|Elliott Sound Products||Guitar Loudspeakers|
The choice of loudspeaker has far more influence over the overall tone of a guitar amp than any other factor. The 'wrong' speaker can make a perfectly good amplifier sound awful to one player, but perfect for another. Much depends on how the amp is used - clean (or relatively 'clean'), distorted or heavily distorted ('crunch') playing styles may require very different loudspeakers, although most guitarists will find (or try to find) something that suits their style(s) so that the same amp/ speaker combination can be used for all their material. Some professionals use different amps for different songs (or parts thereof).
As with many things related to audio, there are many myths around guitar speakers. This is partly because the choice is so personal, but there are many misconceptions and unfounded claims as to what make a good, bad or indifferent guitar sound. Consider that many very accomplished guitar players can actually use any amp that comes their way if need be, and it's their playing style that sets them apart, not the equipment itself. Yes, they will have their preferred setup, but they don't fall to pieces if it's not available.
Many of the claims you'll come across are dubious, and some are downright false. This seems to be an area of great debate all over the Net, with very little agreement and little or no science. Ultimately, the laws of physics determine what any loudspeaker sounds like, even if the exact mechanism is unclear. Some of the most revered speakers around are largely rough approximations of their original models, and with many of them now made in China and re-badged, you absolutely do not automatically "get what you pay for".
There's a great deal said about magnet material, and while some of it sounds plausible, there's a lot more to it than just the type of magnet. While claims abound, there's very little evidence that most have any basis in fact. This doesn't mean that you won't hear a difference, but it's likely that the difference is due to other factors, and is not due to the magnetic material used. A loudspeaker's magnet and voicecoil form the 'motor', which generates the force needed to move the cone. A high magnetic field strength and a voicecoil with many turns creates a strong motor, increasing efficiency. If one speaker is just 1dB louder than another, it will almost always sound 'better', all other things being equal.
Ceramic (strontium ferrite) magnets are by far the most common, despite their relatively poor magnetic properties. The compensation is to make the magnet much larger, and speakers with ceramic magnets are normally just as efficient as those using Alnico or neodymium, but are almost always significantly heavier. The magnetic structure is usually quite different for the different magnet types, but that doesn't mean that there is any significant difference in the magnetic field across the voicecoil gap.
Using a particular magnet type and/ or brand name doesn't necessarily mean anything tangible. Sometimes it's simply a case of 'Famous Person' uses this type of speaker, and people imagine that by using the same driver they will sound just like 'Famous Person'. This only holds true if everything else (including their skill level) is the exact equal of said 'Famous Person', something that is rarely the case. In general, I suggest that you try different speakers until you find one you like. This is actually harder than it may seem, because there could be outside influences that taint your perceptions, such as peer pressure, a sales person's 'persuasion', or the simple knowledge that this is the same speaker that 'Famous Person' uses.
The earliest speakers (as we know them) used electromagnets, because there were no magnetic materials that provided sufficient field strength along with no propensity to demagnetise themselves. This was a vexing problem, but electromagnets can deliver an extremely powerful magnetic field given sufficient turns and current. Prior to the mid 1930s (or thereabouts), a great deal of work had gone into development of suitable electromagnets that could also act as the filter choke for the valve ('tube') amplifiers that were used. The primary usage was for radio (or 'wireless' as it was called at the time), and some very clever designs used a 'hum bucking' coil to prevent the ripple from the DC supply from appearing in the audio. There may have been some very early guitar amps that used electro-dynamic speakers, but I don't have any details.
The first really good magnet material was Alnico (or AlNiCo - aluminium, nickel and cobalt, with the remainder being iron). While it still more than holds its own against newer materials, it's also expensive. Alnico is favoured by many guitarists because it's believed to have that 'vintage' tone. However, the cone material, surround, spider and the mechanical construction of the pole pieces will usually have a great deal more influence than the material used for the magnet. There is little evidence that the magnet alone makes any audible difference. An Alnico magnet speaker from ~1970 is seen in the next photo. Over the years there have been a number of trade names for Alnico, including Alni, Alcomax, Hycomax, Columax, and Ticonal.
Figure 1 - Alnico Magnet
The Alnico magnet slug is the slightly crinkly-looking section at top centre. The conical piece below that encloses the centre pole and the end of the voicecoil so it's not open to outside contamination. It's not apparent how the Alnico slug is attached to the rear or centre pole-pieces. Some (older) Australian and NZ readers will recognise this assembly instantly - it's a Plessey Rola 12U50, a 300mm (12") 50W speaker that was very popular here in the 1960s and 70s. People are still using them, and is should be fairly obvious I have the one pictured (and another the same). Mine were actually 12UX50 twin-cone, but I removed the 'whizzer' cones because they were damaged and are rather dreadful at best, so keeping them wasn't an option. There was another version called the 12UEG ('EG' for 'electric guitar') in the mid 1960s, but I never came across one. They were rated at 30W.
