• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Review and Measurements of Sabaj Da3 Dac compared to Dragonfly Black

AlbyD

Member
Joined
Aug 15, 2018
Messages
8
Likes
5
Hi Amirm,

thanks for your measurements. As an owner of a Sabaj DA3 myself, I read your review thoroughly however not being an expert when it comes to electronics I am having some troubles understanding one of the tests and its practical significance.

I am talking about the THD+N vs OUTPUT test :

index.php



As stated, the DA3 starts clipping at only 1 volt. To me it sounds surprisingly low to be considered a good dac am I wrong? (Your conclusions do not seem to give this too much importance to this result and I would like to understand why)
I would expect any decent dac to be able to put out at least 2v without clipping or rising distortion too much, no?

Another thing I do not understand is: if DA3 clips @ 1volt, why does the first "dashboard" view show it putting out 1.9 volts with a nice, non-clipped signal and distortion at about 0.003%? Should it not reflect clipping and distortion as seen in the output test?

It is very possible, as a noob, that I misunderstood the output test. What practical consequences does this clipping@1v have in real life usage?
What volume levels (given its 0-38 digital levels, 19 being the default) are safe to use into a 32 ohm load to avoid the above mentioned clipping and raised distortion?
Can the DA3 drive a high impedance load such as an ext amp decently without distorting and clipping?

Sorry for the many questions, just trying to understand as I said I own the device and I am wondering if it is worth keeping or not.
Me personally I like the way it sounds, generally, just I always noticed it seems to be very sensitive to noise coming into the usb interface, and in some cases i do hear some "glare" in the treble so maybe that is that infamous clipping/distortion mentioned earlier.

Thanks in advance
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,376
Likes
234,507
Location
Seattle Area
As stated, the DA3 starts clipping at only 1 volt. To me it sounds surprisingly low to be considered a good dac am I wrong? (Your conclusions do not seem to give this too much importance to this result and I would like to understand why)
I would expect any decent dac to be able to put out at least 2v without clipping or rising distortion too much, no?
It is a good question. The answer is that DACs normally drive a pre-amplifier that has far higher impedance than the loads I use for headphone testing. That usually eliminates any clipping.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,376
Likes
234,507
Location
Seattle Area
Another thing I do not understand is: if DA3 clips @ 1volt, why does the first "dashboard" view show it putting out 1.9 volts with a nice, non-clipped signal and distortion at about 0.003%? Should it not reflect clipping and distortion as seen in the output test?
Per my explanation, in dashboard view I use much higher load impedance of 100 k Ohm:

1534358353820.png


So it is an entirely different load than 300 ohm and lower I use to simulate headphone loads.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,376
Likes
234,507
Location
Seattle Area
It is very possible, as a noob, that I misunderstood the output test. What practical consequences does this clipping@1v have in real life usage?
What volume levels (given its 0-38 digital levels, 19 being the default) are safe to use into a 32 ohm load to avoid the above mentioned clipping and raised distortion?
Can the DA3 drive a high impedance load such as an ext amp decently without distorting and clipping?
It all depends on how loud you listen, and how efficient the headphones are.

You can convert the voltages on the graph to watts using: Watts = Volts * Volts / Load Impedance. I have done that on the graph for 33 ohm.

Clipping distortion is usually quite audible. Play something with good bass and listen to its fidelity as you turn up the volume. You should be able to easily hear the bass notes distort. If they do not, and the music is loud enough, then you are good to go.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,376
Likes
234,507
Location
Seattle Area
Sorry for the many questions, just trying to understand as I said I own the device and I am wondering if it is worth keeping or not.
Me personally I like the way it sounds, generally, just I always noticed it seems to be very sensitive to noise coming into the usb interface, and in some cases i do hear some "glare" in the treble so maybe that is that infamous clipping/distortion mentioned earlier.
You asked very good questions and it is my bad for taking such things for granted and not explaining them.
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
Sorry for the many questions, just trying to understand as I said I own the device and I am wondering if it is worth keeping or not. Me personally I like the way it sounds, generally, just I always noticed it seems to be very sensitive to noise coming into the usb interface, and in some cases i do hear some "glare" in the treble so maybe that is that infamous clipping/distortion mentioned earlier.

The "glare" you hear shouldn't be there unless the DAC is actually clipping, and normally clipping doesn't sound like what I think people are describing when they say "glare", because it only affects the peaks and has a very distinctive sound.

The level of distortion Amir has measured when the DAC is not clipping should absolutely not be audible.

When you've noticed glare, have you tried turning the volume down slightly to see if it's still there? If not, the most likely source is elsewhere - either your headphones or the recording itself.
 

