Beis AD24QS 192khz/24 bit ADC kit

[Moby]

I will do the Lavry blue compare test against Beis. When I listened Blue I wasn't that impressed like AD24QS impresed me. But I didn't listened both together  Now I have some more tests so that will be possible. Also I'm still in the comunication with mr Beis and it seems that he is working on the digi sync option for AD24QS.
Here's the latest letter from him...

every now and than I continue to think about the wordclock issue. Currently I have an idea that I want to persue the next days. But that's complicated, code for a complex FPGA has to be written, tested and the hardware has to be built at least for a test.

With the help of a DAC an external synchronization might be possible, too: The DAC's Rx has MClk, LRClk an BClk outputs these signals should easily be usable as master signals for an ADC with the ADC's Tx and ADC-IC in the slave mode (currently the ADC-IC is in master mode). All other settings like M1, M2 must be set correct manually.

It wouldn't be a wordclock, but a synchronization by an other digital audio signal. Maybe that helps.

Should my idea of external synchronization work satisfactory I might continue to follow my old plan of a multichannel ADC (i.e., probably 8 channels, expandable) with ext. sync capability and possibly a MADI output.

It is, BTW, not really difficult to build an synchronizable clock when you simply use a wide-range VCO like you can find in other devices. But all discussions about low-jitter sources become ridiculous when such an RC-based oscillator is used. This is why I aim at a VCXO (voltage controlled crystal oscillator).

Best regards and a happy New Year

Uwe

 

[Moby]

Some more audio clips. Thanks to Dave 
You will gear that Beis is around 9db louder(Dave didn't changed some values in the input stage )I left the file as it is but do some level change before you listen. Cheers  ;D

Original (PT bounce) http://www.mediafire.com/?dzomqmumy1o
Digi ADC                 http://www.mediafire.com/?ikmz5olnozg
Beis AD24QS           http://www.mediafire.com/?mnwwzmiynyz
Lavry Blue              http://www.mediafire.com/?wytnngjc4ez

 

[deuce42]

Moby - I would be interested in buidling an expandable one - ie a 4 or 8 channel version as Mr Beis suggests in his email to you. Thanks

 

[Moby]

Cool, we just have to give some time to Uwe. BTW, we still counting just a few kits  :-[

 

[lofi]

im with deuce, I only need 44.1 though and looking to be able to go to about 32 tracks simultaneous on adat

 

[jdbakker]

Moby - I would be interested in buidling an expandable one - ie a 4 or 8 channel version as Mr Beis suggests in his email to you.

For multichannel better chips are available (at least from a DIY point of view). See this thread for some examples.

I see clocking as one of the major challenges. Herr Beis is right that a VCXO is one of the ways to get better jitter performance than his current design, however that would rule out Vari-Speed operation of the converter as no VCXO can be pulled more than 0.01% (~100ppm) off its nominal frequency.

For multichannel you also likely want ADAT out. Chiips are available, but you need to get a license from Wavefront (Alesis' former chip design division) to be able to buy them.

JDB.
[is the Black Market still the best place for this discussion?]

 

[Moby]

is the Black Market still the best place for this discussion

My initial idea was to organize some group buy . Since Uwe has no any audio samples of his product I decided to add that too. BTW , discussion about Beis ADC exsists here http://www.groupdiy.com/index.php?topic=14299.0 but I don't see any problem discussing here because this is just a feeler  .

For multichannel better chips are available

Better... well, what's better that's a question. For example , AD24QS use same chip as many commercial products but sounds more accurate than most of them. Also, adding adat lightpipe can be option (we can ask somebody here to do such a project ) ,  it's possible to connect all digi via AES  and adat is not so great jitter-vise. Just my 0,2 c 

 

[jdbakker]

I don't see any problem discussing here because this is just a feeler  .

The problem with tech discussions in the Black Market is that BM threads are deleted after a few months of no activity.

[quote author=Moby]it's possible to connect all digi via AES[/quote]

Possible? Sure. Connecting 32 channels (as lofi was talking about) through AES gets very cumbersome very fast, though.

[quote author=Moby]and adat is not so great jitter-vise.[/quote]

What does that have to do with anything? Link and/or receiver jitter only matters if you clock your converter chip straight from the digital receiver, which you never do if you care about jitter performance at all.

JDB.

 

[Moby]

The problem with tech discussions in the Black Market is that BM threads are deleted after a few months of no activity.

Yes, that can be a problem...

What does that have to do with anything? Link and/or receiver jitter only matters if you clock your converter chip straight from the digital receiver, which you never do if you care about jitter performance at all.

If I undestood Uwe sync will be still transfered via AES, but one of them will be master. That's why I'm talking about ADAT as bad clock distribution.

 

[lofi]

i go into my daw via adat (rme cards) they run the clock using bnc across all the various ada's i have. i would like a pure ada to replace the modded ada8000 i currently use but still want to stay adat (lightpipe)

 

[Moby]

i would like a pure ada to replace the modded ada8000 i currently use but still want to stay adat (lightpipe)

Right now we are talking about stereo ADC with possibility to add more synced AD24QS. Madi is planned but I'm afraid that ADAT is a bit obsolete digital transfer because of sampling freq. limitation. BNC Wclock was the first thing I asked Uwe but there is no simple and jitter free solution. At the end, you can use stereo AD24QS for critical things like vocals or some stereo signals like mix-master. 

 

[jdbakker]

Madi is planned but I'm afraid that ADAT is a bit obsolete digital transfer because of sampling freq. limitation.

