[L_u_k_a_s]

Quote:

Originally Posted by ericj /img/forum/go_quote.gif Just got one of these, and i have to say, boy, the quality. Where is it?! Pin 2 is lifted and bent over to pin 3 on the dac. Easy to fix, but, gives one the impression that these weren't inspected at all.

Actually, it is designer's intention. See this reply from seller to my questions about pins 2 & 3 and images attached.

Reply from seller:
hello, i have saw all my dac boards.all the boards are the same .

and i asked the technican .

they tell me that it isn't bad.

it is designed like that.

it is better .

if you don't like that , you can change it by yourself.

 

[ericj]

Quote:

Originally Posted by L_u_k_a_s /img/forum/go_quote.gif Actually, it is designer's intention. See this reply from seller to my questions about pins 2 & 3 and images attached. Reply from seller: hello, i have saw all my dac boards.all the boards are the same . and i asked the technican . they tell me that it isn't bad. it is designed like that. it is better . if you don't like that , you can change it by yourself.

Yeah, I looked at pictures of 3 or so of these without that pin bent over before i posted here.

Also, reading the datasheet, I'm not sure what the idea behind it is. It looks like it just assures that the chip is in SPI mode if anything comes through the control port.

Edit: Also, having read a few threads about this dac, I'd like to know what was 'better' about the network around the opamp serving only to increase the DC offset, and about the spdif input network being totally wonky.

 

[L_u_k_a_s]

Quote:

Originally Posted by ericj /img/forum/go_quote.gif Yeah, I looked at pictures of 3 or so of these without that pin bent over before i posted here. Also, reading the datasheet, I'm not sure what the idea behind it is. It looks like it just assures that the chip is in SPI mode if anything comes through the control port. Edit: Also, having read a few threads about this dac, I'd like to know what was 'better' about the network around the opamp serving only to increase the DC offset, and about the spdif input network being totally wonky.

I guess it enables 96 and 192Khz sampling rates which are not working on boards without pin 2 lifted and pins shorted or modified in other way.

For network around opamp - I skip entire output stage, take unbalanced signal directly from the DAC and use valve buffer. Then you need only +5V DC to supply the board. Or no output stage at all - just 2uF cap in series, then 47K resistor to gnd and 100R resistor in series.

Wonky spdif - you mean loose conection on electrical? RCA is the worst connector for audio - analog and digital. I use 75 ohm BNC, preferably on both ends - source-cable and cable-dac. Or put DAC inside CD player and hardwire it.

 

[ericj]

No, enabling 96 and 192khz is a jumper between the spdif receiver and the dac.

Pin 2 is AD0 or CS depending on whether you're in i2c or spi mode. datasheet states "In i2c mode, AD0 is a chip address bit. CS is used to enable the control port interface in SPI mode. The device will enter SPI mode at any time a high or low transition is detected on this pin. Once the device has entered the SPI mode, it will remain until either the part is reset or undergoes a power-down cycle."

That means that if pin 3 is anything but floating, and pin 2 is attached to it, the control port interface is in SPI mode.

Edit: Except, pin 16 is grounded, so we're in hardware mode - no control port.

For pin 3 it states "in i2c mode, AD1 is a chip address bit. CDIN is the control data input line for the control port interface in SPI mode."

Edit: looking at the schematic in post 33, pin 3 is tied to +3.3v, and if pin 2 weren't lifted and bridged over, it would be tied to ground.

What do you mean by "take unbalanced signal from the DAC"? - the DAC's output is balanced, the opamp is there to convert it to single-ended.

And no, I do not mean the RCA jack. I mean the input network just the other side of the RCA jack and TOSLINK receiver. Some posts on diyaudio complained that the values of the components are completely wrong.

Edit: But it turns out i might mean something else entirely here.

 

[ericj]

Oh, I see what you're talking about with regard to pin 2 being tied high. Except, not really.

See here: diyAudio Forums - CS8416/CS4397 board at 96kHZ? - Page 1

If i read the datasheets right, tying this pin directly to vcc means the dac is in 96khz mode all the time.

Which is most uncool, imho.

 

[adamus]

I have done a side by side with this dac and the cambridge audio dacmagic. The noodle dac has the opa2107 in it and has had its electorlytic output caps removed and replaced with obbligato.

The noodle dac is the better sounding dac to my ears by quite some distance.

for all its problems, it really agrees with my ears.

