[f] AMB on Facebook

γ3 high resolution DAC development

gamma3, gamma24 (plug-in for the γ3)

γ3 high resolution DAC development

Postby amb » May 9th, 2014, 12:36 pm

Updated November 24, 2015

Current status

- γ3 is now complete and released.

Quick links to γ3 DAC major subsystems

- The γ3 (backplane) board and general γ3 development: website, subforum, this thread (development)
- The α24 fully-differential line amplifier (as analog output stage): website, subforum, development thread
- The γ24 high performance DAC core: website, subforum, development thread
- The ζ1 Audio Widget asynchronous USB-I²S module: website, subforum, development thread
- The LCDuino-1 display I/O processor: website
- The ε31 bridge board (optional): forum thread
- The σ11 regulated power supply: website
- The σ22 regulated power supply: website

Note: Since this post had been updated as the project went along, some portions may contain descriptions of future tasks which are now done.

Introduction

A new AMB desktop DAC has been in gestation for a long time now, and very little has happened since I came up with the original concept. I think it's time now to start talking about the project and get things rolling.

I would like this to be a community project. While I've already fleshed out the overall architecture and selection of chips, etc., I would nevertheless welcome any feedback, suggestions and even help (firmware or hardware work) to make this project come to fruition.

In 2013 I posted a wish-list of features:

  • high res async USB Audio Class 1 or Class 2
  • multiple coax, optical and AES/EBU (XLR) inputs
  • I2S input
  • galvanic isolation for all digital inputs and outputs
  • all inputs support up to 24/192
  • upsampling with ASRC (possibly defeatable)
  • dual differential DAC
  • many filter settings to choose from
  • modular analog output stage (opamp or discrete)
  • balanced (XLR) and unbalanced (RCA) outputs
  • LCDuino-1 control of all functions (LCD display, IR remote, etc) with special firmware
Well, I think I'm going to be able to do all of these except one (I2S input), but also add a couple of other features too. The discrete analog output stage may be a later option, I'm going to focus on an opamp-based solution initially, and it will be a good one. Also, I would like to add DSD support over USB.

As I mentioned before elsewhere, I am now in possession of an Audio Precision System 2 SYS2322 Dual Domain audio analyzer with several add-on filter options, and an Altor Audio JKGEN digital signal generator. The Audio Precision was an expensive acquisition but it will be invaluable for future projects too. I will use these, in conjunction with various PC software as tools to create a DAC that will offer world-class performance.

I will go into details about the thinking behind the design in the next posts, one subsystem at a time.

Note: Originally, this thread served as the discussion for the entire γ3 DAC, but it has been split into multiple threads. Notably, the γ24 DAC core now has its own development thread. This thread should still be used for general γ3 system and architecture related discussions, and for the γ3 backplane board which ties the whole system together.

Block diagram

For the γ3, a backplane board will be the platform for all external digital and analog inputs and outputs, as well as serve as the interconnect between the following plug-in modules:

- The ζ1 Audio Widget asynchronous USB-I²S module (for USB audio)
- The γ24 high performance DAC core (the heart of the system)
- The α24 fully-differential line amplifier (one per stereo channel, serving as analog output/LPF stage)
- The LCDuino-1 display I/O processor (as the user interface front-end). Special LCDuino-1 firmware will support the γ3 system.
- The ε31 bridge board (optional, limited availability).
- Power supplies.

Note: separate development forum threads and sites are used for each of the above modules, see the top of this post or links.

The following is a simplified block diagram showing the conceptual building blocks of the γ3 DAC system. You may need to maximize your browser window in order to see the whole diagram. Design details are outlined in later posts.

Image

Here is an assembled γ3 backplane board:
Image

The γ3 backplane board with the ζ1, γ24 and α24 modules installed:
Image

Schematic digram

Schematic diagrams for the ζ1, γ24, α24 and LCDuino-1 are found in the links at the top of this post. The following is the schematic diagram of the γ3 backplane board. It is shown in five pages. Ground symbols with the "IG" notation refer to the galvanically isolated ground for the γ24 and α24.

