Hi Louis,
The digital inputs of the ezDAC are nothing fancy, just three BNC's for coaxial S/P-dif input that can be switched to the ezDAC PCB with three miniature relays. It all passive.
USB input is a nice feature, but if you have a module that outputs I2S, then you can bypass the CS8416 and send the I2S straight to the ASRC chip. It's a bit of a detour to let the USB module convert I2S to S/P-dif, and two inches further down the path have the ezDAC's CS8416 do the exact opposite. You would need to switch the three I2S signal lines then, but with the proper relay, that shouldn't be a problem.
Regards,
Ray
Regards,
Ray
The digital inputs of the ezDAC are nothing fancy, just three BNC's for coaxial S/P-dif input that can be switched to the ezDAC PCB with three miniature relays. It all passive.
USB input is a nice feature, but if you have a module that outputs I2S, then you can bypass the CS8416 and send the I2S straight to the ASRC chip. It's a bit of a detour to let the USB module convert I2S to S/P-dif, and two inches further down the path have the ezDAC's CS8416 do the exact opposite. You would need to switch the three I2S signal lines then, but with the proper relay, that shouldn't be a problem.
Regards,
Ray
Regards,
Ray
Thanks Ray, your response sent me to do some research.
You raise a good point re the double conversion when using a ready made USB-S/PDIF converter.
So far I interpret the need to use a USB receiver that provides the I2S outputs (such as using the PCM2706), with the I2S then connected directly to the ASRC (AD1896) would subsequently be subjected to the up-sampling and re-clocking process? ... am I off on this?
A consequent thought comes to mind; while operating with the USB to I2S conversion going directly to the ASRC chip, would it be necessary to disarm (kill) all connections to and from the CS8416 receiver chip (power and interconnects to the ASRC)? ... this could become a major task re-routing of bus connections!
Also, what about the interdependency between the CS8416 and AD1896....does the re-clocking circuit need the CS8416 to be powered up and interconnected?
Your help is much appreciated!
... and glad to learn you have expanded the family.
I have 10 grandchildren that search me out when they visit... still enjoy them at 71.
Regards,
You raise a good point re the double conversion when using a ready made USB-S/PDIF converter.
So far I interpret the need to use a USB receiver that provides the I2S outputs (such as using the PCM2706), with the I2S then connected directly to the ASRC (AD1896) would subsequently be subjected to the up-sampling and re-clocking process? ... am I off on this?
A consequent thought comes to mind; while operating with the USB to I2S conversion going directly to the ASRC chip, would it be necessary to disarm (kill) all connections to and from the CS8416 receiver chip (power and interconnects to the ASRC)? ... this could become a major task re-routing of bus connections!
Also, what about the interdependency between the CS8416 and AD1896....does the re-clocking circuit need the CS8416 to be powered up and interconnected?
Your help is much appreciated!
... and glad to learn you have expanded the family.
I have 10 grandchildren that search me out when they visit... still enjoy them at 71.
Regards,
One of the best DIY-jobs so far 
The ASRC and receiver chip work independent of each other. The incoming data from the CS8416 is just received by the ASRC, not 'clocked'. The 8416's PLL does lock on the clock that's present in the S/P-dif signal, but the ASRC doesn't need it. The 8416 only acts as a S/P-dif to I2S converter. The resampled data is clocked out of the ASRC by the 24.5MHz on-board clock. Any I2S data that you provide at it's inputs will be resampled.
You have to disconnect the three I2S line between the two chips if you want to connect another I2S source. This can be done by simply removing the three 100R resistors. Power can stay on. I tried it with a Marantz CD72 and connected the ASRC directly to the SAA7310 decoder: works perfectly. If you use a rotary switch with three center contacts, you can connect multiple sources if you like (don't forget to insert the 100R resistors). Only thing you have to be aware of is that the ezDAC ASRC is 3.3V logic. If you connect a source-chip that runs on 5V, you need a simple voltage divider in each line to reduce the voltage.
Regards,
Ray
The ASRC and receiver chip work independent of each other. The incoming data from the CS8416 is just received by the ASRC, not 'clocked'. The 8416's PLL does lock on the clock that's present in the S/P-dif signal, but the ASRC doesn't need it. The 8416 only acts as a S/P-dif to I2S converter. The resampled data is clocked out of the ASRC by the 24.5MHz on-board clock. Any I2S data that you provide at it's inputs will be resampled.
