Support for Botic Linux driver

[nautibuoy]

I used 7Zip to decompress and win32diskimager to load image.

Regards

OK, no offence intended with the question. Did you manage with the build following twluke's suggestions.

When I get a chance I'll post an updated image file, with the latest patch.

Ray

 

[smanz]

@twluke. Thanks for your help.

OK, no offence intended with the question. Did you manage with the build following twluke's suggestions.

Your question does not offend at all. I'm very clumsy for Linux. I got nothing.

When I get a chance I'll post an updated image file, with the latest patch.

I wait impatiently. Many thanks.

Regards

 

[nautibuoy]

@twluke. Thanks for your help.

Your question does not offend at all. I'm very clumsy for Linux. I got nothing.

I wait impatiently. Many thanks.

Regards

A new image is uploading at the moment and should be available in about an hour. Here's a link to it;

https://www.dropbox.com/s/vb0jh2fwcqudi6p/BBB-eMMC-28929.img.gz?dl=0

I hope it works for you.

BTW, you might care to try this tool, it'll unzip and image all in one and is very easy to use;

Etcher by resin.io

Ray

 

[nautibuoy]

The new image seems to have uploaded successfully.

Just a note to make people aware that the image will flash your eMMC at first boot so you may want to back up any current image that is of value to you.

Having copied the image onto your uSD, insert it and boot from the uSD (hold the boot switch down on the BBB for a few seconds initially). The image will be flashed to the eMMC (takes about 10mins or so). On completion, power off, remove the uSD, and restart.

Login/password are debian/temppwd.

Ray

 

[smanz]

Thanks Ray. Finally, your image runs on my system.

My first impresssion is that HQ-NAA-BBB/Botic no is better them Foobar push files to BBB/Botic thru UPnP/DLNA.

At the moment, "Mister NO"* player is not for me.

I use this Foobar component foobar2000: Components Repository - UPnP/DLNA Renderer, Server, Control Point

Regards

*"Mister NO" player. Is my alias for HQ Player. NO Mp3, NO CUE, NO ape, NO friendly interface, NO user manual, NO free, NO ...................

 

[nautibuoy]

Thanks Ray. Finally, your image runs on my system.

My first impresssion is that HQ-NAA-BBB/Botic no is better them Foobar push files to BBB/Botic thru UPnP/DLNA.

At the moment, "Mister NO"* player is not for me.

I use this Foobar component foobar2000: Components Repository - UPnP/DLNA Renderer, Server, Control Point

Regards

*"Mister NO" player. Is my alias for HQ Player. NO Mp3, NO CUE, NO ape, NO friendly interface, NO user manual, NO free, NO ...................

No problem, each to their own.

I would probably agree with your assessment if using HQPlayer/NAA as an alternative to UPnP playback of media 'as is' but to me the improvements I perceive with HQPlayer's upsampling/filtering, especially to high data rates, are worth having.

Incidentally, I 'drive' HQPlayer via Muso;

http://klarita.net/muso.html

Thanks for confirming the image works too.

Ray

 

[twluke]

After the latest patch by miero, the BBB/Botic/NAA for HQP became well stabilized. However, Even after this remarkable improvement, there still remained a problem.

I wrote before:

Of course, there is one problem. After getting started with playing music, a kind of background noise also appears. This is faint and similar to the sound of wave at the shore, probably associated with the process of DSD locking. This noise soon disappears but reappears at some point, though there was no regularity. The situation is common through both B3 and B3SE.

I noticed this faint background noise is still persistent, particularly when using the Hermes/Cronus with 45/49 clocks for DSD512 output to the B3SE with its DAC clock still configured as it is. However, I also noticed that this noise, probably related to the DSD locking, became totally absent when listening through my customized BBB/Botic/NAA system with 22/24 Crystek clocks, that had no reclocking function like the Cronus.

On the assumption that the noise may be associated with this function, I assembled another customized BBB/Botic/NAA system with 45/49 Crystek clocks connected to B3SE via Teleporter as shown below. As expected, this system well eliminated the noise and I could listen to very clear, almost transparent DSD512 sound, probably better than with 22/24 clocks.

I'm well aware that the situation may be different if the external clocks are used as master, but, if not, the BBB/Botic using Hermes/Cronus with 45/49 clocks may be less suitable for NAA, though I can not explain why.

 

[francolargo]

However, I also noticed that this noise, probably related to the DSD locking, became totally absent when listening through my customized BBB/Botic/NAA system with 22/24 Crystek clocks, that had no reclocking function like the Cronus.

On the assumption that the noise may be associated with this function, I assembled another customized BBB/Botic/NAA system with 45/49 Crystek clocks connected to B3SE via Teleporter as shown below. As expected, this system well eliminated the noise and I could listen to very clear, almost transparent DSD512 sound, probably better than with 22/24 clocks.

