This review page is supported in part by the sponsors whose ad is displayed above
|Hervé Delétraz Interview
Recently I had the remarkable opportunity to combine business with pleasure during a trip to Europe and spend a day meeting with Hervé and Serge in Geneva. This proved to be one of the highlights of my trip. Having never been to Switzerland, I had no idea what to expect. Regardless of your education and experience, we are all saddled with preconceived notions. Mine was the vision of the Swiss as cool, polite and somewhat aloof. Boy was I wrong. Hervé and Serge were warm, open, down to earth and both gifted with a great sense of humor. They welcomed me into their homes and lab and through this interview, offered an insight into their wonderful driving obsession. The conversation bounced between history and technical theory and mixed with some refreshing honesty as you will see. During my various meetings and discussions with Hervé, a single word stood out prominently in my notes: "passion," defined by Webster as intense, driving, overmastering feeling or conviction. That certainly defines Hervé [lower right].
Q: Tell us a little of your background and your beginnings in audio.
A: In 1984, I received my diploma in engineering and began to design a digital amp. I wanted to make an amplifier of the simplest possible design. I built a prototype amplifier in 1999.
Q: A digital amp?
A: No. It was the darTZeel Model 0, a big monster - big case but pretty empty inside. I designed it for myself and friends, not as a commercial product. The problem was that my friends did not want to give me back the prototype because of the sound. My friends encouraged me to make a series because the sound of the prototype was that good. So I was encouraged to continue.
Q: (to Serge) How long have you and Hervé worked together?
A: I have been part of darTZeel from the beginning, when the first circuit board exploded as I turned on the Variac. Today, I handle sales and marketing and manage other areas. Hervé and I are a good complement. I have total trust in his design genius. On the marketing side, you have to focus all the ideas and imagination to make a product successful. The original amp was built for Hervé personally. He wanted the best for his system. The only purpose was for it to be loved by a listener but eventually it became Hervé's job and he involved me in this adventure. When you translate that to a commercial product, priorities change. You cannot wait 16 more years for a new product. The challenge was, how do you translate the passion to a commercially successful product?
Q: What was so special about the prototype?
A: The problem was in fact that I was not worried about harmonic distortion because in instruments, you already have a lot of natural harmonics. Most musical instruments have 20 or 30 percent natural harmonics, so limiting harmonic distortion is not necessary. If you have only 1 percent harmonic distortion compared to 20 or 30 percent of natural harmonics, you don't hear a difference. So my quest for this amp was more for speed, less phase shift and simplicity of the circuit. The new 108 has only six silicon junctions from input to output for the signal path. There is no feedback at all from the output, the schematic is very simple - but it is not simple to make it work properly. I tried for years and finally developed the circuits of the 108 which are patented worldwide because they are, in fact, new designs. No one had a design like this although some might look similar. Now the story of the darTZeel company begins. I showed the amp at the New York 2000 HE Show for the first time. At that time, I had one distributor, Jonathan Tinn in the United States. In a short time, we were distributed worldwide, including Europe, Japan and Asia. It was not easy building this network with one product because the preamp was released much later than anticipated.
Q: What about the unusual name?
A: (Serge) The name is an anagram of Hervé's last name, Delétraz (darTZeel). In addition to the meaning, the name has visual impact and creates the image of technology and circuitry.
Q: With the growth of home theater and the decline of two-channel systems, why pursue such a small niche market?
A: We have two ears, two eyes and I think we can pretty well reproduce music with only two speakers. At a concert, the musicians are in front of you, not behind you. You can have the sound reflections which give a rear sound without the need for additional speakers. I do not like rooms with too much absorption.
(Serge) I have known Hervé since we were children because we are cousins. So I can add an explanation. I think that home cinema is associated with TV or visual images and both Hervé and I are aficionados of sound and music. Home theater was originally dedicated to the image first. It is only recently that theater sound is being addressed by the high end audio companies. Our goal is to produce the ultimate in music reproduction. I won't say that we will never make a multi-channel system but not today. The sound tracks in movies are at times much exaggerated. However, with well reproduced music, you just close your eyes and you can feel the experience. For Hervé and me, it is all about the music.
Q: Do you design for a particular sound?
A: No! Generally you can associate a sound with a brand. Designers make a family sound but that was not my goal. My goal was to accurately reproduce the musical signal as close as possible. I do not think about making it sound like tubes or transistors. I just wanted to make a good amplifier.
Q: Describe your design process.
A: I design first using a computer simulation to make sure that the electronic design works. We then build a bread board prototype. We listen intensely and after we are satisfied, we move to a circuit board and listen again. Listening is the most important part in the design process.