You have to disconnect the three I2S line between the two chips if you want to connect another I2S source. This can be done by simply removing the three 100R resistors. Power can stay on. I tried it with a Marantz CD72 and connected the ASRC directly to the SAA7310 decoder: works perfectly. If you use a rotary switch with three center contacts, you can connect multiple sources if you like (don't forget to insert the 100R resistors). Only thing you have to be aware of is that the ezDAC ASRC is 3.3V logic. If you connect a source-chip that runs on 5V, you need a simple voltage divider in each line to reduce the voltage.
Regards,
Ray
Last edited:
ezDAC
Good work!...your performance was by EVIDENCE "heavenly"!
... and now the ezDAC:
good info, thank you! ... now have a pretty good feeling for what I will try to implement.
The plan is to work all chips at same 3.3V level, but from separate supplies with a common power transformer. The panel input selector will power-feed break-before-make relays that complete short parallel wire runs from alternate (3) I2S sources to the resistor set at the CS8416. Will likely be including the optical S/P DIF input option as contingency.
Will be posting as I progress during the next few months... takes me that much time, but hey I am retired and no longer chasing rabbits!
Will peek in periodically to see if any progress with you or other builders.
Regards,
One of the best DIY-jobs so far
Good work!
...your performance was by EVIDENCE "heavenly"!... and now the ezDAC:
good info, thank you! ... now have a pretty good feeling for what I will try to implement.
The plan is to work all chips at same 3.3V level, but from separate supplies with a common power transformer. The panel input selector will power-feed break-before-make relays that complete short parallel wire runs from alternate (3) I2S sources to the resistor set at the CS8416. Will likely be including the optical S/P DIF input option as contingency.
Will be posting as I progress during the next few months... takes me that much time, but hey I am retired and no longer chasing rabbits!
Will peek in periodically to see if any progress with you or other builders.
Regards,
It is a bit inelegant, but do you think there's any actual degradation going through the SPDIF conversion? I know I2S is intended for very short (inter-chip) paths, and I'm reluctant to bring it out to a rotary switch (or a relay).
But assuming one were to go direct, any suggestions for a USB-to-I2S chip?
USB receiver/I2S output
I have the same concern re loss of "bit/timing fidelity" when reconverting a USB input to S/P DIF (via a I2S step) and back to I2S...
Maybe the multi-sampling/re-clocking function at the AD1896 chip would smooth out such added "jitters" and bit losses?
The PCM2706 receiver chip would provide the I2S outputs.
The PCM 2704 would strictly provide the S/P DIF signal as used by the HagUSB adapter ( Hagerman Technology LLC: HagUsb USB to SPDIF Audiophile Converter )
Regards,
Louis
I have the same concern re loss of "bit/timing fidelity" when reconverting a USB input to S/P DIF (via a I2S step) and back to I2S...
Maybe the multi-sampling/re-clocking function at the AD1896 chip would smooth out such added "jitters" and bit losses?
The PCM2706 receiver chip would provide the I2S outputs.
The PCM 2704 would strictly provide the S/P DIF signal as used by the HagUSB adapter ( Hagerman Technology LLC: HagUsb USB to SPDIF Audiophile Converter )
Regards,
Louis
USB to I2S conversion
I came across this one page diyAudio thread:
http://www.diyaudio.com/forums/digital-source/125817-please-explain-direct-i2s-usb-dac.html
Post #4 appears flawed by insinuating that the clocking and I2S cannot be disjoined when receiving USB with PCM 27** chips. The I2S information is output by the PCM2706. Would this not be the case... the I2S can be used as is by the re-clocking taking place at the AD1896 chip?
Regards
I came across this one page diyAudio thread:
http://www.diyaudio.com/forums/digital-source/125817-please-explain-direct-i2s-usb-dac.html
Post #4 appears flawed by insinuating that the clocking and I2S cannot be disjoined when receiving USB with PCM 27** chips. The I2S information is output by the PCM2706. Would this not be the case... the I2S can be used as is by the re-clocking taking place at the AD1896 chip?