Forgive me for not following more closely. Could you clarify the clocking scheme of your 'custom' system? I assume the Crystek clocks are feeding the BBB directly, correct? ...but nothing else? ...and frequency divider?

TIA,

Frank

 

[nautibuoy]

Thanks for the update twluke.

I've been listening at DSD512 quite a bit (via headphones) and have experienced similar results to you.

With the latest patch the BBB/Botic/NAA is very stable and I don't recall a single dropout even though everything is being pushed hard, especially the network connection.

But I am getting occasional spells of a low level static-like noise (like the sea on the beach as you describe). It doesn't last long, is infrequent and, I believe, is always at the start of tracks, which tends to reinforce your suggestion that it is related to locking. SQ without the noise is excellent, very organic.

Unlike your setup, my reclocker (90/98MHz oscillators) is syncing to the B3SE (it's onboard oscillator isn't powered). I think that we can conclude that the issue isn't related to this.

I was wondering if the ES9018 jitter reduction might be a factor and was going to find B3SE config guide and try turning it off but your observations suggest it won't. Still, I will try it anyway and report back.

Need to think on this some more.

Ray

 

[nautibuoy]

Forgive me for not following more closely. Could you clarify the clocking scheme of your 'custom' system? I assume the Crystek clocks are feeding the BBB directly, correct? ...but nothing else? ...and frequency divider?

That's my reading of twluke's custom build, he's replacing the BBB's own onboard clock with high quality clocks, selectable for 44.1/48 rate data families.

As the BBB is essentially (re)generating the PCM/DSD data streams from the network data stream that seems to imply that the PCM/DSD should be good with respect to jitter? The B3SE will be doing jitter reduction too. So what are you losing by removing Hermes/Cronus - just isolation as far as I can see - maybe the simplification outweighs this. Maybe I'm talking out of my hat!

Ray

 

[nautibuoy]

...I also noticed that this noise, probably related to the DSD locking

Well, I think I can make the noise happen at will, I seem to get it every time I switch from a 44.1K source file to a 48KHz one (and vice versa). Always on the first, and only the first, track at that rate.

This suggest it is related to data locking?

I'm only getting it at DSD512 (but haven't tried PCM384).

Ray

 

[nautibuoy]

...Always on the first, and only the first, track at that rate...

Not quite; no sooner had I posted that earlier than I had the noise on a track that wasn't the first one and it has happened again since.

Ray

 

[nautibuoy]

I tried disabling the jitter reduction on the B3SE (SW2-5 moved to 'off'); the result is no music, the lock LED goes out and the mute LED turns on.

Moving SW2-5 back to 'on' and everything is back as it was.

Ray

 

[Signalyst]

*"Mister NO" player. Is my alias for HQ Player. NO Mp3, NO CUE, NO ape, NO friendly interface, NO user manual, NO free, NO ...................

Excuse me, but every HQPlayer copy comes with PDF user manual... Cannot say the same about lot of other software.

I could also have my list of NO stuff for foobar. Apart from horrible GUI, NO multi-source aggregate library views, NO regular expression searches on the library, NO DSD convolution engine, NO DSD speaker configuration (distance/level), NO complex DSD routing and mixing, NO things included with the base player but instead need to hunt for gazillion of shady plugins, NO official performance testing with various DACs, NO free DAC design optimized for the player, NO active filter/modulator design, NO GPU acceleration for DSP algorithms, NO .......


P.S. HQPlayer is about quality, and there's no quality whatsoever in MP3 - one of the most crappy lossy codecs on earth. NO reason to support something that is too bad to begin with especially when the patent license costs minimum $15k per year: mp3licensing.com - Royalty Rates .

 

[nautibuoy]

...if not, the BBB/Botic using Hermes/Cronus with 45/49 clocks may be less suitable for NAA, though I can not explain why.

I've but a few hours into listening today and have reached some working conclusions.

My implementation doesn't use Hermes/Cronus but another 'brand' isolator/reclocker but they essentially offer the same functionality with similar hardware (though my oscillators are 90/98MHz) so I think my conclusions map across well I think.

I think BBB/Botic with Hermes/Cronus should make an excellent NAA solution for PCM up to 384KHz and up to quad-rate DSD (DSD256).

It is also DSD512 capable but experiences short bursts of a static like noise when used with the B3SE DAC.

I think the next step is to try and determine whether the static noise is being generated by the BBB/Botic/NAA/isolator/reclocker or by the B3SE.

As we are seemingly using the ESS9018 right on the edge of its capability (some seem to think it is only DSD256 capable) that is where I cast the eye of suspicion. I think the next step should be to try something instead of the B3SE - any ideas? Personally I think I'll push on with a No-DAC low pass filter project to try with DSD512.

 

[twluke]

Forgive me for not following more closely. Could you clarify the clocking scheme of your 'custom' system? I assume the Crystek clocks are feeding the BBB directly, correct? ...but nothing else? ...and frequency divider?

