8 channel AD converter idea/discussion

[Rochey]

Based on the feedback so far from MCS and Designmaster - these changes are to be made to the Schematic

Changes to make

• Add 10uF Tant Capacitors to "high current" driver devices.

• Convert the 4way output pins to 10way ribbon connectors using a DGND connection for every other pin. That will also allow me to send another clock derivative.

• Add a "clock source select" section - allowing the user to select a 44.1KHz crystal circuit or 48KHz circuit.

I'm also summarising a lot of this thread on the blog in my signature -- although i'm no expert blogger quite yet :wink:

 

[tmbg]

re your comment on the blog about CPLDs, they're well within the reach of DIY, and I can probably help you with that part.

 

[Rochey]

For those of you who are CPLD fans, I've had a look at the Xilinx range of CPLD's - and I must agree, for more advanced boards (i.e. a next generation board of this type?) then a cpld is a definite possibility.

Programming them looks relatively simple (you need a parallel port on you're PC from the looks of things, and the right cable that comes with the $50 starter kit).

Anyway, I'm hoping to sit down tonight and do the updates discussed.

cheers.

R

 

[Guest]

Rochey et al.
Most of this is way over my head - but as is the case with most of the stuff here - now and again I actually cotton on.

Just out of interest - to make sure I am on the right page - and apologies for interrupting.. but..

If I read this right I can use this Project to take 8 balanced outputs and send them via ADAT to a ADAT in of a 8 track multitrack (thus avoiding the unbalanced inputs on the multitrack)

and... if you guys can figure it the other eay around
In a DIY passive monitor controlling box - have a SPDIF in coming from a DAT/CDRW etc... machine instead of via analogue cable.

 

[Rochey]

[quote author="uk03878"]
If I read this right I can use this Project to take 8 balanced outputs and send them via ADAT to a ADAT in of a 8 track multitrack (thus avoiding the unbalanced inputs on the multitrack).[/quote]

Yes The initial idea is to create a ADAT card that will connect to analog to digitial and digital to analog "daughtercards".

Next step (i'd like to do) is to create standalone stereo ADC's and DAC's with SPDIF outputs.

You then have the building blocks to make your own Analog interface with either SPDIF or ADAT.

A possible offshoot of the project will be an S/PDIF DAC (now that TI is releasing a 19KHz S/PDIF reciever)

 

[Rochey]

update to schematic -

changelog:

-10uF Tant Caps added for decoupling
-Driver Buffers Changed to sn74ahc541
-SIL Resistors changed to DIL for ease of layout.

http://www.tendolla.com/design/diyinterface/ADATWC%20-%2025nov05.pdf

 

[Rochey]

changelog:

Onboard crystals added, with jumpers for switching - can also be connected to switches for a frontpanel.

http://www.tendolla.com/design/diyinterface/ADATWC%20-%2025nov05v2.pdf

Feedback appreciated

cheers

R

 

[designmaster]

What are you doing regarding ESD?
The wordclock is an external BNC (I guess) input, I saw no protection on the schematic. You might add a 75 Ohm Resistor and a small Cap. Better solution is a full ESD protection diode, but not so easy to get.

Regarding PCM4202:
There might be a misstake on the datasheet as well as on the Eval. board.
The full input range is 6.0Vpp, the Vref output is therefore 3.0V but with 5V voltage supply this will not work or am I wrong? The supply should be 6V. Since you have good connections to TI can you clearify that issue?

DM

 

[Rochey]

[quote author="designmaster"]What are you doing regarding ESD?
The wordclock is an external BNC (I guess) input, I saw no protection on the schematic. You might add a 75 Ohm Resistor and a small Cap. Better solution is a full ESD protection diode, but not so easy to get.
DM[/quote]

this requires some research...
Should the 75ohm resistor and cap be in a "low pass" configuration?

. ---[===]------- WCLK
.                i
.             -----
.             -----
.                i
. ------------------Ground

Any recommendation on cap value?

