iFi iDSD Diablo Review and Measurements

Earlier impressions on the Diablo can be gathered here: https://www.superbestaudiofriends.org/index.php?threads/ifi-audio-idsd-diablo-impressions.10461/

I wanted to get my hands on the Diablo because of a singular use case: a transportable AIO for full-sized headphones, specifically the ZMF Verite Closed (and to a lesser extent the HD6xx). I'm sure people will have other use cases which they want explored. I will do my best to cover them if asked; however I won't make my usual extended efforts because I lack time these days.This unit will be going on loaner, so sign up if there something about the Diablo that you want to evaluate that I did not cover sufficiently.

The measurements will come later. I feel doing subjective testing under controlled conditions and an extended period of time first is better than measurements first. Doing measurements first can result in confirmation bias and cause me to do dumb things like crank up the volume with the HD650 at 105db and proclaim that I hear -88db THD+N distortion from the amp, when in fact I'm hearing the -23db THD+N distortion from the headphone.

The best transportable amplifier I have ever heard was the ALO Audio RX MK3 from 2012. Does the Diablo match it? Not quite, it's very close. Hard to say because the Diablo doesn't have an external input for an apples to apples comparison. However the RX MK3 being the "best" comes with a number of huge qualifiers: 1) physically unwieldy; 2) does not come with a DAC - we need yet another brick; 3) noisey with even today's medium sensitivity IEMs; 4) no longer available!

Here is solid proof on how I use it. It's on a hotel nightstand and that the hotel phone.


I was visiting Austin to scout out some meet sites near downtown, so I decided to drop by the Capitol building. This is a painting of Davy Crockett and friends at the Alamo in the state senate chamber.

 

The Diablo is a balls-to-the-walls design with unessential features eliminated. This means no audio inputs, no iFi match, no measure vs. listen filter, no 3D, no Xbass, etc. I always chuckle when I see the measure vs. listen vs. Bitperfect. In this case, iFi has decided to choose the filter for us and I am glad they did. I am certain it's their Bitperfect filter (or listen for DSD) and not the measure filter. I hope in the future, iFi just gets rid of the measure filter. Obviously one sounds better than the other from iFi's designers.

There's a USB A for input, USB C for charging, SPDIF input, and balanced output on the back.


On the front, there's standard 1/4 phono out, 4.4mm balanced out, and the Eco, Normal, and Turbo switches, which affect gain. I am not certain if these switches affect more than gain (which will of course affect voltage and thereby power output).


Note the following constraints for the purposes of my evaluation:

1. Turbo Mode: After level matching, I felt Turbo Mode sounded the best, the most involving, the most lively. Normal Mode didn't sound bad either, but was Turbo was just a bit better. Supposedly, the Diablo doesn't not employ global negative feedback and utilizes local feedback loops. However, changing gain can still make a difference. The downside is that on the higher gain (will measure to see how much), we have usable area to play with on the volume knob. For me, this is not particular to the Diablo - it's a problem for everything. I listen to stuff on high-gain all the time, including with high-sensitivity Grados, so I deal with the a super sensitive volume knob that is always under the 10 o'clock position.
2. I used the 1/4" output. Yes, the balanced output as moar power. TBH, I just want to keep my life simple and not have to buy three cables or two cables and three adapters for every headphone that I own. Personal opinion: I wish this balanced craze would just die, like how MQA is dead. That being said, I did try the balanced outputs and will report on this later. Let's just say the balanced outputs is different rather than better.
3. I used it 75% of the in battery mode without the power cable plugged in. I didn't think there was a difference in SQ either way, but I wasn't trying hard to discern the differences. Quick take is that battery sounds great.
4. I used it 80% of the time with the ZMF Verite Closed and HD650 as I found these headphones to have the best synergy.
5. A Google Pixel 2 (yeah, I need to get a new phone because MS Outlook and Teams runs like a dog) was used as the source via USB. Worked like a charm no drivers. USB Audio Player starts up right away without a hitch.

 

Now this is where I have some constructive criticism for iFi. iFi has always prided itself on providing accessories galore, especially on their higher-end more expensive products. Over time, I have become accustomed to iFi providing me with everything I need in the box without having to hit Amazon and wait another day or two. I found two things lacking:

1. A USB C to USB A female cable so that I could connect to the USB of my phone. I strongly feel that in 2021, folks are not only connecting to laptops, but also to phones and tablets. Likewise offer something similar for Apple users.
2. A 4.4mm balanced to 4-pin XLR adapter.

To iFi's credit, everything else needed was included: USB-A audio cable, a USB-A to USB-C charging cable, and adapter to connect headphones with a 3.5mm jack to the 6.3mm single-ended output.

 

So let's get down to it, how does it sound?

