ES9018 DAC or NFB-11 Story PDF Print E-mail
Tuesday, 26 April 2011 18:41 | Written by Administrator

Sometimes waste of money is the only way to learn something new and valuable and that knowledge is priceless. If you didn’t waste a fortune, you are absolutely a winner. Getting similar experience through regular education is either impossible, or will cost 10-100 times more than you wasted (if you lucky to get dedicated and passionate mentor).

As new technology arrives, it is not always about forefront brands that make a cut. ESS Technology was founded by Fred Chan and Forrest Mozer in 1984 and employs barely 500 people worldwide (source: Wikipedia). Initial goal was to promote speech synthesis system (ESS – Electronic Speech Synthesis) for early computer systems like Commodore 64. Later company sold chips for computer’s audio cards, than – portable multimedia players.

First time I learned about existence of company named ESS from Twisted Pear Audio website and their Buffalo DAC based on ES9008 ESS DAC chip. As Buffalo DAC matured, ESS introduced new chip – Reference ES9018. After decent amount of R&D new Buffalo DAC arrived. I passed as availability and price were unacceptable and I was happy with moded and lampized ZERO DAC, investing time and effort into new tube amplifier. First commercial high-volume brand to use this DAC chip was OPPO with Audiophile BDP-95 Blu-ray player.

It is time now to switch attention to the hero of this story. Or is it villain? My curiosity and excitement grew as new articles about ESS ES9018 DAC kept coming with more and more sources calling it “the best DAC chip ever”. Mr. Kingwa from Audio-gd just came up with new integrated DAC and headphone amp with ES9018 inside – NFB-11. It also has USB interface delivering I2S signal to the DAC chip – this was a deal breaker. Price was promotional and offer promising, I had to try.


Order was placed in October and unit arrived in… February following year. Mr. Kingwa was struggling with number of orders and requests at first. As promised shipping date approached, new delay: USB module had problems syncing with ES9018 that works at 80MHz clock speed. Eventually everything was resolved and unit arrived at my house. Not expecting much from this unit without modification I got… not much. Lifeless performance and flat dynamics with hint of new details. It was like somebody placed pillow over artist mouth and was trying to suffocate music. As this might be a little exaggeration, DAC was playing like any $30 CD player from Wal-Mart with little more detailed sound. At this point even availability of USB interface could not save this baby. I could not take that slow suffocation of DAC chip any longer, so first action was to cut off all analog stages and plug amplifier directly to the DAC chip. Frankly, ES9018 can handle such intrusion; it will switch to voltage mode and give enough juice to drive amplifier. Result of such mode was not groundbreaking albeit promising. Still unlistenable, some clearness was breaking through with even more details. ES9018 is still suffering, badly suffering and it needs help. Some analysis of schematic:

1.  Insufficient power filtering and conditioning from rectifier down to power lines feeding DAC. Separation of analog and digital DC voltages is not enough for a good sounding DAC.

2.  Regardless of claims, capacitors are bad all around.

3.  Without second thoughts, existing analog stage is bad. With or without capacitors, two hundred transistors will not do a good job to deliver clear results. Even Mr. Kingwa’s ACSS, “Diamond” output stage and non-feedback amplification could not save the music from monster grip of semiconductor’s forest.

4.  PCB board quality and routing is not the best, signal traces not optimized. Overall, channels cross-talk is surprisingly high and reducing stage and depth of music.

5.  2 million transistors phone amp goes to scrap.

NFB-11 Inside

NFB-11 inside: usable part of PCB in red.

