Martyn Maynard tells you how to make your very own DI box.
Martyn Maynard of Audiogenic Ltd discusses the theory and practice of DI boxes
I can think of no area in the modern studio that has been neglected for so long as the design of Direct Injection boxes. It has always been felt that DI boxes are something that the maintenance engineer can knock up, and in general little thought has been given to their design. Few companies have bothered to apply their technical expertise to this area, probably because of the small financial return from the sale of such units.
It may be stating the obvious, but to recap: 1. using a DI box, one can connect directly to the instrument without any acoustic connection. 2. It is ideal where, for example, the sound output from the amplifier does not do justice to the instrument, or where one would like to feed the sound directly to the mixing desk. 3. It is also useful in a small studio that is bass-heavy, where it may be convenient to DI the bass guitar and not use an amplifier at all, to prevent the bass from farting around the studio and getting into everything.
When one thinks of DIing instruments the bass usually springs to mind, but any electric instrument may benefit: from acoustic guitars with Barcus-Berry type pickups to electric pianos. Often a combination of DI and microphone can be used, the DI giving clarity and depth, the microphone reproducing the raunchy sound of an amplifier, or the sparkle of an acoustic guitar. By combining both sounds and adding a little reverb some very pleasing sounds can be obtained. Not forgetting, of course, the connection of electronic instruments such as synthesisers to the desk.
Before going into the design of DI boxes, we must consider the pickup that we would like to connect to them. Two types are in general use, the first of which is the piezo-electric variety. This works by mechanically stressing crystal 'sandwiches' between two plates, producing a voltage which is fed to an amplifier via the wires connected to the plates. The whole assembly is fixed to the soundboard of the instrument or forms part of the bridge over which the strings are tensioned. The other type is a magnet around which a coil is wrapped, placed in the vicinity of the strings (or tines in the case of an electric piano). Movement of the strings upsets the magnetic field producing a current in the coil. The magnetic pickup likes to see a high input impedance; the lower the impedance the more current is drawn, so the voltage drops. This drop in voltage is proportional to the resistive and reactive components of the pickup, and generally the lower the input impedance the more the top frequencies sulfer. A typical pickup is shown in fig 1. C1 + R1 provide a treble roll-off with R2 controlling the output level.
The piezo-electric pickup also likes to see a high input impedance; low impedances tend to cause a loss in bass frequencies. The input impedance to a recording desk is typically 600Ω so the obvious and most simple way of building a DI box is to match the impedance by using a transformer. Using a turns-ratio of 10:1 we could make a DI box with an input impedance of about 50-100K, and by using a higher turns-ratio, higher input impedances may be obtained, (fig 2). But nothing unfortunately, is ever that simple: firstly, the higher the turns-ratio the more unstable the transformer becomes. Nasty distortions like ringing may occur. Secondly the 'insertion loss' increases as the turns ratio increases and with a 10:1 transformer an insertion loss of 20dB is likely, which means we need 20dB more amplification. Consequently there's a lot more noise! The problems are further aggravated by the circuit used in the instrument, as can be seen in fig 1. When the volume is adjusted (R2) the impedance changes and this in turn changes the load conditions presented to the DI, resulting in a change of tone with volume due to the mismatch with the transformer.
One way to overcome these problems is to use an 'active' DI box (fig 3). There is a clever little device on the market called a FET which offers an extremely high input impedance. With a handful of components an amplifier can be built having an excellent noise performance. Alternatively an IC 'op-amp' can be utilised. By using a 10MΩ input impedance we can remove any loading problems because the pickup sees an almost open-circuit. The DI output can be arranged to give a low impedance suitable for driving the desk, and a transformer can be used having an acceptable turns ratio. Also, additional low impedance unbalanced outputs can be arranged to drive amplifiers, pedals, etc. The power for the amplifier can be obtained from an internal battery or from the standard microphone phantom-power supply. An input impedance of 10MΩ is easily achieved with excellent stability, and the unit need not have any insertion loss — in fact it may have gain.
It is often necessary to insert the DI box between the instrument and the amplifier, a clean feed being sent to the desk while also driving the amp. By using an active box in this fashion with a 10M input impedance the instrument will not 'see' a load. (It often improves the sound to remove the instrument's internal controls altogether, in other words, turn volume controls full up and top-cut controls off). The output is sent to the desk and an auxiliary low impedance output is used to drive the musicians' amplifier, which has the additional advantage of reducing the amount of noise the amplifier generates (as this is proportional to input impedance). This effect pleases musicians and generates comments like 'I've never heard it as clean and as tight as this before'. This is purely because such a unit becomes 'transparent' when connected in this fashion, unlike a simple transformer which will affect the quality of sound and have a noticeable insertion loss.
In conclusion I can only say that in my experience, the active DI box, although a little more expensive to produce, offers a much cleaner and more dynamic sound than a passive unit, although in some applications like interfacing with electronic instruments with lower output impedances, transformers alone are cheaper and quite satisfactory.
Feature by Martyn Maynard
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!