Measurements and Review of SMSL iDEA and AudioQuest Dragonfly Black DACs

[sejarzo]

Output actually looks pretty clean overall. Despite the higher noise floor its still mostly under -120db. So it shouldn't be audible?

Yeah, but to me, it clearly sounds inferior to my original version ODAC.

 

[Jimster480]

Yeah, but to me, it clearly sounds inferior to my original version ODAC.

That definitely could be the case, but at $40 for a DAC+Amp I would like to see something really able to beat it at this pricepoint!

 

[amirm]

I will review the Fiio K1 separately and will find others to test against it.

 

[Jimster480]

I will review the Fiio K1 separately and will find others to test against it.

There are a couple other items in your list to compare against the FiiO K1, like the SMSL X4, Ingerik Direct PCM2704

 

[amirm]

I do indeed!

 

[jaytrinitron]

Ok, I dug into this some. I think I see the issue.

I made my measurements at maximum level. Here is what it looks like there with and without load:

View attachment 10963

As we see it is clipping with 32 ohm load. Archmiago used 20 ohm which would make this even worse.

I then dialed down to where there is no clipping and now the loaded output is very similar to unloaded:

View attachment 10965

Now the output impedance measures 0.4 ohms.

I was careful in looking at the waveform when I first started but then I stopped doing it. My bad. I will correct the original post. This also explains the difference in some of my previous measurements if the device was clipping with 33 ohm load.

Thanks for bringing this to my attention.

Hello Amir and friends. I have some questions about this:

What other reviews have measurements taken where the device was under 33 ohm load? Anyone know?

And would this significantly change the results/conclusions of past DACs?

 

[amirm]

What other reviews have measurements taken where the device was under 33 ohm load? Anyone know?

And would this significantly change the results/conclusions of past DACs?

I only make such measurements for devices with headphone amplifiers. I do not perform them for DAC-only products. When I do test headphone amps, I test with 33 ohm. You should be able to see that in all of my reviews. Here is the latest example:

Probably in the oldest reviews I did not test this way as I did not have my dummy load built then.

 

[jaytrinitron]

I only make such measurements for devices with headphone amplifiers. I do not perform them for DAC-only products. When I do test headphone amps, I test with 33 ohm. You should be able to see that in all of my reviews. Here is the latest example:

View attachment 12767

Probably in the oldest reviews I did not test this way as I did not have my dummy load built then.

Ok thank you, I am new to this whole audio thing and not sure about the various terms and such. So if a headphone amp you tested before clipped at 33 ohm, it wasn't actually clipping? Would that significantly affect your review results/conclusions?

 

[jaytrinitron]

I only make such measurements for devices with headphone amplifiers. I do not perform them for DAC-only products. When I do test headphone amps, I test with 33 ohm. You should be able to see that in all of my reviews. Here is the latest example:

View attachment 12767

Probably in the oldest reviews I did not test this way as I did not have my dummy load built then.

Also, I see that you didn't post equivalent # of bits/resolution/linearity (I think these are all the same thing?) for these devices, but on the whole Archimago seems to get much higher numbers than the ones in yours reviews for these metrics. Is that because your test equipment is more sensitive/resolving, or maybe you are calculating the metric differently? Archimago says this in his Audioquest Dragonfly Black review:

The Dragonfly's resolution is around 16.5-bits, slightly better than CD resolution. Similarly, Stereophile's measurements described the Black as "the less expensive Dragonfly offers 17 bits' worth of resolution" which I believe is generous as a resolution estimate.

Which means Stereophile's measurements, at least for this device, more or less agree with his. So why do you and him (Archimago) tend to be so far apart in linearity? He has found many cheap devices with 16 bits or more linearity, and you have found none, as far as I can figure. Why is this?

 

[amirm]

So if a headphone amp you tested before clipped at 33 ohm, it wasn't actually clipping? Would that significantly affect your review results/conclusions?

I am sorry but I don't understand your question. In my testing, I drive the headphone amps to clipping (the knee in the curve). They will all clip at some point. If you listen below that level, the clipping does not affect you. That level is headphone/preference related. Sometimes I indicate on the graph the listening levels I used for the headphones I have.

 

[amirm]

Also, I see that you didn't post equivalent # of bits/resolution/linearity (I think these are all the same thing?) for these devices, but on the whole Archimago seems to get much higher numbers than the ones in yours reviews for these metrics. Is that because your test equipment is more sensitive/resolving, or maybe you are calculating the metric differently? Archimago says this in his Audioquest Dragonfly Black review:

It is a different test. He seems to be using the noise level vs output for his computation. I am not sure if he knows to compensate for the effect of FFT Gain. He does not explain.

