Home Projects ADC AK5572 v1.2 available

ADC AK5572 v1.2 available

by nihtila

ADC AK5572 boards are available again, in slightly modified version 1.2. Despite a rather long list of changes there are no differences in performance and no major changes in functionality.

Changes are:

  • XLR-connectors now fit completely on PCB
  • Default maximum input signal is now 4 Vrms (instead of 3.6 Vrms) to align with other boards
    • 4 Vrms possible in single-ended signal as well; in v1.1 input protection clipping prevented this
  • No reset button, just header
  • Screw terminal for supplies
  • Changes in default component population and settings
    • Input AC-coupling is now optional and uses bipolar electrolytics
    • Jumper header is unpopulated by default
    • ADC is I2S Slave without any jumper (opposite in v1.1)
    • Digital high-pass filter is enabled without any jumper (opposite in v1.1)
    • 32-bit I2S option removed from header and now set by 0R resistor
  • Possibility to configure input buffers as instrumentation amplifier and use bottom side trimmer to adjust gain
  • Wee DAC headers added at edges so that Wee DAC power supply board can be used
  • Changes in input protection

Despite this list, for majority of users using the board remains exactly the same. Old feature list from v1.1 still holds:

  • Up to 768 kHz / 32-bit audio
  • Very high (measured) performance, typically:
    • 121 dBA dynamic range / SNR (124 dBA in mono mode)
    • -113 dB THD+N (-1 dBFS, 1 kHz)
    • -117 dB THD (-8 dBFS, 1 kHz)
    • -130 dB crosstalk (10 kHz)
  • Balanced differential input
  • Stereo or mono operation
  • I2S output; Master or Slave
  • HW configurable (supports I2C as well)

This post focuses on differences between v1.2 and v1.1. For basic information refer to the original ADC AK5572 v1.1 post.

Here is a factsheet for this board, and PDF version here.

Design

Here are some of the changes explained in more detail and what it means.

PCB

Board dimensions are the same (5x10cm, 4 layers) and major blocks are almost unchanged. Main difference is that now XLR connectors fit completely onboard and are populated by default.

Input circuit

AC-coupling capacitors are not populated by default but shorted with shorting links on bottom side. If capacitors are soldered, these links must be cut with a knife. When capacitors are in place, they can be shorted with jumper links on 2×2 header. Capacitors are now 10u bipolar electrolytics as opposed to 1u plastic caps in v1.1. Lower 1u capacitor caused odd low frequency noise behaviour on some boards. If you’ve got your v1.1 ADC with 1u cap don’t worry, then I didn’t see any issues with your board. As a workaround some boards were shipped with 33u through-hole bipolar electrolytic. In most cases AC-coupling caps are not needed (hence not populated by default) as digital high-pass filter takes care of small DC.

Input circuit, one channel

Input buffers are still the same by default. However, now there is a possibility to use them in instrumentation amplifier configuration. Furthermore, bottom side trimmer potentiometer can be used to adjust gain. This is all really for people who need this customisation, normally it is not used or needed. They are mostly for professional/studio use where gains may need to be trimmed. There is still a limitation that this gain control mostly allows increasing gain and supporting very high signal level (where attenuation is needed) with adjustable gain may not be suitable. This was more of a proof of concept to see if such trimmer configuration impacts audio performance – luckily not.

Optional input gain trimmers (not soldered here but for illustration only); this could be accessed via holes in rear panel

Differential amplifier stage is the same but gain is now 0.5. As ADC input range is 2 Vrms (1 Vrms per input), this gives 4 Vrms input range to this whole circuit.

There are changes in input protection. Major outcome of this is that 4 Vrms single-ended signal can now be used for maximum ADC input. In v1.1 max 2 Vrms single-ended signal was usable. Input protection is still not as robust as I want so there will still be some changes in future revisions. It doesn’t matter in normal audio use but I’d like the input to withstand almost whatever is thrown at it.

This is an evolving design and gain trimming and input protection will likely see more changes in the future.

Configuration pinheader

Large pinheader with jumper links to change settings is not populated by default now. This along with some other changes (especially use of Shorts instead of 0R resistors) saves time in assembly. To offer better default settings without the header, HPF and Master/Slave settings were inverted. The same settings are still available when the header is soldered and jumper links used. Except 32-bit I2S setting that was removed due to space constraints, and moved to be a resistor setting. After all, it doesn’t matter if I2S is 24-bit or 32-bit; there is no more information.

Settings are now as follows.

