Quiz: Which DAC Is This?

[Graph Feppar]

Did amir reveal that its the SMSL dac?! I dont understand what "Ah you got the clue! Well done." means,is it "yes" or "you are close"?

 

[amirm]

He got it. It is the SMSL Sanskrit 10th edition. The hint was "colorful" as that is one of the few colorful DACs out there. I did not expect anyone to be able to tell from measurements. That would have been an impossible task.

 

[invaderzim]

Given the time of year am I the only one that read the headline "which dac is this" to the tune of "what child is this"?

 

[Ron Texas]

He got it. It is the SMSL Sanskrit 10th edition. The hint was "colorful" as that is one of the few colorful DACs out there. I did not expect anyone to be able to tell from measurements. That would have been an impossible task.

Not so good in the measurements, but that's why things get measured.

 

[GGroch]

This surprises me. I understand SMSL has made some recent errors that resulted in necessary board revisions...but with previous highly regarded models like the M8 and even the earlier Sanskrit 6th I thought they knew what they were doing with DACs. So a significant relapse.

 

[Tatteredmidnight]

This surprises me. I understand SMSL has made some recent errors that resulted in necessary board revisions...but with previous highly regarded models like the M8 and even the earlier Sanskrit 6th I thought they knew what they were doing with DACs. So a significant relapse.

I love me some SMSL (I own an M8, Sanskrit 6th, and SU-8), but every single one of their devices I own has some flaw or issue beyond the norm.

My Sanskrit 6th's USB input died (I even replaced the USB connector to no effect).

My M8's screen burned in so badly that I actually got a refund (and got to keep the device). I also had to mod a USB extension by cutting the power and ground connectors due to digital noise (clearly audible at max volume and high gain on a JDS Atom).

My SU-8 is rev 1.1, you already know the issues here.

The value/performance/feature balance is right on, the QC, not so much.

 

[bigx5murf]

My only SMSL product is a sd192 optical dac. It's hooked up to my CCA, and lost in the rats nest of wires behind my gear. It's been chugging along for years without needing a single reset. Wish I could say the same about the CCA, which seems to need a reset every few months.

 

[AnalogSteph]

Now for the interesting question: Why?

This looks like a massive jitter fail to me. The 1 kHz tone has substantial phase noise sidebands. I can only see either of two things happening:
1. USB clock regeneration is performing badly. This has always been tricky business. If so, there should be a difference between, say, 44.1 kHz and 48 kHz.
2. Bad grounding that lets a ground loop in the measurement setup affect the PLL VCO. These spikes look suspiciously like 60 Hz mains noise.

Either way, a design measuring like this should have gone back to the drawing board during the prototyping stage.

 

[bigx5murf]

Would like to see how it measures on other inputs.

 

[restorer-john]

I did not expect anyone to be able to tell from measurements. That would have been an impossible task.

Perhaps the FFT spectra of each and every DAC could be characterized like a fingerprint as long as the measurement criteria/system setup was identical each time. It is, after all a bunch of numbers that could be stored in a database and compared rapidly with others in order to locate a 'match' or at least a list of possible suspects.

With a large database, patterns would emerge (like the ESS hump) and identification could be performed using pattern recognition. Detective Amir could ID a commercial D/A converter product from just the FFT.

 

[amirm]

The review is out with more diagnostic information: https://www.audiosciencereview.com/...-measurements-of-smsl-sanskrit-10th-dac.5702/