Q: Your speakers are unusual. Tell me about them.
A: I love these speakers because they are very unforgiving. They are by Rehdéko, a French brand. It does not exist any more. Its original designer was a real magician. He died since and I really miss him. I know his son and the father who left instructions that after his death, the speakers would stop production. There were many secrets in the manufacturing of these speakers, almost like a Stradivarius violin. The speaker membrane for example is treated with many different varnishes. Many of the varnishes are aged for maybe seven years before being applied. The result is that the speaker membrane is more rigid than Kevlar. The speaker is very fast and slightly aggressive in the midrange. It is amazing. The designer, a musician, wanted to reproduce instruments in the most natural way. So when you hear a trumpet or a saxophone live, it is very aggressive - these speakers are exactly the same. It is not hi-fi sounding, but very musical. They are highly efficient, about 106dB for 1 watt at 8 ohms.
Q: Are these the speakers used for your design and voicing?
A: They help a lot in the design because if you have a tiny problem in the circuit, you can easily hear it immediately. They are great design tools, but ultimately, they sound like music.
Q: What kind of music do you listen to?
A: Only a small part is classical - I'm too young. But seriously, I like all types, including jazz and rock old rock groups like Genesis, Pink Floyd, Jethro Tull and the Beatles. I appreciate classical, but it is much more difficult to reproduce accurately. Even with the best system in the world, it is extremely difficult. One hundred musicians playing in natural space, 20 to 30 meters from the audience at full volume is difficult to capture on a recording. Jazz is easier to reproduce. Three, four or five musicians could actually fit into the room so with jazz, you are closer to reality.
Q: Once you get to the basic sound you want, how do you tweak the design to meet your goals?
A: Again, we do not particularly tailor the sound to any style. On the power amp, we could tailor the sound. We could, for example, make it rounder. The amp is extremely fast. If we made it a bit slower, you could make it sound more 'tuby'. Depending on the input impedance, you could round the bass but it is not my aim to tailor the sound. I look for accuracy. I will tweak the value of some components however, a higher or lower resistor, for example, to add a bit more current for better speed. The relationship between technique and sound is very close because a good measuring amplifier can often sound bad, but the contrary is more difficult. It is difficult to have an amp that measures poorly and has good sound. There are some, maybe, but it is not easy.
Q: What about tube amps, some measure poorly but are very musical.
A: Some tube amps, yes, but I think that with tube amps, if you keep them for two or ten years, you could end up with fatigue because the sound becomes too polite, too round, too smooth. With the 108, we noticed the opposite effect; the longer a customer listened, the more they liked it.
Q: Discuss the long break-in period for your amps
A: Reviewers have written that the amp needs a lot of burn in to make real music. Agreed. But I also believe that the listener adjusts to the burn in.
(Serge) I think both the equipment and the listener adjust, especially with the darTZeel. The sound presentation with the darTZeel amp is a bit different than other pieces and you might need some time to get adjusted.
Q: How does the circuit design of the new preamp compare to the 108 amp?
A: The circuit design of the preamp, phono section and amp use the same patented design. In fact, the design can be used for any circuit - microphone, preamp, phono or power amp. The schematics are the same, but of course, we don't use the same devices. The amp uses power devices and the preamp, small devices - transistors. We use very low noise matched transistors and there are no op-amps in the signal path.
Q: Your designs focus on eliminating what you call 'temporal' distortion. You indicate this as the most audible type of distortion.
A: Temporal distortion occurs when you have phase shift on the bottom or the top. Reviewers might measure the square wave response of high frequencies. I've never seen a square wave response of a very low frequency. Normally an amp begins to show measurements at 100Hz. To have the plateau tilting indicates phase shift. An amp going down to 10 or 20Hz might have an excellent rise time but the plateau is tilted.
Q: Is your circuit wide band?
A:The frequency response extends from a few Millihertz. We don't have full DC from input to output because if you have DC at the input, you could have trouble. I have a coupling capacitor but we go very low. So the phase response is very good, even at 20Hz you have excellent square waves. You can hear it! The bass is much more precise. It might appear to be a little lean but that is because it is so precise and so fast. The duration is shorter so it might appear less but it is very clean.
(Serge) People are often used to boomy bass that shakes the room. If you critically listen to the bass driven by a darTZeel, you will hear that it goes very deep but with none of the resonant effects because it is exceptionally fast. Bass with other pieces produces monotonic sound. Our circuit design allows you to clearly hear the individual chords. You will hear far more bass detail with darTZeel.
Q: Therefore one of the design secrets to the darTZeel circuit is the square wave response minimizing phase shift?
