XMOS DSD 384 kHz / 32bit USB

summary about DSD

Hi!

Few months ago bought i2soverUsb and TI DAC combo, because i didn't need to play DSD.

Now it's time to go DSD, and i would like to stay minimalistic.
Is it enough to buffer the DSD output and2/3rd order lpf?
Is it crucial to add the external clock?

Thank you all (and Joro too).

EDIT:
what about clik&pop?
New firmware solved that?
 
Last edited:
some questions about the i2soverusb board

Hi joro,
Recently I upgraded my setup big time using your board in combination with a nice linear psu and a RPi running PicorePlayer. I'm currently using sp/dif output, but plans are to use i2s instead and change the program of my Arduino based controller which controls my dual AK4399 DAC to work with the I2SoverUSB board. I have some questions regarding how the board exactly works.

First pin 3, configuration pin: does it have any meaning as an input, as are A0-A2? Or is it only used for configuration?

Second: when exactly is CODEC RESET pin triggered? Is it LOW when e.g. a different fs is output or MCLK is changed?

At the moment, without changed configurations of the picoreplayer, I got for a very short period sp/dif unlock when transition from one track to the next (same fs). I wonder what causes this, is it related to how the board works or must I look for a solution in configuring the RPi?

I'm really enjoying the sound of my setup a lot, thanks to your great design, I think it was well worth the money.🙂 Thanks in advance.
 
I just bought one of these boards because the JLSounds website states "DSD Native ... Linux ...". However, I am being told that JLSounds needs to have the ALSA drivers updated before "DSD Native ... Linux ..." is really working.

Assuming this is in fact the case, when do you expect the ALSA drivers to be updated?

I currently have an older version of this board installed in my DAC (TPA Buffalo IIISE) and am waiting to install the new board until this issue is resolved.
 
Hi,

... Is it enough to buffer the DSD output and2/3rd order lpf?
I haven’t got direct observations, but from some users of the XMOS board, I have positive feedback for such way to listen DSD.

...
Is it crucial to add the external clock?
For this kind of playing DSD – no it isn’t. The analogue signal is get very close to the outputs 🙂

...
what about clik&pop?
New firmware solved that?
For both new and the old firmware one must use mute (codec_reset) pin7.

Wasn't there a dual 4495 project ongoing?? USB interface all on one board!??!
Yes, there is such project, I’ll describe it in the commercial sector really soon, all test went well 🙂

You can add more details about your dacs?
Those DAC chips (R2R DAC) are no longer produced and there are a lot of fake ones on ebay. My projects are really old based on those chips and they need serious “update”.

First pin 3, configuration pin: does it have any meaning as an input, as are A0-A2? Or is it only used for configuration?
Pin3 is only for configuration and it works as described, for changing output protocols. Pins A0-A2 are used for both configuration and outputs. When starting the XMOS board, all pins described above are inputs after that they are configured as an outputs. This state remains till next turning off.

Second: when exactly is CODEC RESET pin triggered? Is it LOW when e.g. a different fs is output or MCLK is changed?
Codec_reset (mute) is “0”, when we must mute the output; the pin is 3.3V when playing. There is a buffer of 5ms, i.e. you will need relay with switching time less than 4ms.

At the moment, without changed configurations of the picoreplayer, I got for a very short period sp/dif unlock when transition from one track to the next (same fs). I wonder what causes this, is it related to how the board works or must I look for a solution in configuring the RPi?
Could you please tell what is your current version of the firmware, you could answer on pm.

There is a solution for the Linux. I’ve never worked on linux, that’s why I asked Lyuben to check it, even he is not familiar with linux 🙂 lintweaker was very kind and reacts really fast, but for a pity our reaction wasn’t so fast.
Today Lyuben finally get DSD up to DS256 native on Linux, the praise is for lintweaker and the author of the wonderful HQplayer (without a player which supports native DSD, the things won’t happen).

Native DSD is working and tested on Win with the evaluation XMOS driver (firmware version JLsounds 3.15). This feature is not written, because I don’t have a driver, which annoys. The XMOS driver which I have is licensed by XMOS and the native DSD is blocked, but works fine on DoP. The important thing is that on the outputs of the XMOS board (independently from the transport way) there is raw DSD and could drive every DAC, which is capable to play DSD.

Regards,
Joro
 
There is a solution for the Linux. I’ve never worked on linux, that’s why I asked Lyuben to check it, even he is not familiar with linux 🙂 lintweaker was very kind and reacts really fast, but for a pity our reaction wasn’t so fast.
Today Lyuben finally get DSD up to DS256 native on Linux, the praise is for lintweaker and the author of the wonderful HQplayer (without a player which supports native DSD, the things won’t happen)

Thanks to you & Lyuben for staying with this and getting it working!!
 
The 4490 DAC seems to be available at €50 now.

Anybody tried it yet? Will it be a step up over the 5102 board?

Can I just drop the 4490DAC in instead of my 5102? Do I need additional power supplies?

I just figured the €16 LCdisplay board is required to do the settings (filters + volume).

And there are more options (DRV603 and the LME49710 filter board)

THX
 
Last edited:
Very very close. Lyuben, Joro and I will be having a discussion tomorrow about final steps.

Just a bunch of question:
What will be the SQ/performance differences between Hummingbird and single 4490 DAC board?
Does the latter have balanced output?
Could the output stages of both DACs be replaced by a se/balanced tube buffer/line stage amplifier?
If it could, what should the target buffer's input impedence be?
Thanks in advance...