Hi Frank, you are always welcome and thanks for this question.

The clocking scheme is quite accordant with the miero's instruction, in which the signal lines from 44.1k/48k clocks first go to 2:1 MUX controlled by the external master clock switch from P9_24 and then the output goes to P9_25 (a grey wire in the last picture I posted) just for feeding as you wrote. There is no frequency divider in my system. BTW, the PCM/DSD signals are sent to the B3SE via an isolator(ISO7640FM) with different Vcc.

Regards,

 

[twluke]

I think BBB/Botic with Hermes/Cronus should make an excellent NAA solution for PCM up to 384KHz and up to quad-rate DSD (DSD256).

I totally agree with you. And also, let me add octa-rate DSD (DSD512) in some setting like mine to your list (in a modest attitude ).

I think the next step is to try and determine whether the static noise is being generated by the BBB/Botic/NAA/isolator/reclocker or by the B3SE.

As we are seemingly using the ESS9018 right on the edge of its capability (some seem to think it is only DSD256 capable) that is where I cast the eye of suspicion. I think the next step should be to try something instead of the B3SE - any ideas? Personally I think I'll push on with a No-DAC low pass filter project to try with DSD512.

So, when based on your latest comments, Cronus' reclocking function may not be the source of this locking(?) noise. Then the next question will be thrown on the stability of bit clock (or DSD clock) rate actually used by B3SE.

In case of B3SE in slave mode like yours, I think the BCLK deriving from the external clocks is going to be used by B3SE but if the DAC clock of B3SE is still working as a master like mine, then the final BCLK is going to be formed by the B3SE's 100MHz DAC clock handled by ES9018. Thus it can be said that this difference may finally appear as the difference in the amount of audible noise. Just my humble speculation.

BTW, I suppose ES9018 had long been considered as DSD512-capable as seen in this link.

Regards,

 

[miero]

We might not be able to push these modules to higher frequencies, because of the EMI issues. The wavelength of 100MHz is ~3cm.
Check this: LearnEMC - PCB Layout

 

[nautibuoy]

@twluke, thanks for the link about the ESS9018 being DSD512 capable, though there do seem to be different views on what it is actually spec'ed for.

In case of B3SE in slave mode like yours, I think the BCLK deriving from the external clocks is going to be used by B3SE but if the DAC clock of B3SE is still working as a master like mine, then the final BCLK is going to be formed by the B3SE's 100MHz DAC clock handled by ES9018. Thus it can be said that this difference may finally appear as the difference in the amount of audible noise. Just my humble speculation.

Do we agree that, with your Hermes/Cronus and my alternative isolator/reclocker arrangements, we seem to be experiencing similar patterns of the static-like noise with DSD512, regardless of whether the B3SE is using its onboard oscillator or is getting it's clocking from the reclocker (and the noise fades away and is absent most of the time)?

So, when based on your latest comments, Cronus' reclocking function may not be the source of this locking(?) noise. Then the next question will be thrown on the stability of bit clock (or DSD clock) rate actually used by B3SE.

I am not suggesting that the isolation/reclocking is not the cause of the noise; the evidence we have (in the form of your prototype build in post #1811) suggests that it may be. My reasoning for wanting to try a no-DAC low pass filter is simply that they have no data rate limit and don't use clocking, thus eliminating some of the variables.

Considering your prototype build, it would be interesting to compare the quality of the data signal output from your prototype versus the data signal from your Hermes/Cronus; maybe Cronus is reclocking data that doesn't actually require it and so might just be adding complication (and noise)?

 

[twluke]

Do we agree that, with your Hermes/Cronus and my alternative isolator/reclocker arrangements, we seem to be experiencing similar patterns of the static-like noise with DSD512, regardless of whether the B3SE is using its onboard oscillator or is getting it's clocking from the reclocker (and the noise fades away and is absent most of the time)?

Yes we do. Actually, I found that the similar noise did occur for a tenth of second, very short, at the beginning of track even on my prototype system with either 22/24 or 45/49 clocks, which has been otherwise mostly immune to this noise for long playing.

I am not suggesting that the isolation/reclocking is not the cause of the noise; the evidence we have (in the form of your prototype build in post #1811) suggests that it may be. My reasoning for wanting to try a no-DAC low pass filter is simply that they have no data rate limit and don't use clocking, thus eliminating some of the variables.

Okay, I understood what you are trying to address.

Considering your prototype build, it would be interesting to compare the quality of the data signal output from your prototype versus the data signal from your Hermes/Cronus; maybe Cronus is reclocking data that doesn't actually require it and so might just be adding complication (and noise)?

This comparison is interesting but for now I have no idea how I can set up a protocol necessary for this experiment. What I have now is only an analogue scope...

Anyway, we might expect some comments on this issue from Russ.

Regards,