DAC-END R (ES9018)

[audiodesign]

Here the last modification on shunt power supply.

I have found a little error on the pcb about the negative shunts on CCS.

The 3.3V (VANA) and the 1.2V (CORE) should have 12 ohm 2W on CCS instead of 22ohm to support higher current.

All the shunt are now stable without capacitor compensantion and work very good.

The reference of voltage has been connect on output to reduce the noise to zero and it is impossible measure it.

I am considering to add on output stage a BUF634 to drive an headphone.

 

[quanghao]

today, i have PCB for test this DAC.
If all ok, i tell you more information!
thanks

 

[audiodesign]

I have ready the first release of firmware to control the ES9018.

The my idea is to implement the following configuration in the image.

The firmware now run on a 16F877 20MHz with 16x2 LCD and I have used the P86B96 chip to convert the I2C level get a bidirectional 5V to 3.3V.

For the final design I will use a 16LF877 10MHz and 3.3V 20x2 LCD.

 

[quanghao]

I have finished Test all PCB DAC-END R!

Today, I have finished DAC-END R, My work nearing completion.The test PCB is finished. I am very glad

Here are the pictures.
Please see.

 

[quanghao]

Hi all!

QuangHao really excited and happy with the sound quality, particularly measured parameters of the DAC-END R:

1. I used the tester Tektronix AM700
Signal to blance
Output balnce

2. Measurement parameters:
Frequency range of 0 - 80KHz, at 0 dB, very flat from 0 to 35kHz.
Frequency range from 0-10kHz, 0 dB, very flat

Harmonic distortion almost did not see the.

3. Conclusion:
With this measure, the sound of the DAC-END R will be very honest, this proper design criteria of Andre and me. I am very pleased and surprised at the actual measurement results, as well as listen to.

Invite you to view images
thanks a lot!
Quanghao

 

[audiodesign]

Here the first firmware to control the ES9018 with a 16F877 microcontroller.

1) Input selector I2S, Spdif1, 2 and 3 as image
2) Volume control
3) Mute function
4) Optional XMOS module check
5) All function set by remote control

For the test I have used the my AD1955 controller
http://www.audiodesignguide.com/HiResolution/usbdac2a.html
+ P86B96 chip to convert the I2C level get a bidirectional 5V to 3.3V.

 

[audiodesign]

Here the new firmware and some photos of my test environment.

At this time I am using the XMOS module by Lucian but I will try soon also the Amanero module (Amanero Technologies).

Using XMOS module the actual firmware can display USB status and sampler frequency.

 

[quanghao]

the ground plane looking back towards the dac should not be blocked. avoid routing on the ground plane near the dac outouts!

 

[jmlow]

Hi Quanghao,

Is your AD1865 kit at $247 still available?
I sent an email to your web shop but have not received any reply.

 

[qusp]

the ground plane looking back towards the dac should not be blocked. avoid routing on the ground plane near the dac outouts!

yes thats correct, they made this recommendation way back with the 9008 and you have numerous blockages, it doesnt look like you have a solid ground plane to block...?

 

[quanghao]

yes thats correct, they made this recommendation way back with the 9008 and you have numerous blockages, it doesnt look like you have a solid ground plane to block...?

yes!
With DAC, the layout is very important
I spent about the same time to research, test, design, to have a good layout as possible!
Thanks

 

[audiodesign]

Last weekend I did a listening test comparing my second AD1955 DAC output stage hybrid with the ES9018 DAC output stage with OPA627 + OPA2134.

The detail and definition seemed more good with ES9018 but the sound was too dry and not very musical.

I think the problem on the ES9018 DAC is the high feedback of op-amp so I decided to test it with a vacuum tube output stage.

The my choice is the Raleigh Audio Line Stage clone, a Differential parafeed line stage push-pull with a pair of CCS (constant current source) to isolate the power supply from the signal.

http://www.audiodesignguide.com/HiResolution/Raleighaudio1.jpg
Chapter 4
USB DAC 2 - Hi-Resolution System

Using a pair of ECC86 for each channel is possible to get a good result.

Here the first measurements of the output stage only.

To get 3Vrms on output it is necessary use 20 ohm as I/V instead of 10ohm.

 

[audiodesign]

At this time I am using the XMOS module by Lucian but I will try soon also the Amanero module (Amanero Technologies).

Using XMOS module the actual firmware can display USB status and sampler frequency.

I am testing the Amanero USB-I2S module.

 

[audiodesign]

Here my test environment, now I am waiting my friends for a listening session.