A: Ideally, the amplifying circuit should go 10 times lower and 10 times higher than the audio band to have correct phase shift. Our circuits go 50 times lower and 50 times higher. Square waves are a very helpful design tool. You can see a lot of things. At 20Hz we are quasi flat; the phase shift is less than one degree at 20Hz. In the high frequency, we extend to about one megahertz - 50 times more than 20 kilohertz to attain about one degree of phase shift at 20 kilohertz.
Q: And the effect of this on the sound?
A: In the audio spectrum you are free of any temporal distortion.
Q: There are many amps with wide bandwidth. Some solid state amplifiers with very wide bandwidth such as Spectral and Halcro do not sound anything like your designs. Why?
A: Normally, when you want to reach very high in frequency, you need to use feedback. This, in my opinion, is the holy grail of amplifier design - overcoming the dilemma of high bandwidth requiring feedback and low feedback designs producing low bandwidth. This was the struggle of my design. How do you make a high-bandwidth no-global-feedback design work? The only feedback we use is a small local symmetrical feedback loop for the voltage gain. There is no feedback at all at the output or input. It is extraordinarily difficult to make this circuit work properly so to answer your question, other designers use some type of feedback to achieve bandwidth. I do not and hence the amps 'sound' very different. Also, I use only one pair of transistors as output devices. When you parallel a lot of output devices, you will have fuzziness and you will lose focus.
Q: With the small number of output transistors, how do you handle heat dissipation and stress on these output devices?
A: Thermal stress with only two output devices requires a very sophisticated monitoring circuit -- kept outside the signal path -- which can sense actual power dissipated in the devices. The NHB-108 is the only amp in the world to offer 100 wpc (and much more under 4 ohms) while other amps with a single pair usually offer 50 watts at best. Each output device amplifies a slice of the music and the signal might not arrive at the exact same time from all of the transistors, even if they used matched transistors. Two or three pairs can give good sound, but with more than that, it is very difficult to do. The best sound is with one pair. In our future monoblocks we will use two pairs for additional power but to do this I will have to redesign the schematic.
Q: A scoop? Are monoblocks in the works?
A: No, not currently, but they will be in the future. I envision a 300-watt amp but the challenge to keep the same sound will not be easy.
Q: Since the circuit is similar for the preamp, I assume the design goals were also similar.
A: When we started to design the preamp, Serge and I decided that the preamp would be the ideal companion to the amplifier. It took time to scale down the amplifying circuits for the preamp. Also, the amp uses no switches, relays or fuses in the signal path. For a preamp, this is much more difficult to implement. You have inputs. How do you implement them without switching? Yet we did it. All of the inputs are selected without switches or relays.
Q: Therefore, each input has its own gain circuit?
A: No, although we have a circuit card for each input. We turn on the card when we want to play that input - all without switching. In fact, we apply power to the card using a proprietary light system. The idea was to use light to energize components and with this approach, we can select inputs without the degrading effect of switches or relays. The preamp is completely dual mono. We have two batteries per channel. The grounds are fully separated from left to right and from the chassis. It is essentially two machines in one. The volume controls do use potentiometers for convenience but the signal does not go through it. We do use a resistor ladder, but not in a conventional way.
Q: What is wrong with conventional volume controls?
A: You lose something. I use a system that varies resistance with light. However, until all the patents are in place, I would prefer not to specify the details.
Q: Visually, the preamp's internal designs look more like a PC than a traditional audio component.
A: It is similar to a computer. Each input has a board but without slots because I don't like contacts. The sound connections for each card are hard-wired. The cards contain the proprietary light switching circuitry which enables or disables the specific input. As with the computer analogy, each card is connected in parallel to the motherboard. There is also one dedicated phono board per channel. We use small buffers with four transistors, the topology being the same as our amp - both sharing essentially the same circuit design.
(Serge) The volume control is a special design that has no tracking error. The worst case is no more than a 0.2dB difference, even better than Levinson's laser trend volume control. In its useful range, we have about .1dB of difference. Again, this was done with standard components, not exotic parts for long term maintenance. The machine can be repaired in 10 or 20 years using standard components. It is possible to make an outstanding machine using standard components. It is just the way you use them.
(Serge) This is a very simple, elegant design that was extremely difficult to execute. As an example, I had a conversation with a potential distributor who told me that he also built an amplifier without global feedback but it physically exploded in a few days. My answer was, "Now you understand why it took 16 years to make this work correctly."
Q: Moving to the preamp, when did you begin the design?
