You won’t get clipping by running the DragonFly’s amp to the rails into a proper amplifier. The input resistance on the amplifier being driven is too high to draw enough current to cause clipping.
In general, though, running amplifiers close to/at their voltage rail limits means more distortion and noise. Similarly, most amplifiers exhibit more distortion as current demand increases. The 9601 in the DFR and DFC is no exception (based on the minimal data ESS make available publicly).
Ahead of sound impressions … some other thoughts …
As claimed, the Cobalt is a bit smaller than the Red and Black models. The finish is quite a bit darker than the official pictures seem to show show. The package is the usual form-fitted plastic tray inside a box. And you get the normal DragonFly slip-case that accompanies the other models with the (welcome) addition of a USB-C equipped “DragonTail” cable/adapter/extender.
Unlike the Black and Red models, this unit rattles if you shake it and this is clearly due to the PCB, to which the USB and 3.5mm connectors are directly mounted, moving around in the case. You can easily see and hear them wiggling if you shake or move them.
I don’t know if the original versions were potted, were just a more tightly toleranced fit, or had some additional support, but both the Red and Black models feel a lot more solid than the Cobalt in this regard, which is surprising given the price difference.
This has been reported elsewhere, also. It seems more prevalent than not (i.e. there are some units that don’t have any play in the connectors). About half the people that contact AudioQuest about it said that they were told it was a defect, the other half appear to have been told its normal.
Unsupported TRS connectors for headphones, especially on mobile devices, are something I personally consider to be a potential problem.
If your current headphones play well with the Red, from a power perspective, they’ll play well here too.
This is not surprising as the Cobalt uses the same output stage, the same voltage rail ratings as the Red and consequently drives headphones with about the same ability. If you’re looking for more power/grunt or to drive more demanding headphones than the Red then you will not find it here.
I was surprised how well it drives the Rosson Audio RAD-0 … but harder to drive stuff is not better off than it was before:
Something I’ve never seen on the Red or Black models, but ran into immediately with the Cobalt, relates to how you connect it up. On the Red or Black it never seem to matter if I plugged the headphones in, or out, with music playing, while being driven from the iPhone.
On the Cobalt, several times now, swapping headphones on the fly has triggered the “This accessory uses too much power” message, resulting in having to disconnect and re-connect everything. In normal use, just connecting your headphones before connecting to the iPhone will avoid this - so shouldn’t be too much of an issue.
Proper sound impressions to follow …
Despite my early reservations about minimum-phase, slow roll-off, filters, I really rather like the DragonFly Cobalt. And it is definitely the best sounding of the current DragonFly trio.
It is a clear, across-the-board, upgrade from the Black (and with significantly better drive capability/power), and a qualitative upgrade over the Red**.
The only other similarly sized, similarly capable, DAC/amp I can even think of, that will actually work properly from an iOS device, as well as the other major mobile and desktop platforms, is the NextDrive Spectra X (you would trade MQA for DSD there). The rest of the “competition” is either 4x the size (or more) and won’t run off most phones anyway due to high power-draw, or doesn’t measure up on the sound front.
At a high level, the Cobalt is more resolving than it’s siblings and, indeed, the Meridian Explorer 2 (the only other MQA-capable DAC/amp in this form-factor … and even that is 5x the size of the Cobalt). Detail is excellent, timbre natural, and the overall rendering is more vivid than its predecessors.
The early high-frequency roll-off makes for a signature that is easier to listen to without encountering fatigue issues. In fact, so far, I haven’t had any fatigue with the Cobalt. This does rob the thing of some sense of air and space, and music with an otherwise highly vivid projection of the space/acoustic of the venue suffers some (Cowboy Junkies’ “Mining for Gold”, “The Trinity Session” is a good example). And it can rob some pieces of their sparkle. Not all, just some.
A more specific breakdown/narrative in terms of “bass, mids, treble, dynamics, transients etc.” will come after more listening.
If you don’t need the extreme portability, broad platform compatibility and drive capability, then there are cheaper ways to get better sound, more options/configurability/flexibility, and better drive capability.
For example, the iFi Audio Nano iDSD Black Label has broader capabilities, including DSD support, can also render MQA, has an internal battery, lots more power, S-Balanced output, and is $100 cheaper … but this comes at the cost of it being 12x bigger than the DragonFly Cobalt.
A Modi 3/Magni 3, SDAC/Atom or Topping DX3 Pro would also give you much more flexibility and better sound. But you can’t run those combinations off a USB port nor run them from your pocket.
That does put a big of a question mark in the “value” category. Quality in miniature, in any endeavor, tends to cost more than the same thing in larger packages, but the actual value proposition here is going to be quite personal.
For something I can run off my laptop, my tablet or my phone, without extra power, and drive everything from IEMs to non-pathologically-demanding full-size cans, in an enjoyable and engaging fashion, there’s not really much I can name as serious competition.
*More detailed impressions/descriptions will follow when I get time to listen more and write things up properly; for now I am shooting for the “highlights”.