Page 1
Image

Page 2
Image

Page 3
Image

Page 4
Image

Page 5
Image

PDF version of this schematic diagram

PCB layout

The backplane has two copper layers, the following image shows both layers and the top silkscreen, but without ground fill for clarity. Click for a higher resolution image.
Image

Here are the two copper layers shown with full ground fill:
- Top layer
- Bottom layer

3D rendering

Here is what the γ3 backplane looks like. The oscillators, fiber-optic receivers/transmitters, pulse transformers and module pin headers are shown as not populated.
Image

Parts list

Please refer to the γ3 website for the latest parts list.

This post will be updated as needed, until general production release.

Photos of AMB's reference build

AMB's γ3 is a dual-chassis configuration (DAC, power supply) featuring Par-Metal 20-series aluminum enclosures and custom-machined front and rear panels from Front Panel Express. Two σ11 and one σ22 regulated power supplies are used. A ε31 bridge board is used to connect between the γ3 backplane and the LCDuino-1 boards (in lieu of wire harnesses). The backplane is configured to have the following inputs and outputs:

  • USB input
  • AES/EBU (XLR) digital input
  • Coaxial (BNC) digital input
  • Coaxial (RCA) digital input
  • Optical (Toslink) digital input
  • Coaxial (BNC) digital output
  • Optical (Toslink) digital output
  • Balanced (XLR) analog outputs
  • Unbalanced (RCA) analog outputs

Image

Image

Image

Image

Image

Image

Image

Image

Image

Image

Image

Image

Image

Image

3 people like this post.
User avatar
amb
Site Admin
 
Posts: 10378
Likes: 175 posts
Liked in: 482 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Re: γ3 high-resolution DAC development

Postby daniele86 » May 9th, 2014, 12:45 pm

Nice amb! Amazing!
I'm just waiting for this!!
daniele86
 
Posts: 164
Likes: 13 posts
Liked in: 11 posts
Joined: June 28th, 2012, 11:53 pm
Country: Italy (it)

Re: γ3 high-resolution DAC development

Postby jmc » May 9th, 2014, 4:11 pm

Great news!! This is going to be exciting!
I always thought you'd call it Gamma22 :D
jmc
 
Posts: 29
Likes: 10 posts
Liked in: 2 posts
Joined: November 27th, 2012, 6:39 am
Location: Lisbon
Country: Portugal (pt)

Re: γ3 high-resolution DAC development

Postby muskyhuntr » May 10th, 2014, 2:22 am

Been waiting for this news for over a year! Can't wait.
muskyhuntr
 
Posts: 110
Likes: 11 posts
Liked in: 15 posts
Joined: October 22nd, 2012, 11:59 am
Country: United States (us)

Re: γ3 high-resolution DAC development

Postby mikeg88 » May 10th, 2014, 2:32 am

This is awesome, I am so excited! :D If you need any help testing/evaluating please do not hesitate to ask, even if that means testing/giving suggestions on a specific subsystem.

I like that you are sticking to the WM874x chip, it sounds great IMO and there aren't many dual-mono implementations of it out there yet, as far as I know.

The main questions that pop into my head are:

1) Will this be a single-board design?
2) What are your initial intentions for supplying power (i.e. on-board/off-board regulation or combinations)?
Last edited by mikeg88 on May 12th, 2014, 8:50 pm, edited 1 time in total.
User avatar
mikeg88
 
Posts: 239
Likes: 9 posts
Liked in: 1 post
Joined: June 13th, 2011, 11:12 pm
Location: Mountain View, CA
Country: United States (us)

γ3 overview

Postby amb » May 10th, 2014, 4:49 am

Design goal

First, a few words about the goal of the γ3. It's not meant to be a replacement of the γ1/γ2. The γ1/γ2 are excellent DACs that were designed to fit in a particular case and is small and portable while offering features uncommon in DACs their size. By going to a larger desktop form factor (and being only AC-powered), γ3 is no longer portable, but is free from the confines of the small casing to allow more features and capabilities. I hope the result is a DIY high-resolution DAC that would compare favorably against high-end commercial offerings, yet with a build cost that would be quite palatable.

