Review and Measurements of Sabaj Da3 Dac compared to Dragonfly Black

[flipflop]

Where are you seeing the 0.7 db difference?

Bottom right corner.
Ch1: 90.401 dB
Ch2: 91.106 dB

 

[palamudin]

He is talking about THD graph @amirm

If i might chime in with a "reasonable" deduction. Its a dual DAC layout. Each channel is processed by one dac. That provides some benefits and some pitfalls as no 2 chips will perform with exact same result. Edit: Bonus thing is that xmos is multicore, if by programming separate cores are feeding particular dacs that could also be where fault might be found. Tho im just shooting in the dark.
"SINAD is a new measure: it shows what amount of the signal is distortion and noise free. We see about 90 dB of usable dynamic range that is free of noise and distortion "
Considering thats 0.7 dB of headroom difference per channel id say its negligible. If @amirm is not bored he could do a test with the "usb" fixer to see if data packet noise would be leveled out (or other benefits).

 

[palamudin]

Ultimately we would need a teardown and analysis of someone competent to peek deeper down the development pipeline of signal path. Possibly ripping out programmable data on chips.

 

[garbulky]

Where are you seeing the 0.7 db difference? Here is the relevant part of the dashboard:

View attachment 13230

On top the two waveforms are on top of each other. On the bottom, the levels show just a 0.003 volt difference. That is .01 dB difference if my math is correct.

Out of curiousity Amir, how did you do that math?

 

[jaynyc]

were both devices tested using the same USB port?

 

[flipflop]

"SINAD is a new measure: it shows what amount of the signal is distortion and noise free. We see about 90 dB of usable dynamic range that is free of noise and distortion "

Sounds like the linearity test, but here the Sabaj Da3 measured 108 dB

 

[amirm]

Bottom right corner.
Ch1: 90.401 dB
Ch2: 91.106 dB

Those are distortion metrics, not levels. They are the signal relative to the sum of noise and distortions. Channels rarely match completely when it comes to distortion. So a bit of variation there is fine.

 

[amirm]

were both devices tested using the same USB port?

They were side by side on my laptop. You can see that in the picture actually.

 

[amirm]

Sounds like the linearity test, but here the Sabaj Da3 measured 108 dB

Linearity tests use filtering so some of the distortion is taken away there unlike the dashboard. Also, the test frequency is different.

 

[amirm]

Out of curiousity Amir, how did you do that math?

It is 20 * log10(V2/V1). Substitute 1.980 for V2 and 1.983 for V1 (the channel outputs respectively) and you get -0.013 dB.

You can also use online calculators. The best one is here: http://www.sengpielaudio.com/calculator-gainloss.htm

 

[amirm]

Ultimately we would need a teardown and analysis of someone competent to peek deeper down the development pipeline of signal path. Possibly ripping out programmable data on chips.

I am not sure this one can be opened as the back is full of silicone. If it can be opened, I will do a teardown.

 

[Timbo2]

I am not sure this one can be opened as the back is full of silicone. If it can be opened, I will do a teardown.

In each corner of the back there are screws. It should be possible to cut just the corners of the silicone and gain access to the screws.

I actually have pictures of the board, but they aren’t my photos so I don’t feel comfortable uploading them. The solder job is really rather poor. I’m pleased it doesn’t seem to impact the device. Perhaps it got better in later production.

 

[amirm]

Good deal. I will do a teardown then.

 

[Timbo2]

Good deal. I will do a teardown then.

Somebody posted lower resolution versions of the board in the middle of this head-fi thread.

 

[garbulky]

I appreciate the revised wording in the linearity test. Here you present two values so a bit more balanced imo. Also if you were to tell me a few years ago that USB powered dacs could have a noise floor of less than 120 and 20 bits of linearity, you could have just knocked me right over. I still don't know how that's possible! Were you using the power from a regular laptop or a specialized USB power supply?

 

[amirm]

These are from laptop, powered by its AC mains. At the same time that it is powering those two DACs on one side, it is powering a Topping DX7 that I use for music listening while I work. And all the traffic to and from the Audio Precision analyzer over its USB port. The new AP software is hugely CPU intensive too, causing my laptop case to significantly rise in temp. All in all, this is creating tons and tons of activity in my computer that well designed DACs nicely ignore.

And yes, I am with you on passage of time. I bought my DAC for $7,000 or something like that nearly 20 years ago because I saw linearity tests in Stereophile that said it was so good. To see a $100 product do the same is remarkable.

That said, let me be clear that linearity is a pure test of levels. It actually uses filtering to eliminate much of everything else including distortion. So by itself doesn't mean the DAC is producing that much excellent performance. It is done so with respect to that one tone only, not the full audio bandwidth.

 

[Elix]

Such a feast of reviews! Thanks, Amir!!!

 

[amirm]

My pleasure.

 

[Raimei Templar]

Hello, Amirm, thank you for this review!

I have been eyeing this DAC for some time and wondered if you or someone else here could answer a quick question.

Is it possible to put this unit into a line out mode, meaning the AMP portion of this combo is bypassed and it its just the DAC signal.

If so this seems like it would make a fantastic desktop DAC as it could be connected to a better amp.

I also wonder if you can put the balanced output into line out mode, if so one could get a balanced 2.5mm to RCA cable and a balanced DAC on the cheap.

 

[maxxevv]

Hello, Amirm, thank you for this review!

I have been eyeing this DAC for some time and wondered if you or someone else here could answer a quick question.

Is it possible to put this unit into a line out mode, meaning the AMP portion of this combo is bypassed and it its just the DAC signal.

If so this seems like it would make a fantastic desktop DAC as it could be connected to a better amp.

I also wonder if you can put the balanced output into line out mode, if so one could get a balanced 2.5mm to RCA cable and a balanced DAC on the cheap.

No idea about the former but for the latter, you'll have to ask someone to customise /modify one for you. I have looked for the reverse as in a 2.5mm TRRS socket to either dual XLR or dual RCA, they don't exist. There are 3.5mm jacks to dual RCA, which can be modified to use a 2.5mm jack.