Sound on Stage
Interface 3: Front end Razzmatazz
Front End Razzmatazz — stage mixers under the spotlights
The interface between console guts and signals arriving from microphones and DI boxes on stage is one of a handful of places where things go badly wrong in even the most sophisticated systems. Conversely, good — or fortuitous — front end design is (along with fortuitous EQ design!) the secret of a good sounding console at all levels. Indeed, as this area is so important, we'll look at it in two parts, with some practical conclusions towards the end.
The elementary requirement at the front end is to attain an accurate and well controlled energy transfer. Accuracy implies routine avoidance to counteract the introduction or accentuation of noise and sundry hums and buzzes, not to mention distortion; controlled energy transfer is a more specific case of fussing over phase/frequency response, keeping an eye on ultrasonic energy and worrying about the matching/damping/tuning of transducers, wires and other copper hulks.
Within audio equipment, it's normal to voltage match each stage to the next. Pure voltage matching is a sci-fi concept involving infinitesimally small currents: it's 'all done with volts'. The source (say the output of a desk) would exhibit a zero output impedance, ie. regardless of the load impedance it looks into, or the current demanded by the load placed across its output terminals, no voltage drop occurs. Current is made redundant because the load (let's assume it's a power amplifier) would exhibit an infinite impedance, ie. steals no current from the source. Here, current flow is transcended by electrostatic pressure.
With Bi-Fet op-amps having Giga ohm input impedances, and MOSFET line drivers being available with current sourcing abilities measured in tens of amps, this isn't really so sci-fi after all. However, it's not very practical and it's also perhaps something of an overkill. What it achieves is a precise transfer of voltage — to within 0.00.. (many pages of noughts).. 001%. In other words, accuracy! Moreover, with suitable design, this finesse can be held over a wide range of frequencies despite the machinations of cable capacitance and other vile reactances. Real life audio equipment subscribes to a lesser (economy!) form of voltage matching. Here, the impedances/current capabilities of load and source respectively are scaled down on the proviso that if no more than 1% to 10% of the signal voltage is lost in the transfer, the matching is adequate. Hence the rule of thumb 'The load impedance should be at least ten times greater than the source impedance'. Hidden in this aphorism is the assumption that no reactances lie hidden between the source and load, which is reasonably valid for couplings between a pair of op-amps on a PCB, but becomes a nonsense whenever external connections are involved.
Getting down to microphones specifically, the first of several quandaries is whether voltage matching is valid for transducers per se. Magnetic cartridges, guitar pickups and tape heads are examples of other audio transducers which are not matched in any remotely 'pure' sense. This is largely because mechanical and reflected electrical parameters exercise a host of trade-off mechanisms. Nevertheless, the broad truth about all of these devices — including dynamic microphones — is that each can be designed to suit almost any (arbitrarily chosen) impedance. Magnetic cartridge designers have settled, by and large, upon 47k as a standard; for moving coil microphones (the same basic mechanism in a different box!) load impedances responsible for eeking out optimum results are (a) less well defined and (b), not governed by any universal standard. Microphones originating from the USA tend to prefer to see some 200 to 600 ohms, and are essentially power matched, ie, the output (source) impedance of these microphones is also circa 200 ohms. Power matching has no magic virtues here, and certainly doesn't make the signal 'more powerful'. In fact, pure power matching involves a 6dB voltage loss! Rather, the 200 to 600 ohm load is a fairly arbitrary figure to which the manufacturer has 'tuned' the microphone in terms of most even frequency response and the damping of resonances. The reason for the unfortunate choice of optimum load impedance is largely historical, being tied up with some faulty reasoning in the pioneering work of Western Electric and Bell Telephone in the 1920s.
European microphone manufacturers have been less fettered by traditional axioms, and accordingly their microphones tend to be optimised for higher load impedances circa 1k to 2k.
The microphone's source impedance is usually around 200 ohms, so they're not exactly voltage matched, but that's the basic trend. More important, noise performance isn't compromised any more than necessary in contrast to the 5 to 6dB voltage drops suffered when USA mics are used correctly, ie, power matched. Also, as most British and European consoles exhibit a 1 to 2k input impedance, microphones of USA origin suffer a mismatch. In practice, this means a higher than advertised output level, and more top response — or treble colouration if you prefer to be pessimistic! To an extent, this irregularity can be taken care of by wiring a resistor in parallel with the offending models. Something around 750 to 1000 ohms plonked across the mic's XLR connector should suffice.
Having enthused about voltage matching as a wondrous means of accurately transferring audio signals, and then shown that microphones aren't designed to operate at their best when pure voltage matched in any case, we'll now go on to look briefly at a further obstruction to this ideal.
For all active devices — be they valves, J-FETS or Bipolar op-amps — there's a certain magic source impedance with which noise can be greatly diminished. This point, which involves a balance between voltage and current noise sources, is in the 100k to 10M region for J-FETs and BI-FET op-amps, around 5k-20k for Bipolar op-amps and as low as you want for parallelled or multiple emitter Bi-polar transistors. This parameter is known as the optimum source impedance. It has nothing to do with the input impedance or the intrinsic noisiness of the device — rather, it's the impedance the input wants to 'see' for minimum noise.