Ceramic magnets are more common and much cheaper than Alnico, and (at least in theory) there should be no difference in 'tone' provided all other factors are the same. This means the cone material, voicecoil construction (and material) and even the basket (the speaker's chassis) must be close to identical. The same applies to neodymium magnet speakers. These are the most recent, and 'neo' magnets are far smaller, lighter and more powerful than any previous material.
Figure 2 - Ceramic Magnet
In reality, it can be pretty much guaranteed that there will be very little equivalence in cone and voicecoil construction, and the basket will be different for the simple reason that the different magnetic materials have different needs in terms of mounting. The basket alone may change the sound, although probably not by a great deal. It's commonly claimed that Alnico magnets are more easily (temporarily) demagnetised by the flux from the voicecoil, supposedly giving a 'softer' compression characteristic than harder magnetic materials. 'Hardness' refers to the ability of a material to retain magnetism, technically known as remanence.
The theory is that with Alnico magnets, as the voice coil exerts a magnetic field in response to the input signal, this magnetic field tries to demagnetise the magnet. As its effect lowers the available magnetic field of the Alnico magnet, the speaker becomes less efficient, the voice coil moves less, etc. There is no doubt whatsoever that the voicecoil's magnetic field affects the field strength across the magnetic gap, but evidence (i.e. measurement data) as to the extent which it affects the magnet itself is very difficult to find (which is to say that I found absolutely zero evidence, only claims and anecdotes).
The physics of it is (supposedly) that the small magnetic domains near the surface of the magnet poles begin to change state or 'direction'. The result is said to be smooth compression, similar to the operating curve 'compression' that occurs in a valve amplifier. When the voicecoil's magnetic force is removed, an Alnico magnet will return to it's normal value - at least that's the theory [ 1 ]. While this is a very popular opinion, there's no evidence (which is the important part of any claim).
Alnico 5 is a a popular speaker magnet alloy made up of 8% Aluminum, 14% Nickel, 24% Cobalt, and 3% Copper, with the remainder made up of iron. The cobalt is the ingredient that makes Alnico expensive. Most of the world's supply comes from the 'copper belt' in the Democratic Republic of the Congo, Central African Republic and Zambia. These countries control the market, and cobalt is primarily used in the manufacture of industrial (and/ or military) magnets, wear-resistant, and high-strength alloys. Guitar speakers are well down the line in terms of 'need'. Cobalt currently sells for about US$60/ kg [ 2 ].
The development of Alnico began in Japan in 1931. Tokushichi Mishima discovered that an alloy of aluminium, nickel and cobalt that had high ferromagnetism. The first Alnico alloy had a magnetising field strength of 400 Oe (Oersted, the old unit for coercive force). The SI equivalent is about 32 amperes per metre (A/m). It had double the magnetic field strength of the best magnet steels in existence at the time [ 4 ]. Modern alloys have significantly higher coercive force, with Alnico 5 being around 51 A/m.
Alnico is a very hard material. It's difficult to machine, so most of the time it's cast or sintered into the desired shape and carefully heat-treated to get the desired magnetic properties. It was the first really powerful magnet material, and even today is only bettered by samarium-cobalt (expensive and uncommon) or Neodymium Iron Boron (NdFeB aka 'Neo') - the most powerful permanent magnet material known so far. However, neodymium magnets will disintegrate if exposed to the air, so they are always heavily plated to protect the magnetic alloy.
However, this article is not about magnetic materials in detail. While the materials are different, the magnetism itself has basically the same physical properties regardless of the material. This includes electro-magnets that were common in very early loudspeakers because useful permanent magnet materials weren't available or were too expensive. The magnetic field in the gap has no 'knowledge' of the magnet material, and the same field strength can be obtained by many different materials and geometries. There is some degree of flux modulation with all magnet and polepiece materials, and high powered speakers will always have (much) greater flux modulation when pushed to their limits.
There are many on-line videos that purport to demonstrate the difference between ceramic and Alnico magnets. What is not disclosed is whether the magnetic field strength, voicecoil, cone, surround, spider and dustcap are identical or not. If not, the comparison is between the different speaker configurations rather than the magnet. This doesn't mean there's not a difference of course, but without full disclosure the demos let you hear the difference between complete loudspeaker drivers, rather than the magnet materials. Advertising material rarely (if ever) describes the entire motor and cone structure. It is a mistake to assume that these are exactly the same for different magnet materials.
It's worth noting that there is a limit to the magnetic induction (measured in Tesla) across the gap of a loudspeaker. The limit is mainly due to the steel used for the pole-pieces, and generally ranges from 0.8 to 1 Tesla, with 1.8 Tesla being about the limit for speakers using mild steel polepieces. Use of exotic alloys can boost that up to around 2.4 Tesla, but at considerable added cost. Most speaker designs saturate the polepieces to allow for manufacturing inconsistencies and to ensure that the 'static' magnetic field is difficult to modulate.
