Beis AD24QS 192khz/24 bit ADC kit

[playboss]

the only thing that makes sense with an ADC in mind is to make it have wordclock out. Wordclock in makes no sense cause you need a decent clock for conversion , not one recovered using pll trickery.. I have no idea how you guys want to have multiple channels, Im using a lynx aes card, so the new ESS chip is a godsend hehe. I can spend more money obtaining clean clock , instead of buiding X channels of spdif/aes transmitter.

 

[Moby]

the only thing that makes sense with an ADC in mind is to make it have wordclock out. Wordclock in makes no sense cause you need a decent clock for conversion , not one recovered using pll trickery.. I have no idea how you guys want to have multiple channels, Im using a lynx aes card, so the new ESS chip is a godsend hehe. I can spend more money obtaining clean clock , instead of buiding X channels of spdif/aes transmitter.

I don't see any problem running master clock from Lynx card. I also own Lynx and Rme so what's a problem to sync AD24QS with them?Then, if you are not happy with that clock you can invest money in something better. But this converter sounds good enough via AES cable, to be honest better than Lynx (I can say that because I spent hours listening the damn thing) so I can just imagine how it gonna sound with good external sync 

 

[jdbakker]

the purpose of the wordclock sync is to give a clean clock signal to the destination , so the destination(lets call it PC) doesnt have to use the "recovered clock" that is recovered out of the SPDIF / AES cable.

Not quite. A PC, DAW or any other digital-only endpoint cares very little about jitter, as long as it's not so bad that the receiver loses sync. Also, there's nothing about wordclock-over-BNC that is inherently lower jitter than AES-3 WC (especially unmodulated AES-3); it's just that BNC WC is easier to distribute over a larger installation, either through distribution amplifiers (preferred) or a daisychained setup. Treating BNC WC as a kHz-bandwidth signal will actually introduce more jitter, as the signal's slew rate vs system noise is the dominant factor in interconnect jitter for an unmodulated sync signal.

JDB.
[one of the issues with BNC WC is the total lack of standardization]

 

[playboss]

yes i want to clock the lynx (slave) via the ADC (master) and the DA as ASRC.

 

[playboss]

the purpose of the wordclock sync is to give a clean clock signal to the destination , so the destination(lets call it PC) doesnt have to use the "recovered clock" that is recovered out of the SPDIF / AES cable.

Not quite. A PC, DAW or any other digital-only endpoint cares very little about jitter, as long as it's not so bad that the receiver loses sync.
JDB.
[one of the issues with BNC WC is the total lack of standardization]

i dont see it as an issue . I 'll see which solution results in a stable DAW, thats all what matters , and wordclock sync is just an option. As we know jitter matters at the conversion so a nice local clock is the first priority. 

 

[Moby]

the purpose of the wordclock sync is to give a clean clock signal to the destination , so the destination(lets call it PC) doesnt have to use the "recovered clock" that is recovered out of the SPDIF / AES cable.

Not quite. A PC, DAW or any other digital-only endpoint cares very little about jitter, as long as it's not so bad that the receiver loses sync. Also, there's nothing about wordclock-over-BNC that is inherently lower jitter than AES-3 WC (especially unmodulated AES-3); it's just that BNC WC is easier to distribute over a larger installation, either through distribution amplifiers (preferred) or a daisychained setup. Treating BNC WC as a kHz-bandwidth signal will actually introduce more jitter, as the signal's slew rate vs system noise is the dominant factor in interconnect jitter for an unmodulated sync signal.

JDB.
[one of the issues with BNC WC is the total lack of standardization]

Wow, now I'm totally confused but I still have my ears   This AD24QS AES distribution sounded good to me and I would be Happy with just option to run more than 2 channels.

i dont see it as an issue . I 'll see which solution results in a stable DAW, thats all what matters , and wordclock sync is just an option. As we know jitter matters at the conversion so a nice local clock is the first priority. 

That;s what I'm talking about. Yes, that WC is good to have but ... lets motivate the designer a little :-\

 

[digichild]

I started this after contacting the Uwe Beis designer of this (and maybe only one DIY ADC kit). Unfortunately he can't offer any kit because some of the crucial parts are not available in small quantities. I decided to start this (with Uwe's approval) .
So, let's see how many people want to fly in something like this.  ;D
http://beis.de/Elektronik/ADDA24QS/AD24QS.html

the sound examples are impressive, i would be interested to diy one

 

[Moby]

the sound examples are impressive, i would be interested to diy one

Great, one more member from Germany  I didn't counted how many kits we have for now now but I will be in for 16 channels as soon Uwe find some sync option.  ;D Mathematicians can you count 

 

[digichild]

the sound examples are impressive, i would be interested to diy one

Great, one more member from Germany  I didn't counted how many kits we have for now now but I will be in for 16 channels as soon Uwe find some sync option.  ;D Mathematicians can you count 

16 channnels?
not right now for me...sorry
but if there is a project for stereo ADC...

 

[Moby]

16 channnels?
not right now for me...sorry
but if there is a project for stereo ADC...