Regards
Great result of 5 min of DIY construction
Congrats!! You have every reason to be proud.
Good work!
Thanks guys! I'm a Very Proud New Dad! But also a bit tired... late night diapers, you know
Ray
Post nr. has a point:
"1) Any conversion can increase the chance of distortion, so if its unnecessary, don't use it."
As far as I know, the problem with USB is that the clock or I2S signal (essentially the same) is generated by the PC. If the PC is jittery (most are...) the I2S is jittery. This is then sent straight to a DAC chip. The PC is the master, and the USB DAC is the slave.
With an ASRC in between, the input data is resampled and clocked out to the DAC chip with a NEW ASYNCHRONOUS clock. A zillion new samples are inserted between the samples of the input data and interpolated. The ASRC's input port that receives the I2S signal is set to slave, but the output port that's connected to the DAC chip is set as master, and is timed through a fresh on-board XO or some external clock. The output samples are clocked-out to the DAC chip by the new local clock.
So whether you use a CS8416 or a PCM27xx as I2S source shouldn't matter.
There's a great thread about ASRC here. The AD1890 datasheet is also very useful.
Regards,
Ray
"1) Any conversion can increase the chance of distortion, so if its unnecessary, don't use it."
As far as I know, the problem with USB is that the clock or I2S signal (essentially the same) is generated by the PC. If the PC is jittery (most are...) the I2S is jittery. This is then sent straight to a DAC chip. The PC is the master, and the USB DAC is the slave.
With an ASRC in between, the input data is resampled and clocked out to the DAC chip with a NEW ASYNCHRONOUS clock. A zillion new samples are inserted between the samples of the input data and interpolated. The ASRC's input port that receives the I2S signal is set to slave, but the output port that's connected to the DAC chip is set as master, and is timed through a fresh on-board XO or some external clock. The output samples are clocked-out to the DAC chip by the new local clock.
So whether you use a CS8416 or a PCM27xx as I2S source shouldn't matter.
There's a great thread about ASRC here. The AD1890 datasheet is also very useful.
Regards,
Ray
Last edited:
The question for me is whether to convert USB to S/PDIF, or to I2S. I'm designing a USB converter board for the ezDAC and have assumed (well, up until now) that I would use S/PDIF. I don't agree with the "any conversion can increase distortion" - these are digital signals, remember. I think the ASRC can probably handle any introduced jitter. Still, I may go with I2S. My goal is to have both S/PDIF and USB inputs to the DAC.
That's where the snag is. S/P-dif itself is not a digital signal, only the data embedded in the signal is. Transport is analog, over a coaxial (or optical) cable, with BMC modulation. If the cable is not properly terminated, reflections and interference occur, which create jitter.
Less conversions means less jitter.
Sidestep: the data coming off an audio CD is also not digital
. The HF signal inside a CD player is pure analog. That's probably why the sound is affected by certain modifications, like damping the CD or transport, and improving the power supply to the servo's.Furthermore, the ASRC does not eliminate jitter, it attenuates jitter. So less in, means less out.
So I'd say go USB > I2S.
Ray
Last edited:
Hey, guys. First, let me say Congrats! to Ray. That's awesome. Hope you get a little sleep in...
Just thought I'd chime in with my two cents (or maybe less, you decide). If one is thinking about designing a USB->SPDIF converter, is it really going to be any better than a commercially available converter? There seem to be several available, and it would save a lot of time (and probably money).
There are a few available; the ones I remember are the Hagtech and Twisted Pear. The Hagtech is fairly expensive, certainly more than mine will cost me. Also, mine will have a relay on the board to switch between USB and external S/PDIF. The Twisted Pear modules are great, they seem like good guys running that site. Their converter board (PCB only) is $14US. But once you start buying a few of their modules, the price goes up quickly. I've rigged up 'other people's boards' with the necessary other parts and pieces and it's never quite as nice as a custom solution. Using the ezdac PCB with a custom companion PCB will work nicely in my preamp.
If you count my time as $$, sure, they would be cheaper. But then I wouldn't be doing any DIY at all.
