yggdrasil technical measurements

[atomicbob]

  @atomicbob, will you please repeat these measurements as the Yggy warms up?

As I can fit it into my schedule, I intend to repeat in 7 to 14 days and incorporate the other measurement requests as well.
 
I know this will be requested but I will not be measuring higher sample rates for some time. As @purrin mentioned in one of his posts on the preview of yggdrasil, I don't think any of us have really heard what 44x16 pcm can deliver until now.

 

[mikoss]

Great job @atomicbob - a great contribution and we appreciate this info!

 

[7ryder]

BTW, they're starting to discuss your measurements over at computeraudiophile  
 
yggy

 

[preproman]

 

[coli]

ringing and overshoot.
 
Edit: I would expect the overshoot to make things sound more exciting. Probably compensate for tube amplifier the most.
 
As a point of comparison, NOS DACs don't ring.

 

[johnjen]

  Careful John. The results are spectacular, but the numbers displayed are usually for a single frequency. One really needs to look at the graphs. The crosstalk reduces to approx 100 dB between 100 and 200 Hz for Right to Left and approx 95 dB for Left to Right in the same frequency range. Still exceptional but refer to the graphs for exact values by frequency.

Yes!
 
Still, to even reach those numbers, let alone the aggrigate of all of them, paints a suitably revealing glimpse into what this dac is capable.
 
Giant killer comes to mind…
Well along with DigitalVinyl ®™©  

 
JJ  :thumb

 

[jcx]

the pre/post ringing is Gibbs Phenomena - a mathematical consequence of wanting steep filter slope to have 20 kHz flat response and attenuate 24 kHz image by 96 dB for 16/44 CD digital audio
 
the time domain overshoot/ringing can be moved around with "apodizing filters" to avoid the pre-ringing but then doubles it up in post ringing: https://www.ayre.com/pdf/Ayre_MP_White_Paper.pdf
 
Mike has ruled out the slower filter roll off that Ayer uses from his comment on edging up his magic "bit perfect" filter's corner frequency closer to Nyquist than industry standard
 
another side of the mathematical box "audiophile" products sometimes push on is to not fully filter the above Nyquist image frequencies - need more extended plots to see if this is the case

 

[evillamer]

Can you do a measurement of pre-warmed up Ygg VS warmed up(6 or 12 or 24 hours) Ygg?

 

[Pidgeon]

I think that a new measurement with Yggdrasil fully warmed up (a week) could be interesting for a comparison with your previous measures.

 

[jcx]

 really doubt anything being measured will show a unambiguous trend - otherwise the cause would be well known to EE who build much more sophisticated critical and scrutinized Industrial Scientific and Medical instrumentation
 
such designers do know about some capacitor dielectric changes with bias and time, some semis and other component do stress relive with mild thermal cycles - but the effects where known aren't big at audio frequencies
 
 
and if you are a guru fanboy shouldn't you be disappointed if he couldn't design out the burn in/warm up if he understands the cause?

 

[firev1]

Actually in the prototype which Purrin handled there may be some differences in 0 hour vs warmup changes that are very measureable and have considerable effect in audio frequencies. I know the Geek Out v1 can be measured for such differences easily.

 

[jcx]

you can build circuits which are 1st order temperature sensitive - and you do need to wait for them to stabilize to operating temp - 1st 10 minutes gets most equipment close to steady state - certainly anything of reasonable construction and desktop size is as warm as it is going to get in an hour - of course that is a delta over the room/air temp and will change as the external temp varies
 
you can see warmup/drift specs for serious measurement equipment - but DC parameters are the most affected - not generally audio frequencies
 
the measurements shown in the 1st post aren't likely to resolve the expectedly small objective effects in a audio DAC with competent SS circuits wihich are built with measures to reduce temp effects
 
very much less so the few hypothetical effects that might justify days of operation to warm up to
 
one is that operating warmer than the local air temp drives off some absorbed moisture - water has e_r of ~80 and can be seen increasing measured dielectric loss in nonpolar plastic dielectrics as good as even polystyrene even @ the <0.1% they can absorb

 

[atomicbob]

At 408 hrs I repeated the measurements in post #1 of this thread. They were all substantially similar to the 3 hr measurements except one. The jitter spectrum, which was already remarkable at 3 hrs, improved significantly. The yggdrasil has achieved a milestone in jitter performance, in my opinion.

 

[skeptic]

At 408 hrs I repeated the measurements in post #1 of this thread. They were all substantially similar to the 3 hr measurements except one. The jitter spectrum, which was already remarkable at 3 hrs, improved significantly. The yggdrasil has achieved a milestone in jitter performance, in my opinion.

 
Thanks for confirming this!  Sounds like the oscillators really are at the heart of why warm up matters - which makes perfect sense given how sensitive they are to thermal changes.

 

[jcx]

except that jitter at sub nanosecond levels hasn't been shown to be audible in controlled listening tests - the audiophile jitter reduction sub industry hasn't published new controlled testing verifying their claims - its all hand waving, anecdotes and FUD
 
published peer reviewed jitter audibility thresholds are from decades old papers that show 100s of nanoseconds for random jitter, 10s of nanoseconds for correlated or tonal jitter
 
jitter so low it is characterized as frequency drift does improve in xtal oscillators with aging - nowhere near audio significant frequencies
 
I expect most correlated jitter is built into the circuit - electrical and mechanical construction, parasitics coupling noise to the clock
 
I don't see how sideband tones can improve with aging - you really have to have lots more measurements to show correlation with on time/aging vs other electrical environment changes - what other equipment is plugged in, loading, line noise even what other sw may be running
 
very common is day vs night - how much else is on, actively polluting the electrical environment with EMI