What's wrong with the raspberry pi I2S output? There are of course a lot of other similar small Linux cards these days.
Also not sure why I2S has suddenly become such a "must have". It is OK as a connection between chips on the same circuit board, but it was never intended as a standard for interconnecting external devices. As long as your DAC has a decent front-end, the actual physical interconnection doesn't matter.
A raspberry pi is a godo and cheap way to try things, and if you go for something like squeezeplug you don't need to know much about Linux.Humm, based on the answers above ,I believe i putt a hand in RaspB. world... not too much expensive... while re-put the hand in Linux is not my cup of tea ! We do a lot of effort for our Hifis
Especially considering that you can't buy any music in that format...
Julf,
why I2S, etc ! And why a better layout !
I believe to try to interconnect diy pcbs comming from the multibit dac chips world
while not wasting the plus they give without adding some problems of matching impedance or RF pollution or isolation between the close I2S signals. When I see a board with a I2S output with 3 signals vias and only with one gnd via near it, I say myself than the designer has not do his max !
So as some devices exist, we continue, same than the sipdf/toslink : device is already existing : a great example is the Subbu : why limiting it with spidf... answer: because this is usefull (SqueezeBox,etc !)
I can illustrate it : I play in thi smoment with some DACS lego construction.
AYAII 2014 from Audial (uf-l embeded for I2S for experimenting) + PCMtoI2S from Ian (so 4 uf-l links + 1 uf-l link from the MasterClock + inputs for the pcb streamer
Idem with a board from Painkiller member with AD1862 (monos chip...) + the same I2StoPCM boards.
So you know already the answer : the more open, the more users ! But I must say than here with both the FIFO and the isolator.... a lot of effort were made to limit the issues of the input stage ! No ?
why I2S, etc ! And why a better layout !
I believe to try to interconnect diy pcbs comming from the multibit dac chips world
while not wasting the plus they give without adding some problems of matching impedance or RF pollution or isolation between the close I2S signals. When I see a board with a I2S output with 3 signals vias and only with one gnd via near it, I say myself than the designer has not do his max !
So as some devices exist, we continue, same than the sipdf/toslink : device is already existing : a great example is the Subbu : why limiting it with spidf... answer: because this is usefull (SqueezeBox,etc !)
I can illustrate it : I play in thi smoment with some DACS lego construction.
AYAII 2014 from Audial (uf-l embeded for I2S for experimenting) + PCMtoI2S from Ian (so 4 uf-l links + 1 uf-l link from the MasterClock + inputs for the pcb streamer
Idem with a board from Painkiller member with AD1862 (monos chip...) + the same I2StoPCM boards.
So you know already the answer : the more open, the more users ! But I must say than here with both the FIFO and the isolator.... a lot of effort were made to limit the issues of the input stage ! No ?
Last edited:
The aproach for generating multiple of 44.1khz with the Pi is not undispuded for audio purposes: https://hifiduino.wordpress.com/2014/03/01/
But maybe with the FIFO buffer of the Soekris DAC this is not of importance.
The topic seems to be straying a long way from its intended purpose of documenting Soekris DAC implementations?!
I'm planning to use the Soekris DAC in conjunction with a network renderer.
For what it's worth when I progress my Soekris DAC build I'll be fronting it with with an EDEL-NMR as my first choice, positioning it right next to the Soekris DAC and connecting via very short I2S hook-ups. The Edel can handle PCM and DSD. An advantage of the EDEL board is that it has a 32bit software volume control and I believe removing that function from the Soekris DAC will help if you plan to run filters (I plan to run two Soekris DACs with crossover filters). The EDEL NMR is a UPnP renderer.
Network Audio Renderer | ABC PCB
If I couldn't use an EDEL, or I wanted to run something like squeezelite I would use a Beaglebone Black, running miero's Botic linux distribution, coupled with something like an Acko SO3; this combination will provide external clock signals to the Beaglebone(for 44.1KHz or 48KHz data to retain native sample rates - no resampling required) and will play upto 384KHz sample rates and automatically switch between PCM and DSD. I believe some people have successfully used squeezelite on this platform. Here's info on miero's botic distribution;
http://www.diyaudio.com/forums/twisted-pear/258254-support-botic-linux-driver.html
Just my penn'orth.
