Review and Measurements of Totaldac d1-six DAC

[KMN]

....
For all these variables one cannot test.
....
You cannot test all situations under all circumstances.

Yes absolutely, complete testing would be presumably prohibitively expensive and time consuming for the most part even though we'd all love to see that data.

Though I also wonder if the imperfect data we do have would allow us any valid conclusions regarding emc immunity.For example would a ranking of the apparent results within the scope of Amirs standard test suite make sense? ie possibly the analysis of whether there is presence or absence, or size differences of these grid frequency disturbances , among the various DUT's, within the fft scans which may be useful data to us somehow?

For example is it a valid thing to say that the X-Sabre proved more immune, at least within this test, than the TotalDAC and so therefore may be considered a better bet EMI immunity wise, out in the wild? or is this also not doing a proper apples to apples comparison?

 

[solderdude]

You cannot say anything about EMC immunity from Amirs current tests.
I wasn't there when tests were done so have no idea of how things were connected with which wires and how they were routed and what was powered in what way.

If one wants to know this completely other test equipment and specific tests are needed.
Even then one only knows if and how much immunity and emission there is for certain circumstances.
Using another USB cable or interlink or wiring it another way can totally change everything.

I say just connect everything the way one wants to and if there is no hum, other noises, drop outs or stutters then all is fine.
When plagued with 'nasties' all I can say is good luck with trouble shooting.
'Reduced sound quality' isn't one of the typical ways in which this manifests itself.

 

[KMN]

Yes absolutely. The only variable must be the DUT., or invalid test! ....but what if the only variable being the DUT could be confirmed? ie amir could easily attempt to control for this in his testing if he thought it was important. Valid test, then? Not a complete emi characterization test per any kind of established standard per se but still valid comparison of relative difference, no?

 

[amirm]

Yes absolutely. The only variable must be the DUT., or invalid test!

I can't control the temp during testing so invalid test.
I can't control the outside weather so invalid test.
I can't control the incoming AC voltage so invalid test.
I can't control what song is being played by local stations so invalid test.
I can't control the layout of the cables in identical ways (due to varying equipment size, shape and connector location) so invalid test.
I can't control the PC activity at the time of the test so invalid test.
I can't control other sources of sound in the house so vibration can get into device differently so invalid test.

Question: how do you control all of these factors when you listen? Which one of you measures ground loops and considers it a factor in playback when no hum is present?

We haven't lost common sense in all of this, have we?

 

[amirm]

No, Amir. You gave us suggestion that a leak in a DAC transformer could be a cause instead of ground loops. Not proven as you didn't investigate it any further. I do agree to the point that it could, but in addition to the ground loops. A critical moment for a proof that you had ground loops in the system was in the moment when you noticed both 50Hz and 60Hz components in FFT. For any electrical or electronic engineer with experience (not neccessary a degree) it is a proof and an indication that ground loops must be removed before starting any measurements.

Why? What other measurements do ground loops impact?

Remember, I did NOT ding the device on ground loops, or mains leakage. Problems were elsewhere backed by exact nature of how this DAC operates.

Also, keep in mind that I did 100 times more than what any other tester would do. I varied the input AC frequency and with it, isolated what is coming from the source, versus what exists elsewhere (analyzer, etc.). Beyond that, there was no reason to investigate more because there was not a problem of note there. That is, tiny mains leakage occurs in testing frequency, and possibly far more in your use of the gear. It is not material to far larger issues in the design of this DAC.

Your arguments seem to be aimed at misleading someone not technical with buzz words like "ground loops." Don't try it on us. We know the topic and it has no impact on the rest of the tests or conclusions thereof. After all, the designer of the box would hang his hat on that argument if he could and he did not. What makes you more knowledgeable than him, if not us?

 

[amirm]

It could be as simple as an additional Y-capacitor that loops with other Y-capacitors, just giving an example of a very common cause. Ideally it should be only one Y-capacitor in the testing gear (including a DUT).

This DAC uses a transformer so doesn't have a Y capacitor. That is typically used in switching power supplies.