Figure 3 - Loudspeaker Motor Assembly (Ceramic Magnet Shown)
A 'typical' motor assembly is shown above. The important parts are labelled so you can see what goes where. If an Alnico magnet were used, it would be located on the back polepiece, directly under the centre pole, and the magnetic circuit would be a different shape to accommodate the magnet. As seen in the photo above (Figure 1), the magnetic circuit for an Alnico magnet may be closed with a 'U-section', while other designs use a pressed steel cup. The exact mechanism doesn't matter, provided there's enough steel to support the required flux density. If it's too thin, it will saturate at a lower flux density, reducing the flux across the gap. Ceramic magnets are (almost) invariably assembled as shown in the drawing, though finer details differ.
Reducing the field strength reduces efficiency, and it also allows the speaker to 'do its own thing' - it is not as well controlled by the amplifier. This is particularly true around resonance, where a lower field strength increases the total resonant Q of the driver (called Qts in Thiele-Small parameters). There are two factors at play - the flux density ('B') and the length of wire in the gap ('L'), giving the 'BL' product you'll see referred to in many brochures. A high BL product gives high efficiency and good amplifier control of the speaker. The 'L' factor only applies to the voicecoil wire that is within the magnetic field of the gap.
However, if the BL factor is too high, the speaker will be overdamped, and may be thought to have poor bass response. Many guitar amps provide 'compensation' by way of having a higher than normal output impedance (ZOUT). Typical valve guitar amps have a ZOUT of between 4 and 16 ohms, and it's common to use current feedback with transistor amps to achieve the same end. If you look at the ESP Project 27 guitar amp design, you'll see that it uses current feedback. ZOUT is typically about 20 ohms, but it can be set for any value the builder prefers. I did my very first transistor guitar amp using this technique in around 1968, and every instrument amp I've designed since then has done the same.
The cone material is of great importance to the sound of a guitar speaker - probably more than any other factor. However, the voicecoil, surround, spider and (believe it or not) the dustcap can also have a profound effect on 'tone'. Most guitar speakers are relatively low power, up to 100W or so is common, although a few are higher. They also have comparatively high efficiency, with around 100dB/W/m being common. This means a light cone, and most have a modest excursion. The resonant frequency is very important, because that defines the 'bottom end' the player (and the audience of course) hears.
During the 1940s through to the ’60s, guitar speakers were rarely rated higher than 15 to 20 watts, but there were a few exceptions in the later years. Most early guitar amps rarely put out more than 30 watts or so, but the 40 watt Fender Twin (using 2 × 6L6GC valves in the output stage) changed that, and later amps from many makers were typically 80-100W. Low powered speakers were fine when used singly in small venues or recording studios, and in multi-driver boxes (such as the 4 × 12 cabinets ('cabs') that became common during the 1960s). When pushed hard, the speakers started to 'break up', adding speaker distortion to the amp's own distortion when played loud.
A number of speaker makers have used metal dustcaps (usually aluminium), commonly glued to the cone rather than the end of the voicecoil. While some guitarists like the extra high frequency 'bite', most do not. I was once flown from Sydney to Melbourne to find out why an amp sounded revolting in a recording studio. The problem was solved by removing the aluminium dome dustcap and replacing it with a piece of felt. The problem was that the dustcap radiated strongly above around 4kHz, with a very distinctive 'hard' and 'metallic' sound (no-one told me about the aluminium dome before I got on the plane). Reproduction of frequencies over 7kHz is generally considered harsh, and most guitar speakers are designed to roll off above 5kHz. Although it does vary somewhat, the resonant frequency of most guitar speakers is around 70-75Hz.
Figure 4 - Corrugated Paper Surround
Surrounds are normally corrugated paper as seen above, which is often the same paper that the cone is made from. Some speakers use a corrugated cloth surround. A non-hardening material commonly known as 'dope' is used to make this region flexible and ensure it's airtight. Many people are used to seeing roll rubber or foam surrounds on hi-fi speakers, but these are unheard of for guitar speakers. The surround (and the overall suspension including the spider) is generally much stiffer than you might expect. One of the reasons is to ensure a reasonably high resonant frequency, and the other is to protect the speaker as a whole from excessive excursion. Guitar speakers are generally not expected to move the cone more than a couple of millimetres, and much of the movement is involved in creating cone 'break-up' - chaotic movement where the cone does not act as a simple piston.
Cone breakup effects would be very difficult to design or model, and I suspect that most cones are designed empirically (i.e. by trial and error) or use tried and known materials and processes to get consistent results. This is very important, as no-one want to buy two or more supposedly identical speakers that sound completely different. Fortunately, this doesn't appear to be an issue. Ultimately, the only thing that really matters is whether players like the sound or not - people don't buy speakers that sound rubbish (well, mostly they don't, and not on purpose).