The Diablo very much sounds like an upgraded iFi Micro iDSD BL (reviewed here: https://www.superbestaudiofriends.o...-micro-idsd-bl-review-and-measurements.8842/_ It follows much of the iFi house sound with the following differences:

The presentation is more mature, refined, with smoother timbre. The Micro BL could sound a bit raw at times, the Diablo is more polished and more engaging. I've often said that portables and transportables never sound as good as well implemented desktop amps, even entry level desktop amps. The Diablo could be an exception here.

With respect to tonal signature, the Diablo is neutral with no emphasis or lack of emphasis in any area. Very few solid state amps meet this criteria. It seems many are colored in some way or another. For example, the Lake People G111, Schiit Asgard 3 and Magnius, Topping L30 I considered colored. It's not that coloration is bad, some great amps like the ECBW and Luxman amps are colored, but it's done in a good way. What I wanted to emphasize is that this is one of those solid-state amps that doesn't impart much, and that also includes the horrible solid-state nasty steely sound often associated with "wire-with-gain"*.

That being said, the Diablo imparts a small dose of inner warmth, even euphony, wetness. Small dose. Especially after initial impressions. I found that I needed to burn-in both outputs 1/4" and 4.4" separately overnight a few days for the Diablo to open up. Hey, don't blame me, blame Sony's guidance for needing to burn-in those fancy SMT caps!

The bass is punchy, dynamic, although not the fastest or most articulate. It's neither slow nor mushy. It should be noted that the squarish high contrast sound doesn't seem to be what iFi Audio subscribes to. The highs are extended and present, but not overly so with air in my face (AKM Velvet). The upper-mids are honest, not laid-back. This could explain why I felt such a great synergy with the Sennheiser HD650 and ZMF Verite Closed. However, the JAR600 sounded truthful here as well. I'll get back to this in a later post with respect the balanced and single-ended outputs and gain switch position.

With respect to resolution, I do wish it was better. It's definitely an improvement from the Micro BL. Those BB179* chips are getting kind of outclassed by newer stuff (RIP AKM). However, they do tone and timbre right - so perhaps that is the most important. Who wants to listen to microdetail when it comes with a constant wave of Jom Gabbar pricks. I'd say resolution about on par with a modded Modi Multibit.

Finally, in order to get more play on the volume knob when used on high-gain, I simply set the output from USB Player to have -10db gain on the Pixel 2.

*Just my personal experience, every time I've heard someone claim "wire-with-gain (check your source)" as an excuse for a solid-state amp, it's usually because that amp sounds like needles into the ears, bright and sizzly, etc.

 

So it is worth $999?

That's a tough one.

The Diablo is an amazing achievement that is the first transportable which I feel can compete with high value well implemented desktop amps. It blows away any Topping stack regardless of price (then again, what doesn't). It trades blows with 789, with 789 doing that blackness and articulation thing better, but Diablo killing it with better tone, timbre, and dynamics. The Diablo provides infinitely more soul and correct tone than Magnius / Modius or Heresy / Modi 3. The Diablo just doesn't quite match what a $700-$999 desktop system with well curated components can do.

No, it's not $700 worth of desktop sound quality. A decent but not revolutionary improvement over the Micro BL. However at $999, there's nothing else quite like it that is transportable and can make the ZMF Verite Open (or HD650) headphones sound good, that is more than good. We talking good enough that would make distract me away from work or make me pursue to listen to it in the evenings by my bedside.

By the red, that red isn't that garish in RL. For some reason, it just comes out super garish in photos.

 

With respect to the balanced outputs, I may need to do more listening here. Initial impressions are that the balanced outputs are more upper mid forward than the SE 1/4" jack output. I needed to set the gain setting on the normal to calm things down. This has the effect of decreasing the microdynamics, the liveliness of the presentation a little bit. However, this was offset by greater microdetail. Balanced is always +6db louder, so moving the gain setting one notch lower would be necessary anyway.

I'll switch back and forth in the few days to get a better idea which one I like better, although I'm already pretty sure I prefer the increased microdynamics (SE high gain) vs. greater microdetail (balanced med gain). The thing about balanced is that is often means cancellation of the second harmonic, and that may not always be a good thing or too much of a good thing*

BTW, I thought I should mention this again: the Diablo slams like any good desktop amp (regardless of price) that can slam (noting that it's tonal signature is even and not emphasized in the bass).

* If that sounds like I am trying to put a complete end to balanced designs (other than inherently balanced via transformers), that's because I am. Again, just my opinion, but I'm getting sick of options that force us to buy more cables. Consumers need to understand that balanced isn't always better and in most cases an inefficient use of parts to get moar power. I blame all this on Tyll and HeadRoom, of which he accepts the blame with a middle finger towards me.

