Just a quick update on the "bclk_ratio_int_div" noise issue.
I checked my kernel sources. Clives patch that resolved the issue earlier was still there. Hmmh.
What changed from last time I verified that patch???
"Kali" - - I still have the Kali attached.
Just removed it. Guess what happened?
Right --- No more issues.
On the first glance It seems that the Kali is not as transparent as I thought it is.
That also explains why e.g. Tim has no issues on Moode 3.0 without Kali.
Let see if we'll figure it out. (Clive for sure won't support anymore. That he made very clear in a PM to me yesterday)
I checked my kernel sources. Clives patch that resolved the issue earlier was still there. Hmmh.
What changed from last time I verified that patch???
"Kali" -
- I still have the Kali attached.Just removed it. Guess what happened?
Right --- No more issues.
On the first glance It seems that the Kali is not as transparent as I thought it is.
That also explains why e.g. Tim has no issues on Moode 3.0 without Kali.
Let see if we'll figure it out. (Clive for sure won't support anymore. That he made very clear in a PM to me yesterday)
@Vasilis
Back to the LPF you recommended.
I'm now looking into it. Just downloaded the datasheet.
Just to recap.
The datasheet says 0R. You recommend 10R.
From my perspective 0R should mean -- bypassing your R shouldn't harm the
DAC from an ESS perspective!?!? This way I would have the R out of the way.
But you say 10R sounds better, even though we're facing slight "coloration" by
that 10R resistor. Is this correct?
Thx.
Back to the LPF you recommended.
I'm now looking into it. Just downloaded the datasheet.
Just to recap.
The datasheet says 0R. You recommend 10R.
From my perspective 0R should mean -- bypassing your R shouldn't harm the
DAC from an ESS perspective!?!? This way I would have the R out of the way.
But you say 10R sounds better, even though we're facing slight "coloration" by
that 10R resistor. Is this correct?
Thx.
Just to let you know.
I finally finished my SW setup.
My SW base is piCorePlayer 3.02.
* customized 4.4.y-simple kernel (numerous config changes)
* stripped down, platform dedicated, squeezelite binary with dynamic libraries (requires extra packages under pCP)
* slight overclocking
* several system/environmental mods
* external Gbit Realtek ethernet dongle - required a new kernel module ->300MBIT/s
Even with Kali in place, above customized setup still makes a nice difference.
Some might ask why am I supporting Tim and not using Moode myself.
Pretty easy. There are not very many OSes out there that come with squeezelite and Clives kernel. Most systems even lack slightest tuning measures.
Now people will have the choice to go squeezelite or MPD on one system and on top
that new kernel and some performance modifications have been introduced.
As nice as it sounds. That's also part of the downside.
Tim, same as Volumio, Rune, etc., provide multipurpose operating systems.
Multipurpose systems come with certain limitations and compromises by nature.
It's like comparing a family van with a Formula 1 race car.
If speed is the target, you won't go for a multipurpose family van.
Bottom line.
I'm more than pleased with the overall progress I made. I might try some more minor HW mods and than that'll be it for the time being.
Enjoy.
I finally finished my SW setup.
My SW base is piCorePlayer 3.02.
* customized 4.4.y-simple kernel (numerous config changes)
* stripped down, platform dedicated, squeezelite binary with dynamic libraries (requires extra packages under pCP)
* slight overclocking
* several system/environmental mods
* external Gbit Realtek ethernet dongle - required a new kernel module ->300MBIT/s
Even with Kali in place, above customized setup still makes a nice difference.
Some might ask why am I supporting Tim and not using Moode myself.
Pretty easy. There are not very many OSes out there that come with squeezelite and Clives kernel. Most systems even lack slightest tuning measures.
Now people will have the choice to go squeezelite or MPD on one system and on top
that new kernel and some performance modifications have been introduced.
As nice as it sounds. That's also part of the downside.
Tim, same as Volumio, Rune, etc., provide multipurpose operating systems.
Multipurpose systems come with certain limitations and compromises by nature.
It's like comparing a family van with a Formula 1 race car.
If speed is the target, you won't go for a multipurpose family van.
Bottom line.
I'm more than pleased with the overall progress I made. I might try some more minor HW mods and than that'll be it for the time being.
Enjoy.
The RPI have some emphasis in hi freq , special in very expensive systems sound a little bit hard , with 10 ohm the effect almost gone , general the 10 ohm soften a little bit the result ,
but this is a design way by my ,, and not the rule ,, any user may play with different think ( up of 12 ohm not recommend ! ) , about the caps don't try to chance the LPF with this way ,, really bad result .. keep the 4700pf and play with the resistors ..
about the colour , even with 0 ohm .. you paint the music nice
not recommend remove the resistor ,, the DAC chip come unstable in some situation
Nothing its never nothing .. LOL
Try 0 ohm carbon , ant then metal film ,, o yes !!! big different in theory zero ..
simple thinks in audio ( again LOL ) look the cable , all its a BIGGG ZERO in measure
serious now ... even the zero ohm in place work like damping for the LPF , and not the unknown follow ( cables & next stage ) circuit , and stay stable ..