Careful amputation yielded 1/3 of original circuit board and few components like power transformer, switches and nice shiny aluminum knob used later for digital input selector. Rebuild started with new power supply. Since existing transformer doesn’t need to power phone amp anymore, it should be sufficient for the DAC with extra reserve needed for flawless operation. Ultrafast rectifier diodes (NOT MUR860, they are too powerful and will not handle low voltage gracefully. My choice was BYW29E) and almost 10,000 uF of filtering capacitors (all Nichicon PW) will suffice. Very critical update: 8 voltage regulators, one for each power pin of DAC, XO oscillator and USB chip. Voltage regulators are all new low dropout, high precision, high rejection and can be loaded with very low ESR filter capacitors. These new regulators are result of some serious research of impedance, filter capacitance and frequency. Parasitic resonance and variable impedance of old regulators (like 78XX or LM317 used in NFB-11) creates high frequency noise and disturbances unacceptable for DAC like ES9018 working with 80MHz clock. Final capacitors are Oscon SEPC placed almost on DAC power pins. DAC is very happy now.

DAC Board

Green PCB to the right - voltage regulators module. Notice little board with oscillator, fix for USB issues.

Time for analog output buffer. After careful consideration, Russian NOS 6N16B (6Н16Б) super-miniature tube was chosen. According to Lampizator, it is the best sounding choice, but doesn’t give the “pride” of owning equipment with sticking huge valves. Well, it is good with me, I need the best sound! Ten times oversized power supply with ultrafast rectifier diodes, chokes and huge capacitors bank should make the tube happy. And happy it was! SRPP with differential input was not only perfect as buffer, but also provided enough gain to act as high-end preamp.


Tube buffer SRPP dual 6N16B with differential input. Notice humongous Obligato output capacitors.


Lampizator - different point

Different view of tube stage. DAC power supply to the right.


Tube power supply

Tube buffer oversized power supply. Large Hammond power transformer.

The rest was just careful design, placement and wire routing. Whole DAC fit into par-metal aluminum enclosure that earlier hosted my prototype of LM3886 tubed amplifier. Input selector (TOSLINK, Coaxial and USB) and power switch in the front created simple design that was all about music.

Finished look

All cables plugged, it was a listening time. From the first notes I could tell it was like nothing I ever heard before. Dynamics, frequency response, tonality, lightness, punch, stage were incredible. Equipment disappears completely: listener is transported into live concert. Now I can’t get enough of live concert recordings. Any kind of music is reproduced accurately and precisely. And details! Every knock of fingers on the strings, every breath and chair squeak. Not exaggerated and not too bright, but very natural and crystal-clear. I’m sold on ES9018 and ESS.

Further tests confirmed medical precision and sonic performance of tube buffer. One thing that I wanted so much and now ended up being disappointed with is the USB interface. Don’t get me wrong – it is a nice sounding source of digital data. But sound comes out flat, uninspiring and nothing like coaxial or TOSLINK. And I’m not the only one with such perception. Lukasz Fikus (Lampizator) just introduced new sophisticated USB interface to his DAC, and was left with same impression. There is definitely large and unexplored field of direct I2C feed from computer or microcontroller-based player to DAC. And it shapes like extremely interesting and tempting new project.

Allow me to summarize experience with NFB-11 DAC from AUDIO-GD and rebuild results. Original untouched DAC doesn’t stand up to expectations, not even close. Flat performance and $30 CD player sound crosses it out of high end equipment list. It is not the best candidate for rebuild as USB interface disappoints and there is nothing more to take out except DAC chip and some peace of circuit board to use as a base for new DAC. But at least you have everything to build new DAC instead of starting R&D work on ES9018, digging through tons of documents and finding best schematic to bring forth immaculate performance of ESS DAC. And then pain of trying to solder something that small together without investing into PCB CAD system and paying PCB house to get decent boards. Maybe Buffalo from Twisted Pear is a better choice, but everything comes with a price and waiting. Lampizator successfully added tube buffer to Buffalo II DAC and results were astonishing. But, according to studies, shunt voltage regulators on that board might be doing more harm than good. There is theory and then – practice. Listen, compare and make-up your mind. Be prepared for investment as this level of DAC technology costs serious money. NFB-11 and other from audio-gd are not on my list.

Good luck and thank you for reading.

Story of DAC...

Last Updated on Thursday, 28 April 2011 16:29