**Caveats here are simply that if you have tonally darker-tilted headphones or significant age-related high-frequency hearing loss, the Red might be a better bet.
This from an AQ product manager in the Netherlands - who emailed the response to my questions. Somehow he seemed to replace my Lyr 3 with a Magni 3…
————
1: there is no advantage in using the Cobalt connected to a JitterBug, since some of the JitterBug technology was used inside the Cobalt. However, you can still optimize your digital audio from a computer, car stereo, or other hardware with more than one USB port by adding a JitterBug in Parallel (only) into a different USB port.
2: there is no line out mode, but could you consider the Cobalt as your DAC/pre-amp and your Magni 3 as the power amp. You will have to find a balance between the volume settings on your iOS device, your app and on the Schiit for best sound / maximum volume range without distortion.
They’re a lot better at stating “the bleedin’ obvious” than they are at providing any actual, useful, information!
Notwithstanding the lack of clarity in how the volume control is actually implemented on the Red or Cobalt models, this is standard fare for DAC/amps feeding an additional amplifier. And without knowing what’s really going on, you’re just left “feeling your way” and potentially getting the “worst of both worlds”.
Absent their use in a portable context, where they’re great products and make a lot of sense, I do not really understand the desire to use them in a desktop system. That is unless one can only have one DAC/amp and must share it between portable and desktop systems.
There are less fussy, better sounding (to me), more flexible options at lower cost for desktop use with, or without, built-in amplification and with proper line-output options.
Which is not to say I don’t think the Cobalt is a nice little device. It is. But at least for me, the bulk of the appeal is in what it can do in its form-factor rather than raw features or performance. There’s a price-premium for that in such a compact form, and that makes sense, but take away the need for that compactness and the value proposition is not nearly as compelling.
Agreed. I’m using the iFi xDSD in line out mode to drive the Lyr 3 and/or the Stax headphone amp.
When I read about the “line out” mode I had hoped that there might be an alternative, especially considering future purchases and two homes.
So disappointing to find that this is just bogus marketing. Why do they even mention it?
$… €… £…
Comparison of 3 Dragonflies with measurements and subjective impressions.
Subjectively I think it’s a better sounding (i.e. more enjoyable) unit than the Red, though as I mentioned earlier that’s going to depend somewhat on taste and what cans it is being paired with.
Fidelity-wise, it’s clearly worse than its predecessors simply due to the choice of filter (other issues not-withstanding). There’s a reason I bitched about the minimum-phase filter and slow-roll off model before ever hearing this thing. You can’t roll-off that early and claim “hi-fidelity”. And “natural sound” is purely subjective anyway.
Still, it is enjoyable to listen to, directly driving appropriate cans, when it’s not at the limit.
I suspect that with most of the headphones that things like this get paired with, a bit of top-end roll-off is probably quite desirable (for me that’d be pretty much anything carrying a Grado or Beyer label, for starters).
No one that’s pairing it with reasonable headphones is going to be running this to the rails, so that’s likely a moot point, but it definitely overshoots it’s spec there and not necessarily with good results. Though it’s behavior at “max volume” certainly isn’t what one wants to see.
I’ve gone on like a broken record, in other threads, about not using these things as pure-DACs into other amplifiers, which is where the “behavior at max volume” is really going to be an issue.
And I still maintain that unless you can only have ONE DAC/amp, then having a different desktop stack, even at lower cost, will get better results, and still leave money on the table for a pure-and-tiny-portable option. Value is a big question mark at this level.
Anyway …
For a bit of fun I fired up the cheap analyzer again … (I could break out the “big LeCroy” but using $900K+ instruments, with no audio-centric-automation, in multi-million-dollar isolation facilities, to measure $300 devices is more of an ePeen-waving exercise than anything truly useful).
So here’s the Red vs. the Cobalt in “inferred” jitter (pure sine wave at 1/4 fs), Red first (excuse me not resetting the X-axis range for the Cobalt):
The differences here are so far below “audible” that they’re ridiculously irrelevant, even though the Cobalt does have a smaller noise band and lower noise artifacts.
And then the same setup for an actual J-Test (1/4 fs primary tone, with the LSB toggled every 250ms - so those spikes at 250ms intervals are supposed to be there. The Cobalt is cleaner and more correct here. Again, Red first:
And, as a final comment … I am returning my Cobalt (a normal, personal, “hell, it’s same day” Amazon purchase). Not because of audible issues. I still think it is subjectively more enjoyable than the Red with what I use “on the go”, but because the “rattling” and movement of the connectors/PCB is pissing me off …
Yeah, I would have like to see archimago give some numbers with the volume dialed back a bit.
Here is my opinion based on not yet having received my Cobalt via UPS. Therefore, I cannot be accused of objectivity or subjectivity.
