• Guitar Amplifiers
• Powering Up
• Bass Amplifiers
• Keyboard Amplifiers
• Odds and Ends

---

Guitar Amplifiers

Most guitarists will say that, with a given amplifier, a certain volume level is required to obtain the right sound. While this is to a point psychological, there are very good technical reasons to support this statement. The "right sound' usually means a fairly high level of harmonic distortion, both from amplifiers and speakers - distortion which gives much needed assistance to an electric guitar by providing the missing harmonics and increased sustain. A certain amount of sustain is obtained by acoustic feedback, particularly with semi-acoustic guitars, however the majority is the result of the amp being driven into "clipping", or distortion. This has the effect of maintaining the level at a (more or less) constant level for much longer than would normally be the case (Figure 1).

If an amplifier is overdriven, the sound produced is usually considered objectionable, so special measures must be taken to obtain the desired characteristics.

Firstly, there is the bandwidth restriction of the speaker cabinet. For the low-end this may be achieved by using an open-back box, to allow bass cancellation, or to contain the speakers in an enclosure of limited internal capacity. High-end roll-off is usually a characteristic of the speakers themselves: speakers - without "whizzer" cones or aluminium domes will generally sound smoother because of their limited high frequency response. A broad peak in the speaker's response at 2kHz to 4kHz adds "presence" and bite, and the response should roll off beyond that point. It is preferable that frequencies above about 7kHz are not reproduced at all, as this will exclude the 7th harmonic (which is discordant) of all strings of the guitar. Damping or venting of guitar speaker enclosures is usually avoided like the plague because damping will lower colouration, a desirable feature in musical instrument amplification. Without colouration all plucked string instruments, for example, would sound much the same.

• Vented boxes are avoided because they emphasise (or allow to be reproduced) the lower frequencies - these are usually not desired, since they tend to muddy the sound.  This is especially true for "heavy metal" playing styles, since severe amp overdrive is commonly used to achieve the sound, and even small amounts of low frequency material (such as string handling) will cause large amounts of bass "waffle".
• The majority of guitar speaker boxes (as noted above) are either open backed, or are relatively small sealed enclosures.  There are naturally exceptions, but even 20 years after this article was first written, are still uncommon.

In fact in a valve amp the clean or "no distortion" state exists only when there is no input. Distortion is present even at low levels, and increases with the input signal.

The above situation is in contrast with the majority of transistor amps, which generally have very low distortion up to the point of clipping, after which distortion rises rapidly. Maximum distortion is the same as for a valve amp - i.e. a square wave, but the transition is such as to be very noticeable; this can be objectionable when the guitarist is playing clean but for the odd note or chord which overdrives the amp, causing a marked change in tonality (Figure 2).

Output impedance and "dynamic output" are other major contributors to the sound of a guitar amp. Output impedance, in this instance, has nothing to do with speaker "ohmage", but is the actual source impedance of the amplifier, as seen by the speaker. In a valve amp, the source impedance may be as high as 200 ohms for a nominal 8 ohm output. This allows the speaker freedom to exhibit its own natural resonances, plus the colouration provided by the enclosure itself. A low output impedance on the other hand, will repress colouration by damping the speaker, in much the same way as would fibreglass in the cabinet. This tends to make the amp sound flat and lifeless, lacking the subtle colouration and tonality which makes a fine musical instrument - and after all, a guitar amplifier is as much a musical instrument as the guitar itself. Transistor amps usually have a low source impedance, typically less that 1 Ohm.

• It should be noted that the power rating for guitar amps is (very) commonly quoted at 10% distortion.  This means the amp is typically about 80W at the onset of clipping (usually taken as the "real" maximum power level).  This is justified on the basis that few guitarists will operate their amps with no distortion at all.  This assumption is not necessarily valid, but has prevailed for a very long time, so is unlikely to change now.
• The source impedance of valve amplifiers is generally lower than that quoted above, and in fact it will generally match the speaker impedance (assuming no feedback is applied). I appear to have goofed in the original article.
• From the very first guitar amplifiers I ever built using transistor output stages, I used current feedback to increase the output impedance of the amp. This restores much of the "valve sound", and provides a simple but effective method of providing short-term short circuit protection (for example while plugging 6.5mm phone jacks into their sockets while the amp is on). This technique has now been adopted by nearly all guitar amp manufacturers. Regrettably, this technique does little to improve dynamic output. Output impedance of my amps was typically 200 Ohms.

