Review and Measurements of SMSL VMV D1 DAC

[bennetng]

Is there a SMPSU inside? They mention a toroidal transformer, so sounds like probably not?

OK, my bad. I think SMSL is honest about it.

 

[gvl]

Based on my reading on it, a-weighting is something you use for speakers or measuring noise levels for work safety. Why in the world one would use it on a DAC?

 

[bennetng]

But in the same 48kHz J-Test, USB unbalanced has 8k and 16k noise (probably not jitter?), toslink doesn't have these spikes isn't it? Also, notice the toslink graph is normalized to 0dB and the one on first page is at about -7dB.

 

[amirm]

Based on my reading on it, a-weighting is something you use for speakers or measuring noise levels for work safety. Why in the world one would use it on a DAC?

It is a very common practice for all audio gear to be measured with a-weighting. It improves the specs. And it sort of attempts to match our hearing sensitivity. It doesn't do a perfect job and hence the reason I don't show it but the precedence in the industry is quite strong.

Now for design verification, you definitely do not want to use it. That is a different situation than marketing specs for a product.

 

[gvl]

At < -130dB those spikes are hardly a concern.

 

[amirm]

But in the same 48kHz J-Test, USB unbalanced has 8k and 16k noise (probably not jitter?)

The amplitude of the sidebands is different so not jitter.

 

[CaptObvious]

Amirm do you want to do a teardown? I don't mind as long as it still works afterwards.

 

[gvl]

It is a very common practice for all audio gear to be measured with a-weighting. It improves the specs. And it sort of attempts to match our hearing sensitivity. It doesn't do a perfect job and hence the reason I don't show it but the precedence in the industry is quite strong.

Now for design verification, you definitely do not want to use it. That is a different situation than marketing specs for a product.

Interesting. I found it suspicious SMSL folks were resistant to turn the HF filtering off here, sounds like it is a known issue to them.

 

[amirm]

Amirm do you want to do a teardown? I don't mind as long as it still works afterwards.

It will be interesting to do. I will of course buy it if I damage it.

 

[amirm]

Interesting. I found it suspicious SMSL folks were resistant to turn the HF filtering off here, sounds like it is a known issue to them.

I find their demeanor so far to be quite open and nice, showing full screen-shots and such (compare that to what Schiit has done). So I prefer that we don't go there.

 

[gvl]

I find their demeanor so far to be quite open and nice, showing full screen-shots and such (compare that to what Schiit has done). So I prefer that we don't go there.

Fair enough

 

[Krunok]

My sense is that it is internal since I see that changing the sample rate changes the signature.

I think they did not catch it because they were always testing with a-weighting. Hopefully they are seeing it now and they can do the proper diagnosis.

I agree with both. Also, although it is hard to tell to me this seems like a design thing and not related to this particular device.

 

[gvl]

@amirm, can you test one leg of the balanced output? If CMR is done inside the unit running from one leg can be a (cheap) workaround for those w/o balanced downstream gear.

 

[amirm]

@amirm, can you test one leg of the balanced output? If CMR is done inside the unit running from one leg can be a (cheap) workaround for those w/o balanced downstream gear.

Here you go. One is the converted balanced to single ended (SE) and the other, straight unbalanced:

The converted signal ended is clean.

 

[Krunok]

Does it mean the problem is in the SE leg of the board? In that case would the U/I converter be "primary suspect"?

 

[gvl]

Here you go. One is the converted balanced to single ended (SE) and the other, straight unbalanced:

View attachment 15413

The converted signal ended is clean.

Nice, thank you. So seems like if one must use this DAC with SE amps a cheap XLR to SE adapter is the way to go.

 

[Krunok]

Nice, thank you. So seems like if one must use this DAC with SE amps a cheap XLR to SE adapter is the way to go.

Or you simply listen to 44.1 kHz.

 

[gvl]

Does it mean the problem is in the SE leg of the board? In that case would it be V/I converter?

I suspect the SE output is taken directly from the DAC perhaps just buffered, and balanced are actually op-amp summed first and then rebalanced again. We've been told differential DAC chip outputs can contain a lot of common mode junk.

 

[gvl]

Or you simply listen to 44.1 kHz.

The problem exists there too as Amir mentioned, the dashboard had 20kHz bandwidth so just not visible there.

 

[Krunok]

I suspect the SE output is taken directly from the DAC, and balanced are actually op-amp summed first and then rebalanced again. We've been told DAC chip outputs can contain a lot of common mode junk.

It does, but it can be done properly.

The problem exists there too as Amir mentioned, the dashboard had 20kHz bandwidth so just not visible there.

It does when you measure it like that but RBCD files won't have those higher frequency.