I am leaning towards the I2S output, depending on how nicely it works out to interrupt the signal on the ezdac PCB, and whether my chassis layout will allow me to place the boards closely. S/PDIF is intended for longer distances than I2S and I would bet it has better noise immunity on its input. You certainly wouldn't want to run I2S between enclosures, not a concern in my case. My primary input will be S/PDIF from a CD player, but I want the flexibility to bring in a signal from my music server PC as well.
If you count my time as $$, sure, they would be cheaper. But then I wouldn't be doing any DIY at all.
I am leaning towards the I2S output, depending on how nicely it works out to interrupt the signal on the ezdac PCB, and whether my chassis layout will allow me to place the boards closely. S/PDIF is intended for longer distances than I2S and I would bet it has better noise immunity on its input. You certainly wouldn't want to run I2S between enclosures, not a concern in my case. My primary input will be S/PDIF from a CD player, but I want the flexibility to bring in a signal from my music server PC as well.
Last edited:
I looked a little further and noticed the voltage swing for S/PDIF (at least according to wikipedia) is around 0.5 to 1V, probably similar to video signals. I2S has a logic voltage swing of 3V or so. So, if you're running it inside a chassis and not on a long cable, I have to say that I2S looks better.
If I ever calculate the labor cost of my DIY projects, a big chunk of will be rework. Looks like I'll be pulling the PCM2704 out of my design and putting in a PCM2706.
I can post more information on the board as it comes along if there seems to be interest. Or, if ezkcdude prefers I not pollute this thread, I can start another.
If I ever calculate the labor cost of my DIY projects, a big chunk of will be rework. Looks like I'll be pulling the PCM2704 out of my design and putting in a PCM2706.
I can post more information on the board as it comes along if there seems to be interest. Or, if ezkcdude prefers I not pollute this thread, I can start another.
.... went with Ray's suggested reviews of the AD1890 Data Sheet, and the
discussion at this thread re. the ezDAC's approach to converting bits to analog:
http://www.diyaudio.com/forums/digital-source/28814-asynchronous-sample-rate-conversion.html
Now I really know how deep I can not go, and knee-deep will do me fine for effort to assimilate/assemble a DAC for computer audio output.
So, will have to continue leaning on the experts hanging onto this thread!
Anxious for get my hands on the boards, and to set a BOM.
Louis
discussion at this thread re. the ezDAC's approach to converting bits to analog:
http://www.diyaudio.com/forums/digital-source/28814-asynchronous-sample-rate-conversion.html
Now I really know how deep I can not go, and knee-deep will do me fine for effort to assimilate/assemble a DAC for computer audio output.
So, will have to continue leaning on the experts hanging onto this thread!
Anxious for get my hands on the boards, and to set a BOM.
Louis
Thanks Evan! She's so beautiful.... and so noisy!
Not much sleep last night (we just woke up), I guess mum ate something she didn't like. It's a learning curve...Actually, that's not a problem, and a perfect way to eliminate two I2S > S/P-dif (and back) conversion steps. There's PS Audio for example, with their Perfect Wave transport and DAC. They use HDMI connectors for the I2S interface. Very clever, as this is suitable for high resolution signals and widely available.
It would be my choice if I had plenty $$$
Ray
Last edited:
I've got a couple of quick questions;
I got my transformers in the other day - 2x12 & 2x15.
I'm going to use 1 of the inputs & outputs from the 2x15 to power the opamps. What should I do with the other set of input/outputs from the transformer? I guess I should cap them and tie them off.
What about a fuse for the ezdac /ezdual? I've got 2 of Ray's 9v power regs and the ezdual. I've looked at Ray's BOM and he's using I think 100ma. (not sure if that's a slow / fast blow fuse). Might I need a fuse that can draw more power without blowing?
Thanks,
JG
I got my transformers in the other day - 2x12 & 2x15.
I'm going to use 1 of the inputs & outputs from the 2x15 to power the opamps. What should I do with the other set of input/outputs from the transformer? I guess I should cap them and tie them off.
What about a fuse for the ezdac /ezdual? I've got 2 of Ray's 9v power regs and the ezdual. I've looked at Ray's BOM and he's using I think 100ma. (not sure if that's a slow / fast blow fuse). Might I need a fuse that can draw more power without blowing?
Thanks,
JG
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.