UpTone Audio EtherREGEN Listening Impressions

[JohnSwenson]

1 hour ago, Puma Cat said:

Its a matter of semantics, I guess, but I think my meaning was clear; regarding the "smearing" of the digital "fingerprint/whatever" due to clock phase noise.  

 

Apologies to John if I misquoted...

I'm still in the process of writing the paper that goes into great detail about all this. I wanted to include some actual measurements of systems with and without the EtherREGEN, but I'm having some trouble getting measurements made. The super sensitive jitter analyzer I recently acquired cannot be directly connected to the circuits, I have to build an interface circuit and the first one did not work, I've designed a completely new one which should not have the problems of the first one, but the board has not come yet. With the holiday it won't be here until Tuesday, then I have to solder all the parts in etc, that will take a little while, then try and get it to work properly with the test equipment. Its all taking much more time (and money) than I had hoped.

 

So I'm not sure whether I should just finish the paper and post it or wait for real measurements.

 

John S.

[JohnSwenson]

3 hours ago, cat6man said:

I'm in the January build but have 3 configurations under consideration at this time.

 

My current configuration is:

1.  hqplayer under audiolinux running on NUC

2.  NAA mode in opticalRendu (oR) to DAC

3.  NUC-->router/rj45-->opticalModule(oM)==>15m fiber==>opticalRendu(oR)

4.  all units running on LPS-1.2 (oM) or SR-4 (oR)

 

I'll re-configure when I get the etherRegen in January but it is not obvious to me if the oM becomes superfluous or not.

When I get the etherRegen(eR), I see three options:

==> (optical ethernet connection)

-->  (copper ethernet connection)

 

A.  NUC-->router/rj45-->etherRegen(oR)==>15m fiber==>opticalRendu(oR)

B.  NUC-->router/opticalSFP==>short fiber==>opticalModule(oM)-->etherRegen(eR)==>15m fiber==>opticalRendu(oR)

C.  NUC-->router/rj45-->15m copper ethernet-->etherRegen(eR)==>short fiber==>opticalRendu(oR)

 

Note that with the 3 new cases, the etherRegen is running Bin/Aout while the oM is also running Bin/Aout also (which is the reverse of the Ain/Bout configuration with my current etherRegen-less setup).

 

Has anyone tried any 2 or 3 of these options and have any listening impressions to share?

 

 

"B" won't work, the B side port of the EtherREGEN is only 100Mbs, and the RJ45 port of the oM is only gigbit.

 

John S.

[JohnSwenson]

I don't know about the Emo, but I have looked and measured the Baaske extensively. The Baaske is just a transformer, that's it. This gives some extra common mode attenuation, and the bandwidth of the transformer is low, just enough to barely pass an Ethernet signal. This can cut down on some frequencies of differential noise, but it also degrades signal integrity significantly. So putting one between the B side jack of an ER and an endpoint is almost guaranteed to make things worse. On the A side it may may have some effect, but it could be worse or better.

 

My gut feeling is to not use these at all.

 

John S.

[JohnSwenson]

17 minutes ago, Albrecht said:

Hi Alex,

Thanks very much for commenting. And I do apologize, - I may have missed some details in the threads surrounding the topology of the etherRegen. 

 

I suspect that this may have been asked before, - but, - if you use the "B" side as an "input" will the "A" ports be forced to be 100MB or will they/can they be gigabit?

 

Cheers,

When going from B to A, the CONNECTION characteristics of the RJ45 A ports don't change, they are still auto-negotiated to 10-100-1000, BUT since the B side is 100 the overall throughput is limited to 100. So if you connect to an A port at gigabit, the bits in a packet are traveling at gigabit protocol, but there is more time between packets.

 

The A side SFP interface is ALWAYS gigabit only.

 

John S.

[JohnSwenson]

11 hours ago, HumanMedia said:

Thermal issues, or non-issues, aside, do people find that specific voltages sound different/better? From the same supply even?

I seem to remember John mentioning that different voltage/power supply combinations could sound better. Maybe that was a microRendu thread? Or was it the EtherRegen, or both?

 

One poster indicated a feeling that a 7 volt Farad supply had a slight edge over the 12v Farad. Has anyone else noted a audio quality preference for difference voltage inputs?

