Measurements of Sonore microRendu Streamer

[amirm]

Oh wait a sec.... the darn thing is over $100!

https://www.amazon.com/Black-Box-Et...&keywords=Self+powered+Ethernet+Opto-Isolator

 

[Blumlein 88]

Come on Amir, we need to get to the bottom of the 8 db increase before going further.

And yes this is a John Swenson design. It pretty much has a Regen board built into it. JS suggests adding another Regen is unnecessary.

 

[amirm]

Come on Amir, we need to get to the bottom of the 8 db increase before going further.

I see the problem already but need to verify tomorrow. I have so many projects piling up that I am having trouble keeping up with them all!

 

[BobShermanEsq]

Come on Amir, we need to get to the bottom of the 8 db increase before going further.

Operator error. Widows media player... Volume control... hmm

 

[Werner]

Come on Amir, we need to get to the bottom of the 8 db increase before going further.

Operator error. One does not assess RMS levels by eyeballing the apparent peaks in the plotted output of an FFT.

If the mRendu really had 8dB of digital gain it would run into massive problems with just about any commercial recording out there.

 

[Thomas savage]

Umm...releasing stuff with obvious errors... You can take the man out of Microsoft but you CANT take Microsoft out of the man

 

[amirm]

OK, I found and resolved two issues that were leading to level differences. Good thing I did not post last night as then I had only caught one of them.

I recently updated to Windows 10 and didn't realize that when you play a file directly and then through "cast" (DLNA) interface, you are dealing with two completely different media players and audio subsystems. The former had two volume controls that were not set to 100%. First was the one inside the "Grove" media player that is the default (rather than WMP). This I caught last night. But second is the system sound, i.e. master volume control.

This also was not set to 100%. The "Cast" interface has its own media player (?) and it had its volume control set to 100% already and not subject to any system volume control changes.

With both volume controls now set to 100%, the level difference in the main tone disappeared:

The two peaks are on top of each other so you can't see that they are the same but I turned one on and off and I could see them being the same.

While I own this mistake, note how easy it is to get thrown off in comparisons without instrumentation. One could easily think that they levels are the same when in reality they were not.

Anyway, while I was in there, I did another measurement to tease out the undulations in the noise floor created by microRendu. I zoomed in both in frequency and amplitude:

We now see pure tone distortions correlated with mains line frequency in US of 60 Hz. The distortions start at 60 Hz and climb up with multiples of it. I have highlighted the third harmonic for example in the above graph.

Note that this is with an aftermarket, ifi Power supply that is designed to better what comes out of USB port of a system. Yet when using the DAC connected to my laptop with tons of stuff running on it, we don't see these. Clear step backward in performance.

If you look at the previous graph above, we see that the are waves travelling at higher frequencies in addition to above. Eyeballing them, they are around 1.5 Khz or so. It is harder to diagnose the source of these.

In summary, we see that objectively we have made the output of the DAC dirtier at lower frequencies. At 60 Hz for example, our noise/distortion floor is a whopping 30 db higher than without microRendu! At higher frequencies, the overall level is lower but the noise is modulated by another frequency which is not a good thing from perceptual point of view.

One of the hallmarks of high-end audio is attention to such detail when cost of the product is secondary to its performance. Yet we have an inversion of that logic here where the product as a system is poorly engineered in the context of high-end audio. The first aim for any high-end tweak should be "do no harm." But harm is done unfortunately.

Thanks to those of you who questioned me on the level differences. Please keep it up! And again, I welcome any comments and corrections from the manufacturer. I will update the original post once everyone reviews this new data.

 

[amirm]

Umm...releasing stuff with obvious errors... You can take the man out of Microsoft but you CANT take Microsoft out of the man

Sad but literally true in this case.

 

[BobShermanEsq]

OK, I found and resolved two issues that were leading to level differences. Good thing I did not post last night as then I had only caught one of them.

I recently updated to Windows 10 and didn't realize that when you play a file directly and then through "cast" (DLNA) interface, you are dealing with two completely different media players and audio subsystems. The former had two volume controls that were not set to 100%. First was the one inside the "Grove" media player that is the default (rather than WMP). This I caught last night. But second is the system sound, i.e. master volume control.

This also was not set to 100%. The "Cast" interface has its own media player (?) and it had its volume control set to 100% already and not subject to any system volume control changes.

With both volume controls now set to 100%, the level difference in the main tone disappeared:
View attachment 2125

The two peaks are on top of each other so you can't see that they are the same but I turned one on and off and I could see them being the same.

While I own this mistake, note how easy it is to get thrown off in comparisons without instrumentation. One could easily think that they levels are the same when in reality they were not.

Anyway, while I was in there, I did another measurement to tease out the undulations in the noise floor created by microRendu. I zoomed in both in frequency and amplitude:

View attachment 2126

We now see pure tone distortions correlated with mains line frequency in US of 60 Hz. The distortions start at 60 Hz and climb up with multiples of it. I have highlighted the third harmonic for example in the above graph.

Note that this is with an aftermarket, ifi Power supply that is designed to better what comes out of USB port of a system. Yet when using the DAC connected to my laptop with tons of stuff running on it, we don't see these. Clear step backward in performance.

