Boss call it an 'Unbalanced to balanced convertor', yet the primary purpose behind this, like any DI (Direct Inject) box is that of buffering and isolation.

By converting instrument signals to a low impedance, balanced feed, control-room grounds can be separated out from the muso's own at the flick of a switch (the idea being to skip the ground-loop buzzes and related hash). Top-end losses can be avoided, even if the cables are lengthy and reach back to the desk.

The secondary function of DI Boxes is as signal conditioners. They can set up a comfortable level for the desk's mic inputs, filter, and invert polarity. Overall, the role of any DI box is relatively simple and we should expect a modern design to process the signal without prejudicing the sound's purity.

Functions

Looking at the block diagram (Figure 1), the music signal entering the Boss begins by encountering a parallel 'Link out' jack, which is convenient whenever we need to break into the line or speaker level interconnects, rather than feed the DI off the dedicated DI outputs found on some backline amps.

Next in line is a 3 position attenuator graded in 20dB steps; 0, -20 and -40dB. Relative to a nominal operating level of -18dBu (100 millivolts), at the output, the Boss DI can handle a maximum of +22dBu (-18dBu to +40dBu). That comes to about 9 volts which is insufficient to link up the DI to the output of (for instance) an instrument amp, unless it's a practice combo of under 15 watts. Otherwise, it's fine for all other usual sources.

Having set the level, the signal passes to the 'Unbalance to balance' convertor, which is nothing more elaborate than a discrete line-driver circuit with a balanced output. The idea is akin to a bridged power amp; in particular in that it omits any sort of transformer.

The 'Phase' switch comes next. Like the switch on your console's front-end, it's available to strictly invert the polarity of the signal.

In the 'Normal' position, the balanced output's pin 3 is hot. This polarity is contrary to the standard XLR wiring pattern, where pin 2 is always taken to be hot. Of course, many UK consoles, even new ones, are still wired to the obsolete USA standard (pin 3 = hot), but if yours is wired pin 2 hot, transposing the pin 2 and pin 3 connections on the DI1's XLR socket will help prevent possible subsequent confusion, not to mention foul language.

The action of the vital groundlift switch is evident from the block diagram (or is it?). If like Dr. Marlowe, Boss use the chassis-ground symbol to indicate signal ground, the question is whether the Boss DI's chassis is floating, (ie. connected to anything in particular). In fact, our examination confirmed that the DI box chassis is insulated from both the input and the output grounds, and also that the ground-lift disconnects XLR pin 1 from the signal ground on the input side, as anticipated. This is good news for avoiding inadvertent hum loops, but screening against RF interference would be greatly enhanced if the chassis were tied to the input ground. The irony here is that the only time a Japanese maker has bothered to isolate the jack sockets from the case, it's all wrong... In any event, pin 1 isolation has important ramifications for the phantom power convertor, for in case you hadn't guessed, the DI1 can be powered off the mic lead if your console has phantom power, but not if you use the groundlift! Boss mention this at the bottom of their instructions, but it could lead to some puzzled looks if you relied on the phantom power exclusively, and had only just got around to experimenting with the ground switch.

Powering is otherwise by PP3, and here Boss have incorporated their proven battery-saving skills. First, the DI's internal battery will only turn on when a jack plug is hooked up to the input.

Second, by selecting the 'Auto' position on the power switch, the unit will turn off, but only when the input signal has disappeared, and 15 minutes have elapsed. And what's more, the DI will turn itself on again, should a signal subsequently re-appear; probably when the pub has closed!

Evaluation

The Boss DI's performance was tested under simulated working conditions with a low impedance plus a length of paired cable hung on the output. The frequency response figures are satisfactory, but in studying the relative differences between the 0.1 dB and 1dB breakpoints for the upper and lower ends of the spectrum, you can see that the bass is dropping off quite rapidly. This indicates a considerable low-end phase-shift, capable of offending the absolute purity of some bass instruments. A lower bass rolloff would be desirable, but don't run away with the idea that competing DI boxes are any better.

The sensitivity table goes to confirm that headroom on a 9v battery is limited. In particular, note that the maximum output level is around 20dB - lower than you'd expect (at -20 dBu), should you switch the attenuator down to -40dB. Boss quote the distortion at under 0.05% without any other qualifications, but our figures are at least 10dB higher. Nonetheless, the average 0.2% distortion returned is typical for all sorts of budget recording gear especially when tested with a realistic output loading, and is certainly no worse than the many consoles stuffed with TL072 chips. Also, the absence of a dramatic rise in distortion at low and high end frequencies is a good sign.

Andy and Rick at The Chapel proceeded to evaluate the box in three sessions with a DX7, electric bass and guitar. They rated the Boss box well above those in their existing collection; this one didn't crack up at the top-end, and the bass wasn't hard or distorted.

On the practical side, DIs should ideally exhibit marathon ruggedness. This is because unless someone bothers to gaffer them down, their light weight soon sees them landing on the floor, just as soon as someone tugs or trips on the wrong lead. Once 'safely' on the deck, it's inevitable that the poor, hapless DI will be trodden on from time to time. The unlucky ones even get flightcases dropped on them, or squashed under a roving Steinway foot.

Boss have obviously developed some awareness in what befalls little boxes placed underfoot of musos from their experience with foot pedals, and so accordingly, the DI1 is housed in a rugged aluminium tube section. This is excellent, but the profile format has nevertheless been ill chosen because it places the jack sockets horizontallly, which makes them much more prone to accidents involving wandering feet. More serious, if the DI is held to the floor on purpose (as if a heavy weight were on top of it), an inserted jack plug can provide enough leverage when trodden on to bend the panel outwards and possibly crack the board inside. This is the penalty for soldering the jack sockets (usually adviseable) to a fragile, bonded-resin paper PCB. In the long run, Boss could try using the rugged fibreglass PCBs that UK DI box makers have been using for a decade, but an excellent short-term remedy is to unsolder the jacks and simply hard-wire them. Even a pair of support bars welded across the upper half of the case could have the same effect. This would prevent the panel bending outwards, which in turn relieves the PCB. Provided you take care in the meanwhile not to stand on any inserted jacks, I'm happy to report that the remainder of the construction was to a uniformly high standard. In particular, all the fixing screws are locked and access for servicing is quick and easy.

Conclusion

The Roland DI is an above average DI unit, but is also relatively expensive and not 100% fool-proof. It costs around 25% more than some possible competitors with a rock 'n' roll stage pedigree, so the DI1's sound qualities must be very good indeed. When buying, this means a doubly careful audition, cross-checking on the most difficult instruments such as guitar and electric bass. This is where most DIs fall down, but I don't think Boss have over-priced their first DI for this very reason.