AlbyD

Member
Joined
Aug 15, 2018
Messages
8
Likes
5
Per my explanation, in dashboard view I use much higher load impedance of 100 k Ohm:

View attachment 14802

So it is an entirely different load than 300 ohm and lower I use to simulate headphone loads.

Ohh I see. I had missed that detail, I get it now thanks.

So all in all we could say the DA3 is a very good dac when used to drive a preamp or amp, while being a decent one when used to drive headphones assuming you don't push it too hard into clipping (by maybe using a combo of low sensitivity hp and/or loud volume).

Another aspect that could be added to your review is : I recall reading about the sabre 9018q2m having an embedded amp stage that uses a "current output" mode vs the classic voltage mode (which apparently is still available to the SOC as a fallback mode, if I am not mistaken the current output can only be enabled if the implementation has a low enough output impedance).

Could you maybe, in the future, test this feature or explain the differences between these two output modes?

----

As I have noticed there is quite a bit of interest around the Sabaj DA3 (rightfully so), I thought I'd write my own unofficial review/general impressions of the device.

I bought one back in march '18, I was impressed with its sound quality upon first listening. Eventually I noticed some variability in sound quality between different systems (desktops and laptops, xmos driver, same OS and settings).
Like I said before, it sounded "brighter" with some systems while more full bodied and musical with others.

My diagnose after further testing was twofold : first, being USB bus powered, it seems to be sensitive to how clean the power coming from the computer is. Furthermore, I found out it is also affected by the Signal Integrity over the data lines.
These two aspects are not the same but related.

I noticed a change in SQ swapping the factory cable (generic thin usb2.0, similar quality to what you'd find bundled with a smartphone to charge the battery) in favour of more robust, double shielded and USB-IF certified cables.
I personally decided to employ a certified USB 3.1 Anker cable, type C to A (1mt).
Using a quality cable made the DA3 sound less harsh in the treble and more full bodied in the mids and bass.

The reason behind it could be, going by what I understood so far, that as Signal Integrity degrades (A poorly designed cable could do that I guess) the PHY usb interface needs to turn on pre-processing logics that generate noise inside the dac itself, affecting the analog output too.

Like I said before, this dac seems to be particularly sensitive to power quality, my guess is because of its small footprint there was no space to include any filtering to speak of, so its noise rejection capabilities are marginal.

To make it sound more consistent between my systems, I employ an externally powered hub only dedicated to the dac. The power supply brick that powers it has relatively low noise (at least compared to bus power) so at least it makes it perform consistently accross different systems.

One could also use something like the ifi idefender to power the dac with an external power supply though, no need for a hub.

Again, using cleaner dedicated power cleaned that glare even more (to the point there is basically none unless you really push the volume up to the highest settings) and created a darker background and more natural organic sound.

Hope this will be helpful to other owners/potential buyers
 
Last edited:

AlbyD

Member
Joined
Aug 15, 2018
Messages
8
Likes
5
The "glare" you hear shouldn't be there unless the DAC is actually clipping, and normally clipping doesn't sound like what I think people are describing when they say "glare", because it only affects the peaks and has a very distinctive sound.

The level of distortion Amir has measured when the DAC is not clipping should absolutely not be audible.

When you've noticed glare, have you tried turning the volume down slightly to see if it's still there? If not, the most likely source is elsewhere - either your headphones or the recording itself.

Hi, yes turning down volume (lets say around the default 19) did reduce the digital glare, but I found particular relief using cleaner power supply and good quality cable over the generic cheap factory one ( with the goal to improve signal integrity and let the PHY turn off pre processing thus creating less noise inside the dac itself.)
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
Hi, yes turning down volume (lets say around the default 19) did reduce the digital glare, but I found particular relief using cleaner power supply and good quality cable over the generic cheap factory one ( with the goal to improve signal integrity and let the PHY turn off pre processing thus creating less noise inside the dac itself.)

Fair enough. I'd still be pretty shocked if USB cables and hubs etc. could make an audible difference with this unit. Amir's measurements used the stock cable and his laptop's USB port. You can see that the noise is 130dB down, and harmonic distortion is more than 90dB down and confined to lower order harmonics - all of this is absolutely inaudible.

I've also been through a phase in which I experimented with USB cables, power conditioners, etc, and I did think at that time I was able to hear subtle differences. I don't mean to discount your experience, but having in more recent years seen a lot more measurements and having come to understand better what is actually going on inside a DAC and in various other audio components, I'm now pretty certain that any shortcomings I was hearing were the result of other components of the system (room and speakers primarily), and that any differences I thought I heard switching cables or adding power conditioners etc were either because I was using mind-bogglingly substandard gear in the first place (which the Sabaj DA3 seems not to be), or the result of powerful illusions on my part.
 

