Bricasti M1 > M5 > M12: Network Player - Member REVIEW - w/OEM Response from Brian Zolner President Bricasti

[barrows]

12 minutes ago, One and a half said:

For example, the Innuos server doesn't quite gel with the MC-3+USB, the Aurender N100 I think it was, didn't work with the Berkerley Alpha USB, let's not talk about DSD512 just yet either. There are DACs that don't work well when connected to an Ethernet to USB out, the process is not that simple. Rarities? Something to obtain a guranteee from the vendor.

That is a USB issue, not an Ethernet issue!

[firedog]

3 hours ago, barrows said:

I am well aware that Ethernet is not limited in sample rate.  This is exactly my concern.  I work with Sonore and our Ethernet products can serve up 768 PCM and DSD 512.  My point was to ask why the Bricasti interface is not able to do higher sample rates.  I am a bit confused though, you are suggesting that the M5 is more capable than its specs show:

 

http://www.bricasti.com/en/consumer/m5.php#m5_specs2

 

That would be a good thing if true, I hope that it is.

 

Additionally, as you seem to have some technical information at your disposal, when you say that Bricasti uses an ARM board, do you mean to suggest that they purchase a commercial ARM processor board complete from a typical computer parts vendor, or that they have actually developed their board and just use an ARM processor on there?

 

I just bought an M5 and have been corresponding with Brian about it. He confirmed that present SW/firmware maxes out at DSD 128, so I’m not sure where the reports above are coming from. Maybe it works for rates above 128 in some setups and not others, so it isn’t an official spec.  It is SW/firmware upgrade able, so maybe that is a future planned upgrade, I have no idea. I will write him and ask. 

 

My understanding is that they developed their own motherboard and buy the ARM processor. 

[barrows]

1 minute ago, firedog said:

 

I just bought an M5 and have been corresponding with Brian about it. He confirmed that present SW/firmware maxes out at DSD 128, so I’m not sure where the reports above are coming from. Maybe it works for rates above 128 in some setups and not others, so it isn’t an official spec.  It is SW/firmware upgrade able, so maybe that is a future planned upgrade, I have no idea. I will write him and ask. 

 

My understanding is that they developed their own motherboard and buy the ARM processor. 

Yes, the picture published in the review noted above has an internal pic (thanks) it clearly shows that the board is developed by Bricasti.  It also shows details of the power supply, but some details are to hard to make out.

Weird that there appears to be some controversy on sample rates available.  It makes (some) sense that the SPDIF/AES output might be limited, but not USB.  Certainly DSD 256 is pretty common these days, both for download purchase and oversampling in software.  And 352.8 is fairly available as well.

[firedog]

7 minutes ago, barrows said:

Yes, the picture published in the review noted above has an internal pic (thanks) it clearly shows that the board is developed by Bricasti.  It also shows details of the power supply, but some details are to hard to make out.

Weird that there appears to be some controversy on sample rates available.  It makes (some) sense that the SPDIF/AES output might be limited, but not USB.  Certainly DSD 256 is pretty common these days, both for download purchase and oversampling in software.  And 352.8 is fairly available as well.

PS is supposed to be the same one as in the M1 DAC, I think pictures of that are also available. 

[One and a half]

2 hours ago, matthias said:

 

I would not stick to USB. Read the review here:

http://www.the-ear.net/review-hardware/innuos-zenith-se-mkii-std-network-audio-server

Jason Kennedy prefers the Ethernet outputs of the SE. Then you do not need devices like the Mutec.

 

Matt

I measured the differences in potentials from a small sample network of two switches and a PC. Measuring from the frames of the RJ45 very simply with a DMM on low ohms mode.

 

Not surprisingly none of anything resembled anything close to 0 and in one case 12V ac something. If any device has a potential difference to another stray currents will flow and keeps propagating through the signal cables this is normal.

 

So the Ethernet cable is picking up noise currents like USB, RS485, RCA analog. Using stp cables the potential differences were 0.

 

I wont subscribe to say Ethernet will be the future,but part of it will be with all the legacy interfaces. Ethernet needs as much care for signal routing like anything else.

[barrows]

19 minutes ago, One and a half said:

I measured the differences in potentials from a small sample network of two switches and a PC. Measuring from the frames of the RJ45 very simply with a DMM on low ohms mode.

 

Not surprisingly none of anything resembled anything close to 0 and in one case 12V ac something. If any device has a potential difference to another stray currents will flow and keeps propagating through the signal cables this is normal.

 

So the Ethernet cable is picking up noise currents like USB, RS485, RCA analog. Using stp cables the potential differences were 0.