A: (Serge) Four years ago. Hervé has thousands of design ideas in his head. One of my jobs as partner is to keep focus. Originally, Hervé probably conceived of the design at the same time as the amplifier.
Q: Describe the journey to get to the release of the new preamp.
A: I originally wanted to share my design with the most people before it became a commercial product. I wanted to share the circuit with the DIY community, share it with other enthusiasts because I knew that the machine would be very expensive to make. There is a lot inside. The preamp is even more expensive because there are more than 1,000 components in the design six hundred components for the motherboard and four hundred components for the phono preamp. Again, it is difficult to make this simple design work well. You need the correct power supply, the expensive 'mystery light' components for switching and volume, the DC offset has to be controlled and many other things. Although the signal path is simple, there are a lot of components to make it work properly.
Q: As to the design, why the unique gold and red case color?
A: It was a big controversy. All other brands are either gray or black. People said, "I can't put your amp in my system; it does not match." Now with the preamp, we have two matching components. Maybe when we launch a source component, like a CD player or integrated amp, the entire system will match. I wanted something different and fun. The 'eyes' and 'power nose' are on the amp. The preamp has a 'pleasure control' for volume and 'enjoyment source' for input selection. I believe that sometimes the audiophile life is taken too seriously. People listen for bass, midrange, highs, soundstage etc. but I just listen for music. I wanted to make my machines more fun, pleasing and human.
Q: You must listen for some specific sound qualities.
A: It is all about the music. It doesn't matter if the recording is good or bad; I don't care. If you have a good recording, it is better, but I prefer to listen to the music.
(Serge's answer to continue) Emotion! At the London Show, we spent a half day setting up the room. Our first listening session had no emotion. Maybe it was our exhaustion or not being used to the new room. However, the next day, goose bumps. We had three demo records to play. Every time we played the same tracks, I had the same emotional experience - even three days later. People would come back again and again. You watch them sit and unconsciously move to the music. They are touched by the emotion.
Q: I am fascinated by the mysterious volume control light system.
A: Mark Levinson, for example, uses a chip with trim resistors in a ladder configuration and they switch them with FET transistors. The problem with FETs is that you have a silicon junction and the signal is altered by the switch. You can use relays but the contact remains the problem. There is an electrical potential difference between two contacts so here we use a similar resistor ladder but we switch using our light system instead of relays. The entire process is digitally controlled with sophisticated software. The processor is always idle except when you touch a button so there is no active clock in the machine to disturb the sound.
Q: Keeping with simplicity, you did need a remote?
A: Yes, but only volume and mute since there are no digital readouts on the preamp. When the remote is activated, the LEDs change from red to green to show that the preamp is receiving the remote signal.
Q: No input changing by remote?
A:: No, generally when you change the source you have to get up and change the CD or record. Keeping with the original design, I did not want a remote but it was conveyed to me that this was necessary by our US importer.
Q: Let's take a quick tour of the preamp.
A: The entire component is dual mono beginning to end. The first board is the phono preamp, with over 200 SMD (Surface Mounted Devices) per channel but again, there are only eight junctions in the signal path from input to output. The balanced input card is a bit different because we put a transformer on the card. All inputs are selected between RCA or our BNC 50-ohm proprietary links. There are three BNC outputs for amplification (to allow tri-amping) and each output has its own dedicated buffer. There are no jumpers. We use tiny pads which are soldered or not soldered to make the switch for phono impedance and gain.
Q: Can the phono preamp settings be changed by the end user?
A: We can do it at the factory but if the end user can use a soldering iron, he can change it himself.
Q: What about phono gain?
A: The maximum phono gain is 66dB plus an additional 13dB in the line, so you have a lot of gain available.
Q: Resistive phono loading?
A: We have four resistors that combine for many loads from 100 ohms to 47k. There is no capacitive loading because the new MM cartridges do not require it. We do have an optional capacitor to kill RF interference.
Q: Again, is the resistive loading user changeable?
A: The end user can do it but it is better that we set it up. The standard configuration for MC is 60dB of gain and a 330 ohm load - a good standard.
Q: Speaking of cartridges, what is in your personal system?
A: We have a special EMT cartridge modified by Van den Hul. It is like a Colibri, just amazing but less hard on the sibilants and very sweet. He made 23 modifications, changing the diamond, cantilever wire, capacitors and more. It sounds fantastic.
Q: Let's discuss the batteries. A few designers have tried them with success but it is certainly not in the mainstream. The trend in high end preamp seems to be separate components with massive power supplies, high filtration and regulation with tons of current capabilities. Why battery power?
