DAC burn-in/break-in? My Gustard X16 measurements

[solderdude]

Here is the chart. My meter initially not stable, then then after initial second or so, pretty stable at ~0.277V. Whole thing took ~20+ seconds.

View attachment 115664

This means you are sweeping at -17dB. To get a better S/N ratio you can go up 14dB in output level.
Obviously the overall level is tiny bit (0.2dB) temperature dependent. Nothing to worry about.

 

[Pdxwayne]

This means you are sweeping at -17dB. To get a better S/N ratio you can go up 14dB in output level.
Obviously the overall level is tiny bit (0.02dB) temperature dependent. Nothing to worry about.

I see. No need to worry about Solo flashing red then? Initially I was using minus 3db in REW and the Scarlett Solo was flashing red.

Also, I am trying to check near a level that I typically listening to music, not at DAC's max voltage. I am thinking that should more closely reflecting what I could hear during listening tests.

The db for the REW captures today is more like +0.2db louder than those I captured yesterday, not just 0.02db. No concern?

Thanks again!

 

[solderdude]

The X16 is linear over the entire range. The lower your signal is the lower the S/N ratio becomes. That's all. A DAC does not sound different when peak levels are near -3dB, -10dB or -20dB.

What you need to know is if other X16's have a similar level difference at the same output level over the same temperature range.

That's why it is important to know the absolute levels and temperatures.
When other X16's show similar behavior then that would be considered normal behavior.
When other X16's show much more steady levels then something could be wrong in a power supply or reference circuit.
The fact that both channels do the same thing points towards it being 'normal' behavior or a common 'error'.

You just need someone else to do a similar test or buy a second one and check that one.

 

[Pdxwayne]

The X16 is linear over the entire range. The lower your signal is the lower the S/N ratio becomes. That's all. A DAC does not sound different when peak levels are near -3dB, -10dB or -20dB.

What you need to know is if other X16's have a similar level difference at the same output level over the same temperature range.

That's why it is important to know the absolute levels and temperatures.
When other X16's show similar behavior then that would be considered normal behavior.
When other X16's show much more steady levels then something could be wrong in a power supply or reference circuit.
The fact that both channels do the same thing points towards it being 'normal' behavior or a common 'error'.

You just need someone else to do a similar test or buy a second one and check that one.

In https://www.audiosciencereview.com/...my-gustard-x16-measurements.20679/post-690674, the dB difference between channels is much lower (0.02 vs 0.2db), but the graph shape is different....but at least someone is checking.

@lizhuoyin, if you can, would you please help check with same dB level in REW sweep and continue to check the output of x16 at cold fresh boot, 1 day continue powered on, and 2 days continuous powered on and see you observe any shift in power level from both channels?

It seems mine can shift more than 0.2db in less than one day.

Thanks!

 

[Jimbob54]

In https://www.audiosciencereview.com/...my-gustard-x16-measurements.20679/post-690674, the dB difference between channels is much lower (0.02 vs 0.2db), but the graph shape is different....but at least someone is checking.

@lizhuoyin, if you can, would you please help check with same dB level in REW sweep and continue to check the output of x16 at cold fresh boot, 1 day continue powered on, and 2 days continuous powered on and see you observe any shift in power level from both channels?

It seems mine can shift more than 0.2db in less than one day.

Thanks!

Without knowing your temperatures, or theirs, that doesnt help. Temp, not time, seems to be the issue here.

 

[lizhuoyin]

In https://www.audiosciencereview.com/...my-gustard-x16-measurements.20679/post-690674, the dB difference between channels is much lower (0.02 vs 0.2db), but the graph shape is different....but at least someone is checking.

@lizhuoyin, if you can, would you please help check with same dB level in REW sweep and continue to check the output of x16 at cold fresh boot, 1 day continue powered on, and 2 days continuous powered on and see you observe any shift in power level from both channels?

It seems mine can shift more than 0.2db in less than one day.

Thanks!

I'll try my best. I have less confidence in my Zoom H6 though.

 

[Pdxwayne]

Without knowing your temperatures, or theirs, that doesnt help. Temp, not time, seems to be the issue here.

Temp of outside shell of DAC?

For me, when brought in from garage, 61F.

Room temp have been stable around low 70s.

4 hours later from powered on, low 80F, as I recalled.

Now, almost two days later, ~84 to 85F, depending on the area of the top. Don't seems to change much from 4 hours powered on.

I have this thermometer

 

[Pdxwayne]

I'll try my best. I have less confidence in my Zoom H6 though.

Thanks for trying!

If you have temperature gun, try checking temperature too.

Appreciate your help!

 

[Pdxwayne]

Measured at ~2 days powered on. DAC temperature ~85F.

It shifted up by 0.3db vs powered on 2 hours. But channel mismatch still about the same at ~0.2db.

Top two lines measured at 2 days. Bottom two at 2 hours.

Nothing else stands out.

I am starting to think that the DAC did not actually burn-in/break-in. Maybe over the weeks, I just gradually adjusted my head a little to the right to compensate for the 0.2db difference. That slight head adjustment create the illusion that the DAC became more balanced.