The following sections provide an overview of the block diagram posted above.

Digital Inputs

Starting from the left side of the block diagram, you'll find four "Digital Inputs". Each of these can be build-time configurable (by populating different connectors and parts on the PCB) to be coaxial S/PDIF (BNC or RCA), optical (Toslink), or AES3 (XLR). The coaxial and AES3 configurations have galvanic isolation via pulse transformers. Optical is of course inherently isolated. All these inputs support sample rates up to 192KHz.

This scheme allows the builder to choose any combination of these three types for the four inputs, based on their preference or need. These four inputs are connected to the Texas Instruments SRC4392 chip, which is a combination of DIR (digital interface receiver), DIT (digital interface transmitter), and ASRC (asynchronous sample rate converter).

Audio Widget USB-I2S

The γ3 also has a USB input, which is integrated on the "Audio Widget" USB-I2S module. This module features an Atmel AT32UC3A3 32-bit microcontroller, running the Audio Widget firmware, and supports asynchronous mode (rate feedback) transfers in either USB Audio Class 1 (UAC1) or USB Audio Class 2 (UAC2). The USB class is user-selectable. The UAC1 mode supports sample rates of 44.1K and 48K, and does not require any special software device driver on Linux, Mac or Windows. The UAC2 mode supports all "standard" sample rates from 44.1K to 192K, it requires a software device driver for Windows, but is plug-and-play on Linux and Mac. I will provide more detailed information about the Audio Widget in a separate posting.

The output of the Audio Widget is an I2S bus. When playing DSD files, the same wires on the I2S bus will carry the DSD clock and data lines, and the Audio Widget will provide indication to the MCU (LCDuino-1) that DSD mode is active.

NOTE: DSD support is not yet implemented in the Audio Widget firmware. I hope to drive the effort to get this done.

All interfaces from to/from the Audio Widget are via digital isolators for galvanic isolation. This includes the I2S bus, I2C bus and other control/status logic. The digital isolators also perform logic level shifting between LCDuino-1's 5V and the 3.3V logic elsewhere.

UPDATE: AMB has released an implementation of the Audio Widget, the ζ1.

Digital Outputs

The SRC4392 provides two digital outputs. We use one of these to provide an optical (Toslink) output, and the other one for either a coaxial S/PDIF (BNC or RCA) or AES3 (XLR) output. The latter is build-time configurable by populating different connectors. The user will be able to select whether this output is taken directly from the selected input (i.e., loop-out), or taken after processing by the ASRC.

SRC4392

We utilize essentially all of the functionality provided by the SRC4392. This chip contains a number of internal multiplexers (switches) to allow user control of signal routing to and from its inputs, outputs, and the ASRC. Of particular interest is the chip's two bidirectional "audio serial ports", each one could be set up as a I2S master or slave. We use one of these as input from the Audio Widget, and the other as output to the DACs.

Texas Intruments describes the SRC4392's ASRC implementaion as "based upon the successful SRC4192 core" (which is what's used in the γ2 DAC). It goes on to say that it's "further enhanced to provide exceptional jitter attenuation characteristics helping to improve overall application performance". As in the γ2, we will upsample everything to 96KHz before sending the data to the DACs. In the γ2, a 176.4KHz or 192KHz input stream will be downsampled to 96KHz. In the γ3, the user can choose the same behavior, or select to change to upsample to 192KHz so that no downsampling would occur.

In DSD over USB mode, the SRC4392 is bypassed, and the data/clock are routed directly to the dual WM8741s.

The SRC4392 must be software-controlled, and the LCDuino-1 will provide such control via the I2C bus. I will write more details about this chip and how it will be used in the γ3 in another posting.

WM8741 DACs

Two Wolfson WM8741 DAC chips will be used in "differential mono" mode, one per stereo channel. This improves performance (increases dynamic range by 3dB). These DACs will also operate in software controlled mode (in contrast to the γ2), provided by the LCDuino-1 via the I2C bus. Running the DACs in software controlled mode makes available many features not accessible in hardware controlled mode.