I would like to have the comparison testers have their Dragonsfly all wrapped up in black tape, wound enough that you could not tell them apart. Have them numbered 1,2,3 or just 2,3 and do the testing and listening without knowing which one they had. Maybe find a unit that didn’t rattle…
I do wonder about manufacturing variation. And like @Torq, I would have liked to have seen them run maybe 3 db down from max. If you’re going to all that trouble of making with the pretty graphs, why not make with the pretty measurements FIRST to ensure that you are really testing the same volume/voltage/etc.?
I don’t like the reports of all the rattling around with the Cobalt. Hope I can shake the dickens out of the one they deliver…wait, UPS will do that for me in advance.
It’d certainly be interesting to do that with these three units*.
While in general I find the differences between competent DACs are very small, often require practiced sighted-listening prior to being able to tell them apart in blind testing, and have stated many times that reviews always tend to exaggerate the differences (mine included … it’s a function of trying to express those differences) … I don’t think hearing the differences between these three particular units blind would be particularly hard.
They do, after all, have easily measurable differences that actually occur at audible levels (in addition to measurable differences that should not be audible).
Doing the same between, say, a Topping D50 and an SMSL SU-8, on the same filter settings, would be much more challenging.
Some people seem to have gotten units that don’t rattle. Again, it’s hard to know if that’s a manufacturing issue or a tolerance issue. If it’s the latter then it’s going to be a “luck of the draw” thing. If the former, it could be that one batch didn’t get potted, or just happens to have both the case and the PCB at the extremes of acceptable tolerances.
*I do, in fact, do blind, properly level-matched, comparisons when reviewing/comparing DACs, a process I’ll have to describe in more detail than I’m going to in this post.
Yes, I do expect there to be audible differences. Blind testing when there are known differences will test the ears and brain for compliance and reliability. 
Either my unit has an issue (i.e. a fault) or I’m starting to wonder if AudioQuest aren’t going to have to release some kind of update for this thing.
Since first mentioning it, I have run into the “The attached accessory requires too much power” message several more times with the Cobalt on my iPhone (Xs, latest iOS revision) … only now it is also happening even when nothing is playing and the headphones are already connected. And more than once while already listening to music (and not changing anything), at levels well below maximum.
Makes me wonder if this is related to the elevated output power over the Red (and the accompanying increase in distortion) - which might be addressable in firmware (heck, it might even be a bug … all that’d take is an off-by-one error somewhere).
And then, while I’m rambling … interestingly, every other “review” I’ve read of this so far is universally positive - hell, “gushing” even … to the point where some of them are calling this a “game changer” and “on a whole new level”.
I definitely can’t agree with that with a straight face.
That I like the sound, subjectively, more than the Black or the Red, as a DAC/amp, and think that overall sound-quality (not fidelity) is preferable here - I can say, but - overall I’m increasingly disappointed with the Cobalt - especially at $300.
Finally, since I’m returning it now due to two issues then, unless I decide to get another Cobalt in the future I won’t be posting a more detailed “review” of the thing from a sound perspective.
@Torq watching the Cobalt hype train…
I must say that I enjoyed that more than I probably should have.
We’ve received e-mails from customers basically saying the same thing. Of course, they are moving so fast that we can’t seem to keep a unit in the office for use right now. I am genuinely curious about it because @andrew and I tend to always have a number DFRs in our travel bags but the DFR does have it’s quirks as well.
It’s sounds like there’s definitely a QC issue happening with the Cobalt. I almost want to order one now to see if I hit the non defective lottery! 
How would you compare the Cobalt (or any Dragonfly for that matter) with the Grace SDAC-B; especially (but bit exclusively) when the Grace/Drop DAC is used in balanced mode?
There are only two scenarios where I would personally consider using one of the DragonFly models over the Grace SDAC-B (even in single-ended mode):
As a portable DAC/amp and not just a DAC - since obviously the SDAC-B is just a DAC.
If you can only have one DAC, period, and must share its use between both a portable and another system.
I would not, personally, buy any of the DragonFly units for use purely in a desktop environment. The permanent DAC/amp model (with no dedicated line-out) is suboptimal there, and none of them measure up to more capable DACs that cost less money.
Objectively, i.e. just by the measurements that actually matter (i.e. we’re not differentiating based on a minuscule change in noise or jitter performance that’s already well below any possible audible threshold), they’d rank as:
On sound (for my preferences), well, I’d swap the positions of the Red and Cobalt units, above, and otherwise leave it as is. I think that’s mostly because while the Red is the more technically accurate of the two, it’s treble has some minor issues that can becoming fatiguing over time (for me, with some headphones/IEMs) that the Cobalt doesn’t and there isn’t much difference beyond that.
For mixed use, i.e. if you needed to cover both portable and fixed usage scenarios you would objectively be better off with a DragonFly Red and a single-ended Grace SDAC than with a DragonFly Cobalt, and it’d be slightly less expensive too.
Again, thanks for the detailed reply. I likely should have been more specific as I’m unfamiliar with the Dragonfly models and was asking strictly about DAC performance.
I didn’t realize the DF models’ DAC was of necessity linked to its amp and thought perhaps the amp could be by-passed to use the DF as a DAC-> line out-> to external amp.