Figure 5 - Spider, Tinsel Leads And Terminals
The above photo shows the spider, as well as the tinsel leads and the terminals. The spider has a significant effect on the sound, because it's part of the suspension and is partly responsible for the resonant frequency. The surround is the other major influence. The combination of suspension stiffness and cone mass (including the voicecoil, former, dustcap, air load etc.) set the resonant frequency. The use of a light cone and stiff suspension means a high resonance (in this case it's 84Hz, but that would fall a little after the speaker has been used for a while). To reduce the resonant frequency, the suspension can be made 'looser' (more flexible), or the cone (plus voicecoil etc.) made heavier. The latter reduces sensitivity.
Another influence on the sound is the length of the voicecoil relative to the magnetic gap. For speakers requiring low distortion and reasonable excursion (Xmax), either the voicecoil is longer than the gap (called an overhung design) or the gap is longer than the voicecoil (underhung). These are common for hi-fi speakers, but less so for very high efficiency drivers. Either way, the efficiency is reduced because either part of the voicecoil or part of the gap is 'unused'. For maximum sensitivity (at least at low input power), the voicecoil and gap should be the same size. Of course, when the cone travels even a small distance, some of the voicecoil will be outside the gap and the instantaneous efficiency falls. This causes distortion, which will (nearly) always be a mixture of predominantly third harmonics, with some second harmonic due to suspension nonlinearity.
Because of the fall in overall (instantaneous) efficiency, there will be some degree of 'compression' as well as distortion, both of which many guitarists like because they help to increase sustain (causing notes to last longer) and add harmonics for a 'richer' sound. There are so many different factors that it's impossible to try to characterise them all, because each acts in combination with the other variables. Some differences will be very audible, while others may go almost un-noticed. In some cases you may even find that the things you most expect to make an audible difference, may in fact make barely any difference at all. I'm not even going to try to quantify what affects the sound in any direction, because I don't have a wide variety of speakers to play with, nor do I have the facilities to try to test every combination.
Almost all guitar speakers share some common properties though. Lightweight cones, nearly always paper, with a corrugated surround (as opposed to roll or foam surrounds). The suspension is generally quite stiff, and the speakers have a fairly high resonant frequency (typically 70-80Hz). Most are efficient, at 98-100dB/W/m, but don't expect the same efficiency at (say) 50W that you get at 1W. I ran a basic test on this, and used 1W, 10W, 20W and 30W at 120Hz into a guitar speaker box in my workshop. The test speakers are a pair of low power guitar drivers (they've been in the box for so long I don't recall what they are), and as near as I can recall they are rated at about 25W each.
At 1W, I measured 83.5dB, rising to 93.5dB at 10W (as expected). Above 10W, there was little change in the SPL, and it only managed 96.8dB at 20W and 96.9dB at 30W. As the power increased above 10W, distortion was audible, and at 30W it was very noticeable third harmonic (as expected). I deliberately used a fairly low frequency, because as the frequency increases there's less cone travel. The test was to see how much efficiency was lost as the voicecoil started to move further out of the gap. You'll find that this effect is rarely mentioned (I've not seen any mention of it when discussing guitar speakers).
However, if you want to see some of the best info that I've come across, a very detailed analysis by Kippel [ 3 ] examines voicecoil displacement, flux modulation, suspension non-linearities and just about every other problem facing traditional moving coil loudspeakers. It's not about guitar speakers, but the concepts and issues are common to all types, from hi-fi to concert sound.
An area where you can expect reasonably good 'equivalence' is for the top end. It's common to get a peak at around 2-3kHz, with response falling rapidly above 5kHz. This is quite deliberate, and anyone who's tried using a wide-range speaker for overdriven guitar will tell you that it sounds pretty bloody awful. The overall response of the speaker is one of the most influential in terms of its sound. A small difference in efficiency or frequency response can make a huge difference to the sound. These will far exceed any difference due to the magnet material (real or imagined).
On the Eminence [ 4 ] website it says (and I quote verbatim) ...
"What differences will I hear between ceramic, alnico, and neodymium magnets?"
"Each material, of course, has different magnetic properties and cost. Neodymium seems to be the wave of the future, especially with reduced weight and overall costs coming down. It produces the most magnetic flux per ounce, making it ideal for use in multiple speaker cabinets to maintain performance while reducing handling and shipping weight. Alnico is a composite of aluminum, nickel, and cobalt. It is the most rare and most expensive. Alnico is commonly thought to produce the most 'Vintage' tone and has a reputation for sounding compressed. Ceramic is the cheapest and most common material. If you are comparing speakers that have the same magnetic flux, but generated from different magnet compositions, you probably won’t notice a difference in tonality. Differences in tonality that are often attributed to the magnet material probably have more to do with the positioning of the magnet and resultant differences in magnetic flux within the motor structure. There in lies the mojo!"
This is in agreement with the comments I've made above. It is important to be careful with references, because a great many are not based on engineering, but are from the 'marketing' department. The principle of marketing is to tell you what you want to hear, whereas engineering tells it as it is, regardless of whether it's what you want to hear or not. Most 'reviews' leave out nearly everything you need to know - I saw one video where completely different speakers (they were even different brands!) were used to 'demonstrate' the 'difference' between ceramic and Alnico. All it did was demonstrate the difference between two different speakers - the magnet is immaterial if there are any differences in the other factors. Any conclusions drawn from the demonstration are based on a completely false premise and are irrelevant.