 

[amirm]

Let's keep in mind that all of your systems are susceptible to such vagaries as far as mains leakage, ground loops, etc. Indeed if you use an RCA connection, you are guaranteed to have ground loops. You can plug them all into the same outlet and still have ground loops! As long as you don't hear it though, it is not a material concern.

I can certainly do measurements at an EMC lab with EMC certified gear but how would that ever apply to real world situations you all live in? That is putting aside the huge cost of such endeavor.

Let's remember that our hearing sensitivity is quite poor in low frequencies. This is why, despite almost constant presence of mains leakage/ground loops, we don't hear it in our system:

60 Hz is just past 0.05 in that graph. As you see, the threshold of hearing is 50 dB higher than baseline at 3 kHz where we are most sensitive. Put inversely, if you listen at 110 dB, then mains noise needs to exceed -60 dB for it to be audible. At -110 dB or whatever we see here, it simply is not an audible problem.

 

[sajunky]

Also, keep in mind that I did 100 times more than what any other tester would do. I varied the input AC frequency and with it, isolated what is coming from the source, versus what exists elsewhere (analyzer, etc.).

This is exactly what I wrote a moment ago. You did 100 tests more than many others, but not many with AC power conveter, did you? You changed frequency to 60Hz, but changing frequency could not eliminate ground loops, there were still there.

 

[solderdude]

The only variable must be the DUT., or invalid test! ....but what if the only variable being the DUT could be confirmed? ie amir could easily attempt to control for this in his testing if he thought it was important. Valid test, then? Not a complete emi characterization test per any kind of established standard per se but still valid comparison of relative difference, no?

Why would one want to test audio equipment to EMC standards ?
Merely to see if a DUT is susceptible to certain signals ?
It would say little to nothing about normal usage or in the thousands of different combinations/situations in which it is used.

In the rare cases a groundloop exists that leads to audible hum, noise or weird sounds there are always solutions for it.
I am quite certain most high end equipment is not tested for EMC.

It is too expensive and there won't be huge financial claims if a hifi amp has some emission or has poor immunity.
This is different for medical, industrial or public devices or mass produced devices like phones etc.

Nah, for audio we just need to check the audio performance and if one finds safety hazards it should be mentioned as well.
In case of the TotalCRAP a little mains hum is the last thing we should worry about in the measurement results.
They don't become invalid because some very low level hum shows up in a plot.
Its all the other aspects in which it performs poor (granted except linearity).
One should be worried about the actual performance and the price tag instead.

 

[KMN]

I can't control the temp during testing so invalid test.
I can't control the outside weather so invalid test.
I can't control the incoming AC voltage so invalid test.
I can't control what song is being played by local stations so invalid test.
I can't control the layout of the cables in identical ways (due to varying equipment size, shape and connector location) so invalid test.
I can't control the PC activity at the time of the test so invalid test.
I can't control other sources of sound in the house so vibration can get into device differently so invalid test.

Question: how do you control all of these factors when you listen? Which one of you measures ground loops and considers it a factor in playback when no hum is present?

We haven't lost common sense in all of this, have we?

I hope not it was just some questions I had and only wanted to hear what others were thinking about it. I did not mean to imply anything whatsoever. .Apologies if I had missed the relevant prior discussions and the fact I probably haven't done appropriate due diligence in trying to understand your testing environment. It is definitely useful to understand the things you feel you can't control for well that may be significant.

 

[graz_lag]

... You can do it right next time when you measure other European equipment. I am an optimistic, looking forward that opportunities are not wasted.

Sure, just be patient, the next TotalDAC's item will get to @amirm soon ...

 

[amirm]

This is exactly what I wrote a moment ago. You did 100 tests more than many others, but not many with AC power conveter, did you? You changed frequency to 60Hz, but changing frequency could not eliminate ground loops, there were still there.

How do you know they are ground loops and even if so, why do you care? And how do you know a transformer eliminates them?

 

[solderdude]

Oh my... it would appear the Vidar measurement must have a groundloopⓇ somewhere.

My eye spots some mains leakage... must be a Y capacitor or other groundloop or Schiit doesn't know how to measure or.... (let your imagination run wild)

 

[Thomas savage]

There's definitely Something loopy afoot.