Higher input voltages produce lower currents which produce smaller voltage drops across cables and connectors, which MIGHT give slightly better performance. The lower current can also significantly affect the power supply output, but exactly what happens is going to be very dependent on the particular supply.

 

So again as with pretty much everything else, no cut and dried universal rule.  If you have the option, try it your self and don't worry.

 

Just remember that say you want to buy a model X 12V supply, but it costs many hundreds of dollars, so you try a model Y 12V supply, if Y sounds better is not a guarantee that the 12V X will also sound better. 

 

John S.

[JohnSwenson]

9 hours ago, Energy said:

On a side note, I want to ask Alex a question..

It's been said that atomic clocks with their long term drift (in ppb) does not matter as much as having lower phase noise, but when you speak of the 10MHz clock drifting in frequency over the years, how much drift is too much to where the frequency causes suffering to the circuit when it can no longer be locked (due to the frequency drifting too far) or at what ±% does it start to potentially affect an oscillator's operation?

The digital circuits used in audio will WORK (properly transport data) over a very wide range of master clock frequencies , somewhere in the several parts per thousand range for sure. Even an uncompensated oscillator such as the 575 will meet that over a very long time. To get outside of that range takes drastic measures like pounding with a hammer, putting it in the flame of a torch etc. Not "normal" usage.

 

ANY OCXO is WAY WAY WAY better than that!

 

John S.

[JohnSwenson]

 

On 6/11/2020 at 3:01 PM, MB said:

Wanna say I have the utmost respect for John Swenson and his design and creativity ability.  I remember him from the squeezebox forum years ago.  Always loved to read his posts.  I had a Touch and John was mentioning he was designing a network player that would be better than the Touch, if I remember correctly.  I suspect that player turned out to be the microRendu, or so I am surmising.

I spent a LOT of time and money on that project. I have several prototypes that actually work. It turned out the hardest part was laying out the back panel, there were 16 jacks on the fairly small back panel!!

 

There were two XLR and two RCA, each pin fed by its own DAC channel, so it was actually a 6 channel DAC. Yep, you could have done a 5.1 system with it. This took some custom drivers to make it work.

 

I even had a connector on the front of the board so you could take off the front panel and screw in a 7" LCD touch screen. I never built that but I did have the connector for it.

 

We had someone ready to build and sell it for a reasonable price, but then the whole project was derailed by legal issues that I cannot discuss. I was so bummed by the whole thing I just dropped out of the forums.Every now and then I pick up the test boards and look at them and hold them in my hands and wonder how it would have gone otherwise. I probably would have spent the rest of my life supporting it!

 

The microRendu was a completely separate design. 

 

John S.

[JohnSwenson]

11 hours ago, sgb said:

It was the combination of the switch with SFP to oM to the uR and HQP that caused dropouts. It disappeared when I bought the oR and the eR.  

HQP uses low level pause frames to implement flow control. The pause frames need a switch to see the pause frames and stop the stream for a period of time. The OM is not a switch so it doesn't know what to do with pause frames so over-runs can happen with HQP.

 

Other protocols deal with flow control differently so using the OM is only an issue with HQP and NAA.

 

The eR is a switch so it deals properly with pause frames so NAA and eR and oR works fine.

[JohnSwenson]

1 hour ago, Mikey26 said:

Any recommendations to mitigate incoming noise from the Coax RF cable that goes into the cable modem? I’ve tried a RF cable isolator and is has helped. My question is are there better RF isolators? Can you use multiples in a chain for further improvement? My goal is to lower the incoming noise as much as possible from the coax RF cable from the cable company.

This isolator you have is NOT an RF noise filter (if it was an RF filter you wouldn't get any signal, since the signal IS RF). These devices are a broadband transformer that lets the RF through, BUT blocks ground noise which can get developed over the miles of cable on the pole or underground.

 

If your cable modem doesn't have one of these builtin (some do) then the massive ground current from the cable can get into your system.

 

There is no "better filter" since what you have already blocks the ground current about as good as it can be done. The only issue is if the transformer is fully passing the RF frequency range used by the cable modem. Most of these were designed for TV and FM radio and may not fully pass the frequencies used by the cable modem. I don't know if the one you have fully does or not. If it works, it probably does!