Regarding PCM4202:
There might be a misstake on the datasheet as well as on the Eval. board.
The full input range is 6.0Vpp, the Vref output is therefore 3.0V but with 5V voltage supply this will not work or am I wrong? The supply should be 6V. Since you have good connections to TI can you clearify that issue?

I'll check on this and report back.

 

[designmaster]

. ---[===]------- WCLK
.                i
.             -----
.             -----
.                i
. ------------------Ground

Right. I would use something around 10pF. It is always a kind of trateoff. Higher values are better regarding ESD but have an impact on signal integrity.
You can also use the BAR66 from Infineon in font of the low pass filter.

 

[Rochey]

So, I have what may seem like a dull question...

a 48KHz square wave would be made up of 48KHz Sine wave + multiple harmonics. Wouldn't a low pass filter take away some of those harmonics?

In which case, what if we put a schmitt trigger on the input after the filtering to sharpen up the clock a little? just a thought

Thanks for your input on this DM, it's appreciated.

cheers

R

 

[designmaster]

Hi Rochey,

the frequency of the signal is not relevant, it's the slew rate and jitter. If your low-pass filter frequency is very deep than you will end up, even with schmitt-trigger input with a high variance of high and low level detection (e.g. jitter).

I have no specification regarding the wordclock interface/driver capability so it is hard to say what's best.
Finally you have to play on the running system with the best values, so add the parts in the PCB and optimize later. You can get Caps in 0805 from pF to nF without a problem. But nF are to high, thats for sure.

DM

 

[designmaster]

Hi Rochey,

one thing a forgot to mention. If nothing is connected on the wordclock input (JP1). IC1 Pin 2 is a floating input. So add a pull-down with 4K7 to prevent from oscillating.

DM

 

[Rochey]

feedback on the 6.0Vpp for the PCM4202:

The input swing is 3.0Vpp per input pin. It is common to double this number to 6.0Vpp when measured differentially across the two input pins (VIN+ to VIN-). This is the standard approach for diff input audio ADCs (Cirrus, AKM, Wolfson, and ADI do this, too).

cheers
R

 

[designmaster]

Hi Rochey,

Thanks for the feedback. But what's the supply voltage of the ADC is it 5V or 6V?

I would expect 6V but in the data sheet 5V in mentioned.
When it comes to protection of the input pins it will be a nightmare if the supply would be 5V. Than you have to play with cascaded diodes to protect against voltage overshoots.

DM

 

[Rochey]

it's a 5V part

typically, you'd do the protection in the frontend before your gain stage (e.g. using the OPA1632).

i need to finish the wordclock card before we get any further though

thanks again

R

 

[Mrosso]

Hi,

Any update on your A/D converter project? I'm interested in how it turns out. Ideally, I'd love to do a DIY project of 1 to 8 channels of pro-quality ADC. I have a great signal chain (Nice mics, SCA pres, G1176 compressor) up to the DAW end. Currently I have a MOTU 828mkII. Would love to have an affordable alternative with Apogee/Lucid like quality. Doesn't anyone make a simple single channel A/D converter sans pre's?

 

[TobWen]

We're currently working on a high-quality A/D (96 KHz will be supported).

We're also working on a mainboard, which converts S/PDIF to ADAT.
You can connect 8 of our ADCs to the mainboard and you'll get an ADAT stream.

The DAC will follow soon. It's pretty much reading in ADAT-documents.

 

[hejsan]

[quote author="TobWen"]We're currently working on a high-quality A/D (96 KHz will be supported).

We're also working on a mainboard, which converts S/PDIF to ADAT.
You can connect 8 of our ADCs to the mainboard and you'll get an ADAT stream.

The DAC will follow soon. It's pretty much reading in ADAT-documents.[/quote]
Hi, sounds awesome.
Does ADAT yet support 8 x 96kHz? My motu traveller's adat connection can be used as either 8 x 48kHz or 4 x 96kHz.

regards,
:thumb: Bjarni

 

[TomWaterman]

Any chance of AES connectivity please?

Sounds very very interesting...with a DAC to follow?

Cheers Tom