Topping D90 Balanced USB DAC Review

[amirm]

@amirm,

Is it possible to test Hi-res BT 5.0 and direct connection to Apple devices?

Are those two separate requests or one? I can test BT from Android. It is a lot more painful to do so with Apple products.

 

[RichB]

Are those two separate requests or one? I can test BT from Android. It is a lot more painful to do so with Apple products.

If BT can stream hi-res, then that is good enough to know. Apple is supposed to support it.
Direct connection to the USB port using the Apple digital camera adapter is my plan-b.

Both of course, if it is not terribly inconvenient.
A statement by the manufacturer also works.

Thanks,

- Rich

 

[amirm]

Thanks Amir for the review.

I was determined to buy a DX7 pro and seeing this review. I had a doubt ¿Which do you think may be the best DAC of these two models? I don't need the headphone amplifier of the DX7 pro. It seems that the D90 has very good values but others not as good as you commented, THD + N versus frequency and 90 kHz bandwidth.

Thank you.

Are you going to use it on a desktop? If so, I like the rotary volume control on DX7 Pro better. If it is in a stereo system with remote controlling it, then I would go for D90. It will be better for bragging points on measurements. .

 

[amirm]

If BT can stream hi-res, then that is good enough to know. Apple is supposed to support it.
Direct connection to the USB port using the Apple digital camera adapter is my plan-b.

- Rich

I don't have a camera adapter. Would need to buy one (I have an iPhone 6S I think).

 

[slingshot]

I thought it would be sensational. Mine is coming next week. Interestingly they said at the RME Blog that the 4499 chip was meant for $3,000 DACs. I guess they were wrong. Their updated one uses a 4493 for some reason. The Monoprice which is sold out is only $500 and this one is $699.

 

[BaaM]

I also believe that companies sending Amir "golden units" is akin to conspiracy theory to me.

Being suspicious is not being conspiratorial.
This is common practice and that is why many independent testers buy the products themselves to avoid this problem.

 

[amirm]

RME would have to operate on much larger margins than Topping given the western operation. So their statement kind of makes sense.

 

[Rigel]

Are you going to use it on a desktop? If so, I like the rotary volume control on DX7 Pro better. If it is in a stereo system with remote controlling it, then I would go for D90. It will be better for bragging points on measurements. .

In principle yes, with a PC. I sleep on it. The DX7 pro would take a month to send and the D90 a week.

Thank you very much for answering and your time.

 

[restorer-john]

@amirm looks like you may have to get into BT testing. I'm sure AP will be able to hook you up with a nice inexpensive module...

 

[ElNino]

RME would have to operate on much larger margins than Topping given the western operation. So their statement kind of makes sense.

This is true. The AK4499 itself costs more than $65 more than the AK4493, and you also need separate I/V, and Topping is using 4 OPA1612, so that's another $12 just in parts. It would be hard for RME to use the AK4499 and reach what they need in terms of margins.

 

[amirm]

@amirm looks like you may have to get into BT testing. I'm sure AP will be able to hook you up with a nice inexpensive module...

They only want $6,100 for the Bluetooth module! No Joke.

 

[Celty]

They only want $6,100 for the Bluetooth module! No Joke.

Good Lord! Gouging or for some reason, justified?

 

[veeceem]

@amirm I think you should test its BT and DX7 Pro's, too! It's super convinience, I just sit on my bed, turn phone's BT on, then the DX7 pro also turned on, lead to amp also turn on, I touch nothing when listening to music!
But most important factor is: its Bt sounds so good, but why no measurements :'( thanks

 

[Max]

The linearity is perfect now, good work Topping (and amirm, too !)

 

[VintageFlanker]

If BT can stream hi-res, then that is good enough to know.

Spoiler: it can't.

 

[tw99]

Good Lord! Gouging or for some reason, justified?

Good test equipment is always expensive.

 

[Robin L]

The panthers want to take yoga classes. I keep telling them they don't have the body for it but they won't listen.

No, no, no— when it comes to asanas, Cats rule. Trust me, felines are natural yogis/yoginis, don't need no stinking' lessons.

 

[pkane]

This is a review and detailed measurements of the Topping D90 Balanced USB DAC with Bluetooth. It was kindly sent to me by the company. The D90 costs US $699 with free shipping from Amazon. The D90 is one a few DACs using the latest DAC chip from AKM (AK4499).

NOTE: Given some recent comments that manufactures may be sending us "golden samples," Topping has offered that I also purchase a retail DAC to compare to this one at their expense. Actually they have offered that I re-buy everything they have ever sent me at retail at their expense! Personally, I have no concern about Topping's ethics in this manner so I don't want to wast time and expense doing so. If you feel otherwise, comment in the thread and I will decide what to do based on that.

The D90 comes in a premium packaging that despite its rather small size, is very heavy:

View attachment 43472​

I am guessing that it is milled from solid chunk of aluminum, much like very high-end audio products. While this doesn't do anything for sound, it does make the unit sit there solidly and not be tugged by cables.

I controlled the D90 with one of my existing Topping remote controls (one comes with it as well). I had not noticed this before so may be a new feature but you can independently control whether the output shows up on RCA, XLR or both. Speaking of that, here is the back panel:

View attachment 43473​

It is all the features you expect for a premium DAC in this era including the increasingly standard Bluetooth input.

