audio-talk

I stumbled across this blog entry on the Volumio site, thought it might be of interest as it suggests a number of possible solutions to getting network audio to an i2s dac.

http://volumio.org/raspberry-pi-i2s-dac-sounds-so-good/

Ray

Schweet! I'm playing with an ES9023 DAC feeding it I2S from a USB I/f off eBay. Liking the little piggyback board on the Pi and removing a layer interconnection, I'll have a play one day when I'm braver!

I wonder if the RaspPi would do any better with I2S than the SB Receiver?

If the I2S output works with squeezelite:

http://code.google.com/p/squeezelite/

then it would be a nice cheap way to add headless SB clients and get I2S too. I guess it's all about timing of the word clock. Does anyone have a feel for how good the implementation is on RPi?

Cheers,
Tony

Hi,

This is the I2S solution that I have been trying to get to work on my Raspberry Pi for about a month. The DAC is a TDA1541a (I don't have the hifiBerry board) and whatever I try it does not work. I have tried all the different versions of Volumio as 1.1 definitely has a bug. I also read how to configure the modules manually using SSH, which I have done and when you run aplay all looks ok, yet when I connect the DAC I get nothing.

The DAC definitely works by the way as it is in daily use connected to my CD PRO2M CD transport.

I need to find my scope leads to see if I can detect the data, BCK and WS from the Pi. If they appear ok I might try a buffer between the raspberry pi and the DAC.

If someone gets Volumio to work with I2S I would be very interested to hear how it is configured.

Cheers
Ian

Frustrating. I wait with interest.

No way I am able to as I don't have the P5 connector, so shall have to buy a new rPi. But regarding successful general i2s connection, I am watching first for announcement of bug fixes and successes from real people like yourself, rather than just claims and sales talk from people with something to sell.

Once it is a common experience to use any i2s board through any dac on the rPi I'll be ordering my new P5 equipped rPi.

I am sure the day shall arrive.

Tony Moore wrote:I wonder if the RaspPi would do any better with I2S than the SB Receiver?

If the I2S output works with squeezelite:

http://code.google.com/p/squeezelite/

then it would be a nice cheap way to add headless SB clients and get I2S too. I guess it's all about timing of the word clock. Does anyone have a feel for how good the implementation is on RPi?

Cheers,
Tony

Do you know that is possible to derive i2s signals from a SB Touch?

The image shows a hookup to an i2s send module as used by my RAKK DAC. I'm sure there is more info out there.

Ray

I'm looking to derive an i2s feed from a RPi for my RAKK DAC but also want a conventional spdif output for other dacs. I'm pretty sure that successful hacking of linux audio drivers will be beyond me at the moment so my aim is to piggy-back on a working solution in the shape of the HiFi Berry Digi.

I got this information from HiFiBerry;

You can stack another I2S board as seen here: http://www.crazy-audio.com/2014/01/hifi ... pberry-pi/

However, this works only, if the second board is configured as I2S slave and does not need ANY configuration from Linux. This won't work with most I2S DACs! It works with our HiFiBerry DAC. In that case, only the Digi board must be configured in Linux. You can't select to send the output to only one of teh card. Both will work parallel.

I reason that if the Digi board works with the RPi (which it is supposed to do with Raspbian Kernel 3.10) then I can pick up i2s from the Digi board. For 30Euros it's worth a punt and the least I get from it is an RPi with spdif output.

Bear in mind that there is no Master Clock signal from the RPi, not a problem with my DAC.

Ray

Yes, it seems the sensible solution. I'll await your findings with interest.

Do you know that is possible to derive i2s signals from a SB Touch?

Hi Ray,

Yes, I've been using an SB Receiver in my DAC for a number of years. I started with SB3 but the I2S was justified the wrong way so it wouldn't work without a CPLD and help from Nick with the coding but I did get it working in the end. Then along came SB Receiver and that had I2S that the TDA1541 would accept without butchery. The 1541 doesn't have the same flexibility as modern DACs to accept the different formats of I2S so I was glad when I found out that SBR used the compatible format. The connections to the SBR were tricky as the traces and pad were so microscopic. I managed to get a header glued down for support. The board on the right is my reclocking registers driven by a Tent clock underneath. The DAC board (Pedja Rogic's TDA1541) is down below, together with the output stage of the MP DAC Mk3.

Here is the old thing:



No changes have been done to it now for probably 7 years, it's about time I gave it an update!

Cheers,
Tony

A major achievement Tony.

Those transformers seem large for low signal stuff.

Excellent Tony and sorry for the eggs!

Ray

Those transformers seem large for low signal stuff.

Yes, they do look a bit over the top.

Two of them went with the MP DAC's driver stage so I didn't spec them (it was a kit) and the other two are for the Pedja Rogic board and I went with the recommended VA ratings. The DAC board uses a lot of different supplies, I could've got a custom tx made but opted for separates in the end.

Excellent Tony and sorry for the eggs! Laughing

You're way ahead of me on the whole RPi I2S thing and modern DACs. I've been out of the DIY thing for a few years mainly due to arriving at a sound that I am happy with and no need to fiddle further and also to other demands on time. The RPi thing has me intrigued though, I do have some RPis already so I could do some experimentation. (The start of the slippery slope again....argh!)

Cheers,
Tony

I finally got the scope connected upto my RPI to see if I could diagnose my I2S issues.

The BCK pin emits a constant clock output, which I assue is correct.

The data out appears only when a track is played, which again seems fine.

However, there is no sign of LRCK at any time. Which I imagine is the reason why my DAC doesn't work! I have resoldered the header pin and this does not help. Looking a the PCB traces I cannot see an issue, though the track is really fine so I might have missed something.

So I think I have diagnosed the issue but dont know how to fix it. I assume its not a software issue, so there must be something wrong with the pi. I will have to sleep on it and try and think of a solution.

Cheers
Ian

Make sure you have the timebase slow enough to show you the LR clock, it will be 48 or 64 times slower than BCLK.

Hi Nick,

When I connect the scope to the LRCK it doesn't even trip the timebase so there appears to be just no signal.

Cheers
Ian