A: The problem with external power supplies is that you still have the umbilical connecting wire so the impedance in the preamp is higher. You need to add regulation inside it to lower the impedance of the power supply. No matter how much you filter the power supply, you still have residual noise. With batteries, you don't have any noise because there is no magnetic field.
Q: I am not an engineer but I have heard that batteries while low in noise, suffer from their own impedance problems.
A: That is true when you use a battery to power the circuit directly. Manufacturers sometimes think, "Okay, we use batteries so we don't need to regulate because there is no hum or alternating current." But it doesn't work like this. We have a twelve volt battery regulated at ten volts. The regulators lower the voltage because when a battery is full it outputs thirteen volts and declines as it discharges to about eleven volts. This is not good for the circuit if the voltage is not constant. Also the impedance of the battery is lowered so everything is happy.
Q: Describe the operation under battery power.
A: Like the amp, once it is broken in, you do not have to leave the preamp on or warm it up for best sound. It needs only fifteen minutes of warm up. Therefore, we recommend that you turn the preamp off when not used. When it is off, it is always charging the batteries.
Q: What is the playing time with batteries?
A: The playing time with batteries is 15 hours. After fifteen hours, the preamp will automatically switch from battery to AC and simultaneously charge the batteries. There is a small but noticeable difference in sound while it is charging. This is another reason not to leave the preamp on all of the time. When you are ready to listen, the batteries could be in a charging cycle.
Q: What about battery life?
A: The batteries, of course, are the highest quality Japanese Yuasa or American Genesis batteries. If you leave the machine on continuously, battery life would be about three years due to the constant cycling. If you turn the machine off after listening, the batteries will not be completely drained so the charge can start with maybe one half power reserve and you can expect perhaps four to six years of use.
Q: Does the preamp have to go back to Switzerland for battery replacement?
A: No, local service repairs in each country can change the batteries.
Q: What type of battery is used?
A: Twelve volt, five-amp lead acid because these are the only batteries which do not have a memory effect. They can be charged at any time at any level of discharge.
Q: What is your recommendation for after-market power cords?
A:: I tried some but I am not a tweaky guy. Tweaks will change the sound but I am not sure that they will improve the sound - different maybe but not necessarily better. If the incoming power is good, the cabling will have a lesser effect. You can get better results from just cleaning your contacts.
Q: Tell me about the Zeel links. Why create a non-standard connection?
A: Our system is single-ended. I really don't believe in balanced. With balanced, you either have to use transformers or twice the electronics. The preamp has balanced inputs and outputs with transformers. I prefer them because they are immune to hum. They are there to be compatible with the professional standard but for me, single-ended is much simpler. In addition to the single-ended and balanced sockets, we have the fifty-ohm input and output.
Q: If you were happy with single-ended connections, why invent another type?
A: There are too many differences among audio interconnects. These differences are not marginal. I examined the problem and wondered if it is really the cable itself or only a problem of mismatch. I conducted an experiment at the Geneva Engineering School and it showed that if you are not matched, you will have some signal reflections. Although these 'echoes' occur above the audible threshold, I found that when you are matched, the sound is much cleaner because you do not have these reflections. It is like the integrity of phase or speed. You do not have any timing errors.
Q: What is the maximum length of these Zeel links?
A: One kilometer! Yes, one thousand meters of cable! There is no loss of highs or transients because the signal is matched. When you have a reflection in the cable, the signal bounces back and forth. Square waves become rounder not because of capacitance or inductance but because the interaction of the reflected signal causes the square wave to round over the edge because each return step is added together not in phase. The sum of these reflected steps is a rounding edge of the square wave. When you are matched, there are no reflections and it works.
Q: Can you hear the difference on a short run of, let's say, seven meters?
A: I can tell the difference with a one meter run. I compared my old Kimber Reference Cable, which I loved, and while less than longer runs, it is still clearly audible - less glare.
Q: Do you supply the 50-ohm cable?
A: While any fifty-ohm cable will work, we use specially made Swiss cable - small, thin, very good isolation, silver-coated, and they sound slightly better than standard cables. Since they are very thin, they can be run under carpet or near the walls.
Q:: What does the future hold for darTZeel?
A: In the very high end, we will make a monoblock amplifier and possibly a source component like a CD player, but coming back to our 'sharing the pleasure' philosophy, we will then make an entry-level product such as an integrated amplifier. This is a real challenge - where to cut costs while keeping the darTZeel sound. Our goal will be to make a product available to 95% of the community. Sometime in the future, we might make a 'cost is no object' piece, maybe only ten pieces just for the fun of it. The success of darTZeel up to this point has been accomplished without any advertising. We go to shows and let people hear the amp. The best advertisement is to let people listen to the products.