Lost In Translation
It’s important to understand that even though transformers are passive devices, they are actually quite complex with internally distributed resistance, capacitance and inductance.
Thus they still exhibit varying electrical and performance characteristics, just as active circuits do. They are, in fact, not dissimilar to loudspeakers – except there are no moving parts.
Earlier, we referred to CMRR (common-mode rejection ratio), which is one of the most important performance characteristics that audio product designers and system integrators look for when specifying transformer-coupled interconnects.
But an impressive value of CMRR does not fully define a transformer’s overall performance. It must also get high marks in each of the following characteristics if the end result is to be clean and transparent sound quality.
Distortion. Like an active circuit, a transformer will inevitably introduce a measure of distortion. In a high-grade design, distortion will be quite low, but present nonetheless. Transformer distortion is a function of level and frequency; the lower the frequency and the higher the signal level, the more distortion. In particular when operating limits of the transformer are exceeded (high level at low frequencies), and core saturation is eminent, distortion will increase rapidly. Core saturation occurs when a transformer’s iron core cannot absorb any additional magnetism, thereby “clipping” the signal.
Linearity. Rarely published as a specification, linearity describes how the other parameters (frequency response, distortion, phase response and transient response) will stay stable (or not) under a range of input levels.
Level. This specifies the maximum input and/or output levels at a specified frequency (normally 20 Hz or 50 Hz) before saturating and becoming non-linear.
Frequency Response. Like any other audio device, there is an upper and lower response limit. Within those limits the response may be perfectly flat – or it may deviate a little – or a lot.
Phase Response. This parameter describes any deviation from a flat phase response within the specified frequency range. Even among the leading transformer manufacturers, phase response data is not always available. When it is, it’s usually presented as a graphical plot, often with the frequency response depicted on the same graph, as they are proportionally related. In low-grade products, usually very little data is supplied.
Transient Response. Specifies how fast the transformer can respond to a short signal burst (or the leading edge of a continuous signal), and how quickly it stops emitting energy after the applied signal has stopped. Like phase response, this data is not always available, but can be extrapolated by examining frequency response and self-resonance data (pulse transformers for digital audio may delve into this specification more deeply).
Transformers that generally exhibit good frequency response and effective CMRR can still color the sound, sometimes dramatically, due to a slow initial response and significant overshoot at the tail end. Sometimes this can be pleasing to the ear, providing a sense of “warmth,” but most times it is not. In any case, it’s an inaccurate representation of the signal.
It’s better to use a plug-in, or a signal processor intended for such an effect, rather than to infect the system with a sub-optimal transformer. As with everything else in audio, when evaluating a transformer, it pays to spend time listening.
Many microphone types utilize transformers, either for step-up (typically ribbon transformers) or step-down (typically tube microphones), or 1:1 for dynamic microphones.
Some manufacturers, such as Cascade, offer upgraded transformer options for many of their ribbon mics. The company even specifies the transformer brand and model, which indicates how important the transformer’s contribution is in achieving optimal sonic quality.
Another common stage source is the direct box, or DI. For many years, all DIs were transformer based, and with good reason. An instrument plugged into a DI has one ground reference through the instrument amplifier, while the sound system has a different ground reference. As a rule, a transformer-based DI will solve a ground-loop problem faster and easier than an active DI.
Transformer-based mic splitters incorporate specialized transformers, with one primary and one, two, three, or more secondaries. We mentioned transformers as a form of insurance policy earlier. Nothing equates to more “audio insurance equity” than a well designed and built transformer-based mic splitter.
Mic preamps are another key application. While many brands offer only electronically balanced inputs, a good number of premium products are either transformer-based from inception, or offer transformer versions as an option.
Moving along the signal path, some high-end consoles employ transformers, and/or offer transformer options on line outputs. Considering that the console is where all signals come together, this is a good place to consider specifying transformer options, when available.
The console, in turn, feeds signal processing of all types: loudspeaker management systems, outboard equalizers and limiters, and/or banks of self-powered loudspeakers. All are candidates for transformer usage, especially in situations that vary from day to day, when there’s little time available to solve induced EMI or ground loop problems.
Present day audio transformers are appreciated and revered for their unique problem solving capabilities; they are perhaps more valuable now than ever before. Today’s audio systems have become incredibly complicated, with many interconnected devices comprising even a small system, and a staggering number of devices in large-scale systems.
In the past, transformers were the only way to get things done; now, in many cases, they may be the only right way.