Try 0 ohm carbon , ant then metal film ,, o yes !!! big different in theory zero ..
simple thinks in audio ( again LOL ) look the cable , all its a BIGGG ZERO in measure
serious now ... even the zero ohm in place work like damping for the LPF , and not the unknown follow ( cables & next stage ) circuit , and stay stable ..
No. It's a driver/overlay related parameter.
There's a README in /boot/overlays. That tells you exactly what parameters are available
for your DAC/driver/overlay.
The only way to figure out if it is active is to look at your /boot/config.txt - afaik.
I was browsing around deciding whether to buy the Kali now or put off the purchase a little while longer and then I stumbled upon this:
https://volumio.org/product/g2-labs-odroid-dac-4452-3-5″-touch-display-remote/
Haven't seen any discussion about it on the web get really. Thoughts?
https://volumio.org/product/g2-labs-odroid-dac-4452-3-5″-touch-display-remote/
Haven't seen any discussion about it on the web get really. Thoughts?
Looks interesting. G2 Labs built some nice DACs in the past from what I heard. I never owned one myself though.
However. I'd rather love to see a properly implemented AKM PI-HAT.
Not sure why people build boards for Odroid.
Than it's 12V.
And it's 185Euro.
Would be nice if these AKM chips could be controlled by a driver/gpio/alsamixer, like the PCM5122, instead of using a remote and display.
Beside that, this would make the package much cheaper.
However. I'd rather love to see a properly implemented AKM PI-HAT.
Not sure why people build boards for Odroid.
Than it's 12V.
And it's 185Euro.
Would be nice if these AKM chips could be controlled by a driver/gpio/alsamixer, like the PCM5122, instead of using a remote and display.
Beside that, this would make the package much cheaper.
I finally integrated a PI3 (with iFi Ipower-Rifa buffered)+Mambo+Kali ( both on LiFePO4+MravCla regs)
into my main system. (This won't be the final power supply setup though.)
It's the first time that I have the Mambo on the main system (PIO Sound Falcon speakers + Anaview AMS 0100 amps)
On the Mambo I replaced the output capacitors with SilverMica caps, as suggested by Vasili.
I also bypassed the 10R resistor (I'm now running according to datasheet).
I also removed the RCA jacks. I used the RCA jack holes for mounting the SilverMicas.
First impression: Major disappointment. ( nothing unusual though )
After all the efforts I put in, I was expecting a step up. Down it went.
My final broken-in and heavily modified HifiBerry DAC+ Pro setup -- without Kali -- had been much better.
I think Greg talked about a similar experience.
However. I loaded a burn-in track and ran it 3 hours. After that things were not that bad anymore. Still, I wasn't happy.
I than realized that I was still running LMS based resampling to 352.8/384. That's been my preferred choice on PCM51xx DACs. Bypassing the crappy PCM51xx on-DAC filters had been a relieve.
I switched back to native samplerates. Afaik the Sabre does some internal oversampling anyhow. There'd be no use to resample the data twice. (Perhaps somebody else has a more profound opinion about that)
Anyhow. I had the impression that things further improved after this change.
There's also much less load on the system going native. Usually this also accounts for slight improvements.
Current status.
I'm still in a situation where e.g. "applause" sounds kind of artificial and congested.
I do also miss a bit of energy in the upper range (cymbals/percussion etc.).
3D and stage is not as it used to be.
I'd probably also have to check if the Sabre DAC output stage and the Anaview amp impedance cause certain issues.
Bottom line. This proves again to me that every single system requires it's own project.
Results from one system (e.g. my desktop test system) can not just simply expected to be applicable for another system.
Anyhow. I won't give up that quickly. I'll give it some more time to break-in.
Enjoy.
into my main system. (This won't be the final power supply setup though.)
It's the first time that I have the Mambo on the main system (PIO Sound Falcon speakers + Anaview AMS 0100 amps)
On the Mambo I replaced the output capacitors with SilverMica caps, as suggested by Vasili.
I also bypassed the 10R resistor (I'm now running according to datasheet).
I also removed the RCA jacks. I used the RCA jack holes for mounting the SilverMicas.
First impression: Major disappointment. ( nothing unusual though )
After all the efforts I put in, I was expecting a step up. Down it went.
My final broken-in and heavily modified HifiBerry DAC+ Pro setup -- without Kali -- had been much better.
I think Greg talked about a similar experience.
However. I loaded a burn-in track and ran it 3 hours. After that things were not that bad anymore. Still, I wasn't happy.
I than realized that I was still running LMS based resampling to 352.8/384. That's been my preferred choice on PCM51xx DACs. Bypassing the crappy PCM51xx on-DAC filters had been a relieve.