 

I wont subscribe to say Ethernet will be the future,but part of it will be with all the legacy interfaces. Ethernet needs as much care for signal routing like anything else.

Hmmm...  Ethernet is differential signaling, and with differential signaling, potential differences matter not.  Now perhaps, if one makes the mistake of using say CAT7 cables with metal shelled plugs connected to the shield, you could induce some problems into the receiving end...  I always recommend customers to not use Ethernet cables with grounded shields, as this does allow for potential problems.

The real advantage of Ethernet is two fold, every interface is isolated, and it is a data transfer protocol with error correction and no critical timing (in an audio sense).  So all critical timing is accomplished after the signal is received, hopefully by a well designed (by an audio company, not a commercial computer company) Ethernet receiver.  (Referring here to an Ethernet DAC).

Additionally, it is not all that difficult to implement an optical interface for Ethernet, which would render electrical noise transmission issues irrelevant. 

[One and a half]

1 hour ago, barrows said:

 

Additionally, it is not all that difficult to implement an optical interface for Ethernet, which would render electrical noise transmission issues irrelevant. 

And combat the noise from the SMPS in the FOC units. 

[fsmithjack]

5 hours ago, barrows said:

So the Ethernet interface is the same across the board in Bricasti products.

Yes it is the same. Great conversation you guys have going! Thanks I enjoy reading it, 

[barrows]

40 minutes ago, One and a half said:

And combat the noise from the SMPS in the FOC units. 

I am not sure what you are suggesting here?  What I am talking about is a DAC (or Renderer) with an optical Ethernet interface.  Not adding an optical section with adapters.  So no worries about noise at the DAC.

[fsmithjack]

23 minutes ago, One and a half said:

And combat the noise from the SMPS in the FOC units. 

Yes I tried this route and got some great benefits here as well but there are the trade off's of course as well. Myself I tried to remove all the SMPS everywhere and keep the rest as simple and as clean as possible. Not saying its better but active approaches have their own effects I guess. 

[matthias]

@One and a half

Empirical Audio at WBF states that Wifi is for SQ the same or better than wired Ethernet. 

There is a nice article from Chris as well:

 

Matt

[One and a half]

3 hours ago, barrows said:

 

Additionally, it is not all that difficult to implement an optical interface for Ethernet, which would render electrical noise transmission issues irrelevant. 

Sorry I didn't respond correctly last post. If there is an optical interface from the source to a DAC or DDC that has the same connector, then yes, the problem of noise is redundant. The longevity though of the optocouplers is another consideration. Do you know of Optocouplers or laser devices/receivers that have a lifetime to match the cables, at least 25 years?

[One and a half]

1 hour ago, matthias said:

@One and a half

Empirical Audio at WBF states that Wifi is for SQ the same or better than wired Ethernet. 

 

From the numerous hints offered by Uptone, it would appear that they are developing a box that blocks wired common mode noise/high impedance leakage currents for Ethernet networks. This device would ideally be placed ahead, in the topic here, the M1 or M5 Ethernet input. I'm not sure if they are offering a buffer for data, but there's maybe no need to. Now this does make some sense and provides a solution to rubbish on the network.

 

@barrows Differential signalling is only as good as the transmitter and receiver properties to reject high impedance leakage currents. I don't buy the Ethernet interface is galvanically isolated to such an extent, that the transformers can block differential noise completely, since regular transformers pass this info without any resistance. Unless of course they are tuned to a specific frequency to let data pass and block everything else.

I know the data gets through, but the noise keeps going until it's heard.

[matthias]

Guys,

I think we should stay on topic.

The "ethernet noise issue", if it exists, has been discussed at lenghts here:

 

Otherwise someone can start a new thread about "ethernet noise".

Thanks

 

Matt

[firedog]

Just got word back from Brian about the M5 playing DSD 256 and 512:

"Yes it should run at those speeds off the USB, we have not fully tested it so its not on the specs."

[barrows]

10 hours ago, One and a half said:

From the numerous hints offered by Uptone, it would appear that they are developing a box that blocks wired common mode noise/high impedance leakage currents for Ethernet networks. This device would ideally be placed ahead, in the topic here, the M1 or M5 Ethernet input. I'm not sure if they are offering a buffer for data, but there's maybe no need to. Now this does make some sense and provides a solution to rubbish on the network.

 

@barrows Differential signalling is only as good as the transmitter and receiver properties to reject high impedance leakage currents. I don't buy the Ethernet interface is galvanically isolated to such an extent, that the transformers can block differential noise completely, since regular transformers pass this info without any resistance. Unless of course they are tuned to a specific frequency to let data pass and block everything else.

