But you are being very helpful, Thanks!!
Immediately it looked like what I was looking for, but I think it only can provide 2 channels of DSD unfortunately. You can choose between multi channel I2S or 2- channel DSD as far as I can see from the application note on the CT7601 EVM .
Please correct me if I am wrong, because I really want to be wrong here ;-)
Do you like this eval. board? Looks like it is very easy to implement. Any comments on sound quality?
Hi Zoran.
Perhaps you are referring to their SRC eval. board? CT7302?
That seems to be a workable solution. What about the sound quality of this board?
Perhaps you are referring to their SRC eval. board? CT7302?
That seems to be a workable solution. What about the sound quality of this board?
If I was to try out the sound quality of the AK4137 would I have any succes with this:
AK4137 I2S/DSD Sample Rate Conversion Board Supports PCM/DSD Interchange Supports DOP Input for hifi amplificatore G8 002|Amplifier| - AliExpress
Sorry Zoran
I wasn't really awake , I think.
It is of course the eval board for the chip I started this thread about 🙄
I have actually a quote for this board from ComTrue...
I must have been sleeping Sorry
Unfortunately, not its best sound quality. There are a number of issues with each of the three basic styles of Chinese AK4137 boards. I have 2 of that particular style. Both of them have been modified in different ways, neither one fully.
Problems are:
- DSD clock inversion (can be okay for AKM dacs, may still work with other dacs but sound quality questionable -- can be fixed by I2C register programming)
- Onboard reference clock jitter too high
- 3.3v voltage regulation for AK4137 too noisy
- Clock voltage regulation too noisy
- Board firmware allows direct conversion of 16/44 to DSD256 (just don't use that mode for best sound quality)
- It will still be sensitive to input I2S jitter, if other than low jitter source used.
- Multiple I2C bus register settings not necessarily optimized for best sound quality
- Best upsampled DSD sound quality requires pre-upsampling of PCM before AK4137, then use AK4137 only for conversion to DSD256 (CT7302 is not clean enough to do the pre-upsampling, I tried it)
Only good implementation of AK4137 I have seen for DSD conversion is on my modded AK4499 evaluation board. PCB around AK4137 was modified to allow input of externally upsampled low jitter PCM into AK4137. AK4137 reference clock input was driven from same low jitter clock driving the dac chip. Since the PCB is already 4-layer and 1st stage voltage regulation for AK4137 is already provided by external buffered LT3042, no other modifications were needed. Also, had to program I2C registers myself.
Last edited:
Thanks for that clarification. Maybe the difference in observations of the sound quality between the AK4137 and CT7302PL is that you have solved the issues with the AK4137 that is there on the evaluation boards and compared with a CT7302PL evaluation board and Andrea/QuaHong has solved the CT7302PL issues in their implementation and compared that to an AK4137 evaluation board??
Anyways, the points you set up:
Inversion of the DSD clock would that be an issue with the No-Dac type of DSD Direct converters (DSC types and others)
Can the eval board accept slave mode for clk. My DSC 2 has onboard low jitter clock.
3,3 V regulation... No problem I always use LifePO4 batteries for 3,3V supplies
Covered by slave mode clk.
See later
Amanero USB board slaved to low jitter clk?
That is beyond me. I have no way to change that.
No problem. I can get J-river to do pre-upsampling
So is it at all possible, without your skills, to get the AK4137 to sound good?
When you ask if 'the eval board' can accept slave clock, are you referring to AK4499 eval board? If so, it doesn't have a master clock. MCLK has to be supplied by the user or generated by AK4118 SPDIF reciever.
Amanaro and or JL Sounds USB board can use an external master clock. Amanero requires 22/24MHz. JL Sounds requires 45/49MHz. AK4137 reference clock can come from anywhere, from a dac board is fine so long as it is still low jitter by the time it gets to AK4137.
However, AK4137 has what AKM calls master and slave modes, but that is for output-I2S clock master (BCLK, LRCK). Best not to use that slave mode as it is more jittery than using a MCLK reference. Where the reference MCLK comes from doesn't matter, just has to be the right frequency (22/24MHz).
Not much can be done with CT7302 that I didn't already try. However, if they did happen to do as you suggest then lets hear about it.
Amanaro and or JL Sounds USB board can use an external master clock. Amanero requires 22/24MHz. JL Sounds requires 45/49MHz. AK4137 reference clock can come from anywhere, from a dac board is fine so long as it is still low jitter by the time it gets to AK4137.
However, AK4137 has what AKM calls master and slave modes, but that is for output-I2S clock master (BCLK, LRCK). Best not to use that slave mode as it is more jittery than using a MCLK reference. Where the reference MCLK comes from doesn't matter, just has to be the right frequency (22/24MHz).
Not much can be done with CT7302 that I didn't already try. However, if they did happen to do as you suggest then lets hear about it.
Last edited:
No not the AK4499 board.
The chain could be : amanero (or Mini DSP x-mos multichannel USB streamer) clocked from the DSC board - AK4137 also clocked from the DSC board - DSC - analog out.
You could do that. I would choose Amanero over Mini DSP. Not sure if Mini DSP will resample your audio or not, but sure that Amanero won't. Modding one of the better Chinese AK4137 is possible, although it would be some work and involve some tiny soldering. For example, a couple of pics below of freeing up the I2C bus on an AK4137 board like the one you linked to previously.
Attachments
The SMD soldering is no problem. I did all the SMD work on my DSC 2 board including smal very fragile Potato Chips. But I2C is not something I can do...
I agree with you.
I am currently developing a cape for BeagleBone. There will be reverse reclocking on quality generators and two AK4137s. The first is upsampling to 384, the second is PCM384 to DSD256. And equalizing the output jitter using flipflops.
For fun, I would like to point to Wolf, at the ASR forum..
He went to dig deeper into the structure of a NuPrime Dac9 SE unit - - and he had found that:
It is using a CT7302, as upsampler on the input;
It is using an AK4493 (very probably in bypass mode) as a final modulator;
Then it goes into a Muses active volume control chip.. 🙂 🙂
So, if You wonder about the sonic capabilities of a Star- like structure, look up the reviews on that box... 🙂 🙂
Ciao, George
(i imagine the dac9 is the same, but with a 4490..)
He went to dig deeper into the structure of a NuPrime Dac9 SE unit - - and he had found that:
It is using a CT7302, as upsampler on the input;
It is using an AK4493 (very probably in bypass mode) as a final modulator;
Then it goes into a Muses active volume control chip.. 🙂 🙂
So, if You wonder about the sonic capabilities of a Star- like structure, look up the reviews on that box... 🙂 🙂
Ciao, George
(i imagine the dac9 is the same, but with a 4490..)
Last edited:
It is not a problem because You will get better results with MCK clocking then with BCK... And If You want clocking with BCK You can recklock BCK with MCK and 2 x FlipFlop and get BCK from -Q out. But it deserves min. 2x Frequency of MCK, like 45/49MHz for max sample rate BCK.
Last edited:
I've almost finished designing the circuit. It takes a few more days to design the PCB. In the near future, a new topic will be opened on the forum on this device.
Yes of course. 🙂
Not. I find it more convenient to use Linux to control multiple GPIO and I2C signals. Therefore, for now, the project will be exclusively with BeagleBone. Maybe later I will do the same for Amanero.
Last edited:
The firmware supports all known audio protocols: Roon, HQP, DLNA, AirPlay, LMS.
There will be no direct playback from network folders. This violates the concept of a network client.