AlbyD

Member
Joined
Aug 15, 2018
Messages
8
Likes
5
Fair enough. I'd still be pretty shocked if USB cables and hubs etc. could make an audible difference with this unit. Amir's measurements used the stock cable and his laptop's USB port. You can see that the noise is 130dB down, and harmonic distortion is more than 90dB down and confined to lower order harmonics - all of this is absolutely inaudible.

I've also been through a phase in which I experimented with USB cables, power conditioners, etc, and I did think at that time I was able to hear subtle differences. I don't mean to discount your experience, but having in more recent years seen a lot more measurements and having come to understand better what is actually going on inside a DAC and in various other audio components, I'm now pretty certain that any shortcomings I was hearing were the result of other components of the system (room and speakers primarily), and that any differences I thought I heard switching cables or adding power conditioners etc were either because I was using mind-bogglingly substandard gear in the first place (which the Sabaj DA3 seems not to be), or the result of powerful illusions on my part.


Yeah, I know what you mean.

Thing is, I have nothing to prove that there was actually an objective improvement, hell I don't even have any measurement gear or tools to setup proper A/B testing ( nor have the real will to do so, probably ).

All I can go by is my honest impressions after multiple tests, in any case I figured that a good certified cable couldn't hurt in any case so I'd still use it even if I couldn't hear any difference. Same for using cleaner power.

I am aware of all the endless arguing on cables, reclockers etc etc if they make any difference or not, and I for sure will not be the one to give definite answers cause I don't have the technical knowledge nor the tools to.

However, it seems logic to me that IF cable quality was to make a difference it would be more on a unit such as the DA3, rather than more expensive/better engineered devices.
Reason being, the DA3 is built and produced with portability and cutting costs in mind so I'd be surprised if thoughtful filtering was implemented (as an example).

By the way, have you ever heard/read what that guy that engineered the chord dacs says?
He swears that even reducing noise down to -350db made still an audible difference, and that is why he creates his dacs with that in mind.

I don't know how much of that is just marketing crap, but apparently he is considered reputable in the audio engineering field.

I do however think that the human hear and brain has a sensitivity that goes beyond what we currently believe.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,376
Likes
234,507
Location
Seattle Area
Another aspect that could be added to your review is : I recall reading about the sabre 9018q2m having an embedded amp stage that uses a "current output" mode vs the classic voltage mode (which apparently is still available to the SOC as a fallback mode, if I am not mistaken the current output can only be enabled if the implementation has a low enough output impedance).

Could you maybe, in the future, test this feature or explain the differences between these two output modes?
I have not seen this mode exposed anywhere. There is usually an output stage after the DAC so it is up to the designer to select such a mode and not a user specified thing.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,376
Likes
234,507
Location
Seattle Area
By the way, have you ever heard/read what that guy that engineered the chord dacs says?
I have not only read, but also sat through one of Rob's presentations. At the end, I asked him how he has confirmed that his belief in distortions at incredibly low levels is audible (-180 dB and such). Specifically whether he had done any blind testing and he said NO. And that he doesn't believe in them. I am confident he has confused himself through improper listening tests of efficacy of what he designs.

You can hear me asking him this at his talk at around 56:50 mark where I first correct him on his filter specs (which are heavily inflated) and then the listening test:

 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
However, it seems logic to me that IF cable quality was to make a difference it would be more on a unit such as the DA3, rather than more expensive/better engineered devices. Reason being, the DA3 is built and produced with portability and cutting costs in mind so I'd be surprised if thoughtful filtering was implemented (as an example).

By the way, have you ever heard/read what that guy that engineered the chord dacs says?
He swears that even reducing noise down to -350db made still an audible difference, and that is why he creates his dacs with that in mind.

I don't know how much of that is just marketing crap, but apparently he is considered reputable in the audio engineering field.

I do however think that the human hear and brain has a sensitivity that goes beyond what we currently believe.

I think you might be underselling the DA3 tbh. If there's anything demonstrated by the kind of measurements @amirm does, it's that good engineering is not necessarily expensive, and that expensive engineering is not necessarily good. But if I were in your position and swore I could hear a difference, I would make the same choices you've made for sure.

The idea that you can hear more than 130dB below the signal is crap IMHO. No research has shown that the human ear has a dynamic range of greater than that, and the 130dB figure is a best-case theoretical limit (the ear sensitises itself to the sound levels it is exposed to at any given moment, so the dynamic range in practice is significantly less than the theoretical maximum).

I actually tend to think the human ear and brain has a level of sensitivity lower than what a lot of people in consumer audio believe or claim. And that what we often take to be incredible abilities to discern artefacts in the sound are in fact much better explained by psychological factors that, unlike many beliefs and superstitions in the world of audio, are in fact well understood by science.

