Asynchronous I2S FIFO project, an ultimate weapon to fight the jitter

[qusp]

AR2, no benefit to syncing titan it wouldnt work anyway as there is a delay for the buffer filling that varies in length depending on sample rate. fifo output is asynchronous wrt the input clock too.

correct, if using the spdif board as well, the only way to get i2s in is via the backdoor mod shown earlier in the thread.

sync mode vs async? a matter of taste really, the jitter is so low from the output of fifo that the pll will not be doing much work anyway. i've been unable to do proper testing to see what I prefer as I havent removed the clock on mt ackodac as yet and I cant run high enough clock speeds on my pre GB 1 fifo board to get fully stable lock in sync mode, stable but every now and then I get a glitch and given i'm running DC coupled amps and digital volume glitches are not so fun. it works leaving the clock unpowered, but i'm thinking the combination of too slow clock speed (22.576 etc) and the fact the onboard unpowered clock is still in parallel, thus messing with the clock edge shape

I would wait a little before buying the clocks, Ian is about to release his Si570 variable mode clock, which will allow 90.3168/98.304MHz. fifo latest version I think will allow 384, perhaps with a little heatsink, Ian can cofirm but thats my understanding. otherwise yes those speeds you mention are correct.

afaik most status LEDs and switching are available on the J9 output header

selecting button is pin 1-2
coax LED is 3 + 4 -
optical LED is 5+ 6-
TTL LED is 7+ 8-
spdif locked LED 9+ 10-

for using external button i'm not sure other than I know it can only cycle through the selections, no direct selection, so you would have to press to cycle just like the onboard, or program multiple presses for one press with arduino etc.

[vertigovh72]

Originally Posted by ggking7

I didn't realize this was being offered fully-assembled. Can I implement this with very basic soldering skills or is it experts-only?

The kit / modules are shipped fully assembled, depending on your specific need, basically no soldering is necessary.

[qusp]

well it would be VERY unlikely that none were needed since the output clock connections are unlikely to match any dac exactly, the closest to a match is the ackodac and even it needs the w.fl rather than u.fl connectors. the units themselves are ready/finished, but the connection to the dac still needs making at the minimum

[iancanada]

Originally Posted by AR2

Ian, before I get my hands busy, a few questions if my understanding is correct. My configuration: Lightharmonic USB to I2S board, ES9012 DAC, FIFO board with S/PDIF board and dual clock.

1. If S/PDIF board is used, does using I2S signal from my USB board, is possible only through additional I2S input port on S/PDIF board?

2. My assumption is that DAC would work the best if used in synchronous mode with FIFO? My USB board also allows to be used as slave ( in master mode uses two clocks with automatic switching) Is it possible to send clock signals to both DAC and USB board? For slave mode I would need to send both clock signals? Or I would be better to leave USB board and just to sync DAC?

3. If I am to use front panel mounted input switching, do I replace your switch with any momentary switch? I need to allow for switching between two sources - S/PDIF and I2S from the board. I will not be sending any S/PDIF signal out.

4. What are the specs for LEDs or your board if I am to replace them with front panel mounted LEDs? I am thinking of taking off LEDs from the board and run wires to the front panel mounted LEDs for input markings.

5. Lastly, I am thinking of buying clocks in the following speeds for the dual clock board - 45.1584 MHz and 49.152 MHz, are these OK for ESS 9012? I am aware that FIFO board doesn't cover 384 KHz, but it goes up to 192KHz.

Here are few images of USB board and front panel board that already has switches and LEDs, that I would like to utilize. That board was preexisting and I am able to utilize it withs some modifications. That is explanation for questions No. 4.

Thank you

Vladimir

Originally Posted by qusp

AR2, no benefit to syncing titan it wouldnt work anyway as there is a delay for the buffer filling that varies in length depending on sample rate. fifo output is asynchronous wrt the input clock too.

correct, if using the spdif board as well, the only way to get i2s in is via the backdoor mod shown earlier in the thread.

sync mode vs async? a matter of taste really, the jitter is so low from the output of fifo that the pll will not be doing much work anyway. i've been unable to do proper testing to see what I prefer as I havent removed the clock on mt ackodac as yet and I cant run high enough clock speeds on my pre GB 1 fifo board to get fully stable lock in sync mode, stable but every now and then I get a glitch and given i'm running DC coupled amps and digital volume glitches are not so fun. it works leaving the clock unpowered, but i'm thinking the combination of too slow clock speed (22.576 etc) and the fact the onboard unpowered clock is still in parallel, thus messing with the clock edge shape

I would wait a little before buying the clocks, Ian is about to release his Si570 variable mode clock, which will allow 90.3168/98.304MHz. fifo latest version I think will allow 384, perhaps with a little heatsink, Ian can cofirm but thats my understanding. otherwise yes those speeds you mention are correct.

afaik most status LEDs and switching are available on the J9 output header

selecting button is pin 1-2
coax LED is 3 + 4 -
optical LED is 5+ 6-
TTL LED is 7+ 8-
spdif locked LED 9+ 10-

for using external button i'm not sure other than I know it can only cycle through the selections, no direct selection, so you would have to press to cycle just like the onboard, or program multiple presses for one press with arduino etc.

Panel interface schematics just for reference.

Ian

[iancanada]

Originally Posted by ggking7

I didn't realize this was being offered fully-assembled. Can I implement this with very basic soldering skills or is it experts-only?

I couldn't 100% gurantee even if it was soldered by myself with tools and skills not that 'basic' .

To assemble it by yourself, if you have SMT experience, shouldn't be any problem. If no, then you can learn some skills from youtube, it might be OK but need to take some risk .

Ian

[ggking7]

Originally Posted by qusp

well it would be VERY unlikely that none were needed since the output clock connections are unlikely to match any dac exactly, the closest to a match is the ackodac and even it needs the w.fl rather than u.fl connectors. the units themselves are ready/finished, but the connection to the dac still needs making at the minimum

Can my Wavelength Proton USB DAC be easily used with the FIFO board or would I be better off with another DAC?

If I start over with only a computer, what would be the ideal and easiest setup for the FIFO board?

Does the FIFO board provide galvanic isolation so that EMI can not be conducted from the computer to the DAC?

[NicMac]

Originally Posted by ggking7

Does the FIFO board provide galvanic isolation so that EMI can not be conducted from the computer to the DAC?

Does galvanic isolation block EMI?
I would think EMI flow right through just like any AC unless some kind of filter is also implemented.

[qusp]

yep, galvanic isolation would block EMI (mostly carried on power supply and ground) unless its radiated EMI, which is usually quite local. however the fifo itself isnt galvanically isolated unless you use the spdif input board, rather than i2s. if you are using 192kHz content or less then I would recommend just using the optical input, others might have irrational beliefs that somehow i2s input will still be superior even with the fifo in place, but i'm not one of them.

[marce]

What EMI are we refering to? or are you just refering to the general background noise generic to all digital systems (even the FIFO board will create some background noise, all digiatl systems do), and if so what is your concern with this noise (and NO the noise it not just SMPS's, it is created more due to numerous gates switching simultaneous.

[qusp]

i'm not concerned personally, I was just answering the query. i'm presuming hes worried about EMI transmitted on the power supply and ground from the PC

I posed the same q to him here, where you have also noticed the theme

I do however see optical combined with fifo as a pretty good solution to all that, given it cuts down the possible sources of ground contamination of all sorts