At this point, we reach a situation where the impedances that the microphone and input amplifying device want to see for best results are mutually at odds with each other: enter coils of copper wire!
Placing a transformer between the microphone and the first stage of amplification within a mixing desk neatly solves the impedance matching problem, at least on paper. With a turns ratio of between 1:5 and 1:10, the impedances that both the microphone and the input stage device (assuming it's Bi polar) need to see for best results can be mutually attained. For instance, with a 1:6 transformer turns ratio, the 200 ohm mic source impedance is seen by the op-amp as 7.2k; the transformer has multiplied the impedance by the square of its turns ratio (62=36).
And looking in the opposite direction, the reflected secondary impedance (including the secondary termination resistor) is divided by 36, to appear as a 1k3 load to the microphone. Once you've come to terms with the ability of a network to have several different impedances at large simultaneously, there's a bonus: the transformer's turns ratio spells a proportional voltage gain. For a 1:6 transformer, the gain will be some 15dB. This is helpful insofar as it improves the signal-to-noise performance of the console, as the first stage gain can be wound back by 15dB. But then, as we saw in the first part, noise niggles are often academic in PA applications, and extra, untweakable gain is a positive nuisance when we come to mic up the percussion, horns and vocals of the more raucous species of Rock 'n' Roll.
Other sundry abilities of transformer coupling aren't unique, and have to be weighed against the potential of other techniques. Balancing is easily achieved with transformers, but again, in a nasty stage environment, it can be more hypothetical than real, as is attested by the bountiful hums, buzzes and radio broadcasts heard on most PAs. More down to earth is the potential of true balancing (where the primary of the microphone transformer has an accurate centre tap to which the mic's screen wire is connected) to provide 100% isolation (Figure 2). In this situation, voca-protected from live mains appearing in either direction. As to whether the average low cost microphone transformer is tested to withstand mains voltages appearing across each winding is another matter...
Against these facilities, of which the mutual impedance matching trick is the most unique and elegant, there exist a multitude of cynical responses. Good transformers are expensive, heavy and bulky, whilst the desires of sane PA equipment users are economy, levity and compactness. Secondly, transformers are routine whipping-horses whenever transient response is at issue. Whilst there's undoubtedly an audible loss of attack and incisiveness when percussive sounds are passed through poor input transformers, it's important to be aware that (a) most low impedance microphones contain transformers to step up the very low capsule impedance to circa 200 ohms, so the avoidance of input transformers within the mixer doesn't achieve very much, (b) in the real world, transformers merely act to aggravate a problem that already exists, namely the abberations generated by long cables and a coalescence of sundry reactances, and (c), rabidly perfectionist audiophiles who use well-designed transformers to match expensive moving coil cartridges will argue passionately that the damage to transient response wrecked by transformers needn't be audible, providing you pay enough money...
Back in the world of economic compromise, where something between £5 and £15 seems a reasonable sum to spend on each transformer, a console's input stage budget soon begins to look expensive. And for this apparent extravagance, one's rewarded with poor to negligible common mode rejection at radio frequencies. In other words, the line balancing fades away to be replaced by the local CB 'enthusiast' on 27mHz. The rejection of hum at the other end of the spectrum isn't necessarily too good either. Far worse, the microphone/cable/transformer/input stage chain can act together in an unholy unison to generate horrific resonances and phase abberations anywhere between — or beyond — audible limits. Even when the misbehaviour is at supersonic frequencies, the intermodulation and splatter effects can be all too audible. Aside from colouration and 'hard to pin down' nasties in sound quality, exaggerated signal levels in the 30kHz range have been known to blow up fast Bi-polar power amps and tweeters. Part of the problem lies in the large amounts of ultrasonic energy picked up by close miking techniques; if it could be filtered out prior to the transformer, there would be fewer unseen quantities to interact, but again, this raises questions about transient response.
As with loudspeakers, the trouble here is not so much that reactances exist, but rather that console, microphone and cable designers act as laws unto themselves, being prone to ignore the irksome and untidy implications of the stage. For instance, in the cosy confines of the laboratory, any mixer front end can be made to perform admirably when fed from a 200 ohm signal source, down two feet of coax, but this routine 'test' is wholly irrelevant to the environment of Rock PA. Once the microphone/cable/input transformer chain is seen as a group of components that have to be tested together with plenty of empirical fiddling (because the parameters of the cables and mics are forever varying), then it's possible with elaborate and finely tuned slugging to achieve very good results with transformers. If console manufacturers provided the appropriate knobs, perhaps we'd see sound engineers tuning up their transformers and cables for the best snare drum sound alongside musicians. But perhaps on the other hand there are alternatives to transformer coupling? In the next part, we'll look at transformerless microphone inputs and come to some practical conclusions.
Feature by Ben Duncan
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!