 

[BDWoody]

Oh my... it would appear the Vidar measurement must have a groundloopⓇ somewhere.

My eye spots some mains leakage... must be a Y capacitor or other groundloop or Schiit doesn't know how to measure or.... (let your imagination run wild)

I think it's back to plankton... Dammit...not again.

 

[solderdude]

The plankton is hiding in the grass at the bottom of the plot.

 

[BDWoody]

So the plot thickens...

With plankton.

 

[sajunky]

How do you know they are ground loops and even if so, why do you care? And how do you know a transformer eliminates them?

Well, if you don't know, ask around, but I suggest not on the forum, as there is something weird there.

As for the isolation transformer, they are designed slightly different to the transformers installed inside the equipment, where a main goal is to contain radiation inside the core. Maybe start from here:

It doesn't need to be a special one, like for medical purpose. If you buy from Radioshark, find one where primary and secondary windings have separate spools, or at least there is a shield between windings.

 

[KMN]

Why would one want to test audio equipment to EMC standards ?

I do appreciate the safety ratings and whatnot, but personally I would find plots of the scans , of my gear, from the sorts of equipment generally required to perform that sort of testing, fascinating if that data were made available to me. Not a requirement or anything but I'd certainly look it over. Why not?

I believe, as was alluded to earlier, things such as layout and and component selection can have a significant and direct impact on equipment noise immunity/susceptibility when done more or less optimally. Since I've also seen that in the extremes I've put my audio systems into, sometimes perhaps ignorantly, the problems/benefits of these design choices can be audible in good or bad ways presumably depending on how these design choices interact with the reality of existing with my il conceived configurations of equipment. And since problems indicated in the audio band can point to possible problems in other bands as well. possibly even showing up in well gathered emi scans to those who can recognize them. the scans would seem valuable. to me.

 

[KMN]

Question: how do you control all of these factors when you listen? Which one of you measures ground loops and considers it a factor in playback when no hum is present?

Try to make due and be as clever as possible with my far less impressive gear than you get to use, ie what I have to work with. I don't have the type of equipment I could make meaningful direct analysis of ground loop parameters likely to yield any valuable insight to it's expected effect on the systems audio performance but can certainly do indirect analysis on those problems that are gross enough in my system to cause disturbances in the signal large enough to be detectable by my measurement equipment or ears and analyzing those effects. I feel like that approach has allowed me to make steps forward through time, even if not necessarily as proven to the level of objective certainty most here would be seeking, or to the level those used to using the much nicer measurement gear such as AP's might be used to seeing, but close enough to satisfy me. and better than not trying at all in my opinion.

I absolutely consider evidence of potential ground loops , as indicated by some commonly associated hum signature or otherwise, as a warning sign to be on the lookout for noise problems at other frequencies as well. Loops, whether on grounds nodes or other signal/s within the system, are inductive, and therefore mutually inductive to other current loops in the universe and thus a potential ingress path for noise energy. This inductance's efficiency at coupling noise energy onto the audio signal, and at which frequencies it is hottest, depends on the physics of these loops layout with respect to the aggressor noise loops supplying the noise energy. This is all something a designer has direct control over while designing layout and can choose to optimize or de-optimize for performance reasons. IE better designs reduce likelihood of induced noise by design. Perhaps irrationally, or not, I'd rather purchase equipment, all else equal, designed by a designer who understands how minimizing loop area of any mesh likely to be a sensitive noise injection point and isolating that loop via separation from noisy aggressor loops can be an overall design performance win, even if it may not show up in a direct lab analysis of system performance in all sorts of testing...., just like a good designer will understand that minimizing dipole length, separating susceptible nodes by distance from high dv/dt aggressor nodes is the approach necessary for designing for minimizing capacitive noise coupling. Who wouldn't want a better analog design by someone who truly understands the physics well enough to minimize inductive and capactive coupling noise energy ingress to the audio circuit by designing a better layout? I would love to be able to draw valid conclusions about these p[performance aspects of pieces of equipment I might be considering purchasing.

I was only just hoping you'd say you could control the variables well enough we could reliably draw conclusions from your data about these effects. Too bad we can't. That would have been nice.