In contrast, a transistor amp can usually deliver only 55% or so of its rated power into double rated load, and about 30% into four times rated load. These variations may not appear to amount to very much, but they do make a big difference in terms of spectral balance and dynamic performance. Such is the difference in fact (and this also applies to valve hi-fi amps, bass amps, etc) that a 100 watt valve amp will generally sound as loud as a 150 watt transistor amp.

The technical reader with an understanding of dB and the logarithmic response of the human ear may be sceptical about the foregoing paragraph, but differences of 1dB are audible when one has a suitable point of reference; in musical groups, where the balance of instruments is critical, such differences are very noticeable.

---

Powering Up

A further power increase results from the fact that a severely clipped waveform will approach a square wave, which has twice the power of a sinewave of the same amplitude. The result is a 100 watt amplifier delivering an average power of perhaps 170 watts short term (i.e. maximum level indicated on a VU meter), and up to 130 watts averaged over a 5 to 10 second period.

Compare this with an undistorted amp handling program with a peak-to-average ratio of 100:1 (20dB). The long term average power will be in the order of 2 to 5 watts with peaks of 100 watts. With mild distortion, say less than 15%, the average power climbs rapidly to around 25 watts; increase the distortion further and the power climbs again until, with maximum overdrive, a 100 watt amp will be delivering 200 watts!

• An important point was raised by a reader, who said that the above explanation only applies if the power supply remains at a constant voltage with the increased load.  This is very true, and in reality, most don't, so the maximum power is limited by the capabilities of the supply.  Valve amps are usually worse in this respect than transistor amps, and the maximum power at full clipping may only be 150% of the rated power.
• Despite this, the average SPL (Sound Pressure Level) is maintained at a high level, and the effect is that it sounds louder than one might imagine it should.  This is similar to using compression (as done with ads on TV).  The maximum SPL is not increased, but because everything is at maximum, it appears to be louder than it really is.

• Having said that, there seems to be a resurgence of old analogue technologies in all areas of electronic music, and this includes the lowly fuzz box.  There have also been moves to try to emulate the asymmetrical clipping characteristics of valves (some dating back many years).
• One of the most innovative (and complex) attempts to recreate the valve sound with solid state electronics is the Vox AC30 Simulator - a contributed article on the ESP Project Pages.

The addition of a graphic equaliser set, now incorporated in many brands such as the Zoom (above), greatly expands the tonal variation possibilities of an amplifier.

---

Bass Amplifiers

One of the major differences is the power requirements. Whereas a guitar amp will usually be overdriven, a bass amp should not be; the sound will become muddy and indistinct, lacking punch and definition. Bass is, by its very nature, difficult to reproduce. The open E string for example, has a fundamental frequency of 41.2 Hz, a difficult frequency to reproduce even given ideal circumstances, let alone at high power and with enclosures which must be able to be moved around! As a result, few bass systems have appreciable output below about 80Hz - and this includes the large "W" bins used for PA systems. Up to four times as much power may be needed at 40Hz to match the acoustic output at 400Hz (Figure 3).

Because it is undesirable to drive a bass amp into distortion, and the bass should be able to equal the guitar in volume, much more power is required for bass than for guitar. In fact if we assume a guitar amp rated at 100 watts and operating at an average power of 130 watts, and assume a bare minimum bass peak-to-average signal ratio of 10:1 (10 dB), our bass amp would need to be at least 1000 watts to match the guitar amp without producing audible distortion! A certain amount of clipping of transients is permissible, and in a well designed system will not be audible because of the masking effect of the rest of the band. It will, however, be very audible on a bass solo if the same volume is used.

Harmonic distortion in the bass region is usually very high, especially in the speakers. Some are capable of a total harmonic distortion approaching 100%! This "impossible" distortion level is caused by speaker "doubling" - the result of a combination of speaker-cone breakup, voice coils which actually leave the magnetic gap and poorly designed enclosures. The main faults in enclosures are truncated horns, excessive horn flare rates, badly vented cabinets, and poor internal bracing of cabinets which are often simply too small for the speaker. Intermodulation distortion arises from the large excursions of a speaker cone at bass frequencies which cause the voice coil to partially leave the magnetic gap. When this occurs, efficiency is lost and distortion appears at the peaks of the waveform. Any superimposed high frequencies (harmonics) will also be distorted.

• It must be noted that many of these problems have been reduced with modern loudspeaker drivers, whose excursion is often prodigious.  Linear cone travel of +/-10mm or more is not uncommon now, and there are many very efficient speakers that are specifically designed for use in small enclosures.  Coupled with the very high powers that are common today, it is possible to make a bass amp that will faithfully reproduce down to 30Hz in a cabinet that is still portable.