While I agree that MADI is nicer from a technical POV I suspect very few people here have a MADI-capable DAW. Fact is that ADAT is the de-facto digital multichannel standard. I'm no big fan, but it does handle 96ksps and 192ksps (although the latter is a bit of a hack, and for 192k you would be better off with AES-3).

[quote author=Moby]At the end, you can use stereo AD24QS for critical things like vocals or some stereo signals like mix-master.[/quote]

For that purpose the AD24QS still leaves me less than impressed, especially wrt jitter and its analog frontend design and layout. For the same money you'd get better performance from an evaluation board of either the TI PCM4222 or the Cirrus CS5381. Not that I'm trying to push people away from the AD24QS, but IMHO you're hyping it a bit beyond what is reasonable.

JDB.

 

[Moby]

For the same money you'd get better performance from an evaluation board of either the TI PCM4222 or the Cirrus CS5381. Not that I'm trying to push people away from the AD24QS, but IMHO you're hyping it a bit beyond what is reasonable.

I understand what you are talking about but did you listened the sound samples. I like to read but I also trust my ears  Simply, it sounds  . Just ignore the noise issue in the first test because it's caused by +9db at front. Also , the Lavry blue test is on the way so we will be able to hear that diference too.
Also, it will be really cool if we continiue our small "debate" under the http://www.groupdiy.com/index.php?topic=14299.0   
and let the people add their voices for group buy here 

 

[Moby]

I updated the sound samples with Lavry Blue sample  Please let me know what you  think about Lavry-Beis compare  ;D

 

[chriss]

Hi there,

I have built both the AD- and the DA-Convertors from Uwe and I like them a lot. Since I wanted a front-panel-rotary-switch and don't need many of the sample-rates it offers I built a diode-matrix to only have 44.1/48/88.2/96/192 on the front-panel. But I'd love to have a 6th position labelled "ext. Word-Clock" :-\ I mailed with Uwe several times about this subject, since Cirrus introduced some chips that make this "easy" to realize. He already played around with this, but I don't know how it turned out at least. Also I think Uwe is more into Hi-Fi than into Studio-Gear, but maybe - if enough people are willing to help - we can encourage him to update the unit. Things I'd like different would be:
- Sample-Frequ.-switch as a connector not an onboard-switchm maybe with LEDs for the sample-frequ.
- power-supply should be offboard and symmetrical, so that the pcb only carries signal-stuff
- Word-Clock-Input - I find it unacceptable to have one convertor in the setup that claims to be the master... If I want to use two of them there's no chance
- I'm not too happy with the multipin-connectors to carry XLR-Out and Power-Signals both in one connector. I'd like to have a 6-pin connector or two 3-pins for the symmetrical outs. For my unit I had to cut some leads at the plug that was quite complicated and doesn't look too professional...
- switchable output-gain would luxyrious jumperable would be desirable. The exact data could be discussed.

So I think one general drawback is that he intended to make the PCB fit in one special housing with all stuff PCB-mounted, and external PSU and so on. What do you guys think of this?  I'd be ready to work as a beta-tester.

Chris

 

[Moby]

But I'd love to have a 6th position labelled "ext. Word-Clock"  I mailed with Uwe several times about this subject, since Cirrus introduced some chips that make this "easy" to realize.

Unfortunately there is slim chance to have "Word-Clock" if I understood Uwe correctly he is trying to do some other kind a clock. Anyway, check his writing in previous post. But any kind of clock what can achieve running multiple ADC's will be the right thing. Bit closed system but nice sounding for relatively small fee 

Also I think Uwe is more into Hi-Fi than into Studio-Gear, but maybe - if enough people are willing to help - we can encourage him to update the unit.

Yes, but he was thinking about MADi what tells me that he is thinking about multitrack too.  Yes, I started this thread because I had same idea. To encourage Uwe and to bring some freshness to our digital sound  Since you are in Germany maybe you are the best man to poke Uwe directly . This feeler grows slowly but it grows, and I'm happy about. Unfortunately , to make some serious group buy and to give real motivation to Uwe we will need more than 500 kit's but that's maybe possible if he finally develop sync option. So, let's try...
Guys, did you listened the Lavry-Beis samples?

 

[playboss]

FYI a new chip is to be released in mere months with integrated spdif trasnmitter and 1/6 of the consumption of that TI chip ( sorry Rochey ) . Sounds like easy as hell to me 
not clear if you need seperate clocks for the 44-48 clock domains either (cuz their other product , dac , does have ASRC - single clock).

wordclock is easy, just use a cs8406.

http://www.esstech.com/index.php?p=products_ADC

 

[Moby]

wordclock is easy, just use a cs8406.

It's easy but does it sound correct ???

 

[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. Basically, the clock prefers to travel on its own cable , and not to share it with audio data, cause the data might end up modulating the clock itself, making the endpoint PLL's work hard. Thatts all. Wordclock is in the kilohertz range, cable transmission is not problem, especially cause it is terminated with tight impedance BNC connectors (pro) . Digidesign superclock is in the mhz range, cable transmission isnt quite possible. On the other hand both RCA (worst) and AES connectors arent quite suited for digital audio, both require seperated clock transmission IMHO .

 

[Moby]

Yes, thanks for info. I was thinking is it possible to add just a connector for external sync to existing PCB than to generate a wordclock from some other source? For people with Big Ben or similar clock gen that will be the great thing , otherwise most of the soundcards have same option so syncing will be available. I'm not expert for digital audio so i can't suggest anything but maybe some of you guys can.  :-\