Ray
I'm planning to use the Soekris DAC in conjunction with a network renderer.
For what it's worth when I progress my Soekris DAC build I'll be fronting it with with an EDEL-NMR as my first choice, positioning it right next to the Soekris DAC and connecting via very short I2S hook-ups. The Edel can handle PCM and DSD. An advantage of the EDEL board is that it has a 32bit software volume control and I believe removing that function from the Soekris DAC will help if you plan to run filters (I plan to run two Soekris DACs with crossover filters). The EDEL NMR is a UPnP renderer.
Network Audio Renderer | ABC PCB
If I couldn't use an EDEL, or I wanted to run something like squeezelite I would use a Beaglebone Black, running miero's Botic linux distribution, coupled with something like an Acko SO3; this combination will provide external clock signals to the Beaglebone(for 44.1KHz or 48KHz data to retain native sample rates - no resampling required) and will play upto 384KHz sample rates and automatically switch between PCM and DSD. I believe some people have successfully used squeezelite on this platform. Here's info on miero's botic distribution;
http://www.diyaudio.com/forums/twisted-pear/258254-support-botic-linux-driver.html
Just my penn'orth.
Ray
Technology always gets adapted and modified with time. Just because Philips didn't specify long runs in the 1986 documentation, doesn't mean that it isn't usable for such purposes.
This is from an NXP app note published in 2008 titled "Sending I2C-bus signals via long communications cables" http://www.nxp.com/documents/application_note/AN10658.pdf
Last edited:
But wouldn't the connections needed for an external clock cause much more jitter than a native clock?
cooooool, some new device
! thanks.Will I risk a : what about casing this DAC ? Did you plan to putt the trafo in another box, etc ?
my aim is to get the data to the DAC, which can then work its fifo magic, without adulterating the data through resampling or fractional division.
There's more reading here;
http://www.diyaudio.com/forums/group-buys/227502-amanero-isolator-reclocker-gb-170.html#post4193073
Ray
I2S - I2C confusion
Let's not mix up the two terms I2S and I2C.
Wikipedia reads:
"The I2S bus separates clock and serial data signals, resulting in a lower jitter than is typical of communications systems that recover the clock from the data stream. Despite the name, it is unrelated to the bidirectional I²C bus."
But if you for some reason still prefer S/PDIF then go ahead
I²S - Wikipedia, the free encyclopedia
Let's not mix up the two terms I2S and I2C.
Wikipedia reads:
"The I2S bus separates clock and serial data signals, resulting in a lower jitter than is typical of communications systems that recover the clock from the data stream. Despite the name, it is unrelated to the bidirectional I²C bus."
But if you for some reason still prefer S/PDIF then go ahead
I²S - Wikipedia, the free encyclopedia
I am planning on building an 8 channel ncore400 amp with 4 of these dacs and an exaU21 inside the same case. I am wondering how long I can run the i2s output from these dacs to the ncore inputs? I know they are to be kept as short as possible, but realistically, how long can they be without issues?
Convention says that i2s runs need to be kept to max of 10cms though I know some people run them longer without issues.
You can sent i2s data over much longer runs if you use a LVDS implementation, something like Twisted Pear Audio's Teleporter modules.
Ray
your hook up would be:
exaU21 --i2S--> Soekris R2R --Balanced Analog--> Hypex nCore.
Keep the i2s wiring as short as possible between the exau21 and Soekris DAC's. Hypex modules are designed for a balanced analog input signal, so use the balance outputs on the DAC. Cable length for the balanced connection is not critical.
cheers
Paul
Looks simple to make one yourself... and saves a lot of money
That's a bifilar wound toroid. Assuming you can get a core with enough inductance and low enough loss (a bit tricky) then you'll have a large interwinding capacitance. This couples common-mode noise fairly effectively. With a core as large as they've chosen it'll probably still have more leakage inductance than a SC, despite being bifilar.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.