But you'd probably already worked that out from my previous posts ;)
 
Last edited:

maxxevv

Major Contributor
Joined
Apr 12, 2018
Messages
1,872
Likes
1,964
Hmm... I've read it in a paper sometime back that the threshold was 116dB and I believe it has been cited here on these forums a few times?
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
Hmm... I've read it in a paper sometime back that the threshold was 116dB and I believe it has been cited here on these forums a few times?

Yeh, I'm probably slightly off with the figure. I think there was one study that put it at 116dB and another that put it in the high-120s.

In any case, the DA3 has a SNR of about 130dB, so will safely surpass human hearing.
 

maxxevv

Major Contributor
Joined
Apr 12, 2018
Messages
1,872
Likes
1,964
Agreed.

I think based on what has been brought up to date in studies, a 115~120dB level should be suffice to be beyond the reach of the general human ear to discern without difficulty. In that regard, the Da3 is good!
 

AlbyD

Member
Joined
Aug 15, 2018
Messages
8
Likes
5
I think you might be underselling the DA3 tbh. If there's anything demonstrated by the kind of measurements @amirm does, it's that good engineering is not necessarily expensive, and that expensive engineering is not necessarily good. But if I were in your position and swore I could hear a difference, I would make the same choices you've made for sure.

The idea that you can hear more than 130dB below the signal is crap IMHO. No research has shown that the human ear has a dynamic range of greater than that, and the 130dB figure is a best-case theoretical limit (the ear sensitises itself to the sound levels it is exposed to at any given moment, so the dynamic range in practice is significantly less than the theoretical maximum).

I actually tend to think the human ear and brain has a level of sensitivity lower than what a lot of people in consumer audio believe or claim. And that what we often take to be incredible abilities to discern artefacts in the sound are in fact much better explained by psychological factors that, unlike many beliefs and superstitions in the world of audio, are in fact well understood by science.

But you'd probably already worked that out from my previous posts ;)

Ahah, look maybe I got a little bit too emotional with this "clipping at only 1 volt" thingy. Before that I did use to think the DA3 was one the best bargains I got recently (paid about 90 dollars for it) and hell it probably is.

Like I said any DAC i have I'd probably try to get a good quality cable and power anyway (I am talking about reasonably priced cables such as ankers etc nothing expensive or exotic) which could only help and not hurt anyway.

Thanks for the video link Amirm, great input.
 

DrBones666

New Member
Joined
Aug 17, 2018
Messages
2
Likes
0
Hi Amirm , i wonder if you can help me with using the Da3 with my android phone, im not sure which cables to buy which will allow data transfer between the phone(s6) and da3 (the da3 will need power and data transmission). The s6 has a micro usb connection and the da3 a type c and the powerbank type a, i have found a type a splitter to two micro usb and a micro usb to type c converter, do you think this solution will work or are there any better solutions you can think of to try? https://www.amazon.co.uk/gp/product..._title_srh_1?ie=UTF8&psc=1&smid=AXZ3JQ1GVFPIF
https://www.amazon.co.uk/gp/product...c_title_dp_3?ie=UTF8&psc=1&smid=AXZ3JQ1GVFPIF
Thankyou for the great review on this product in this thread btw.
 
Last edited:

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
Ahah, look maybe I got a little bit too emotional with this "clipping at only 1 volt" thingy. Before that I did use to think the DA3 was one the best bargains I got recently (paid about 90 dollars for it) and hell it probably is.

Like I said any DAC i have I'd probably try to get a good quality cable and power anyway (I am talking about reasonably priced cables such as ankers etc nothing expensive or exotic) which could only help and not hurt anyway.

Thanks for the video link Amirm, great input.

They don't do USB cables, but if you haven't heard of them already, this company is a great place to go to for affordable AV cables based on sound scientific principles.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,376
Likes
234,507
Location
Seattle Area
Hi Amirm , i wonder if you can help me with using the Da3 with my android phone, im not sure which cables to buy which will allow data transfer between the phone(s6) and da3 (the da3 will need power and data transmission). The s6 has a micro usb connection and the da3 a type c and the powerbank type a, i have found a type a splitter to two micro usb and a micro usb to type c converter, do you think this solution will work or are there any better solutions you can think of to try? https://www.amazon.co.uk/gp/product..._title_srh_1?ie=UTF8&psc=1&smid=AXZ3JQ1GVFPIF
https://www.amazon.co.uk/gp/product...c_title_dp_3?ie=UTF8&psc=1&smid=AXZ3JQ1GVFPIF
Thankyou for the great review on this product in this thread btw.
The first one is charge-only cable so don't use that. The second one (adapter) should work as I have used them in that context but without testing, I can't say it will for sure
 
Top Bottom