 

iFi Diablo in nutshell:

• Diablo. Great product nomenclature. Please continue with this. I have a hard time explaining to anyone in the family or any other audiophile what the iFi Audio Micro iDSD Black Label Signature actually is. I also hate typing that many words every time I write a review.
• Hefty price tag, but nothing like it. Sucks to be you or me, but alas, this is life. However, not without reason...
• Only transportable in existence that can compete with good entry level desktop setups with respect to full sized headphones. An engaging sound that makes you not want to put the headphones down.
• Tonally neutral, the Diablo's strongest trait. A presentation that isn't too syrupy, warmpoo, lean, bright, laid-back, solid-state nasty, etc. Strong weight, heft, and attack only seen in transportable amps once every ten years.
• Could be a bit more resolving of microdetail
• Could provide more accessories (cable with phone USB3 or lightning) so I don't have to spend another $40 on Amazon
• Balanced outputs sound different, more upper mid aggressive. Can work around with lower gain setting. Not necessarily worse or better and could synergize well with Audezes.
• Probably not ideal with super efficient balanced armature IEMs.
• Body of Diablo gets warm during use. This probably means it sounds good because good audio gear is supposed to get hot.

Will follow up with measurements for those who wish to assess gear based on someone's Noise and Distortion (using a 20kHz bandwidth with a steady state sine wave at 1000 cycles per second at an alarming high 2Vrms into 300-ohms.) Seriously, you don't have to listen to my gibberish this measurement is your thing. (You are going to buy a Topping anyway).

 

I didn't try with Grados. The reason is that I'm very particular about Grados and often find that they sound better without an "amp" or amp with more than a few parts. This means either straight from a portable player or from a SET design. The reason is because Grados take just thousandths' of volts to get moving. Thus I've never felt a powerful amp did any favors for Grados, but got in the way instead. e.g. needing to use unity gain.

My advice is to get the the iFi GO Blu for Grados. This little thing sits on the nightstand next to the Diablo when I want to use Grados: https://www.superbestaudiofriends.org/index.php?threads/ifi-go-blu-review-and-measurements.11523/

The iFi GO Blu has the opposite problem of the Diablo. It sucks for headphones which traditionally have required a bit of oomph and juice to get going. However, synergistically with the Grados, it's amazing. I also happen to really like the Cirrus/Wolfson sound too, at least the way iFi implemented it in the GO Blu. Seriously, my custom x4 45 SET amp and Schiit Yggdrasil LIM isn't that much better than the iFi GO Blu when it comes to powering Grados. I don't know what it is, but every now and then, I'll find a specific synergistic combo that really works, and this is one of them. You can run the GO Blu via USB from a phone or use LDAC bluetooth. I noted no sonic degradation when using LDAC.

 

Possible brain damage from my contracting COVID. Actually no symptoms. I think it was the bad cold or flu that I immediately got afterward when I tested COVID negative.

 

With Audeze's, I suspect none. In fact, I feel it would be synergistic since Audeze's are laid-back sounding and I prefer more upper mid aggressive amps with them. However with HFMs, which can have timbre issues in the highs, it could be problematic.

 

iFi Audio Diablo
1kHz 0.775Vrms into 300-ohms
Low gain, 0dbFS output, volume adjusted for 10dBU (0.775Vrrms)


Very good results. All harmonics are -100db below the fundamental. More importantly a fairly nice pattern with higher second harmonic with successively lower harmonics (the 5th does stick out a bit).

iFi Audio Diablo
1kHz 0.775Vrms into 32-ohms
Low gain, 0dbFS output, volume adjusted for 10dBU (0.775Vrrms)


Doesn't bat an eye into lower impedance loads.

 

iFi Audio Diablo
THD vs output level (Vrms)
32-ohm and 300-ohm loads
1kHz steady state sine
90kHz bandwidth


For 32-ohm and 300-ohm loads, the result is very similar. The sweet spot of the Diablo is between 500mV and 1.5mV. This is ideal for Sennheisers and many ZMF cans which peak near 100db at these voltage levels. At low gain, we don't get output higher than 2Vrms.

SINAD vs output level (Vrms)
32-ohm and 300-ohm loads
1kHz steady state sine
90kHz bandwidth

 

Haven't gotten to it, but it's a good idea. I feel the 32-ohm "standard" is out of date or that something like 16-ohms would be more reflective. Easy enough to do, just need to parallel a few resistors. In the future perhaps. I am getting an LCD-X in, so I just use that as a load!

 

iFi Audio Diablo
SMPTE IMD level sweep


This is a dual tone 60Hz + 7kHz sweep at at 4:1 voltage ratio. Providing this for reference.

 

Out of curiosity, this is high-gain, where I like to listen at. The downside is a twitchier volume knob. Most other times, the distortion is higher too, but this is not the case here! The noise here is higher compared to low gain, but the distortion looks to be a bit lower!

iFi Audio Diablo
1kHz 0.775Vrms into 300-ohms
High gain, 0dbFS output, volume adjusted for 10dBU (0.775Vrrms)