DAC 2496 (AK4393) DAC KIT With CS8416+AK4393+5532

The oscillator has three connections:

INPUT
OUTPUT
GND

So INPUT goes to digital +3.3V,
GND is self explanatory
OUTPUT goes to OMCK (pin 25 on CS8416)? Now pin 25 is connected to GND and via R4 (47K) to SDOUT. Should that be removed? And if yes, what should and what should not be removed?
 
Ooooops, gave you the wrong link - that's a passive crystal...

You need something like this;
10pcs 11 2896MHz 11 2896M Active Crystal Oscillator Timer SMD 5x7 5 7 | eBay
or
5 x 11 2896MHz Crystal OSC 8 Pin Format TTL Enable Option DIY Word Clock | eBay

You have; Vdd, GND, Output - all of these are self explanatory...
OMCK is connected to GND, but SDOUT is connected to GND through a 47k resistor - so no need to touch SDOUT or that resistor - the DAC schematic is drawn a little strange...
Just unsolder the OMCK pin...
 
Last edited:
OK, thanks. The biggest problem will be to find a soldering tip that can desolder the tiny pins of the SMD chip. :rolleyes:

So, after connecting all this, the output of the DAC chip will be muted, if no S/P DIF signal is present? And once the CD player starts to submit the digital stream, it will automatically start to decode and play music?


Haik
 
I used a 2mm wide soldering tip (not that small) and a scalpel with really thin tip to push the pin up...
You can heat up the surrounding pins, just make sure you pull up only the OMCK pin...

The output of the DAC won't be muted, it will be active, however you won't get that annoying hiss when no SPDIF signal is present, it will be silent... And that's all you need...
 
OK, I ordered a tiny (0.5 mm) soldering tip from Farnell (number 418572). It was the smallest which will fit my soldering iron. Let's see if I can lift off the pin.

Good for me, that one DAC was blown (the 3,3V regulator on the digital side went broken due to a power spike on the mains power). So now have a "half working" DAC unit, which I can break and use the parts.....
 
For those who are interested, I post a screen shot of a scope.

The signal you see here is the 11.28960 MHz signal generator, which is sold on e-bay.

Not a perfect sinus, but probably it is not a problem. Next week I am going to play with the SMD IC's.
 

Attachments

  • DSC05730.jpg
    DSC05730.jpg
    833.2 KB · Views: 586
Just been looking at SMD soldering videos on Youtube, which lead me into looking at videos for soldering ics using a flux pen.

I wish I'd looked at these before ordering from Along, as I then would'nt have asked to have the ics soldered for me.

I see there are rosin and non rosin flux pens - which would suit our purposes better of the two types?

These flux pens have to be ideal for soldering op-amps as well. Interestingly chisel tips (my favourite) are used with flux pens.

I've seen a module of two OPA627AU for $20 - does the price sound right/what does 'AU' signify and the module pins seem to have a much larger footprint than an op-amp. Is this an optical illusion (the camera can and does lie). If it is not an illusion can the module pins be bent to fit? Thanks in advance for answers.
 
I've seen a module of two OPA627AU for $20 - does the price sound right/what does 'AU' signify and the module pins seem to have a much larger footprint than an op-amp. Is this an optical illusion (the camera can and does lie). If it is not an illusion can the module pins be bent to fit?

AU simply means it's the SMD version (SO8 package).

AU pins are spaced like all others SO8 opamps (1.27mm).
 
just replaced the 7912/7812 voltage regulators with Salas power supplies (OPA827).

Really impressive improvement, sound is very authoritative now. Removed a lot of harshness (despite the 7X12 being lower noise versions).

I am considering using the final Salas board for the +9v supply. Anyone with opinions as to whether this will make a difference?

jim.
 
Why don't you use directly 3,3V and 5V salas shunt removing the final SMD regulators? You will squeeze most performance this way.

AK4396 can use a 5V rail for both analog and digital sections. Although a separate for digital is recommended.

The CS digital chip needs 3,3V though.

I found a way combating the offset voltage and using no cap. One Ni-Mh + Ni-Cd = almost 2,5V and few mV which won't hurt my amp bias very much.

The problem is, this chip doesn't drive capacitance very well (25pF by datasheet), so should a 1k resistor in series help? I have about 100-150pF of capacitance to drive (Miller and cable).
 
Last edited:
Well spotted....I could use a 5v salas for both the analog and digital power for the AK4396, or even just leave the 3.3v as it is.

My transformer only has single 9v output....is it OK to have that driving two bridge rectifiers (the 7809, and the 5v salas) ?

not sure what you mean by the capacitance / offset issues.
 
Last edited:
Hello all,

finally i got my Dac kit already soldered shipped form singapore.
It seems that pin 27 and 28 from the AK4396 chip are connected together via a small spot of solder.

Now before using the dac and applying the BOM, i wonder if this happenend maybe accidentally by the supplier.

When I look at the shematic posted by dario it seems no problem, because pin 27 and 28 of the AK4393 are connected both to ground anyway.

But the shematic refers to the AK4393.
Is this maybe the same with the AK4396 or should i try to desolder the connection?


regards Knorst
 
hi clave.
1. i'm going to attempt this board with assistance from my father!
which would you recommend more 4393 or 4396?
2. been meaning to give it a go for some time and have your bom etc. not sure if it's the final one, could you provide links again please - can't find it in search.
3. does the shunt make much difference at all? only, being a novice noob, i'm guessing i'll be content with the sound without the shunt.

many thanks