Of course that stereo is possible. I just expressed my needs  ;D

 

[mhelin]

Why not to build a combiner module which uses I2S for interconnects, resamples everything to 96 kHz using two SRC4184's and sends the data out via AL1401AG (2x)?  The clock used as master for ASRC could obviously output as WC and/or superclock for DAC or PC.
E: Not so simple as some circuit for S/Mux would be also needed..

 

[Moby]

Why not to build a combiner module which uses I2S for interconnects, resamples everything to 96 kHz using two SRC4184's and sends the data out via AL1401AG (2x)?  The clock used as master for ASRC could obviously output as WC and/or superclock for DAC or PC.

That's the question for Uwe... I'm affraid 

 

[mhelin]

Just use this box:
http://www.rme-audio.de/en_products_adi_192_dd.php

 

[playboss]

lynx aes-16 card (the older pci has even linux drivers) plus ESS adc chip is probably the cheapst and most straightforward (with some cheap pulse transformers , check the Lynx page hi resolution photos for ideas).
I use the SRC4194 with nice multibit DAC in hardware mode ( lvds in between) once a better ASRC comes out, I'll swap the LVDS transmitter PCB (with the ASRC on it)  and thats it .

 

[mhelin]

I think HDMI is also coming sooner or later to pro audio markets. You could already use a chip like Analog ADV7520NK to send 8 channels 24bit/192 kHz audio over HDMI:
http://www.analog.com/en/audiovideo-products/analoghdmidvi-interfaces/adv7520nk/products/product.html

Also Silicon Images already has both transmitter (SiI9134)  and receiver (SiI9135) chips available (http://www.siliconimage.com/products/family.aspx?id=1). For building a PCI(e) interface just connect the I2S outputs of the receiver chip to corresponding inputs on PCI controller (like VIA Envy24, C-Media chips etc. supporting multichannel PCM).

Some people are considering using HDMI cables and connector for delivering audio in custom I2S format (maybe not so good idea because it's not standard). The HDMI standard uses TMDS for decoding the audio data.

 

[playboss]

not sure if you tried to obtain it, but I know someone who has a few of those silicon receivers, had to sign nasty NDA and all.
If hdmi was feasible, they'd already use it for digital crossover and stuff. But there are hefty license fees (several 10.000 dollars) so its unlikely to be opened. I see someone desperately wants to be mobile with his lappy in front of him  ;D then look for USB2 "Audio Device Class 2.0" , they wont include it in windows, but possible with mac or linux. Some ppl already build some sort of streaming interface they say its easy , though custom windows drivers are PITA.

 

[Rochey]

I think HDMI is also coming sooner or later to pro audio markets. You could already use a chip like Analog ADV7520NK to send 8 channels 24bit/192 kHz audio over HDMI:
http://www.analog.com/en/audiovideo-products/analoghdmidvi-interfaces/adv7520nk/products/product.html

Also Silicon Images already has both transmitter (SiI9134)  and receiver (SiI9135) chips available (http://www.siliconimage.com/products/family.aspx?id=1). For building a PCI(e) interface just connect the I2S outputs of the receiver chip to corresponding inputs on PCI controller (like VIA Envy24, C-Media chips etc. supporting multichannel PCM).

Some people are considering using HDMI cables and connector for delivering audio in custom I2S format (maybe not so good idea because it's not standard). The HDMI standard uses TMDS for decoding the audio data.

Mikko,

Similar questions have been discussed with a lot of my customers (Pro Audio Manufacturers). Often, feedback has been that implementation is still quite expensive, and that the jitter performance of audio over HDMI is awful.
Many manufacturers I've met are talking about implementing MADI for their multichannel products. MADI is an open standard that can be implemented in just about any FPGA. As most systems in Pro Audio (at a decent level) use FGPA's for one thing or another - don't be surprised to see more MADI equipment.

Cheers

/R

 

[playboss]

and there is that RME pdf where they brag about their digital pll cleaning the 30ns jitter madi into several ns .  ;D

 

[Moby]

Good news  Uwe is working on the WClock option  ;D . It's still "under construction" but ...
If that really works we will have to organize some decent group buy to make this thing possible. Some 300 stereo kits will be the minimum.  8)

 

[Davo]

Hi Sergio,
Cool progress!  If Uwe gets the wordclock option going, I'm the first in line for that mod!!! ;D

Also, just wanted to point out a couple things about my AD24QS (the one that provided the samples)
..... I upgraded chips, resistors and Electrolytics throughout.

I put in the Cirrus CS5381-KSZ rather than the 61.... and LM4562NA's for opamps rather than the OPA2134... obviously something to play with...  both these were part of Uwe's "flagship" chip upgrade.
I used PRP resistors... although I'm sure the ones Uwe provides are quite good, I always use the PRP's.... just habit.
And Elna Silmic II electrolytics.... which don't fit on the board at all, they're sticking out at all kinds of angles.... Uwe
provides yageo's, but as I've had excellent experience with Elna's, that's what I did, and I'm sure they
provide some of the boards sonic character

Cheers,
dave