Article: New AudioQuest DragonFly Cobalt

[miguelito]

I’m a long-time user of the Dragonfly Red. Just got the Cobalt. It has more punch, ease, air, and better overall presentation than the Red, even with hard to drive headphones. Really nice. I use it mostly with an iPhone XS Max (plus the USB3 camera adapter). Headphones I use: HD600, HD650, AKG 702, AKG 7xx, Denon AH-D2000, Noble Kaiser 10 custom IEM.

 

LED sample rate color coding is different from the DF Black & Red. Initially I thought there was something wrong with the Cobalt as I didn’t expect the color coding to be different from the others. But I realize the Cobalt’s coding makes more sense: Green is 44.1, Blue is 48, Light green is 88.2, Light blue is 96, Purple is MQA rendering. So the greens are for the 44 family, blues for the 48 family, and such. Easy to remember. In contrast the Black & Red have a non-sensical color scheme: Green/Blue/Amber/Magenta/Purple for the same progression.

 

PS: Cobalt’s literature says 88.2 should look yellow. Nah... It is light blue.

[miguelito]

On 7/11/2019 at 10:12 AM, MikeJazz said:

I have a more question about usability. What exactly happens when we are listening to music with the dragonfly cobalt and we have a incoming call and we want to listen to it. Does the DAC (1) silence the music automatically (2) pass trough our sound, if the headphones have mics...

On the iPhone the Cobalt (or the Red) will not be seen at all by the phone when it comes to calls. If you receive a call, you will have to pick up the phone as you would if you had nothing connected to it, the earpiece and microphone in the phone would be the ones used. You don’t need to disconnect the Dragonflys to pick up a call. I would expect the same behavior with all USB-connected DACs.

[miguelito]

On 7/11/2019 at 2:00 PM, Musicophile said:

I'd love to upgrade. Anybody interested in buying my Red?

Definitely an upgrade. Better than Jitterbug+Red (which I use on my Macpro).

[Musicophile]

4 hours ago, miguelito said:

Definitely an upgrade. Better than Jitterbug+Red (which I use on my Macpro).

Thanks. I may just be too lazy to find a buyer for my Red. I’m not good at the reselling business. 

[miguelito]

1 hour ago, Musicophile said:

Thanks. I may just be too lazy to find a buyer for my Red. I’m not good at the reselling business. 

Yeah. Just get it. 

[miguelito]

18 hours ago, Musicophile said:

Thanks. I may just be too lazy to find a buyer for my Red. I’m not good at the reselling business. 

I was as well. Superphonica here has been great, sold two items already - simplifies the whole process. I am keeping my Red probably just to leave on my macpro. 

[miguelito]

I measured the current draw produced by the DF Red and DF Cobalt. I found that the Cobalt draws MORE current than the Red, contrary to AudioQuest claims. Specifically, I found the current draw to be 40mA-60mA for the Red (higher current with higher sampling rate, MQA rendering draws the most), and 60mA-80mA for the Cobalt. Tried both easy to drive (AKG-K7xx) and hard to drive (HD-650) headphones - the headphone did not make a difference in the current draw. Voltage was the same in all cases (5.16V)

[mansr]

4 minutes ago, miguelito said:

I measured the current draw produced by the DF Red and DF Cobalt. I found that the Cobalt draws MORE current than the Red, contrary to AudioQuest claims. Specifically, I found the current draw to be 40mA-60mA for the Red (higher current with higher sampling rate, MQA rendering draws the most), and 60mA-80mA for the Cobalt. Tried both easy to drive (AKG-K7xx) and hard to drive (HD-650) headphones - the headphone did not make a difference in the current draw. Voltage was the same in all cases (5.16V)

I can confirm this. With a 50 Ω load, the current draw goes up even more, but that's just physics. The current supplied to the load has to come from somewhere.

[mansr]

@miguelito What volume level do you find comfortable with your headphones?

[miguelito]

9 minutes ago, mansr said:

@miguelito What volume level do you find comfortable with your headphones?

About 60% of the slider in Audirvana give or take, for all my headphones

[miguelito]

1 hour ago, mansr said:

The current supplied to the load has to come from somewhere.

Exactly.

[HalSF]

@Archimago takes a close look and is underwhelmed. 

 

http://archimago.blogspot.com/2019/08/measurements-dragonflies-audioquest.html

[lucretius]

17 hours ago, HalSF said:

@Archimago takes a close look and is underwhelmed. 

