Dual Mono DAC AK4490/AK4497

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Member
Joined 2015
Paid Member
I hadn't noticed that the 4497 had separate pins for PCM and DSD.

That is very annoying.

Most (if not all) USB interfaces have a single combined I2S/DSD output. That will have to be split in two, and possibly buffered.

Most if not all USB receivers provide a "DSD signal" pin.

for which usage will that be annoying ?
You can connect the combined output of your USB board to the combined input of the 4497 (pins 3,4,5) (same for 4495 and I guess 4490) and switch DAC behavior through the uC upon DSD/PCM detection made by USB boards.

If you want to add a spdif receiver (such 4118) you have to add a mux quad 2-1 on the combined input. Is that what you referring to ?
 
Nope, no dice. DSD requires register programming, same as every other DAC chip on the planet. The internal mathematics are too severe a change to allow for auto-switch.

I've been running a Buffalo III for many years now and it does distinguish by itself whether it is receiving PCM or DSD data (see Status Register (27)). I'm sure that I'm not setting any registers differently for DSD vs. PCM mode.

If you look at the register table, there is no register to specifically select "DSD input".
 
for which usage will that be annoying ?
You can connect the combined output of your USB board to the combined input of the 4497 (pins 3,4,5) (same for 4495 and I guess 4490) and switch DAC behavior through the uC upon DSD/PCM detection made by USB boards.

If you want to add a spdif receiver (such 4118) you have to add a mux quad 2-1 on the combined input. Is that what you referring to ?

I was looking at AK4497's Figure 1 (the block diagram) and came under the (wrong) impression that PCM and DSD signals came in through different pins. I should have looked at Figure 2, where is is obvious that PCM & DSD use combined pins. So, false alarm.

I probably had AK4118 in my mind, which does indeed have separate inputs for I2S and DSD (the annoying part I was referring to).

I do feed my 4490s with either PCM or DSD and select the proper format with my uC.
 
I'm just talking about Japanese DAC BOARDS, mentioned in kukynas post #246.
And idea was: not to use in this combination any USB at all. Receive a kind of internal signal from (FIFO for example) which we will be able to use as drive signal for the DSD/PCM mode switch of AK4495/97 DAC. I mean DSD/PCM detection.

This will be kind of solution for "non-automatic DACs"

thank you)
 
Member
Joined 2015
Paid Member
I was looking at AK4497's Figure 1 (the block diagram) and came under the (wrong) impression that PCM and DSD signals came in through different pins. I should have looked at Figure 2, where is is obvious that PCM & DSD use combined pins. So, false alarm.

I probably had AK4118 in my mind, which does indeed have separate inputs for I2S and DSD (the annoying part I was referring to).

I do feed my 4490s with either PCM or DSD and select the proper format with my uC.

OK. I've seen your progress on your blog. It 's great !.
I'm currently designing an 4497 DAC from scratch and was currently working on input stream with the various existing board. If Amanero and JLS provide DSD detection, I don't know how Waveio does, I'm waiting an answer from Waveio thread. I would like my DAC to be diy friendly : compatible with most USB adapter.
 
No you have different input only if you are in DSDPATH=0 mode
 

Attachments

  • PCM-DSD.JPG
    PCM-DSD.JPG
    41.5 KB · Views: 651
Member
Joined 2015
Paid Member
If the DAC use a digital interface transciever like a DIX9211.
Perhaps it could use the DIX9211 PIN output that indicates if PCM or not.

Those interface adapter (such DIX, AK411x, etc) are for spdif I/F, and spdif has been designed with PCM so it don't handle DSD. A tweak exists : DoP (DSD Over PCM), it transport DSD payload into spdif frame. With DoP your player (typically a software running on your favorite PC, MAC, Arm...) will play DSD and the DAC (over the DoP transport made on USB link) will receive PCM. But it's not our topic as we're talking about DSD input.
So those interface chip (for the ones I know..) doesn't help to detect DSD, they can detect non PCM but non PCM isn't always DSD.
 
The boards there has a separated PCM and DSD inputs. As well they doesn't automatically switch between PCM and DSD, need manual selecting.

why would you think? both of the inputs sharing same I2S lines and are driven by ON/OFF isolators once DSD/PCM signal is provided, you can use both inputs if you can supply P8 header with proper info (DSD/PCM), look at the manual, he is even mentioning usage of Amanero board

ES as well as AK449x has shared set of I2S lines for both modes so the only differentiator is "DSD signal" pin (called differently per each chip) to tell DAC chip which signal (DSD/PCM) is supplied
 
The AK4490/4497 do not support automatic switching between PCM and DSD. You have to know what kind of signal you will be converting and use a uC to set up the DAC accordingly.

I have a AK4495 board with XMOS USB/I2S input and I can feed it PCM or DSD (DOP) without need to change anything on board. I use HQPlayer to switch between PCM or DSD conversion.

BTW, I use direct transformer out for the AK4495 and LiPoFe battery supplies for the 3.2v digital and 6.4V analoge supplies.

I also have a Chinese AK4497 board, but I haven't got time to implement it yet.
 