 

John S.

[JohnSwenson]

17 hours ago, Ricardo007 said:

@JohnSwenson

hello John,

do you think this isolator would work for cable modem (it is from Jensen a famous name in audio for transformers):

https://www.jensen-transformers.com/product/vrd-1ff/

it says has a bandwith of 2MHz to 1300Mhz

 

would it work also for a (magnum dynalab) tuner?

I have not tried that particular one, but according to the specs it should work very well.

 

John S.

[JohnSwenson]

22 hours ago, yyz said:

Well I did see that and I also saw the statement by UpTone that A > B is preferred over B > A. Do you any knowledge on the product to clarify that statement regarding my use case? I was hoping to hear from someone from UpTone on this since they follow these threads.

Hi yyz,

The A side contains a switch chip which drives the 4 RJ45 ports and the SFP cage. The connection across the moat also goes to one of the ports on that switch chip.  The B side just has the one RJ45 port, no switch chip.

 

The isolation techniques are identical going from A to B and B to A. There is one difference, the clock circuitry is on the B side and has to travel through one of the isolation circuits to get to the A side. This isolator is VERY low jitter it adds a little extra jitter to the clock crossing the moat, but not very much. This is the difference in going from B to A rather than A to B, the clock is not quite as good as it is on the B side. This is not a very big difference but it is there.

 

So if you have two DACs that you want equally good connections, connect the B side jack to your network and connect both DACs to the A side. You COULD put the network on the A side and one DAC on the B and the other an the A. The DAC on B side would have a SLIGHTLY better signal than in the first scenario, but the DAC on the A side would be much worse since it does not cross the moat at all.

 

Either way works, if you want really good results from both DACs use the first approach, if one DAC you want the best possible and the other you don't really care that much how good it is, go with the second.

 

John S.

[JohnSwenson]

8 hours ago, PYP said:

a miniature muon generator/accelerator is involved!  Don't tell Alex I told you!  I signed a non-disclosure agreement...

Who told you about muon generators? A company has actually come up with a muon generator using lasers, unfortunately it is still insanely expensive, but hopefully that will change over time. I have spent 40 years designing all kinds of neat things using muons if only we had an inexpensive miniature muon generator. It looks like the that may actually be happening soon! Put that together with carbon nano tubes and look out!

 

John S.

[JohnSwenson]

Since the POWER does not depend on voltage, the CURRENT goes down as the voltage goes up. Getting a supply that can supply all the current needed seems to be far more problematic than getting the higher voltage, so we generally recommend using a higher voltage if you have it available. Not because it is "better" but because you will have a much higher probability your supply will have enough current.

 

John S.

[JohnSwenson]

8 hours ago, Mihaylov said:

This is good for little heat dissipation, but bad for the level of ripple rejection according to the graph from the datasheet.

 

Most of the LT3045/LT3042s have 0.8 volts across them. The core voltages (1.1V) have 0.9V. This is a very careful engineering tradeoff, this was as high as I could go and live within the thermal constraints. I have to prevent the chips from getting too close to their upper die temperature (you never run right at it, you need a significant buffer there) AND I needed to cut down on the total power drawn, we had some hard limits as to total feed power to the sections of the board. The values chosen just meet the thermal constraints while giving quite good performance, I would not call what we achieved "bad", maybe not quite the best possible, but any real design is never about getting one parameter the best it is theoretically possible to get and letting others fall by the wayside. It is always a complex set of tradeoffs optimizing the best overall device within the constraints.

 

We could run the regulators with 3V across each one, but then it would be a much larger board and cost at least 3 times as much and draw 2.5 times the power. Because the board is bigger chips are further apart from each other which could actually wind up with worse electrical performance. I made the decision that this was NOT a good tradeoff.

 

John S.

[JohnSwenson]

4 hours ago, tgb said:

 

😀

but... if you could provide a eye-diagram of the Etherregen ouput, it could help... to understand... "better SQ" 😀

all the best !