There are two metals used for voicecoils, copper and aluminium. Copper is by far the most common, having good electrical conductivity and it's easy to join using solder. However, it's much heavier than aluminium which is a disadvantage. Aluminium is difficult to terminate, so much of the aluminium wire used for voicecoils is copper plated so it can be soldered. While aluminium wire was very popular for a while, it seems to have fallen from favour to some degree. Aluminium appears to be uncommon for guitar speakers. All voicecoil wire is insulated with a high-grade, high-temperature enamel coating to prevent the individual turns from touching each other (causing a short circuit), or from moving (which will ruin the loudspeaker).
In some cases, the wire is rectangular or square instead of round. This allows more wire per unit volume, and this increases the winding efficiency because there are no little gaps between the turns as you get with round wire. Edge-wound rectangular wire is at the extreme end of speaker voicecoils, and isn't common for most guitar speakers. A wire measuring 1.4mm × 0.7mm has a cross-sectional area of 1mm. A round wire with the same area (1mm²) has a diameter of 1.12mm and occupies a physical area of just under 1.5mm². There's roughly 0.5mm² of 'wasted' space, making the voicecoil larger for the same number of turns (assuming the same overall diameter). It's not quite so bad for the second layer if the turns are wound properly.
Aluminium has about 50% of the weight of copper for the same length and resistance. However, the wire must be thicker because aluminium has only ~60% of the conductivity of copper. The net result is that aluminium has a slight overall advantage for weight, but the reliability of terminations still remains a problem. If it's not copper-clad, the only reliable connection is welding, which itself is not trivial with aluminium.
Figure 6 - Loudspeaker Voicecoil Options
The drawing above (somewhat simplified) shows three of the options. The 'conventional' arrangement is the most common for guitar speakers, but the number of layers (and the length of the coil itself) will vary depending on the design choices made. The former requires a strong bond to the wire, cone and spider, and also provides the termination points where the voicecoil winding is joined to the flexible braid (aka 'tinsel') that's used to bring the wires to the terminal block mounted on the basket. The entire assembly is likely to be epoxy impregnated to ensure that the windings can't separate from the former or each other.
The voicecoil former (aka bobbin) has to be strong, light, and capable of withstanding the worst-case maximum voicecoil temperature without failure. Early speakers used paper (actually more like thin cardboard) which is still a very popular choice, but materials also include Kapton (polyimide), Nomex, Kevlar, aluminium, phenolic resin, fibreglass and even titanium. Metallic formers are useful to help disperse heat, but they cannot be a closed circular form because that would create a shorted turn. There is always a very small gap between the ends of the tubular former to prevent a short circuit. The wire is bonded to the former using a variety of different adhesives, many of which appear to be proprietary, so details aren't available. Many will be high-temperature epoxy or polyurethane resins, and many improvements to these have been made over the years. Few will last very long if subjected to temperatures exceeding 200°C, and nor will the enamel insulation on the wire.
The ideal former is very light, strong, and free of resonances. Many proprietary configurations have been developed, and few speakers have formers that are 'inappropriate' in any way. The choice ultimately comes down to cost vs. expected power handling, but each different material has the potential to affect the sound. Whether this is 'good' or 'bad' depends on the listener, and this is never more true than with guitar speakers.
Even though aluminium is light compared to most other metals, it's much heavier than paper or the various plastics or composites mentioned above. It's also important to ensure that it's well damped, as it may have unwanted resonance(s) because of the nature of most metals. By way of example, it's no accident that bells are made from metal - I've not seen a plastic bell, and doubt that it would work well .
There are a few ways that guitar speakers are classified. There's the distinction between 'British' and 'American', with both covering 'modern' and 'vintage'. In reality, these are somewhat arbitrary, and there's no particular reason that (for example) a 'vintage British' and a 'modern American' speaker couldn't have near identical sound. There are many others too of course. In Australia there were several locally made speakers that people quickly discovered were ideal for guitar, some using Alnico magnets, some ceramic. The Alnico magnet shown in Figure 1 is from a speaker that was very popular for some time in Australia during the 1960s. This was a 300mm, 50W speaker that seemed close to indestructible. They were also available in a 'twin-cone' version that was popular for column PA speakers (these were the original 'line array').
Much the same happened all over the world, but most smaller countries probably don't have any viable speaker manufacturers left. There are still a couple in Australia (or there were at last count), and world-wide there must be hundreds of small 'boutique' speaker manufacturers. Whether they get their parts (baskets, cones, etc., or even finished speakers) from China is unknown. There can be no doubt that China now has more speaker manufacturers than anywhere else, but most will be 'OEM' (original equipment manufacturers) and will have their products re-branded to whatever the end customer requires.