 

[ericj]

Yeah, that's why i bought it. Low price + a little labor for great performance.

"Most uncool" is responding to the problem of "You forgot to hook up the 96khz select pin on the receiver to the 96khz enable pin on the DAC" by forcing the dac to only accept signals over 50khz. fwiw there's no way for this receiver to tell this dac to accept signals over 100khz - for 192khz you'd need to lift both pins and attach some switches to manually select that mode.

If you plan on upsampling everything to 96khz, well, that's really not my style, but the "pin 2 lifted and shorted to pin 3" works.

Did you change any of the resistors and caps around the opamp socket?

Stupid question: How is coax vs. optical selected?

 

[fordgtlover]

Quote:

Originally Posted by ericj /img/forum/go_quote.gif snip... Did you change any of the resistors and caps around the opamp socket?

I have changed the PLL filter components (R2, C22, C23) with the values from this post, the R8, R9, R14, R17 (30K) and C32, C33, C34, C35 (150pF) for the output filter from this post, and the output mod from this thread.



Quote:

Originally Posted by ericj /img/forum/go_quote.gif Stupid question: How is coax vs. optical selected?

Coax V optical is selected via the 3 pin header on my board, which is next to the 7 pin power connector.

 

[ericj]

OK. From the datasheet, there seems to be 4 inputs available in hardware mode, but it's clear as mud how they're selected. I'm considering cramming an M-Audio Sonica (like a Transit with no inputs) into the same enclosure, and would like to be able to switch usb/coax/optical.

Any ideas?

Thanks for the pointers, btw.

 

[L_u_k_a_s]

Quote:

Originally Posted by ericj /img/forum/go_quote.gif What do you mean by "take unbalanced signal from the DAC"? - the DAC's output is balanced, the opamp is there to convert it to single-ended.

Just connecting signal directly from R- and L- to the single ended valve buffer. Removed four capacitors after DAC so the whole onboard opamp output stage is not used.

 

[ericj]

So you mean you took the inverted signal. OK.

 

[L_u_k_a_s]

Quote:

Originally Posted by ericj /img/forum/go_quote.gif I mean the input network just the other side of the RCA jack and TOSLINK receiver. Some posts on diyaudio complained that the values of the components are completely wrong. Edit: But it turns out i might mean something else entirely here.

I'm back to post No. 169. What's wrong? I canť figure out whether caps are 0.01 μF, but otherwise, it is very nice - exactly how CS8416 datasheet suggests (chapter 16). For coax, it is C9 in series to RXP0 of receiver and R1 to gnd, for optical C7 in series to RXP1. Very short path. And RXN connected to C24, which goes then to gnd. There are other components connected to RXP5, I believe it is only "logic" for switching between RXP0 and RXP1 (coax and toslink) input (R7 - value 10R and R24 - value 1K are part of this logic).

Yesterday I put this little board inside CD player, it steals +5V DC (opamp output stage not used) from CD player board and has 4 inputs - 3 electrical spdif (from CD, external BNC, external coax) and toslink - conveniently selected with 2 pole 4 position switch, as well as BNC out passthrough. One pole of switch chooses electrical input and other pole takes care about logic electrical/optical. I put it inside smallish player just 27 cm wide and 7 cm high, can post some images if someone interested. I have only this crappy photo from outside. First I wanted to use I2S connection from CDp to DAC, but I'd have to solder directly to pins of chips and possibly break something. So I used spdif as if it were external DAC.

 

[ericj]

I mean the PLL filter network on the cs8416 - R2, C22, C23. The values are from an old version of the datasheet, the new version has a much better configuration. Described here:

CS4397 dac - The Art of Sound Forum

Also, since we're in hardware mode rather than software slave mode, there is no RXP5.

 

[ericj]

wow, I've really had this board 18 months?
 
I finally started reworking it today.
 
Friendly advice: Be careful and mindful when desoldering the smd parts. I accidentally took off R13 and didn't notice that i wasn't supposed to have done that for 20 minutes.
 
I found all sorts of little smd parts on the workbench, but guess what value resistor i still haven't found, over an hour later

 

[ericj]

So, I found my missing resistor: 
 

 
Next up:
 

 
I know the opamp resistors aren't necessary if i use transformer output, but i want to be able to compare the opamp output with the transformer output.
 
The edcors, fwiw, cost $19 shipped for the pair.