The problem here is that eye diagrams are designed for binary signals, Ethernet is not binary (two voltage levels), 100Mb has 3 voltage levels and Gigabit has 5 voltage. What would an eye diagram for that even look like? The test equipment certainly isn't designed to do it, there is no standard on what a 3 or 5 level eye is supposed to include. Gigabit is about10 times worse because each pair has both directions AT THE SAME TIME on the same wires. This results in receive and transmit (5 voltages) overlaying on top of the other, this can beover 10 voltage layers, and how do you distinguish receive from transmit to include in the eye.

 

100Mb might be possible to do (separate RX and TX pairs and"only" three voltages to deal with but it is not going to look like any eye diagram you have ever seen. I'm not even sure eye diagram software on my scope can even deal with it. The three voltage levels are going to confuse the PLL clock recovery system. who knows what that is going to produce. And as is the case with every test instrument made it is going to ignore all the low frequency components which is where most of what we are trying to deal with exists. So even if you could actually make one, it probably wouldn't show much of any different, it's not that the ER doesn't do anything but that all the test equipment out there designed to measure such things deliberately completely ignores the aspects we are trying fix.

 

John S.

[JohnSwenson]

11 hours ago, Markus8 said:

@JohnSwenson Excellent summary why things are far more tricky than they appear to most people out there. 
Understandably there is a desire to measure/explain but I guess that will still take some time until this can be done properly. 

I just had an intriguing change in this. My scope had been having some functionality missing recently and I have been working with the manufacturer on it. They finally concluded it was a conflict between two of the the software options installed. By disabling one I could get the other to work. After we got that working they sent me an option key for a brand new serial analysis option that is way better than what was already on the scope. The manual says it supports PAM3 and PAM4 (but not PAM5) so it should be able to work with 100Mb Ethernet! I don't have it running yet, I have to download the software THEN install the key. But maybe just maybe it might be able to do an eye diagram of 100Mb Ethernet. It will be some time before I can get it up and running.

 

John S.

[JohnSwenson]

On 3/18/2024 at 6:27 PM, R1200CL said:

Sounds like these 10GB eye patterns then only make a selection of a certain voltage level. Correct ?

Why are they only defined in a standard for 10GB, and not for 1GB ?

 

But then fiber shouldn’t need voltage levels and eye patterns, but fiber to my knowledge has eye patterns. So maybe the measurement is taken somewhere else on the PCB ?
 

I assume in the end these voltage levels use same clock anyway. 

Do you have a link to the 10G eye diagram you are talking about? It is most likely the electrical side of an optical link, that IS binary and has a standard eye diagram. You can certainly do the same thing gigabit fiber. The problem is with the copper twisted pair interface. The actual cable pairs have pretty low bandwidth over long distances so they use multiple voltage levels to increase the data rate then use 4 pairs to get it up to gigabit. THAT is the interface that is going to be hard to do an eye diagram for.

 

John S.

[JohnSwenson]

On 3/21/2024 at 8:02 PM, tgb said:

hi,

could you please elaborate ?...

because the way you wrote it, it looks like you have a cheapo scope (70MHz BW / <1kEur)... and it sounds weird...

.. and especialy the "the aspects we are trying fix" : what do you mean here ? (1)

 

Back to subject :

- an 100Mb ethernet signal, from the ER (100MB for the ER is a smart choice)

- it is a easy signal to trigger using any scope, cause signal freq is 125M, thus any cheap 00M BW scope will do the job (3kEur max)

- anyone can tweak a switch or a FMC-bridge and measure the signal using a fairly cheap oscilloscope, and check which tweak is OK or KO vs SQ. So what ?

Ethernet is cheapo to measure vs USB ! lol ! (although I can measure USB 2.0 signal thanks to a "vintage" scope I bought at 3kEur only)

 

Strange exchange... 

an answer to (1) would help....

 

Hi tgb,

this is NOT about bandwidth of a scope, it is about how serial data streams are sent and received and how measuring devices have evolved to analyze these. It will take me a good part of a day to write this up properly and I don't have a good part of a day free right now. I think I'm going to do this as a white paper and put it on Alex's site, it is WAY to much to put on a forum message. This seems to have become a BIG DEAL so I want to cover this fully and put it in a specific place so I don't have to keep on talking about it over and over again.

 

John S.