Ultimately, it doesn't matter what the speaker is called (name, style, model), it's whether you can get the sound you want from it. Consider that many professional guitarists use whatever equipment is provided by the promoter in the countries where they tour. They will have their specific requests of course, but in some cases it's simply not possible to provide the exact equipment listed in the rider (the band's wants, demands and/or needs). In the vast majority of cases, this ultimately causes no problems (BB King once had to use one of my (transistor) amplifiers because the music shop that supplied the gear didn't have a spare Fender Twin - true story). Apparently he rather liked it (but I didn't get to sell him one).
Speakers are mounted in a variety of different configurations, and with different box styles. Most 'combo' amps are open backed, because they require ventilation and the amplifier is in the top section of the enclosure. Airflow is essential with valve amps, but is no less important for 'solid state' transistor amps. The heatsink must have airflow, and having it sticking out the back is not acceptable. Leaving the back open solves this, at least to some extent. The majority of 4 × 300mm (12") type enclosures (commonly known as a quad box) are sealed, with no openings other than those for the speakers.
The sound from open backed and sealed boxes is (sometimes radically) different. There is no 'better' configuration for all players and/ or venues, and the choice is very personal. Open back cabs tend to create more on-stage 'spill', which can make the sound engineer's job that much harder. However, some engineers use a microphone in (or directed towards) the rear of the box to produce the FOH (front of house) mix, preferring the usually 'mellower' speaker rear radiation. Others use a mic front and back so they can mix the two for the desired sound.
There is another class called an 'isolation' cabinet. The speaker is completely enclosed to minimise the SPL (sound pressure level), and these are more likely to be used in a studio than on stage. There is a microphone inside the cabinet, and some have their own speaker while others are designed to accept a normal speaker box. Some are lined with acoustic foam while others have minimal lining. Absorbent foam helps to minimise internal reflections that can create a 'hollow' sound and it also reduces sound leakage to the outside world. Some provide input/ output connectors for the speaker and mic (respectively), while others may use a narrow slit for cables. Attenuation (reduction of SPL) depends on construction.
One thing that is almost universally eschewed is a vented/ ported enclosure. While some bass players like the extra efficiency at the bottom end, most guitarists dislike the sometimes 'woolly' bass that vented boxes produce when driven from amplifiers with a comparatively high output impedance. This applies to almost all valve guitar amps, and a great many transistor guitar amps as well. This is not to say that a vented box should not be used. Like everything to do with guitar speakers, it's a personal choice.
The 300mm (12") speaker has been the guitarists' favourite for a very long time. There are players who use (or prefer) 250mm (10") drivers, which may be as a single driver (usually a combo amp), or 2 × 250mm or a quad box. There are a few smaller amps (typically 'practice' amps) that use either one or two 200mm (8") drivers. These can be used in the studio or even on stage for quieter groups, and some can be surprisingly loud despite their size.
Most speakers these days are wired so that a positive voltage applied to the positive (+) terminal will cause the cone to move out. This creates a compression (an increase in air pressure) in front of the speaker. However, it has not always been like this. For some time, JBL (for reasons that no-one can explain) wired speakers the opposite way, so positive to the positive terminal caused the cone to move in. Today there seems to be general agreement that a positive voltage should cause a compression (cone moving out), and I've not seen a driver for many years that was wired differently. The polarity can be tested with a 1.5V cell. The cone movement isn't great, but it's usually easy to see (or feel) which direction the cone moves with each polarity.
It's important that if two or more drivers are used with an amplifier, all should be in phase. That means they should all move outwards and inwards at the same time with the same polarity. From the amp's perspective it doesn't really matter if they are (all) in or out of phase, since there may or may not be an overall inversion of the signal from the guitar to the speaker socket. There isn't really any convention on this, although most designs do retain 'absolute polarity'. However, there can be significant phase shifts at some frequencies that vary depending on tone settings, and it's unlikely that the polarity is audible. It is well known that some asymmetrical waveforms can sound 'different' if their phase is inverted, but only in an A-B test.
Of somewhat greater concern is that a large array of speakers (think the classic 'double stack' - 2 x quad boxes) is very large compared to wavelength at frequencies above 1kHz or so. This causes high frequency beaming and lobing, where the upper frequencies have a very irregular coverage pattern. There isn't anything much that can be done to reduce this (other than a totally different speaker configuration), and it can create problems - especially for other band members who may have to put up with excessive treble if the guitarist listens off-axis. Even a single 300mm (12") speaker in a conventional small combo box will show this effect. Some amps have the facility to tilt backwards so the speaker can be aimed at the player. Others have a sloping baffle for the same reason, and stands are also available to do the same thing. These do help a little, but they don't solve the problem.
Impedance is important. Valve amps are designed to operate into a particular nominal impedance, and if you use a speaker (or combination of speakers) with a different total impedance, the amp will not perform properly. It's even possible to damage the amp - an excessively low impedance (e.g. less than half the nominal) can cause output valves to overheat their plates, and this (sometimes dramatically) reduces valve life and reduces output power. A higher than normal impedance can cause 'flash-over' at the valve base due to excessive voltages being created within the amp itself, and also reduces output power.