DAC4497-2 DUAL AK4497 by http://www.easyaudiokit.com/

why would you think? both of the inputs sharing same I2S lines and are driven by ON/OFF isolators once DSD/PCM signal is provided, you can use both inputs if you can supply P8 header with proper info (DSD/PCM), look at the manual, he is even mentioning usage of Amanero board

ES as well as AK449x has shared set of I2S lines for both modes so the only differentiator is "DSD signal" pin (called differently per each chip) to tell DAC chip which signal (DSD/PCM) is supplied

Yes, you're right.
we can connect like this:

FIFO-MC ---->CN1---------->CN2
SDATA/DSDR-->PIN1--------->PIN3
LRCK/DSDL--->PIN3--------->PIN1
BCLK/DSDCLK->PIN5--------->PIN5
SCLK/SCLK--->PIN9--------->PIN9
GND--------->PIN2,4,6,8--->PIN2,4,6,8

in case of FIFO+Master clock, but we still have no solution how to recover from I2S signal flow differenciator (DSD/PCM) signal and feed to exact pin of DAC chip WITHOUT using USB to I2S conversion.
 
in case of FIFO+Master clock, but we still have no solution how to recover from I2S signal flow differenciator (DSD/PCM) signal and feed to exact pin of DAC chip WITHOUT using USB to I2S conversion.

If I understand correctly what you are saying, there is no "DSD/PCM selector pin" on the AK44xx DACs. This DSD/PCM signal needs to be routed to a uC which will in turn send the necessary commands to the AK44xx DAC to switch it to either PCM or DSD mode.

I'm pretty sure that Ian's reclocker will provide a way to signal the AK's uC of the signal type. You could ask him.
 
Why do you link SCLK to Vdd (PIN9) an not to the master clock (PIN 7) ?

I tried to find out how works the switch PCM/DSD on the SimpleEntryDAC4490 from easyaudiokit.
No signal from DXI9211 to the DAC Pin and no software detection with the PIC (no link with any clock signal exept PIC clock) found.
http://www.easyaudiokit.com/bekkan/manual/SimpleEntryDac4490.pdf

Any ideas ?

Since the PIC connects to both the DXI and the AK chips through I2C, that's probably how they talk. I'm guessing that the PIC sees DSD on the DXI's input (by checking a register on the DXI) so it configures the AKs to DSD mode.

This is just a guess.. I haven't looked at the DXI's datasheet to see how it works.
 
Member
Joined 2015
Paid Member
Why do you link SCLK to Vdd (PIN9) an not to the master clock (PIN 7) ?

I tried to find out how works the switch PCM/DSD on the SimpleEntryDAC4490 from easyaudiokit.
No signal from DXI9211 to the DAC Pin and no software detection with the PIC (no link with any clock signal exept PIC clock) found.
http://www.easyaudiokit.com/bekkan/manual/SimpleEntryDac4490.pdf

Any ideas ?

Interesting spec and diagram. I had quick look on it and DIX9211. DIX knows nothing about DSD, its only PCM or passthrough. DSD inputs should be connected on Port0 & 1 of the board which are GPIO B & C of the DIX. When switched by the PIC the DIX act as a passthrough from Port0 or 1 to the 4 I2S/DSD input pins of the 4490s.
Two options :
a/ there's no PCM/DSD detection , the port 0 and 1 have to be configured before usage to DSD or PCM, then when you switch to these inputs the 4490 is setup accordingly.
b/ the PIC does the detection by analysing the frequency of BICK connected to port C0 of PIC. Frequency for various fs are well know for PCM and for DSD, this may be analysed via software (I've to do some calculation in order to check if we have differents frequency for each fs and mode). The display on page 14 saying 'Unknown I2S' let me think this is reasonable to admit that PCM/DSD detection is made to BICK analysis. In addition port C0 is connected internally to Timer 1 of the PIC, which may be useful for frequency analysis. If true it's smart ! but those are only assumptions :)
 
Member
Joined 2015
Paid Member
Too fast !

Interesting spec and diagram. I had quick look on it and DIX9211. DIX knows nothing about DSD, its only PCM or passthrough. DSD inputs should be connected on Port0 & 1 of the board which are GPIO B & C of the DIX. When switched by the PIC the DIX act as a passthrough from Port0 or 1 to the 4 I2S/DSD input pins of the 4490s.
Two options :
a/ there's no PCM/DSD detection , the port 0 and 1 have to be configured before usage to DSD or PCM, then when you switch to these inputs the 4490 is setup accordingly.
b/ the PIC does the detection by analysing the frequency of BICK connected to port C0 of PIC. Frequency for various fs are well know for PCM and for DSD, this may be analysed via software (I've to do some calculation in order to check if we have differents frequency for each fs and mode). The display on page 14 saying 'Unknown I2S' let me think this is reasonable to admit that PCM/DSD detection is made to BICK analysis. In addition port C0 is connected internally to Timer 1 of the PIC, which may be useful for frequency analysis. If true it's smart ! but those are only assumptions :)

Went too fast on DIX datasheet reading. It has a DSD Mode aimed at suppressing jitter on HDMI connection. This mode act as a passthrough but doesn't seems to provide any kind of detection. So the problem remains unless a more detailed read highlight some interesting feature...
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.