currently working on not-so-original rendition of the dsc.
"dual mono" with shunt power supply.
uses the Mcfifo as reclocker/isolator/buffer.
differential signaling generated at the source (in my case a BBB).
timing and impedance correct (i hope...)
for the moment i have put the OpenHardware logo on it... but i have to triple check before thinking to release it because uses a well-renowed topology for shunt and i guess it infringes some policy. If i can't i will release the project without the power supply on it.
keep you posted.
attached a very early prototype that needs much more work at the moment, just to tease and start some more inspired people than me to comment on it and maybe shout some suggestions
"dual mono" with shunt power supply.
uses the Mcfifo as reclocker/isolator/buffer.
differential signaling generated at the source (in my case a BBB).
timing and impedance correct (i hope...)
for the moment i have put the OpenHardware logo on it... but i have to triple check before thinking to release it because uses a well-renowed topology for shunt and i guess it infringes some policy. If i can't i will release the project without the power supply on it.
keep you posted.
attached a very early prototype that needs much more work at the moment, just to tease and start some more inspired people than me to comment on it and maybe shout some suggestions
Attachments
Looks great.
Ivan, in his future transformers, at my request, will make a central hole with a thread for reliable mounting. Therefore, it is not necessary to install a pin receptacle for an unconnected transformer contact. Saving ~1$. 🙂
Ivan, in his future transformers, at my request, will make a central hole with a thread for reliable mounting. Therefore, it is not necessary to install a pin receptacle for an unconnected transformer contact. Saving ~1$. 🙂
Last edited:
Thanks "master" 😉
what kind of screw will be mounted ? i guess till M4 we are fine with the already available pad.
just out of curiosity... have you kept the delay track between the ShiftRegister-Clock and the Storage-Clock ?
never experimented with them, but i have a 2.6.2 without them, and a DSC2 with it.
both worked fine.
have you kept them or you have straigt-shorted storage and SR clock (like 2.5/2.5.2/2.6.2) ?
thanks
Last edited:
Ivan just confirmed that there will be an internal nut M3
Right. I also see no reason to delay. Datasheet 74LVC8T595 allows these inputs to be used simultaneously.
You can not worry.
SHCP=STCP works great. This is verified on a real test board. Signalyst DSC1
SHCP=STCP works great. This is verified on a real test board. Signalyst DSC1
Last edited:
Good work! I'll follow this, and Pavel's latest development, with interest and will hopefully be able to hear the fruits of your labours in due course.
Love...
Your work here, following with great interest.
BTW, I have found a "problem" with Pavel's DSC-2.6 with Lundahl 1572XLs, or perhaps one could views this as an amp problem! Into Hypex NC-400s and Purifi amplifiers, there's still some "birdies" noise at the speakers. Bruno Putzeys has mentioned that some class D amps have problems with some DACs playing DSD which have a bit too much high frequency residual left on their output.
I wonder if different transformers (less bandwidth) would not have this problem, or if more filtering would be necessary to eliminate this problem.
The problem is subtle, but can be demonstrated with slowly plucked guitar or oud notes, playing alone (where one can easily hear the space between the notes).
Your work here, following with great interest.
BTW, I have found a "problem" with Pavel's DSC-2.6 with Lundahl 1572XLs, or perhaps one could views this as an amp problem! Into Hypex NC-400s and Purifi amplifiers, there's still some "birdies" noise at the speakers. Bruno Putzeys has mentioned that some class D amps have problems with some DACs playing DSD which have a bit too much high frequency residual left on their output.
I wonder if different transformers (less bandwidth) would not have this problem, or if more filtering would be necessary to eliminate this problem.
The problem is subtle, but can be demonstrated with slowly plucked guitar or oud notes, playing alone (where one can easily hear the space between the notes).
This is not a problem with your amplifiers.
This problem of correlation noise is present to varying degrees on some DSC instances. I still could not understand the cause of this interference.
Here are some tips to fix this:
This problem of correlation noise is present to varying degrees on some DSC instances. I still could not understand the cause of this interference.
Here are some tips to fix this:
- You can try thoroughly flushing the DSC board in an ultrasonic bath.
- If you use HQPlayer, switch to the modulator "AMSDM7 512+fs"
Thanks...
It is interesting, I do not hear this noise with some other amplifiers (all other things being equal, same test tracks, etc). But it is present with NC-400 and Purifi.
No HQPlayer right now, using Roon for DSD oversampling, but I am using the 7th order modulator and the "precise linear phase" filter, which I believe is also anodizing and steep, so I think the filter selection is not contributing.
No ultra sonic bath here, but this board was built by you... I could do an alcohol scrub down.
It is interesting, I do not hear this noise with some other amplifiers (all other things being equal, same test tracks, etc). But it is present with NC-400 and Purifi.
No HQPlayer right now, using Roon for DSD oversampling, but I am using the 7th order modulator and the "precise linear phase" filter, which I believe is also anodizing and steep, so I think the filter selection is not contributing.
No ultra sonic bath here, but this board was built by you... I could do an alcohol scrub down.
Could this be an instance of what Pavel Macura found with the NC-400s?
On Class D Amplifiers Measurements | Audio Science Review (ASR) Forum
I decided to stop working on the DSC2 project.
The development of the latest version of DSC2 is completed. But not yet implemented as real PCBs. I hope someone can bring this version to its logical end.
I offer a complete set of project files in the KCad format - http://puredsd.ru/dsc2expra_rel.zip.
The project documentation and source code is published on an “as is”.
The project consists of two parts:
The development of the latest version of DSC2 is completed. But not yet implemented as real PCBs. I hope someone can bring this version to its logical end.
I offer a complete set of project files in the KCad format - http://puredsd.ru/dsc2expra_rel.zip.
The project documentation and source code is published on an “as is”.
The project consists of two parts:
- main - main PCB DSC2 with transformer output and Open Hardware License. Interactive BOM for KiCAD
- output_nazar - active output PCB for replacing transformers. Circuit decisions are taken from open sources. But do not support Open Hardware License Interactive BOM for KiCAD
Attachments
Last edited:
No, Daphile ergonomics are not comfortable for me.
And the quality of PCM-DSD conversion through sox is also worse than HQPlayer.
And the quality of PCM-DSD conversion through sox is also worse than HQPlayer.
this is a real pity... sorry to hear that.
Nonetheless your contribution (and Jussi.. of course) can't be ignored in this application.
I would continue with the "DSC stick" (that is strictly done around my very personal needs...), but maybe, in the future, i will try to finish what you have started here.
Tonight i have such a great load of material to study that i am salivating 😀