 

http://archimago.blogspot.com/2019/08/measurements-dragonflies-audioquest.html

 

@Archimago makes a reference to the "Wowza" hype review articles from Headphonia, Darko.Audio, MajorHiFi, Audio Advice, Hi-Fi+, and The Audiophiliac Guttenberg.  I have never found the reviews from those mentioned to be useful except to point out features (usually found in the manufacturer's press releases) and to provide nice pictures.  If they are not shills, they really should be.

 

[miguelito]

7 hours ago, lucretius said:

@Archimago makes a reference to the "Wowza" hype review articles from Headphonia, Darko.Audio, MajorHiFi, Audio Advice, Hi-Fi+, and The Audiophiliac Guttenberg.  I have never found the reviews from those mentioned to be useful except to point out features (usually found in the manufacturer's press releases) and to provide nice pictures.  If they are not shills, they really should be.

I have found many reviews to be helpful.

[miguelito]

On 8/3/2019 at 5:19 PM, HalSF said:

@Archimago takes a close look and is underwhelmed. 

 

http://archimago.blogspot.com/2019/08/measurements-dragonflies-audioquest.html

Thx @Archimago for this analysis. The clipping at 100% volume is a really bad bug if it is common to all units. Jeez... I compared the Cobalt to the Red with my headphones only - HD600/HD650/AKG K701/AKG K702/Denon AH-D2000/Noble Kaiser 10 custom IEMs. Obviously never hit 100% volume. I found the Cobalt to sound a bit smoother and punchier than the Red frankly, or should I say I found the Red to be slightly more “grainy on the highs” for lack of a better description. I did find - as mentioned elsehwere - that the Cobalt draws more, not less current, so more power, which is in line with it being more punchy I suppose. 

[mansr]

6 minutes ago, miguelito said:

The clipping at 100% volume is a really bad bug if it is common to all units.

Mine clips at full volume only with a load of 5 kΩ or less, so there's apparently some variability. It's still bad.

[firedog]

Archi says if you run it at 98% volume you get around the problem. So not so serious, IMO. The more serious issue is if the $300 version is not as good on only just as good as the $200 one. 

[mansr]

3 hours ago, firedog said:

Archi says if you run it at 98% volume you get around the problem. So not so serious, IMO. The more serious issue is if the $300 version is not as good on only just as good as the $200 one. 

Or the $100 one, in some cases.

[miguelito]

15 hours ago, mansr said:

Mine clips at full volume only with a load of 5 kΩ or less, so there's apparently some variability. It's still bad.

I would regard this as an awful bug. Having said this, it is good to know that with a high impedance load (>5kOhm) it doesn’t clip. In most cases where you’d want to set the volume to 100%, you’d be going to an additional amplification stage, and most those would be pretty high impedance (>>5kOhm). Thx for checking this.

[miguelito]

8 hours ago, firedog said:

Archi says if you run it at 98% volume you get around the problem. So not so serious, IMO. The more serious issue is if the $300 version is not as good on only just as good as the $200 one. 

I think this is a pretty darn awful bug, frankly. WTF???

[mansr]

1 minute ago, miguelito said:

I would regard this as an awful bug. Having said this, it is good to know that with a high impedance load (>5kOhm) it doesn’t clip. In most cases where you’d want to set the volume to 100%, you’d be going to an additional amplification stage, and most those would be pretty high impedance (>>5kOhm). Thx for checking this.

Archimago's sample clips even with a 1 MΩ load. There appears to be some variation in production.

[miguelito]

On 8/5/2019 at 10:41 AM, mansr said:

Archimago's sample clips even with a 1 MΩ load. There appears to be some variation in production.

Saw a firmware update for both Cobalt (v1.10 latest) and Red (v1.07 latest). 

 

Anyone tried this? Any differences?

 

Thx.

[The Computer Audiophile]

4 minutes ago, miguelito said:

Saw a firmware update for both Cobalt (v1.10 latest) and Red (v1.07 latest). 

 

Anyone tried this? Any differences?

 

Thx.

Haven't tried it yet. We will publish two reviews of the Cobalt this week. One subjective and one objective. 

[mansr]

7 minutes ago, miguelito said:

Saw a firmware update for both Cobalt (v1.10 latest) and Red (v1.07 latest). 

No mention of a Cobalt update on the AQ website.

[miguelito]

18 minutes ago, mansr said:

No mention of a Cobalt update on the AQ website.

I run the updater on Mac and it said there was an update to v1.10. This might actually be wrong, as I also run it for the Red and said there was an update from v1.06 to v1.07 - but v1.07 is fairly old (I just realized this), and I have surely updated to v1.07 whenever it came out (I am a serial updater... ).

 

What firmware is reported for your Cobalt on the AQ app?