Power supply is of course included inside the unit which I appreciated. Less clutter on the desk.

For my testing, I focused on XLR outputs only. It sharply reduces the chances of ground loop which can occur more often due to connection to a PC and long chain that creates.

DAC Audio Measurements
As usual, we start with our 1 kHz tone dashboard:
View attachment 43474

Not that it matters audibly but clock accuracy is excellent showing a precise 1 kHz frequency. Output is 4 volts which is what we like to see form XLR output.

The star of the show of course is astonishingly small distortion. It showcases the performances of the AK4499 DAC chip which until now, would lose ever so slightly to offerings from ESS on distortion front. With SINAD of 120.5, the D90 barely misses taking the top spot in our SINAD graph:

View attachment 43475

With best case dynamic range of hearing being 116 dB, you are assured that the D90 is provably transparent. What you hear, is what your source has in it. Not distortion and noise from the DAC.

Using 32-tone test track to resemble "music," intermodulation distortion shows incredibly low unwanted distortions:

View attachment 43477

Someone had asked how my Audio Precision rates in this test. Here it is:

View attachment 43476

Note however that when I test analog products (e.g. amplifiers and headphone amps), I don't use the built-in DAC in Audio Precision. The companion analog generator actually has less distortion than above.

Anyway, the D90 loses a hair to AP in lower frequencies but bests it in higher frequencies. Either way, the D90 is instrument grade digital to analog convert and can be used to test analog audio gear.

Testing intermodulation again but this time against level we get:
View attachment 43478

We see the improvement in both noise (sloping down part of the graph) and distortion (right side where the graph starts to climb up).

Linearity (precision of level) is nailed as well:

View attachment 43479

Dynamic range shows the noise level:

View attachment 43480

Had this been a couple of dBs better, it would have improved the SINAD to take the first place. As it is though, it is exceptional. You can play at 120 dB SPL (live concert levels) and yet have no audible noise from the DAC!

I ran jitter with USB as usual but also added S/PDIF since there has been some concern about its performance:
View attachment 43481

There are some spikes, especially with S/PDIF so it not as perfect as the rest of the performance from measurement point of view. Audibly though, it is more than perfect with spikes way below threshold of hearing. Note that these are with full amplitude 12 kHz tone. You don't have such thing in music. Jitter is proportional to level of the signal so if you lower the 12 kHz tone, they would go down in level proportionally. So no matter which way we look at it, this is not an audible concern. But a slight visual discomfort.

THD+N versus frequency shows good performance but not as good as I expected:
View attachment 43482

This test uses 90 kHz bandwidth to capture harmonics of 20 kHz (the dashboard uses 22.4 kHz bandwidth). So the question is, are we seeing distortion products or other ultrasonic junk? For that, let's fun a spectrum analysis at 500 Hz where there is a hump above:

View attachment 43483

Harmonic products are below 130 dB so that is definitely not a problem. Instead we see a pretty tall spike near the sample rate of the source. That is caused by aliasing due to reconstruction filter not having enough attenuation. Speaking of that, here are the responses for the filter:

View attachment 43484

I was disappointed to not see a very sharp filter. I wonder if there is a firmware bug that is causing two of the settings to be the same? Or AKM doesn't provide it. Back to our previous issue, we see that the two sharpest filters only attenuate 95 dB or so. This means that aliasing components will exist below that level and that is what we saw in the THD+N vs frequency. Fortunately we don't hear 44 kHz and higher and levels are too small to cause issues for speakers/amplifiers. So not an audible issue.

Thermal Stability
D90 performance did not change over the 10 minute test so I got bored and stopped it:
View attachment 43485

Conclusions
The Topping D90 produces lowest noise and distortion than 220 DACs tested so far. It shows that AKM is ready for business to compete with ESS on top of the performance charts. It has excellent build quality and comes from a manufacturer that believes in verification of design with measurement. Its measured performance is nearly flawless. To get rid of last few niggles, would cost you a few hundred dollars more. You decide.

I live for such discoveries where a company cares about giving us the maximum performance they possibly can. And at a cost point that is thousands of dollars if not tens of thousands of dollars than high-end DACs.

For me, I am happy to strongly recommend the Topping D90 DAC.

------------
As always, questions, comments, recommendations, etc. are welcome.

The panthers want to take yoga classes. I keep telling them they don't have the body for it but they won't listen. I looked up the prices for private yoga lessons and they are high! So please donate what you can using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/

Looks like an excellent performer, except for the slow filters. Can you please run a frequency scan up to 1MHz playing a 1KHz sine wave? Curious to see if there are any other artifacts further up in frequency.

 

[JohnYang1997]

Thanks Amir for the review.

I was determined to buy a DX7 pro and seeing this review. I had a doubt ¿Which do you think may be the best DAC of these two models? I don't need the headphone amplifier of the DX7 pro. It seems that the D90 has very good values but others not as good as you commented, THD + N versus frequency and 90 kHz bandwidth.

Thank you.

Get d90. You won't regret it.

 

[saturnaal]

Slowly but firmly entering ADI-2 DAC price levels, and digging their own grave! Considering D90 has no headphone amp its price is above the RME.

I'd think it would be pretty easy to argue the opposite, seeing as you can pair this with one of the class-leading $100 amplifiers and have performance in some metrics besting the RME.