Clock configuration

CS3CS2CS1CS0Sample rate
openclosecloseopen768 kHz
openclosecloseclose384 kHz
closeclosecloseclose192 kHz
closecloseopenopen96 kHz
closeopenopenclose48 kHz
This is the same as in v1.1

When in I2S Slave mode (default), it is recommended to keep all CSx links open for automatic clock setting.

Digital filter

SDSloFilter roll-off type
openopenSharp (default)
opencloseSlow
closeopenShort Delay Sharp
closecloseShort Delay Slow
This is also the same as in v1.1

Other settings

JumperOpenClose
HPFDigital HPF enabled (default)HPF disabled
SlaI2S Slave (default)ADC is I2S Master
HPF and Sla have been inverted here and 32b setting removed.
By default the 2×8 header for jumper settings is not populated; thus default is all links open
Note the Designed with KiCad logo!

In addition, support for Wee DAC edge headers was added. There is not really an application for that yet except that the PSU baseboard could be used as a power supply for the ADC.

Board with jumper header and Wee DAC headers soldered
Board with jumper header and Wee DAC headers soldered

Dual mono operation

Mono operation is still available for creating a dual mono ADC. It improves SNR and DR by 3 dB but comes at a cost as mono board costs almost the same as a stereo board. There is not really improvement in crosstalk as it’s already almost unmeasurable in stereo board due to completely separate circuits for channel 1 and 2. However, dual mono always remains as an interesting design philosophy in my opinion.

Measurements

Audio performance is on par with version 1.1. Based on a test batch of 7 boards there are no statistically significant differences.

Input level

By default the maximum input level that gives 0 dBFS digital signal is 4.0 Vrms. This aligns with other nihtila.com boards that use 4 Vrms level in balanced connections. In v1.1 maximum input level was 3.6 Vrms.

Another difference is that now input works as “true differential” in a way that one can get the maximum level with single-ended signal as well. So the ADC happily takes in 4 Vrms single-ended input signal. In v1.1 this was limited to 2 Vrms due to input protection clamps.

Power consumption

Power consumption figures are slightly changed, typical figures for stereo board being:

  • 5VD: 12mA
  • 15VA: 95mA
  • -15VA: 57mA

This means that the board does get significantly warm.

Files

References and additional information

Version history

Schematics / PCB version history and known errors and bugs

This page version history

  • 5 July 2020 Initial version

See shop page for purchasing instructions.

You may also like

7 comments

João Barbosa July 10, 2020 - 21:23

The amount of skill, work, time and passion that you give to your products is outstanding!
I’m very impressed!

Reply
Kruger July 17, 2020 - 19:22

I received a board from Tomi. Incredible piece of work.
Still waiting an I2S receiver in order to test it. Can’t wait.

Reply
Spencer Lee July 18, 2020 - 10:18

Is it possible to select 44k 24 bit as most recording engineers run in this format.

Reply
nihtila August 14, 2020 - 21:47

This was replied to email a while ago but just for records to mention here, yes 44k is possible. It just means Master Clock to the ADC must be an integer multiple of 44.1k, for example 22.5792MHz, as opposed to 24.576MHz for 48k family of sample rates.

Reply
Dirk August 14, 2020 - 16:28

I received my AK5572 v1.2 (bought on Tindie),
and i have to say that it is of exceptional quality.
Not only the circuit board, hardware etc .. but also the sound quality is outstanding. It doesn’t have to hide behind other “high end” ADCs in any way, the opposite is the case

One or the other will be interested to know that Lyuben from jlsounds.com is planning to release an I2S to USB converter.
(in the same high-quality design and quality as his well-known “I2SoverUSB v.III” converter)

That would be the ideal complement for this ADC,

since the “miniDSP MCHStreamer” doesn’t convince me in terms of quality –> http://nihtila.com/2020/04/11/jlsounds-i2soverusb-and-minidsp-mchstreamer-usb-to-i2s-boards-with-w-dac/

Reply
nihtila August 14, 2020 - 21:39

Thank you for your comments Dirk.

And that is very interesting news regarding JLSounds! I get lots of questions how to connect this ADC to PC, and I’d love to see a good solution by JLSounds to recommend. Do you know if there is more information somewhere? Maybe I could try to contact him directly as well.

Reply
Dirk August 15, 2020 - 14:14

there’s nothing ob his homepage or elsewhere in the web,
but month ago (in May 2020) I asked him directly via email if there is (or will be) an I2S->USB solution from him as well.
He answered that he has in fact plans for “I2S input to PC”, and that this might eventually be ready “after summer holidays” … 😉

Reply

Leave a Reply to Kruger Cancel Reply