I know the data gets through, but the noise keeps going until it's heard.

Well, nothing ever blocks noise completely, there will always be some capacitive coupling.  Just as noise will never be completely eliminated from a system.  But, we can reduce noise to much, much lower levels easily by the transformers at each end of the Ethernet cables.

As for the much touted high impedance leakage currents, this can easily be eliminated (as mentioned by JS).  Additionally, it has never been demonstrated that high impedance leakage currents even cause any problems, it is just a suspicion.  I just power my router with a super good custom supply, and ground the output, then I ground the Mac Mini which acts as my server with Audirvana+, no leakage currents to be paranoid about.  Did these changes make any sonic difference though...  I cannot be sure.  If there was any improvement, it was very, very, very subtle.  In my experience most of the Ethernet tweaks being reported around CA as making for "massive" improvements are ridiculous, very subtle minute improvements, maybe!

But, those improvements will depend on the Renderer in use as well.  Since I have the Signature Rendu SE, and none of those reporting "massive" improvements do, perhaps that is the difference.

Anyway, optical input on the Renderer/DAC for Ethernet would make all this talk moot, although I am sure the tweakers would find another "massive" improvement...

 

I suspect, if you want to put an Ethernet interface in a DAC, the more important thing is going to be the internal isolation and shielding.  An Ethernet receiver requires quite a bit more processing power than an USB receiver.  I have done Ethernet DACs DIYhere, with a very good (swiss made) Ethernet receiver, and a dedicated power supply (own tranny secondary) but not isolated.  My isolated USB interfaces outperformed the Ethernet direct approach.  I think doing a really good internal (in the DAC) Ethernet receiver would require a separate shielded sub-enclosure (Linn does this) and complete galvanic isolation on the I2S lines with re-clocking flip flops right before the conversion stage: this is not hard to do, but it is more than what most DAC manufacturers are doing now.

[One and a half]

7 hours ago, barrows said:

Well, nothing ever blocks noise completely, there will always be some capacitive coupling.  Just as noise will never be completely eliminated from a system.  But, we can reduce noise to much, much lower levels easily by the transformers at each end of the Ethernet cables.

As for the much touted high impedance leakage currents, this can easily be eliminated (as mentioned by JS).  Additionally, it has never been demonstrated that high impedance leakage currents even cause any problems, it is just a suspicion.  I just power my router with a super good custom supply, and ground the output, then I ground the Mac Mini which acts as my server with Audirvana+, no leakage currents to be paranoid about.  Did these changes make any sonic difference though...  I cannot be sure.  If there was any improvement, it was very, very, very subtle.  In my experience most of the Ethernet tweaks being reported around CA as making for "massive" improvements are ridiculous, very subtle minute improvements, maybe!

But, those improvements will depend on the Renderer in use as well.  Since I have the Signature Rendu SE, and none of those reporting "massive" improvements do, perhaps that is the difference.

Anyway, optical input on the Renderer/DAC for Ethernet would make all this talk moot, although I am sure the tweakers would find another "massive" improvement...

 

I suspect, if you want to put an Ethernet interface in a DAC, the more important thing is going to be the internal isolation and shielding.  An Ethernet receiver requires quite a bit more processing power than an USB receiver.  I have done Ethernet DACs DIYhere, with a very good (swiss made) Ethernet receiver, and a dedicated power supply (own tranny secondary) but not isolated.  My isolated USB interfaces outperformed the Ethernet direct approach.  I think doing a really good internal (in the DAC) Ethernet receiver would require a separate shielded sub-enclosure (Linn does this) and complete galvanic isolation on the I2S lines with re-clocking flip flops right before the conversion stage: this is not hard to do, but it is more than what most DAC manufacturers are doing now.

Some great information here, and agreement. For the choice, at the moment, unless distance isn't a problem, then use USB, it's the devil we know, and if distance is the problem, then use Ethernet. On either approach, measures need to be taken to avoid propagation of noise, including the power supply and AC network.

[One and a half]

On 3/5/2018 at 9:46 AM, barrows said:

That is a USB issue, not an Ethernet issue!

Where's the source coming from...Ethernet. USB non working is not the source of the problem. What's the point in having a network renderer, when it can't interface to the 'real' world. that was the point I was making.

[barrows]

12 minutes ago, One and a half said:

USB non working is not the source of the problem

Yes it is, the USB interface has no idea where the data is coming from, whether it is Ethernet or a laptop.  I am well aware of the issues to which you speak of, and the problem is in the USB code.  At Sonore we keep a lot of records on this, and there are some USB receivers, in some DACs, which have issues with linux/ALSA based USB transmission.  The streamers you mention invariably use linux/ALSA, the problem is the USB receivers not being compatible, their USB code developers have clearly only been concerned with OS X and Window compatibility.  Anyone who wants to know more about this needs to contact Jesus at Sonore, as he is the most knowledgeable person on this.