I switched back to native samplerates. Afaik the Sabre does some internal oversampling anyhow. There'd be no use to resample the data twice. (Perhaps somebody else has a more profound opinion about that)
Anyhow. I had the impression that things further improved after this change.
There's also much less load on the system going native. Usually this also accounts for slight improvements.
Current status.
I'm still in a situation where e.g. "applause" sounds kind of artificial and congested.
I do also miss a bit of energy in the upper range (cymbals/percussion etc.).
3D and stage is not as it used to be.
I'd probably also have to check if the Sabre DAC output stage and the Anaview amp impedance cause certain issues.
Bottom line. This proves again to me that every single system requires it's own project.
Results from one system (e.g. my desktop test system) can not just simply expected to be applicable for another system.
Anyhow. I won't give up that quickly. I'll give it some more time to break-in.
Enjoy.
@Vasilis
I just figured or better -- it crossed my mind again -- that you go for 3.3V on AVCC.
That's basically the reason for 1.9Vrms - you mentioned earlier - instead of 2Vrms (@3.6V AVCC) on the output.
On the other hand I read somewhere that the 3.6V option would be the better
choice for that chip. (output swing might hit supply rails @3.3V AVCC)
I'd assume you went for 3.3V because it's easier to implement.
Probably you could explain, why 3.3V instead of 3.6V?
Thx.
PS-Edit: I just stepped over this related and interesting discussion confirming above.
I just figured or better -- it crossed my mind again -- that you go for 3.3V on AVCC.
That's basically the reason for 1.9Vrms - you mentioned earlier - instead of 2Vrms (@3.6V AVCC) on the output.
On the other hand I read somewhere that the 3.6V option would be the better
choice for that chip. (output swing might hit supply rails @3.3V AVCC)
I'd assume you went for 3.3V because it's easier to implement.
Probably you could explain, why 3.3V instead of 3.6V?
Thx.
PS-Edit: I just stepped over this related and interesting discussion confirming above.
Last edited:
I have working the chip even with 5 volt ..
but the most stable and overall solution for any situation its the 3.3 volt , and this is even for AYAZI dac ,
up of 3.3 volt more lively and a little bit open stage ,, in some program ... and in some system ... but ,, in some system its wild and finally tired sound
i speak for AYAZI because i have the ability for any voltage in any point doe to COMPLETE different think
in case of MAMBOBERRY LS DAC , i try make the best with very specific component ,, in a price range so high as high its salami , cheese and some soda for Sunday sandwich ,
so the best and cheaper regulator its only 3.3volt , my point its .... the sabre play well with 3.3v , BUT if 3.6 volt was better choice ,, the mamboberry would have chosen again 3.3volt , just for economical reason
so .. feel free to feed with any voltage up to 3.8 volt ...
( sorry for my bad english yes i now its really tired to read my, but forgive me .. my mind confuse between 4 year France , a little english , and thousand of electronic paper .. )
but the most stable and overall solution for any situation its the 3.3 volt , and this is even for AYAZI dac ,
up of 3.3 volt more lively and a little bit open stage ,, in some program ... and in some system ... but ,, in some system its wild and finally tired sound
i speak for AYAZI because i have the ability for any voltage in any point doe to COMPLETE different think
in case of MAMBOBERRY LS DAC , i try make the best with very specific component ,, in a price range so high as high its salami , cheese and some soda for Sunday sandwich ,
so the best and cheaper regulator its only 3.3volt , my point its .... the sabre play well with 3.3v , BUT if 3.6 volt was better choice ,, the mamboberry would have chosen again 3.3volt , just for economical reason
so .. feel free to feed with any voltage up to 3.8 volt ...
( sorry for my bad english yes i now its really tired to read my, but forgive me .. my mind confuse between 4 year France , a little english , and thousand of electronic paper .. )
It seems you're enjoying quite fancy salami-cheese sandwiches over there in Greece.
(Even without Kali and power supplies on the balance)
Ok. I very well understood what you're talking about:
3.3V is your preferred choice from a soundquality perspective.
However. Things might differ from system to system.
That puts me in the situation to add a 3.6V reg to the system or to leave it as it is.
Or I just prepare a nice salami-cheese sandwich myself and enjoy some nice music.
Thx.
(Even without Kali and power supplies on the balance)
Ok. I very well understood what you're talking about:
3.3V is your preferred choice from a soundquality perspective.
However. Things might differ from system to system.
That puts me in the situation to add a 3.6V reg to the system or to leave it as it is.
Or I just prepare a nice salami-cheese sandwich myself and enjoy some nice music.
Thx.
Last edited:
@soundcheck you own me a diner ...
just almost i burn the night diner ... the chicken its half black and half white
Yep I know - <= 130k. Thx.
Enjoy your biracial chicken.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.