SYSTEM-1 from the outset was always going to be an active 3-way DSP speaker. The reasons for this are explained in our previous blog:
‘Why don’t you just make a passive bookshelf speaker like everyone else?’
Beyond this we set ourselves the following requirements:
- The sound system should perform well in small to medium sized rooms
- Be physically fairly compact
- Capable of playing at ‘realistic’ volume levels with low levels of distortion
- Have an extended bass response
We also wanted to use the very best components and cabinet construction. This meant the bill-of-material (BoM) cost would be high, but we also wanted to make the speaker available to as many people as possible, so we targeted a price of around £6.5k .
The only way to make this possible would be to sell direct, but this suited our philosophy of having a close relationship with our customers, so we decided to proceed.
In small rooms, the speakers often have to sit close to the rear walls or in a corner. The problem with this is the speakers can easily excite room modes. These are specific frequencies predominantly in the bass region that relate to the physical dimensions of the room.
When a bass note in your music is around the frequency of the room mode (there will be several at varying frequencies) it will sound particularly pronounced and can literally sound like a ‘BOOM’. This is a rather annoying and obvious distortion, as it can completely take over the listening experience.
Sealed vs. Ported cabinet:
Ported speakers are particularly bad for activating these modes and the problem is exacerbated by these speakers needing to play at reasonable volumes to correctly activate the ports and sound neutral.
Various types of room modes:
Sealed speakers in contrast sound fairly neutral irrespective of volume and don’t cause booming anywhere near to the same extent. This in combination with good transient response, a gentle bass roll-off and limited re-radiation of energy, makes it the preferred choice.
Whilst we would have liked to make the cabinets out of aluminium or concrete for example, the costs are prohibitive, so we decided on heavily braced deep-route MDF. This is a stiffer, denser form of the standard MDF you can buy in a hardware stores.
We also wanted a dedicated cabinet for mid-highs and another for bass. This has the advantage of limiting the bass frequencies reaching the delicate midrange driver and tweeter and disrupting their movements. Secondly, there are studies that suggest humanoid shaped speakers, with a head and a body tend to sound more neural and we would agree.
CODE Acoustics TOP-2 cabinet:
Cabinets should always be stiff, but it’s particularly important for bass cabinets where pressure levels behind the drive units can be huge.
With this in mind the mid-high cabinet would use 18mm thick panels with a single internal brace. The bass cabinet would use 25m thick external panels with extensive cross-bracing. In fact the maximum unbraced panel area is circa 100mm by 200mm!
CODE Acoustics BASS-2 cabinet bracing:
By doing this you push the cabinet resonant frequency up, but you still need to eliminate this energy, otherwise it’ll be re-radiated into the room. This is a distortion and will give the bass less bite and more overhang.
So we extensively lined the cabinets with visco-elastic damping panels, which have properties of both solids and liquids, being particularly good at turning vibrational energy into heat.
In addition our bass cabinets have two internal voids, which are filled with sand. This has the benefit of increasing the cabinet mass and absorbing some of the unwanted vibrational energy.
CODE Acoustics BASS-2 cabinet:
Furthermore, the cabinets use Melamine foam, which is demonstrably better than egg-box foam typically used in speaker cabinets.
Another issue with many speakers is they use spikes to ‘decouple’ the speakers from the floor. If you use the little spike cups there will be a degree of decoupling, but if you spike them straight into straight into an underlying sheet of ply through the carpet, you’ll actually achieve very good coupling. The exact opposite of what you are trying to achieve!
In this circumstance, the room will actually act like a massive auxiliary bass radiator and significantly change the sound. Again, it will sound ‘warm’, but this is just re-radiated bass energy and a distortion.
What you actually need is a material between the speakers and the floor, which is very bad at transmitting vibrational energy (transmissability). Traditionally you might use rubber, but if you look at the graph above, you can see a material called Sorbothane that transmits less than 1/10 of the energy that rubber will! This material is used extensively in electronic enclosures to prevent vibrations affecting sensitive electronic components, but strangely not used to a large extent in loudspeakers.
We’ll use Sorbothane for two purposes in SYSTEM-1. Firstly it’ll be used to prevent bass vibrations transmitting to the floor of your listening room, but it’ll also be used to prevent the bass vibrations affecting the midrange driver and tweeter.
Components – drive units:
Given we wanted a speaker that wasn’t fussy about room placement, the bass drivers would have to be forward facing. Considering this, and the fact we wanted a fairly compact cabinet, we felt we couldn’t go beyond an 8” driver.
To reproduce accurate bass, we needed a drive unit with a strong motor and light cone. A driver like this lends itself to a small cabinet volume, which opened up the possibility of using 2 x 8” driver units per side. Again this sits well with us as it significantly increases the radiating area and we have a strong belief that to reproduce bass naturally, you need to move a large volume of air gently.
We have long been fans of Volt Loudspeakers, who are based in Dorset England, as they make some of the finest drive units in the world. The BM220.1 looked like the perfect unit.