The obvious solution is to use more speakers in larger enclosures - and more of them - driven by even more powerful amplifiers. An effective alternative is bi-amping; an electronic cross-over separates amp-speaker combinations which may be more easily optimised for the frequency range being covered. This approach has been used with some success, and although it is more costly it will eventually become a standard technique. The advantages of a bi-amped system are lower intermodulation distortion, and greater acoustic output for the same total power as a conventional amp/speaker combination.

Unfortunately there are many trade-offs which must be made, both for financial and physical reasons, and many of the better combinations are precluded by prejudice. For example, the average. bass players' opinion of 300mm (12") speakers is not printable.

A variation on bi-amping is to use separate amps and speakers for a stereo bass guitar. In this arrangement the bass and treble pick-ups on the instrument are brought out separately and fed to separate amps. This technique has been around for many years but as few instruments are wired for stereo, most bass players have not had the opportunity to experiment with it. Any two pick-up bass can be easily modified however, and it is a trick well worth trying.

---

Keyboard Amps

The basic requirements, at least insofar as the bass and midrange are concerned, are much the same as for bass amplifiers. Power requirements are of the same order, and similar problems involving loss of definition may occur due to insufficient power, poor enclosure design and a lack of understanding of the demands which are placed on a keyboard system. These include the extreme dynamic range (if pianos, the vast pitch range of synthesisers, and the special difficulties presented by such instruments as string ensembles, Mellotrons and Leslie type organ speaker cabinets (which are usually not loud enough unless miked).

• Mellotron?? I'm not going to even try to explain this one (you wouldn't believe me unless you were there). Try a web search, you might well be surprised.

Obviously, the power required will depend on the volume at which the musician intends to play, but 100 watts would have been a minimum for all but the quietest bands. In a rock band this may have to be increased to around 1000 watts, with 450 watts on bass, 450 watts on midrange and 100 watts for treble.

An exclusively mid-range enclosure should not use speakers larger than 300mm (12') if a good transient response is to be obtained, and a flat frequency response is desirable.

 

Horn loading is an excellent way of achieving high efficiency, and if properly designed will provide low distortion, good frequency response and excellent transient response. The bass enclosure should not be horn loaded, because a horn which will work properly down to the lowest frequencies will be too large. Instead a reasonable sized vented enclosure should be used.

• A better alternative today would be to use a sealed and equalised enclosure (see the EAS subwoofer project on the ESP Project Pages).  This will give the best transient response, but requires much more power.  On the positive side, the box will be very much smaller that would otherwise be possible.

Passive crossovers are not recommended for high power keyboard systems except at the top end, from horns to "ring-radiators", where they will provide the most cost-effective solution without loss of performance. In general, though, their excessive power loss and the difficulty of optimising the crossover points makes them undesirable. A large keyboard system, then, is basically a very large hi-fi, with colouration being something to be avoided. It is not, or should not, be an extension of the instrument because of the number of different instruments which must be accommodated. Any colouration which is introduced should therefore be controlled by the musician.

---

Odds and Ends

The range of effects available is enormous and growing; the echo, reverb and tremolo of yesteryear have been joined by esoteric devices like automatic double tracking ("ADT") digital delays, phasing, flanging, chorus, true vibrato (frequency modulation), peak limiters, compressors, harmonisers etc and etc.
 

The simplicity of an amp like the classic Fender Twin Reverb (above) still appeals to many musicians. The Twin Reverb is used by both rock and jazz guitarists, and is also popular with electric piano players.

More important now than ever before are leads and terminations. Cannon type connectors are replacing 6.5mm jacks on most equipment, and guitarists are changing over to Cannons because of their greater reliability. Heavy duty, low loss cables are being widely used, providing less noise and lower failure rate than standard cable, which responds poorly to having a 200kg rack of power amps wheeled over it.

• Although there was (at the time of writing) a tendency for guitarists to use cannon ("XLR") connectors, this movement has seemingly died. This is a shame, since they are far more robust and reliable than 6.5mm jacks.  There are, however, a number of professional locking 6.5mm phone plugs and jacks available, that did not exist when the article was written.

Additional notes and comments have been added where appropriate, but for the most part the article is as originally written. The byline above was true at the time, but is now way past it's "use by" date - except for Elliott Sound Products, which is very much alive (although it is part-time at the moment).  Such is life.   The scariest part is that this was over 20 Years ago!

PLEASE NOTE: The inclusion of two photographs of guitar amplifiers shall not be taken as an endorsement of the brands featured, nor shall this statement imply any criticism of same. They were included in the original article, and are thus reproduced here. No more, no less.