Transistor amps don't care if the impedance is higher than normal (including an open circuit), but they get very annoyed if the impedance is too low, and will often fail to show their displeasure. Higher than expected impedance reduces output power, and that can sometimes be used to advantage. A quad box that can be set to 8 ohms or 32 ohms (for example) can reduce the power dramatically, making the system less overpowering on stage, and much easier to manage in the studio (or bedroom).
All speakers in a box should be of the same make and type, with the same impedance. Mixing impedances means that one driver may get the lion's share of amp power and fail, usually at the least opportune time. The imbalance also means that you don't really know if the impedance is alright for the amp unless you know how to calculate the combination properly. Even knowing that does not help power distribution, so the risk of driver damage or failure still exists.
If there are other differences (such as the cone, surround and/ or spider), the speaker with the weaker suspension may be pushed well outside its limits. This won't occur with an open backed box, but closed back systems can develop significant pressure inside the enclosure. This is particularly true if the system is used for bass, which means longer cone excursions and more pressure (both positive and negative). Where different types or sizes of speakers are used within the same cabinet, there will be a divider so that each set has its own sub-enclosure to prevent unwanted interactions.
Most speakers are fairly 'tight' or stiff when new, and may seem bass-shy. After being driven for a time they change character slightly. The surround and spider corrugations begin to loosen up, and the net result is usually a better bottom end and the cone breakup characteristics change. The changes are usually for the better in terms of enhancing the tone for guitar work. Severe overloads will also change the sound, but usually in very much the wrong direction.
Heat buildup in the voice coil has always been very real a problem, particularly as speakers were expected to handle more power. When the voicecoil gets hot, its resistance rises and so does its impedance. This reduces sensitivity, and if the heat is too great it will cause the adhesives and enamels to soften and may allow the voice coil to come apart. Even slight deformation can cause 'poling', where the voicecoil wires rub against the polepiece. This signifies an ex-speaker, and it either has to be replaced or re-coned. The gap between the winding and the poles is only around 0.25 to 0.3mm (roughly 0.010 to 0.012 inch), so it takes very little deformation to cause serious problems. I suggest that you read Power Vs. Efficiency, which covers the issue of voicecoil temperature in detail. Another article that you should read is Speaker Failure Analysis which describes what does and does not cause speakers to die.
Heat isn't a major problem with lower powered speakers - unless they are pushed beyond their ratings of course. As speaker power goes up, the problem becomes progressively worse. For a speaker with a nominal efficiency of 100dB/W/m, 94% of all the power delivered to the voicecoil is dissipated as heat, with only 6% producing sound. A speaker that's being punished with 100W of input power has to get rid of 94W of heat. It may not sound like much, but feel how hot even a 60W incandescent lamp gets if you want some context.
Many attempts have been made over the years to get the heat out of the voice coils, including the use of aluminium dustcaps, regular dustcaps with vent holes, or a vented centre polepiece. These techniques all rely on the cone's movement to create some airflow to pull heat from the voicecoil. Aluminium formers help to disperse the heat more effectively than paper or plastic.
You need to be wary with vintage speakers because it may be possible for the spider to shift. This will allow the voice coil to shift, causing poling. Most vintage baskets were painted (and the paints used were not as good as those available now), so the glue holding the spider may have been applied to paint rather than bare metal. With time and vibration the glue can lift the paint, allowing the spider to move and the voicecoil with it.
The adhesives in true vintage speakers were very poor compared to the epoxies and other adhesives available today. Cyanoacrylate ('CA' or so-called 'super-glue') is particularly strong, but is also rather brittle unless formulated with other materials to provide resilience. There are many different formulations of many different adhesives used in speaker manufacture.
Severe mechanical stress (such as a speaker cab falling off the stage) can cause very serious damage, which often cannot be fixed. I've seen (and heard many tales about) speakers where the entire magnet and polepieces have become detached from the basket after a fall. The deformation of the metalwork is such that it it rarely possible to repair the driver - it has to be replaced. If you can't get the exact same driver and there is more than one, all should be replaced because they have been severely stressed, and will probably sound different from the replacement driver anyway.
The requirements for bass (guitar) speakers are usually very different from those used with guitar. For starters, open 'E' on a 4-string bass is 41Hz (41.204Hz), or open 'B' on 5-string and 6-string basses is 31Hz (30.868Hz). Most bass players want a clean sound, so it's not at all uncommon for the amp to be rated for much more power than the speakers. Up to double the power is normally alright, but there are some significant exceptions. The primary exception is if the player uses 'fuzz' bass - either by overdriving the amplifier or with a pedal. Some bass amps have provision for built-in overdrive, and as an example it's an option for the ESP bass amp project (see Project 152).