What I measure is different in that I test how well a DAC tracks input digital values relative to output. When that deviation is > 0.1 dB, I stop and use the value just before that. The 0.1 db is my threshold (there is no industry standard here). If I use a different threshold, I can get higher or lower resolution. So best to use my linearity numbers within the scope of reviews I do here.

 

[jaytrinitron]

I am sorry but I don't understand your question. In my testing, I drive the headphone amps to clipping (the knee in the curve). They will all clip at some point. If you listen below that level, the clipping does not affect you. That level is headphone/preference related. Sometimes I indicate on the graph the listening levels I used for the headphones I have.

No worries, I think I misunderstood your previous post, the one here:

Ok, I dug into this some. I think I see the issue.

I made my measurements at maximum level. Here is what it looks like there with and without load:

View attachment 10963

As we see it is clipping with 32 ohm load. Archmiago used 20 ohm which would make this even worse.

I then dialed down to where there is no clipping and now the loaded output is very similar to unloaded:

View attachment 10965

Now the output impedance measures 0.4 ohms.

I was careful in looking at the waveform when I first started but then I stopped doing it. My bad. I will correct the original post. This also explains the difference in some of my previous measurements if the device was clipping with 33 ohm load.

Thanks for bringing this to my attention.

So I'm still confused. What exactly was done wrong, and how does it affect the previous measurements?

 

[dc655321]

No worries, I think I misunderstood your previous post, the one here:



So I'm still confused. What exactly was done wrong, and how does it affect the previous measurements?

Current-limiting behavior is causing the clipping. If @amirm was seeing clipping at 32 Ohms for a given output voltage, then Archimago's use of a 20 Ohm resistor would show clipping at an even lower ("worse") voltage because the lower resistance circuit would draw more current. Or, try to draw more...

 

[amirm]

So I'm still confused. What exactly was done wrong, and how does it affect the previous measurements?

Sorry, now I remember. This only impacts measurements of output impedance of headphone amplifiers. The data from this post on is correct for all amplifiers reviewed. For those prior to that post, it is possible that their output impedance is lower than one I reported depending on whether there was any clipping.

Note that this does NOT change any other measurement. It only impacts the one number I report for output impedance. The rest of the review holds.

 

[baxo]

is this DAC same with Sabaj DA2?

 

[Jimster480]

is this DAC same with Sabaj DA2?

Do you mean the chip?

 

[baxo]

Do you mean the chip?

yes and everything else since some say that they are complete clone

 

[duo8]

yes and everything else since some say that they are complete clone

Yeah they're pretty much identical except for the shell.

 

[bravomail]

I had very bad experience with SMSL Idea/Sabaj Da2 which measured very good in Amir tests. I think it simply does not have enough power to drive my AKG K7xx, and it sounds harsh. When paired with Schiit Magni 2 and at default volume setting it cannot provide enough volume. Every time you unplug SMSL Idea - it will reset its volume to default level, hence I tried it at that unfortunate setting. I have Hifime Sabre 9018 dongle USB DAC, with similar form factor to SMSL Idea, which ironically measured bad/so-so in Amir's testing. But, it drives AKG K7xx without issues, it doesn't have useless volume buttons, it pairs with Schiit Magni 2 just fine (you probably won't need it).
All above simply tells me that methodology of testing and the parameters tested need to be reviewed. What is being presented - does not provide the whole picture.

 

[amirm]

I had very bad experience with SMSL Idea/Sabaj Da2 which measured very good in Amir tests. I think it simply does not have enough power to drive my AKG K7xx, and it sounds harsh. When paired with Schiit Magni 2 and at default volume setting it cannot provide enough volume. Every time you unplug SMSL Idea - it will reset its volume to default level, hence I tried it at that unfortunate setting. I have Hifime Sabre 9018 dongle USB DAC, with similar form factor to SMSL Idea, which ironically measured bad/so-so in Amir's testing. But, it drives AKG K7xx without issues, it doesn't have useless volume buttons, it pairs with Schiit Magni 2 just fine (you probably won't need it).
All above simply tells me that methodology of testing and the parameters tested need to be reviewed. What is being presented - does not provide the whole picture.

These reviews are with my older Audio Precision analyzer which made testing of USB devices very difficult and time consuming. As a result, the tests were not comprehensive enough. Specifically, power versus distortion was not measured. This has been a standard in my newer reviews since I bought the new Audio Precision analyzer in spring which can directly control USB devices. There, I show the available power and contrast between devices. I also heavily test for available power in my listening tests. Most people now report that they are stressing the amp less than I do so I think the data is conservative in nature.