I will still continue to observe for a bit longer before I move the x16 downstairs and do a sighted quick RCA switch test vs E30 Maybe it will be pretty easy to tell which is which by keeping my head stationary and check for slight sound stage shift.

 

[lizhuoyin]

X16 0dB. REW -20dB. H6 gain @3.

OK. I did several tests. The conclusions are:
1. The channel diff remains 0.02dB;
2. First test shows 0.02 dB shift for the same channel after cold start all devices with 30 minutes cool off time. After 5 minutes, the shift reduces to 0.01 dB. I don't know if the shift was caused by X16 or H6.

After 40 minutes, the shift is gone.

Room temp at around 65 F. No tool to check device temp but about the same at top when feel by hands.
At this point, I don't think it matters.

The following graph shows the progression. The top line is @10 hours. The second line is @20 minutes after cold restart; The third line is @5 minutes and the last line is @0 minute.

 

[lizhuoyin]

Measured at ~2 days powered on. DAC temperature ~85F.

It shifted up by 0.3db vs powered on 2 hours. But channel mismatch still about the same at ~0.2db.

Top two lines measured at 2 days. Bottom two at 2 hours.

Nothing else stands out.

I am starting to think that the DAC did not actually burn-in/break-in. Maybe over the weeks, I just gradually adjusted my head a little to the right to compensate for the 0.2db difference. That slight head adjustment create the illusion that the DAC became more balanced.

I will still continue to observe for a bit longer before I move the x16 downstairs and do a sighted quick RCA switch test vs E30 Maybe it will be pretty easy to tell which is which by keeping my head stationary and check for slight sound stage shift.

View attachment 115732

I seriously doubt that the shift was caused by Solo and diff was caused by X16.

 

[Pdxwayne]

I seriously doubt that the shift was caused by Solo and diff was caused by X16.

You meant Solo is fine and my x16 is the one that caused the shift?

Note that I also measured E30 and no channel issue. Thus I think Solo is OK....

 

[lizhuoyin]

You meant Solo is fine and my x16 is the one that caused the shift?

My bad. I suspect Solo caused the shift.

 

[Pdxwayne]

My bad. I suspect Solo caused the shift.

I see. I will bring back E30 to my room and also check for any shift. That way I can be sure that it is the Solo that cause all the shifts so far.

Thanks for checking!

 

[Pdxwayne]

My bad. I suspect Solo caused the shift.

Thank you very much again for checking!

Indeed, it was the Solo that shifted. When comparing E30 vs X16 again just now, the difference between them remains the same. It is just that all of them shifted up by about 0.2db.

First day:

Today:

 

[Blumlein 88]

At this point I'm not sure why you are continuing. I can show you a simple way to check relative clock drift if you'd like to include that in your warm up measurements.

 

[Pdxwayne]

At this point I'm not sure why you are continuing. I can show you a simple way to check relative clock drift if you'd like to include that in your warm up measurements.

Please provide the way. Thanks!

Yes, at this point, I am thinking that I will do a sighted listening quick switch test soon. I will see if I can sense the slight soundstage shift or not.

Again, I thank you and many others here for being patience with me during my discovery process.

 

[Blumlein 88]

Play and record a quarter sample rate tone. 12 khz for 48 khz sampling for instance. You can record the E30 into the Solo and get the clock speed difference. You can then record the Gustard into the Solo and get the clock speed difference. And also know the relative difference between the two DACs.

In Audacity the top track is what you get with a 12 khz tone at 48 khz if clocks are in perfect synch. The lower track is what you get if they are out of synch. You can see I have selected a portion which is the trough of that wave form undulation. That is how many samples it takes to go from being in phase, cycling thru the timing difference out of phase and then being in phase again. That lets you figure out the % difference or ppm. In this example I artificially sped one up 10 ppm. I selected visually, and I had selected 100,011 samples. So even this simple graphical selection was pretty close in accuracy. So 1 part per 100,000 is also 10 parts per million. 10 ppm. You can zoom in for more accuracy selecting. Also as the speed difference gets lower the length to cycle between in and out of phase is longer which aids in accuracy of your measurement. Done with care you should easily get 1 ppm accuracy and you can do better just visually. There are ways to do even better still, but rarely would you need that kind of precision.

 

[Blumlein 88]

Almost forgot to tell you how to determine if the device being recorded is faster than or slower than the ADC clock.

Select just inside the left hand trough as shown below.

Then zoom in until you see the individual sample points.

If the tops are sloping downward to the right the source is faster than the ADC clock.

If tops are sloping downward to the left then the source is slower than the ADC clock.

 

[Pdxwayne]

Almost forgot to tell you how to determine if the device being recorded is faster than or slower than the ADC clock.

Select just inside the left hand trough as shown below.
View attachment 115761

Then zoom in until you see the individual sample points.

If the tops are sloping downward to the right the source is faster than the ADC clock.
View attachment 115762


If tops are sloping downward to the left then the source is slower than the ADC clock.
View attachment 115763

Thank you very much!

Regarding the top track which is perfect, you created that using audacity first? Then playback that track to the DAC, am I correct?