There will be a total of five digital filter responses to choose from (compared to three in the γ2), including linear-phase or minimum-phase apodising filters that are not available on the γ2. We are also taking advantage of the WM8741's internal volume control attenuator. The WM8741 provides a 10-bit control range which gives 1024 steps of 0.125dB per step. However, we will implement 256 steps of 0.5dB steps. The WM8741 natively supports PCM as well as DSD. For DSD we will use the "DSD Plus" mode which is only available via software control.

As on the γ2, an "Anti-clipping" mode will be selectable.

Differential analog output stage and analog outputs

The γ3 will feature an analog output stage module implemented as a differential amplifier with buffered inputs. Premium opamps will be used for this section. Both unbalanced (RCA) and balanced (XLR) outputs will be available. This stage will also provide a 3-pole analog low-pass filter (LPF) to remove digital artifacts from the DACs. It will be powered by a dual-rail power supply, and there will be no output coupling capacitors.

UPDATE: This analog stage has been released as the α24 fully-differential line amplifier.

An alternate module based on discrete components may be developed in the future.

Master Clock

The master clock is a low-jitter 24.576MHz oscillator. It will be located close to both the WM8741s and the SRC4392 for best performance. In addition, there are additional 24.576MHz and 22.5792MHz oscillators which are used by the Audio Widget.

Microcontroller and user interface

The LCDuino-1, its LCD display, and the handheld IR remote control provides the user interface to all functions. Instead of Volu-master, the LCDuino-1 will run different firmware specifically tailored for the γ3. It will allow the user to select any of the four digital inputs or USB as the playback source, change the volume, mute the audio, select a digital filter, select digital output mode, etc. A front panel pushbutton switch offers "Learn IR" and power on/off and mute functionalities, and an optional front panel motorized potentiometer gives you a knob to change the volume with. As in the LCDuino-1/δ1 combination, the audio signal does not go through the pot. The volume change and mute control are implemented in the WM8741.

The LCD display will show the current input sample rate (PCM mode), volume/mute setting, selected filter, anti-clipping mode, etc. Big font modes will also be available that shows only the volume.

The Menu setup will offer adjustable fine/coarse volume increments, volume adjustment range limits, user-programmable input names, adjustable backlight intensities, and several other user-selectable DAC settings.

In the γ3 application, the LCDuino-1's real-time clock option will not be supported. All parts related to that function should not be populated on the PCB.

All interfaces from to/from the LCDuino-1 are via digital isolators for galvanic isolation. This includes the I2C bus and other control/status logic. These digital isolators also perform logic level shifting between LCDuino-1's 5V and the 3.3V logic elsewhere.

PCB modularization and casing considerations

The Audio Widget USB-I2S converter (ζ1), the SRC4392+DAC core (γ24), and the differential analog output stage (α24) will each be on its own PCB module, to be plugged-in to the main γ3 backplane PCB. The γ3 is not designed to fit any particular case, so you can choose what you want as long as it fits.

The front panel will contain the on/off/mute/config pushbutton switch, the LCDuino-1 with its LCD display, the IR sensor, and the optional motorized pot. Connections from the LCDuino-1 to the main γ3 board will be similar to how it is connected to the α10 preamp backplane, via Molex KK connectors and wires.

Power supply

Not shown in the block diagram is the power supply. There will be three power supplies, each one with its own transformer (or a special transformer with multiple secondary windings), as follows:

+5V - always-on supply for the LCDuino-1 and Audio Widget (Audio Widget power is enabled only when the unit is "powered up:).
+5V - digital supply for the SRC4392, WM8741 and oscillator (regulated down to 3.3V and 1.8V with multiple onboard LDO regulators).
±10V - analog supply for the differential analog output stage and for the WM8741 (regulated down to 5V by onboard LDO regulators).

These three power supplies are not implemented on the main γ3 PCB. You may use two σ11s (or σ25s with additional heatsinking) and a σ22 to get all of these, or other power supplies of your choosing. None of these require a high-VA rating transformer. You could mount them in the main chassis if there is enough distance and clearance from the DACs and analog output stage, or you can use a separate chassis.