 

[zzzmonster]

Hi, since this is dac end model, will you be trying to use 2 ess chips instead of 1, 1 for each channel - many have said this gives good result (as in the buffalo ii and iii dual mono on twisted pear and diyaudio), also john westlake (designer of audiolab 8200cd) is also using 2 for his upcoming t-dac.

I read somewhere that the 9018 performance is compromised when using it in Stereo mode.

ps: There are also people thay have tried dual mono (ie quad buffalo or in accuphase DC-901 4 9018 chips for stereo mode)

 

[qusp]

Hi, since this is dac end model, will you be trying to use 2 ess chips instead of 1, 1 for each channel - many have said this gives good result (as in the buffalo ii and iii dual mono on twisted pear and diyaudio), also john westlake (designer of audiolab 8200cd) is also using 2 for his upcoming t-dac.

I read somewhere that the 9018 performance is compromised when using it in Stereo mode.

ps: There are also people thay have tried dual mono (ie quad buffalo or in accuphase DC-901 4 9018 chips for stereo mode)

haha yeah stereo is compromised.... if you can call -120dB THD+N compromised, sorry thats ridiculous and can only be part of a sales pitch, how the dac can be excellent using 2, but compromised with 1? running 2 has only a benefit of theoretically increasing DNR by 3db, good luck actually pulling that extra 3dB out of the analogue stage and making it count in your system.

I run both modes in an ackodac and have done so with buffalo II's in the past, the only reason I run dual mono now is for my headphone output just because the extra dac is there, otherwise I use a dac for stereo balanced bass and a dac for stereo balanced highs in a digital crossover feeding 4 balanced mono poweramps (these are nearly done, but not quite complete)

most people I know that have run both are just ticking all the boxes.

 

[audiodesign]

This evening I made a listening session with a my friend to evaluate the differences between the USB-I2S module with XMOS and the new Amanero.
Both the modules are superior to previous models I have tested: M2Tech and Open Source.
The Amanero module sounds more open but also less solid and less smooth of XMOS module. The XMOS module sometimes seems more confused but it also have a larger image of the acoustic scene.
Changing the interconnect cable between DAC and amplifier, from Silver Resolution Reference Analog Interconnects by SignalCable to SilverMIca Zero-One by Voodoo Cable, the sound of Amanero is very close to the XMOS.
I do not understand why Amanero provided for the module power from USB only, I will ask if they can make the change.
The Amanero can paly also 32bit 384KHz and DSD files (SACD format) but I still need to test it.
My test enviromnet is:
- Amplifier End Amplifier End by Andrea Ciuffoli
- Monitor 3 the Monitor 3
I hope it is possible to insert both the connectors on the ES9018 board.

 

[qusp]

what do you mean? it has been pretty clear powering from DC is not a problem at all for amanero, in fact that covers a large percentage of the discussion in the thread. just cut the power and supply at C27, but it still needs ground connected just like any USB. that supplies the whole board and still using the onboard regulators, you can supply clocks and other parts directly by removing the regs, just read the schematic.

 

[zzzmonster]

haha yeah stereo is compromised.... if you can call -120dB THD+N compromised, sorry thats ridiculous and can only be part of a sales pitch, how the dac can be excellent using 2, but compromised with 1? running 2 has only a benefit of theoretically increasing DNR by 3db, good luck actually pulling that extra 3dB out of the analogue stage and making it count in your system.

I run both modes in an ackodac and have done so with buffalo II's in the past, the only reason I run dual mono now is for my headphone output just because the extra dac is there, otherwise I use a dac for stereo balanced bass and a dac for stereo balanced highs in a digital crossover feeding 4 balanced mono poweramps (these are nearly done, but not quite complete)

most people I know that have run both are just ticking all the boxes.

SNR is not the main thing right, with dual or quad mono, many have claimed it sounds more analog - with a single 9018, it sounds very clear and detailed, but dry and uninvolving. I do not believe that the more analog sound simply comes from better implementation alone - cos almost all single 9018 sounds uninvolving

 

[audiodesign]

I have found a little problem to interface the Amonero to ES9018 in the DSD mode.
Two connections since to be inverted, I am checking with Amonero tech.

USB to I2S module -> ES9018
I2S DATA o DSD1 -> PCM DATA2,3,4 e 5 (CH1-2, CH3-4, CH5-6 e CH6-8) oppure DSD DATA 2,3,4 e 5
I2S CLK o DSD CLK -> PCM CLOCK oppure DSD BIT CLOCK
I2S FSCLK o DSD2 -> PCM FRAME CLOCK oppure DSD DATA 1[/FONT]

I will update soon about this item.

--------------

It is possible to use an external power supply for the Amonero module removing the 3.3V regulator or the little choke.