this. Jo Ki feeds his AB1s in parallel with the LS3/5as.
Although both speakers are working in the crossover region, in practice this
has not proved a problem and Jo's system sounds superb. Jo does not place his
LS3/5as on top of the AB1s and instead uses what I have called "The Ki Method" something I have also adopted.
I took a different approach and
replaced those horrible electrolytics with audio-grade polypropolyene capacitors.
Three 80uF polyproploynes were put in parallel in each channel to make up 240uF.
There are two disadvantages to this, the polypropolyne capacitors are much bulkier
than the electrolytics, and they are quite expensive. So I mounted my AB1 crossovers
in an external box away from the vibration
inside the AB1 enclosure. The results are fabulous.
 |
Prior to veneering them I tested
the subs with my LS3/5as to ensure they were working. Special attention
was paid to make sure there were no air leaks. I then veneered and varnished
them. The results are excellent, the subs allowing the LS3/5as to go deeper
and louder whilst retaining the excellent midrange and stereo image of
the LS3/5a. I own a pair of the Grade 1 BBC LS5/9 monitors and was able
to compare the combos with them directly.
|
The bottom end compares very well
with the LS5/9 (although this speaker goes louder) and I find I much prefer
using the LS3/5as and subs.
If you cannot get hold of a pair
of real AB1s, want to save some money, or like me simply enjoy a constructional
project, then these are well worth making. If you have to buy everything from
new, including the B110 drivers, the total cost is around £120.00 for the pair.
Back to Contents
Page