If you look at the previous graph above, we see that the are waves travelling at higher frequencies in addition to above. Eyeballing them, they are around 1.5 Khz or so. It is harder to diagnose the source of these.

In summary, we see that objectively we have made the output of the DAC dirtier at lower frequencies. At 60 Hz for example, our noise/distortion floor is a whopping 30 db higher than without microRendu! At higher frequencies, the overall level is lower but the noise is modulated by another frequency which is not a good thing from perceptual point of view.

One of the hallmarks of high-end audio is attention to such detail when cost of the product is secondary to its performance. Yet we have an inversion of that logic here where the product as a system is poorly engineered in the context of high-end audio. The first aim for any high-end tweak should be "do no harm." But harm is done unfortunately.

Thanks to those of you who questioned me on the level differences. Please keep it up! And again, I welcome any comments and corrections from the manufacturer. I will update the original post once everyone reviews this new data.

Laptop is running on a battery, off the grid. You are running the MR on the grid with an entry level supply. Not an accurate comparison. What else is on your AC line that can introduce artifacts (junk) ? Your computer power supply, etc?

 

[amirm]

Laptop is running on a battery, off the grid.

No, the laptop is plugged in. Even if it were not, it would be running off internal switchmode power supply which is not "audiophile" grade or anything like it.

 

[amirm]

Investigating further, I connected the microRendu to my lab linear power supply. Here are the results:

As before, green is the ifi Dac connected direct to my laptop via USB. Red is the microRendu with ifi power supply.

The graph in yellow is with my bench power supply powering the microRendu. It nicely cleaned up the supply modulations using the iFi power supply.

I will do more testing on the iFi in different articles but for now, it clearly is not as clean as my linear lab power supply. I don't know how they say that it is much cleaner than such supplies.

Second, microRendu is highly sensitive to power supply noise. This should not be the case. It should have its own post regulator that cleans up stream supply noise. Given the tiny amount of space they have, likely they could not include proper filtering in there. Worse yet, it is passing through such noise to USB output that connects to the DAC. Whatever they say they have done to improve USB connection, is not in play here.

 

[BobShermanEsq]

No, the laptop is plugged in. Even if it were not, it would be running off internal switchmode power supply which is not "audiophile" grade or anything like it.

When a laptop is plugged in it still run on the battery, the AC will still be isolated. Nothing to do with audiophile anything, just if you see line junk you are not getting an accurate comparison.

 

[BobShermanEsq]

Second, microRendu is highly sensitive to power supply noise. This should not be the case. It should have its own post regulator that cleans up stream supply noise. Given the tiny amount of space they have, likely they could not include proper filtering in there. Worse yet, it is passing through such noise to USB output that connects to the DAC. Whatever they say they have done to improve USB connection, is not in play here.

All the ancillary comments does not come from an accurate point of reference. "it should have" " "could not include proper filtering " "Worse yet" "is not in play here" Are all basically meaningless comment that do not actually apply. They are your opinions which are not necessarily relevant.

 

[BobShermanEsq]

So there's the potential for micro-revenue's lower noise floor as long as there's a proper proper PSU?

At least they do not charge for paperclips.

 

[amirm]

So there's the potential for micro-revenue's lower noise floor as long as there's a proper proper PSU?

Edit: answer is actually no. Please see this post: http://www.audiosciencereview.com/f...ore-microrendu-streamer.577/page-3#post-16766

 

[amirm]

When a laptop is plugged in it still run on the battery

Nope. Typing this after I removed the battery. It has been this way on every laptop I have owned. Go ahead and try it with yours. Keep it powered on and take out the battery and report if it dies

 

[Thomas savage]

Nope. Typing this after I removed the battery. It has been this way on every laptop I have owned. Go ahead and try it with yours. Keep it powered on and take out the battery and report if it dies

Nice one boss

 

[BobShermanEsq]

Nope. Typing this after I removed the battery. It has been this way on every laptop I have owned. Go ahead and try it with yours. Keep it powered on and take out the battery and report if it dies

Some will run, some do not. Did you run the tests with or without the battery? BTW: Thought you were using your "music server" not a laptop?

 

[amirm]

All the ancillary comments does not come from an accurate point of reference. "it should have" " "could not include proper filtering " "Worse yet" "is not in play here" Are all basically meaningless comment that do not actually apply. They are your opinions which are not necessarily relevant.

The data and analysis add all the relevance needed. Lab power supply fixed distortions/noise that the included "audiophile" power supply did not. Clearly then that demonstrates that microRendu is passing through its power supply noise to external DAC. I explained how it could be fixed which every analog designer would agree with. If you are not one, I appreciate the incredulity.

 

[amirm]

Some will run, some do not.

Some? You were pretty sure about all of them being that way just a minute ago. Tell us which laptop you have and do the test and report back. And which ones you have tested which don't work without battery.

Did you run the tests with or without the battery?

Why you ask?

BTW: Thought you were using your "music server" not a laptop?

Perils of not reading the posts before jumping in. From OP

My test setup was typical of the past. My HP laptop playing a J-test signal connected either directly through a USB cable to ifi DAC or going to microRendu and then the same ifi DAC.

If you thought I was using a music server, why did you comment on use of battery in my laptop?