Bass cabinets are often vented ('ported' if you prefer) to get high efficiency at low frequencies. In the early days of amplified instruments, the speakers were usually just guitar speakers, which themselves were 'general purpose' speakers until speaker manufacturers started to specialise. These days, many makers have a range of speaker drivers designed specifically for bass guitar and/or amplified double-bass. These have a longer 'throw' voicecoil to reduce distortion, and generally have a much lower resonant frequency. This impacts on efficiency, so a 'decent' bass rig will have a lot more power than a guitar amp.
While there's some discussion/ argument as to the 'best' speaker, enclosure, amplifier, etc., etc., it's not quite as polarising as for guitar speakers. One area where there are differences of opinion regards the speaker size. 380mm (15") bass speakers were once the mainstay of bass players, although guitar-style quad boxes with 4 × 300mm (12") drivers were also common. These days, many players seem to prefer 4 × 250mm (10") drivers, sometimes using two cabinets.
As with guitar, the choice of speaker (or speaker system) is personal. It's not at all uncommon for bassists to use a 'tweeter' - usually a compression driver and horn to get that top-end 'bite'. This is especially with slap-bass styles, where the amount of 'bite' expected can be considerable. Using smaller drivers usually means a tweeter isn't needed, but may also mean that there's not enough bottom end. A good combination can be to use a 'stereo' bass, with the neck pickup driving one or two 380mm drivers, and the bridge pickup driving a 250mm quad box - with separate amplifiers of course. This used to be fairly popular, but very powerful amps and speakers rated for silly amounts of power seem to have diminished the need.
It's hard to come to any specific conclusions, other than to state that the selection can be very personal. In reality, most guitarists will be happy enough with most guitar speakers because the amp's tone controls will compensate for response deviations (at least to a degree). There will be exceptions, but this may be due to anything from contractual obligations, simple prejudice or familiarity. Some people just don't like change. If they didn't know that the speakers had been changed they may not even notice (provided the replacements have close to identical frequency response).
There are quite obviously many factors that determine the sound, but of those, the magnet itself is well down the list. At the very top of the list is the material used for the cone, dustcap and voicecoil. Spiders and surrounds also affect the sound, but without a large-scale blind test it's very hard to quantify the audibility of the various components. Reading forum posts and believing what random (and often anonymous) people say is certainly not a useful way to decide on the ideal speaker.
Unless you have listened to a guitar speaker with an aluminium dustcap and found that it produces the sound you are after, I suggest that they be avoided. If you intend to use the speakers in a studio, then you also need to verify that the sound is right when a microphone is used. Mics 'hear' things very differently from the way we humans do, and you may get a nasty surprise if you aren't aware of the potential problems. The same applies to 'unusual' cone materials. Most guitar speakers use paper cones, but some 'universal' drivers may use polypropylene cones that may have a very different sound when overdriven.
Make sure that you have sufficient speaker power to handle your amplifier. A 100W guitar amp ideally needs speakers rated for at least 200W. You have a little more flexibility if you use a valve amp, and you'll usually get away with around 150W speaker power for a nominal 100W amp. 4 × 50W speakers is close to ideal for any 100W amp, but other combinations will work too. There are some playing styles that don't stress the speakers much ('clean' guitar for example), and you can generally get away with less speaker power. However, if the amp is ever pushed hard, then it's worth the peace of mind to know the speakers can handle the full output.
The decision to use Alnico, ceramic (ferrite) or neodymium magnets has nothing to do with the tone per sé. Tonal differences are primarily influenced by the cone and suspension materials as described above. Of course, it may be that an Alnico speaker happens to have the exact sound you are after, and lacking an equivalent using ceramic or 'neo', that's probably going to be the one you buy. It is important not to conflate the magnet material and other parameters - they are separate, despite many of the claims you will hear. That there might be differences is certainly possible, but consensus of designers is that the magnet doesn't affect the sound.
Finally - Beware of all marketing information and 'colour glossies' - they are designed specifically to sell you 'stuff', and convince you that non-existent 'differences' are real.
Please be aware that in common with a lot of material on the Net, any or all of these references may disappear at any moment. I try very hard to ensure that references are current, but this can become very tedious. In some cases, I have used other reference material that may not be listed, but that's mainly for verification of claims made in the references provided. Some claims are simply unable to be verified at all, and as such I tend not to mention things that defy verification.
In all cases, the references are for further examination by the reader. There is no connection between ESP and any of the organisations that are referenced. ESP is completely independent, and does not benefit in any way from citing any company or individual. Opinions in referenced material are those of the company concerned, and are not necessarily endorsed by ESP.
|Copyright Notice. This article, including but not limited to all text and diagrams, is the intellectual property of Rod Elliott, and is Copyright © 2018. Reproduction or re-publication by any means whatsoever, whether electronic, mechanical or electro- mechanical, is strictly prohibited under International Copyright laws. The author grants the reader the right to use this information for personal use only, and further allows that one (1) copy may be made for reference. Commercial use is prohibited without express written authorisation.|