BTW, most of these same USB receivers will have the same problems when used with a linux based (which is a lot of them) music server, without any Ethernet connection involved at all.

 

Sorry for the thread drift, I will leave it at that, although I think discussing optimum Ethernet performance might still be relevant here?  I'll let you decide @fsmithjack

[ChrisG]

Thanks for all of this great information! Now I know one of the reasons my Linn Klimax DSM sounds so good! 

[fsmithjack]

UPDATE:

 

I am extremely pleased with the performance and upgraded sound quality of my Bricasti M1SE's built in Network Player. Such a great upgrade!

 

At the advice of another super helpful CA member who is a fine fellow: ismewor

 

ismewor has helped me improve my system a great deal.  The members here on the CA are more friendly and helpful than any other forum i have been on.

 

Here is the switch I just purchased and that is being delivered today. The Ethernet cable plugs right onto my DAC now from my switch so the hope is that this is going to be a nice upgrade. I will let everyone know my thoughts on it. 

 

 

OCXO AUDIOPHILE SWITCH

 

LINEAR POWER SUPPLY
PROPERITY DESIGN CIRCUIT 
Our Engineer study the switch and design the linear circuit that best provide ultra-low noise power to feed the switch and oscillator module. 

DUAL RAIL PARALLEL OUTPUT DESIGN 
Not usually seen in regular linear supply, it requires additional cost and design in mind. However, this sure to provide constant feed as needed in sudden voltage drop or unstable power environment 

HARD CRAFTED RECTIFIERS DIODE AND SUPER-FAST DIODE 
Which work hand in hand with AC-DC conversion system to provide stable working environment. 

LINEAR POWER SUPPLY - Specifications
- AC 120V, 60Hz 
- 9V Ouput 
- Oyaide DC connectors
- Furutech IEC 

OCXO CIRCUIT 
ULTRA HIGH PRECISION OCXO 
The unit core features an (Oven Controlled Crystal Oscillator) OCXO. This OCXO was designed for high audio quality in mind to provide extreme precision frequency (+-10ppb) and reduces phase noise. 

24K GOLD CIRCUIT TRACE 
The 24K circuit trace on our PCB is needed to match our power supply which able to provide extra stability for the oscillator module. 

ULTRA STABLE REGULATOR DESIGN 
With our new design, we are able to achieve the measure we needed for the oscillator. 

ZERO GROUND CIRCUITRY 
It continuously drives the ground signal and greatly reduces noise (jitter) caused by unstable volts. 

OCXO CIRCUIT - Specifications 
- Military Spec OCXO 
- 2oz Copper PCB +/- 10ppb 
- Ultralow Jitter <1ps RMS 
- 12V DC input 2.1mm x 5.5mm

 

[firedog]

What happened to asking Chris to combine threads?  You just posted the same thing 3 times. 

[dxc]

Hi: 

 

I have a original Bricasti M1 with firmware 1.35. Currently playing from a Mac Mini with Audirvana. Am considering upgrading the Mac Mini for a Aurender N100H.

But now I'm interested in this Network player option in the M1. 

So I guess I could still use the Mac Mini(non modded) as a server. 

How is you set up with the Bricasti M1 with network option?

Would it be better to upgrade my M1 compared to upgrading my Mac Mini to an Aurender N100H?

 

Thanks

[fsmithjack]

On 3/9/2018 at 1:56 PM, firedog said:

What happened to asking Chris to combine threads?  You just posted the same thing 3 times. 

 

I did merge them. I was talking with multiple people on threads about the upgraded switch I was buying so wanted to share with them as well. 

[fsmithjack]

On 3/5/2018 at 7:37 AM, firedog said:

Just got word back from Brian about the M5 playing DSD 256 and 512:

"Yes it should run at those speeds off the USB, we have not fully tested it so its not on the specs."

I forward this info to Brian Zolner and here is his response:

 

"We spec it at DSD128 USB out, but really it should run to DSD256, just that we have not tested it to verify and spec it,  it appears that the reviewer in Hong Kong found it to work, so guess it does, if so we can update the specs on it, I can contact them and find out more.   352.8 PCM when used as DoP is same as DSD128, not quad which would be 2 x the rate in PCM terms or 705.6 PCM.  AES and SPDIF are limited to DSD 64 or equal to 176.4 PCM ( 192k max ).

 

Let me know if you guys have any other questions or comments for Brian.

Thanks 

."