More to come...
User avatar
amb
Site Admin
 
Posts: 10378
Likes: 175 posts
Liked in: 482 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Re: γ3 high-resolution DAC development

Postby kt88 » May 11th, 2014, 9:36 am

Looks very interesting. The user configurable inputs are especially interesting, and a very good idea imo. I will be looking out for updates on this project, so that hopefully it'll replace the y2 with my M3 amplifier, and let the y2 compliment my Pimeta at work.

Once again, thank you for sharing. It's always interesting to read things that go through the designers head while the device is still being designed - I think this is an excellent way for the readers of the forum to learn.
User avatar
kt88
 
Posts: 107
Likes: 3 posts
Liked in: 14 posts
Joined: April 1st, 2010, 12:15 pm
Location: Israel
Country: Israel (il)

Audio Widget

Postby amb » May 12th, 2014, 5:21 am

Finding a USB audio solution

If there is one thing on the γ1/γ2 that could be improved upon, it's the USB audio interface. In 2008 when γ1 was developed, there were few viable solutions for a DIYer other than the Texas Instruments PCM27xx or PCM29xx series of USB audio chips. There were a few others available, such as the TAS1020, but the use of this chip requires significant firmware effort. Chips from C-Media or Tenor were also rejected because they were not available in single quantities through distributors like Mouser, Digi-Key and others. Given this, the PCM2707 was chosen for the γ1 along with its limitations -- it supports only sample rates of up to 48KHz, and it uses adaptive mode which has much more jitter than asynchronous mode (see explanation below). Nevertheless, the PCM2707 was easy to implement in the γ1, was a good fit in the small enclosure, and performs well. In the γ1 full/full++ configurations, the USB audio is converted to S/PDIF and then goes through the CS8416 receiver chip, which performs PLL-based clock recovery and removes a lot of the jitter. In the γ2, jitter is virtually eliminated through the ASRC.

For the γ3 however, we definitely need to look for a better solution. The new requirements are the following:

  • Supports all "standard" sample rates from 44.1KHz to 192KHz, 16 and 24 bits
  • Runs in asynchronous mode
  • Converts from USB to I2S, not to S/PDIF so that the low-jitter characteristic of async mode isn't compromised by the S/PDIF interface
  • Supports all three major operating system environments (Linux, Mac, Windows)
  • Must be readily "buy-able" in small (or single) quantities
  • Must not require a herculean development effort (e.g., writing complex firmware from scratch)
  • Must not require a licensing or royalty fee
  • Must not require signing non-disclosure agreements (NDAs)
  • Open source highly desirable (hardware, firmware, software)
As you can see, designing a DIY DAC is a lot different than designing a commercial DAC -- aside from functional/technical requirements, there are factors to consider that are not normally an issue for a commercial audio company. This of course adds to the challenge of finding the right solution.

A corollary to the 192KHz sample rate requirement is that the interface must be capable of UAC2.

When I started looking for a new USB solution a couple of years ago (with the γ3 in mind), I came across the same set of chips that we previously abandoned and had to do that again for the same reasons. But I also found "newcomers" that warranted a closer look. Here are some of the ones I checked:

  • Gordon Rankin of Wavelength Audio had developed code for the TAS1020 chip to run in asynchronous mode, albeit the maximum supported sample rate is 96KHz. No DSD support. Closed source implementation and a license fee not friendly to DIYers. Used in Wavelength Audio DACs, Ayre QB-9 USB DAC and some others.
  • CEntrance had also developed code for the TAS1020 to support up to 96KHz sample rate, but it runs in adaptive transfer mode. No DSD support. Closed source implementation and a license fee not friendly to DIYers. Used in the Benchmark DAC1 USB, PS Audio PerfectWave DAC, Bel Canto USB Link, etc.
  • XMOS USB Audio 2.0 - Supports UAC1 and UAC2 in asynchronous mode up to 384KHz 32-bits sample rate and depth, DSD support. A reference hardware design and closed source firmware is available from XMOS. A Windows device driver is available from either XMOS or Thesycon, both closed source. The XMOS driver is free but the Thesycon driver has a license fee. There is also a "Evaluation/prototyping" version of the Thesycon driver for free, but it expires after 60 minutes of use, and begins inserting "beeps" into the audio for every 5 minutes of playback. XMOS is used in a number of commercial DACs.
  • exaDevices exaU2I - Supports UAC2 in asynchronous mode up to 384KHz 32-bits sample rate and depth, no DSD support. FPGA core and proprietary, closed source Windows device driver. Was expensive and now no longer available.
  • RigiSystems USBPAL - Supports UAC2 in asynchronous mode up to 384KHz, DSD support. Closed source firmware and Windows ASIO driver. Claims support for Mac, but Linux support status unknown.
  • Amanero Combo384 - Supports UAC2 in asynchronous mode up to 384KHz 32-bits sample rate and depth, DSD support. Closed source firmware and Windows driver.
  • Audio Widget - Supports UAC1 (44.1KHz, 48KHz) and UAC2 (44.1KHz - 192KHz), 32 bits. No DSD support (yet). Open source hardware, firmware and Windows ASIO driver.

Audio Widget

Audio Widget is based on the SDR widget project, which is an open source system of firmware, hardware and programs for HAMs/Radio Amateurs. In late 2010, work was initiated to extract and reuse parts of the project for hi-fi/high-end audio purposes. Both projects implement asynchronous USB audio in the Atmel AT32UC3A3 32-bit microcontroller (MCU). Audio Widget hardware, firmware, software utilities, and the Windows ASIO driver are all open source, developed by people in the enthusiast community.

Based on the list of requirements it should be obvious why I chose the Audio Widget. The decision is also helped by the fact that Audio Widget measured very well on the bench -- it has vanishingly low jitter performance (tests done by Demian Martin, "1audio" on diyaudio.com and quite an expert in the field).

Audio Widget might not have DSD support yet, but we could add that capability. That's the beauty of community-based open source. The DoP (DSD over PCM) open standard defines how DSD data is to be encapsulated into a PCM data stream. The player software performs the "encoding", and the USB audio processor (in this case Audio Widget firmware) detects and "decodes" the DSD data and sends it down the I2S wires, and the DAC's main MCU would then switch the DAC chip into DSD mode. I have joined the Audio Widget development team and would like to work on adding DSD capability in the firmware. This is an area where I could probably use some help, because I am going to be spread thin trying to develop the γ3 hardware and the special LCDuino-1 firmware all at the same time.

Audio Widget supports sample rates up to 192KHz, not 384KHz as some of the other interfaces. Given that neither the SRC4392 nor the WM8741 support rates higher than 192KHz, it's a moot point. Even 192KHz is somewhat excessive, not only does audio file sizes become rather large, there are some other issue with sample rates higher than 96KHz that we must address, and this will be covered when I describe the DAC in more detail.

Adaptive vs. Asynchronous

What is adaptive mode and asynchronous mode? And why is the latter more desirable? I think Stereophile sums it up nicely in the first few paragraphs of its Ayre QB-9 USB DAC review (Note the date of 2009, which puts some of the statements in perspective):
http://www.stereophile.com/digitalproce ... 9_usb_dac/

There are also other sites with good coverage of this topic, here are a couple of examples:
http://www.pearlaudiovideo.com/blog/exp ... -adaptive/
http://www.thewelltemperedcomputer.com/KB/USB.html

Audio Widget status

I've been following Audio Widget development for a long time now, and have been testing and evaluating the Audio Widget as implemented in the QNKTC (now known as Henry Audio) AB-1.1. Internally, it has an Audio Widget USB-I2S module plugged in to a DAC board featuring the ES9023 chip. The person behind QNKTC is Børge Strand-Bergesen ("borges" on diyaudio.com), he is a member of the Audio Widget firmware team (and now, the primary member). Børge has done a lot of work tweaking the firmware over the past couple of years, and I consider the current firmware status as stable and mature. DSD playback is something that I'd like to bring to the table.