Take a second to look at the size of the magnet on the BM220.1 (It’s huge)! It’s likely double the size of the magnets (motor system) on even some very expensive woofers. A motor like this, combined with a low mass diaphram gives you a great power-to-weight ratio and high efficiency. This means the drive unit is far more able to reproduce bass transients (kick drum for instance) accurately than a woofer with a heavy cone and small magnet. This combination is often chosen to give extended bass response, so you get a large amount of low quality bass!
In fact the BM220.1 is so well controlled they only needed 8L of air volume each, meaning we could keep the bass cabinet small, whilst still having good bracing and a decent amount of damping material.
The Volt bass drivers are very well behaved and can be crossed fairly high. This takes the load off the mid-range driver, so we could pick a 4” unit, which is around the optimum diameter in our opinion.
We say this because we really want one driver to handle as much of the vocal range as possible. This is circa 300Hz to 6kHz, with the ears most sensitive range being around 1-5kHz. Our crossover points are 250Hz and 4.5kHz, so we’re not far off!
Humans are so attuned to understanding speech that any distortion in it’s reproduction is flagged up by our brains. It simply doesn’t sound correct. This is why you need to be extremely careful if you have a crossover point in this range as it can cause phase issues. So a crossover point of 2.5kHz as with most 2-way speakers is far from ideal!
Scan Speak 12MU/8731T00:
We have long admired the Scan Speak 4” Illuminator midrange. They are frighteningly expensive, so you rarely find them on even passive 2-way speakers under £10k. In fact there is a well known American company that is using this driver in a speaker that costs over £100k, to put into perspective quite how good it is!
The cone is light, stiff and well damped, with one of the most beautiful motor systems I’ve ever seen. The magnet is neodymium with a high flux density and is sculpted to allow the sound waves coming from the rear side of the diver to move over it, rather than being reflected. Not to mention how flat the frequency response is (below). This means only the slightly amount of DSP correction is required, which is exactly how you want it.
Scan Speak 12MU/8731T00 frequency plot:
Whilst many people are getting excited about titanium and beryllium domes, they quite often have a problem; they have a severe break up node at the top end of their range. Whilst this is in theory outside of our perceived hearing range, in some cases it just doesn’t quite ‘sound’ right. In our opinion it’s a high price to pay for a slightly extended band-width.
Scan Speak D2905/950000:
We prefer to sacrifice a small amount of frequency response and have a well behaved diaphragm. A long-time favourite tweeter is the Scan Speak D2905-95000. This is one of those designs that just sounds ‘right’, and some subsequent evolutions were actually a backward step in our opinion. It’s frequency response is incredibly flat, it has high power handling and a nice gentle roll-off.
This is one of those drive units that raises the question; does the price of a drive unit indicate how good it is? As a very general rule then yes, but there are many gems out there that outperform equivalent units costing many times more.
Scan Speak D2905/950000 frequency response:
Electronics – Overview:
Most active speakers mount the electronics in the speaker cabinets, but as previously mentioned, electronics are highly sensitive to vibration and the one place you can guarantee vibrations is in a speaker cabinet.
CODE Acoustics CONTROL-1:
So we’ve used a dedicated electronics box with Sorbothane feet to further reduce the vibrations. The CONTROL-1 box houses 6 channels of class-d amplification courtesy of Hypex as well as a 6 channel DSP crossover.
The control box would also have a full suite of inputs including AES/EBU, Optical, SPDIF, USB & XLR Analogue.
CODE Acoustics CONTROL-1 inputs:
The amplifiers are the latest generation Hypex n-core OEM modules. There are a pair of dedicated 100 watt amplifier for the tweeters, which have incredibly low distortion in the milliwatt and single digit watt region where a tweeter often operates.
Hypex NC100HF OEM:
Hypex NC100HF THD graph:
In addition to the two dedicated tweeter amp modules we also use a pair of 2-channel 250 watt modules (see pic below) to power the mid and bass drivers. The 8 ohm bass drivers are run in parallel to produce a 4 Ohm load, maximising the amplifier power. So just to clarify, that’s 6 dedicated channels of amplification; each tweeter get’s into own amplifier, as do the mids, while each pair of bass drivers also get an amp.
The left and right channel have their own power supply so there is almost no chance of channel cross-talk. The amplifiers have very high damping factors and being connected directly to the drive units, give fantastic control.
Hypex NC252MP OEM:
Hypex NC252MP OEM THD graph:
DSP and crossover:
The DSP unit also comes from Hypex and can accept high resolution files up to and including 24 bit 192kHz. This is programmed with our bespoke filters, which are L-R 4th order (24dB per octave). These are relatively steep slopes and ensures the drive units only work in their intended range.
Hypex DLCP (DSP module):
So SYSTEM-1 will be an active 3-way speaker, using only the very finest components. It is only possible to sell this quality of product direct (and keep the price sensible) due to the very high material costs involved.
If you would like more info, have any comments or suggestions, or interested in a demo, please get in touch.
Thanks for reading,
Speakers, Loudspeakers, Active Loudspeakers, Code Acoustics, Hifi, High Fidelity, Audiophile, SYSTEM-1
One Reply to “CODE Acoustics SYSTEM-1 (White Paper)”