When starting this company we had to decide between me-too strategy and innovation. We chose the latter, much more risky but also more rewarding when successful. Me too leads to mediocre products selling in the crowded market. We choose to be different.
We did not copy any solutions from anybody. Well, we learned good practices from very old engineers but not from competitors.
One of key factors is that Lukasz Fikus, chief designer – is not electronic educated but electrical engineer. This is a huge difference. Everybody else in this industry is electronic. That means particular style, that is taught at electronic schools. Electrical is different. it is much more primal and basic and it fits tube technology much better. You may disagree but our products speak for themselves – hi-fi from electrical perspective, not electronic. Everything that Lukasz learned other than his own experience and trial and error – was from his old masters from Ukraine who spent their whole life working with tubes and chokes.
What it implies , is that we don’t use opamps, and we use lots of chokes and transformers. We don’t build things as they are supposed to be built, we do what makes sense. We break tens of rules that an electronic engineer would never break but this gets us where we want to get.
What makes our DAC
The DAC which made us famous can be seen as a combination of four factors:
1.Selection of chipset mix (DAC and RECEIVER) that sounds best
2.Creation of immaculate tube analog stage that is an interface between the chip and the amplifier
3.Creation of power supply which makes a huge difference – for both the analog stage as well as for the chips
4.Selection of high-end parts like NOS tubes, silver wiring, copper foil PIO capacitors, iron core chokes and transformers made to our specs from Polish Copper.
Yes. Four main factors. Get every factor perfect and you have a great product. That is so easy that anybody who will spend 6 years times 6 people times 10 hours per day – relentlessly perfecting these issues – will get similar results, providing he has similar music taste.
Our design highlights
1. There is a lot of controversy about chips.
Some people believe in high specs – the chip MUST BE 32 bit and 380 kHz or it is not good. Some people believe in “old = good” and they accept nothing newer than Philips TDA1540. Some people believe that even good chips simply must be multiplied at least paralleling 8 of them per channel to get results. The internet is just full of opinions and testimonies about DAC chips. People simply have that need to pigeon-hole each chip with fast and concrete tags like: dull, fast, detailed, smooth, etc. No one seems to go deeper than skin-deep. And believe me – there is SO MUCH MORE to it than these tag words.
What we found is that the chip-set sound is: 10 % DAC chip choice, 50 % receiver chip choice, 20 % receiver-DAC connection interface and 20 % power supply construction. Yes, the DAC chip choice – from worst case to best case scenarios could affect only 10% on the subjective scale of sound quality.
We took the time and effort to test every chip we could buy or find in old products, it was close to 100 chips that we tested in similar condition and without output stage influence.
That explains why we don’t disclose the DAC chip name. We simply don’t want to be pigeon-holed by ignorant people, based on our choices.
2. ANALOG STAGE
It is more important than the DAC chip selection. Most DACs sound pretty much the same but it is the analog stage that makes the difference.
99 % manufacturers go exactly for the analog stage prescribed by the DAC manufacturer in their data sheet. Thats so embarrassing. They could do better than that. The Data sheet call for two opamps per channel and thats all. Some manufacturers use higher level own designs of opamps – thats not better, well maybe one small notch better. Some manufacturers go for discrete transistor stage – thats better and sometimes sounds great. The worst case scenario – God forbid – is opamps from data sheet PLUS TUBE BUFFER. Yes, it is so popular these days to stick two tubes just after the opamps. A good example is Shanling CD players, Opera Droplet CD, MHZS, Vincent CD6, and many more. We are definitely not in this camp.
99%的厂家严格按照芯片的规格书画出模拟放大级。这太难堪了。他们明明可以做得更好。规格书给每个声道放二颗运算放大器，然后完了。有些厂家用自己设计的更高级的运放电路， 这也不会更好，当然也有可能有一丢丢进步。一些厂家用分立晶体管设计此级。这就好些了，有时候非常好。最最差的剧情–上帝保佑–按芯片规格书搭好运放，再加上胆管做缓冲。山灵CD机就是这样的例子，Opera Droplet CD, MHZS, Vincent CD6。。。很多。(李总不要打俺，俺就是翻译。。。）。我们绝不跳这个坑。
What we do is we design the tube stage with pure triode single ended action into the DAC output and we perform I/V conversion, analog filtering and amplification in one step. The purest, most elegant way.
3. As we know POWER IS EVERYTHING.
We effectively listen to power supply modulated by the signal. So the power supply is even more important than the output stage circuit itself. Thats why 70 % of budget of the DAC goes to the power supply. It starts with selection of sections, how many points we need to supply and with what voltage and what amperage. In basic Level 3 DAC there is 10 sections, going up to 20 and more in Level 5. Every section sub-group has separate transformer winding so we must design custom transformers with many specialized windings. This is an expensive way of doing things but we believe it is the only way.
事实上，我们是用信号来听电源供应。所以电源供应比输出级本身更重要。这就是为什么解码器的70%预算用来造电源。从选择节点开始。我们先规划有多少个需要供电的节点，电压和电流参数。在基本款解码器Level 3中，有10个节点。到了Level 5，就变成20个节点。每个节点的地线支线都需要变压器中的独立绕组。所以我们必须订制有很多独立绕组的变压器。这很贵，但是我们相信必须这样。
For filtering we use different techniques, but lets focus on the most important supply – for the tube anodes. We use TUBE RECTIFIER, not silicon, even in the Amber and Level3 and in every higher level as well. after the tube rectifier we install capacitive output, followed by the big iron choke, followed by more capacitance and another choke. Yes, even lowest Level 3 has CLCLC filter per tube. This technique is RIDICULOUSLY EXPENSIVE. Other companies use CRCRC which has three caps for 1 dollar and two resistors for 1 C each. Our approach is to use film capacitors instead of electrolytes and chokes for resistors and it cost 100 times more. That is our way of doing it.
In every position where a good mil-spec part does the job, we refuse to use boutique parts. We leave boutique parts for the people who don’t know any better. But in some positions parts do matter. In short – we refuse to spend big money on AC power inlets like Furutech, We don’t use tantalum resistors or god forbid – carbon. We will never use black gates as heater filters. But for the digital section electrolytes, for output signal caps and the tubes too – we get fanatical. Also all internal wiring is silver in teflon where it is called for.
In all our DACs we offer USB input option. Our module is of a highest level, with 32 bit engine, capable of 384 kHz speed, compatible with all operating systems of all computers. It is asynchronous and uses two internal clocks for re-clocking. It also is self powered from internal power supply which we build into our DACs. No power is taken from the computer and the noisy computer line is PHYSICALLY interrupted and insulated.
Our USB is not converting to SPDIF – it is much better than that – it converts to i2s and it is connected DIRECTLY to the DAC chip, bypassing all input and receiver stages. There is no better USB solution than this. Our USB is also supplying the DSD signal to the DAC which means you have only ONE computer program and one USB cable and one USB port for both PCM and DSD technology.
Every of our DACS except Level 7 can have balanced output. We mean REAL balanced output, not fake one.
What we mean by real is that we run four independent parallel digital conversion processes in the DAC chips. Each of the four signals (Left+, Left-, Right+ and Right- are amplified by four analog stages and four triodes. all the way from chip to output we run it independently. What others do is that they sometimes do single ended process and at the output they add an opamp to create negative phase. Or they do it by output balancing transformer. Yes we do get balanced signal as a result but not the benefits of balancing which are only possible our way. Our balanced DACs have SE output as well