Bringing together decades of experience with mic preamps, Focusrite's latest innovation is the RedNet preamp – a remote-controlled, DC coupled, capacitor free design that is the most open, transparent and natural sounding pre Focusrite has ever made.
By employing techniques that are seldom seen outside the rarefied territory of the audiophile, the new Focusrite RedNet preamp offers next-generation sound quality for any modern audio recording application.
RedNet designer Simon Jones sat down with Richard Elen to discuss the technology behind the preamp.
There is novel circuit implementation around the op-amp – novel, at least, in professional audio. “There's DC servo correction around the gain block, essentially," says Simon, “so that you don't get DC offsets building up as you are ramping gain, for example." Any amplifier, however good it is, will generate DC offsets. If there is a DC offset on the input of an amplifier, it will be much larger on the output. The idea of the servo is to try and correct that by applying an offset to the input, appropriate to the level of gain, to correct any offset that exists on the output.
In conjunction with this, almost all the signal path is DC-coupled. The only AC coupling in the entire path between the input and the ADC is at the input itself, a feature that is there solely to block phantom power from the input signal. In addition to facilitating the DC servo, this also helps to keep the signal path as clean and open sounding as possible. Neither DC servos nor DC coupling are familiar in most professional audio equipment; they are techniques more commonly found in high-end audiophile gear. But it's techniques like these that help to give the RedNet mic preamp a unique and characteristically open, transparent sound.
The input signal conditioning circuitry in RedNet is a classic Focusrite transformerless design, which ensures that the sound remains as clean as possible. The fact that the load impedance of a microphone can dramatically affect what it sounds like is extremely familiar territory to Focusrite, thanks to years of experience with the ISA range and other designs, and in the case of the RedNet mic pre, the microphone loading was chosen based on this experience to give a neutral sonic balance to a wide range of microphones.
Overall, the intention behind Simon's design for the RedNet preamp is easily summed up. It is, as he puts it, “quiet, clean and flexible." This design was coupled with a systems approach based on Focusrite's experience. “We know how to do mic preamps – we've done them for decades," says Simon, “and we also know how to do good conversion."
RedNet being a networked system, it is essential that the preamp was able to be remote controlled. There are a few major considerations in remote-controlling a preamp.
First, are you going to do it discretely yourself – ie implement all the separate gain steps yourself; second, how small do you want the gain steps; and finally, what is the ability of the system to be able to offer click-free gain changes? Dealing with the latter question first, there are two ways to reduce noise during gain changes: you can either only make changes at zero-crossings, or make the gain changes very small.
Zero crossing detection will generally work very well as a means of avoiding noise generated during gain changes. “The biggest problem we as designers have with any digitally-controlled gain stage is that people will always listen for clicks and other artifacts when there is no signal," says Simon, “which isn't a fair test." Zero crossing works by changing the gain when the signal passes zero volts, within a tolerance band of around +/- 5mV. The problem is that if there is no signal, you never get past that tolerance band. There can also be DC offsets that stop the zero crossing point being reached when there is no signal.
Consequently, there has to be a time-out sensing system that essentially changes the gain regardless of whether zero crossing is not achieved within a certain period. What people hear as they make gain changes in the absence of signal is the system being forced to change the gain after the time-out period – because there's no signal. This makes it a rather unfair test. A better test is to put a relatively low signal in, attempt to ramp the gain, and listen to it then. In fact, in most of the cases where you want to change the gain of a mic preamp (as opposed to a fader) there will be plenty of signal (perhaps you have to turn the gain down a few dB because something is particularly hot) and here, zero crossing works perfectly well.
Remotely-controlled-gain amplifiers are available using this kind of approach, and the one used in RedNet is supplied by That's (best known as the company that took over manufacture of the dbx chips). It's a two-chip solution. The gain control component – a resistive ladder and zero crossing detection part of the equation – is handled by a 5171, while the operational amplifier is a 1570.
After performing a number of listening tests, Simon concluded that the zero-crossing approach provided the most natural gain control for RedNet, sounding “very similar to analogue control rather than discrete steps."