The first test subject was a prototype constructed using simple resistors with a 1% tolerance. The tonal result was promising and very musical, however it was missing a bit of resolution. This was only the starting point for further improvements, which ultimately ended up with the use of the high quality Vishay 0.01% Bulk Metal Foil Resistors. Each unit requires the use of 96 resistors! Since these precision resistors are only available in the two-digit Euro range, one can estimate that the whole unit is therefore not exactly bargain priced. The control functions of the individual resistors for the conversion process are assumed by FPGAs. FPGAs (Field Programmable Gate Arrays) are integrated circuits for digital technology, in which logical circuits can be programmed. These ICs are often used, especially when fast signal processing is crucial. Now I'm about to discuss the many possibilities that are available with the Brient components. 

An interesting side effect of this discretely built converter: The totaldac does not require a current/voltage converter, usual in any conventional DAC. This is usually implemented using op-amps, which in turn also has an effect on the sound. The discrete converter already provides 1.4 volts rms output voltage, so that the tube output stage obviously acts as a kind of buffer to separate the converter from the outputs. More detailed information was not available to us since Brient does not want to show all of his cards. 

For the test, I ordered a version of the DAC with an ECC82 in the output stage of the d1 tube DAC. The units are available in different versions - a version with a transistor output is also possible, as well as a balanced unit. The latter employs two complete converters, in which twice the number of Vishay resistors are required! 

All versions include a volume control, working with 69-bit resolution in the digital domain in the FPGAs. I hear a gasp now among some readers! Among most digital attenuators there is indeed the problem of bit-loss resulting at low volume levels with high damping. The version used here should eliminate this problem of bit-loss - but regarding this topic, Brient is also not very informative. This secrecy is of course understandable due to the great amount of knock-offs coming out of some countries.

I'm quite curious. The whole thing is controlled – in fact exclusively - using a plastic Philips remote control. This machine certainly deserves a more chic version, but it does of course do the job properly. If the remote control should get lost somewhere in the crack of the couch, you won't be able to operate anything at all!

JIn any case, with an output impedance of 420 ohms you can also drive a power amp directly from the tube DAC. Thus, with Vincent's DAC you are getting a simple, but very sophisticated front end. As I mentioned before, the tube DAC delivers a maximum output of 1.4 volts rms, so you should know what voltage is required by the power amplifier to max it out. Commonly it is 2 volts or less.

What other special features does the DAC offer? The remote control can control several functions:

• FIR: the totaldac is designed as a non-oversampling DAC. Experience shows that it will play somewhat  more harmoniously, having a more natural sound. The disadvantage is a -3 dB frequency drop at 20 kHz. This is not a characteristic of the totaldac - this effect occurs in all non-oversampling DACs. For those of us who have ears like bats, this drop can be compensated via a switchable FIR filter. 
• PHASE: Here the absolute phase can be switched.
• EARTH: Using another switch, you can separate the signal ground from the power supply ground. Here you just have to try the best variation.
• BASS BW: A high-pass filter, which lessens the deep bass, listening sessions at the witching hour...

But it gets even better for the connoisseurs among us - Brient offers an addition to the DAC. This includes a digital crossover with a choice of up to three channels. Here, crossover is performed in the digital domain in the FPGAs. Thus, you can select all crossover parameters such as crossover frequency, filter slope, filter order and digital delay. While this is a very interesting solution, it would be beyond the scope of this report and the test machine is not set up for this.