Audio Widget needs no special software drivers when running in UAC1 mode. It is also plug-and-play on MacOS X and Linux in UAC2 mode (minimum Linux kernel version 2.6.37 needed). On Windows 7 and 8, you need to install the Audio Widget ASIO driver, and you need an ASIO-aware player program (such as foobar2000 or JRiver with ASIO plugin). The Audio Widget ASIO driver was developed by Nikolay ("nikkov" on diyaudio.com). I am now also a part of the development team. The driver is also stable and mature, but Børge is working on some enhancements (64-bit native driver when used with a 64-bit player on x64 systems, updated USB library version, a memory-based WAV audio player, etc).

Aside from Børge's implementations, there has been a couple others by Yoyodyne Consulting (George Boudreau, who also helped design the NwAVGuy ODAC), but George's implementation is not modular. His Audio Widget is on the same PCB as the DAC.

There is a discussion thread at diyaudio.com about the Audio Widget:
http://www.diyaudio.com/forums/digital- ... t-170.html

Integration of Audio Widget into the γ3

Since the Atmel AT32UC3A3 is a 144-pin LQFP package (very fine-pitched pins), and there are many other very small SMD parts on the Audio Widget, the overall success rate of DIY soldering on such a thing is probably going to be poor. Thus, it makes sense to offer pre-built modules. Just think of it as a USB-to-I2S "chip", even though it's actually a board with parts on it.

I plan to design the γ3 main board to be pin-compatible with Børge's Audio Widget module. You cannot buy his module separately from his online shop, but I may make an arrangement with him to offer pre-built modules through AMB audio shop. Alternatively, I could implement my own Audio Widget module and offer them instead. Again, open source provides options and opportunities that are not possible with closed designs. As long as all the parts on the Audio Widget remains in production, we won't face obsolescence. You can't say that for some of the other USB-I2S solutions that I mentioned above.

The Audio Widget modules I offer will have AMB-specific firmware (based on the "standard" source code with special compile-time options), with a unique USB VID/PID, different than Børge's or George's implementations. It will identify itself as "AMB USB DAC Audio Widget" to the computer. This is done to differentiate the modules for support purposes. The firmware is user-upgradable via the USB port.

UPDATE: AMB is now shipping its own Audio Widget implementation, called the ζ1. See the first post of this thread for links.

More to come...
User avatar
amb
Site Admin
 
Posts: 10378
Likes: 175 posts
Liked in: 482 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Re: γ3 high-resolution DAC development

Postby tommarra » May 12th, 2014, 7:42 am

Ooooh year --- where can I sign up for the pre-order list of the parts!
AMB Projects built: B22 (Active ground), S22, S25, LCDunio, D1, D2, A20
AMB Projects in progress: A10+B24 integrated amplifier
http://www.amb.org/forum/nishants-beta24-alpha10-integrated-amp-build-log-t2193.html
tommarra
 
Posts: 225
Likes: 2 posts
Liked in: 0 post
Joined: June 15th, 2010, 5:43 am
Location: Georgia
Country: United States (us)

Re: γ3 high-resolution DAC development

Postby fishski13 » May 12th, 2014, 7:50 pm

i'm super happy to see this.

personally, if DSD is possible, that's fine, but i'm skeptical about it as a viable format. i understand the idea that "one must be absolutely modern", but i wouldn't care if the y3 supported DSD or not. open source is good and the Audio Widget seems to be supported without commercial interests.

i'm down with opamps in the analog output stage. i'm not sure what opamp(s) you had in mind, but if i could make a sheepish request - would it be possible to have DIP sockets?
User avatar
fishski13
 
Posts: 282
Likes: 0 post
Liked in: 0 post
Joined: June 6th, 2010, 6:17 pm
Location: MN-Twin Cities
Country: United States (us)

Next

Return to γ3 high resolution DAC and γ24 DAC core module

Who is online

Users browsing this forum: No registered users and 2 guests

Quick Links

AMB audio DIY resources
AMB audio shop