Topping D90SE Review (Balanced DAC)

[Deleted member 43441]

This is a review and detailed measurements of the Topping D90SE flagship balanced USB DAC. It was sent to me by the company for testing and I believe costs US $899.

The D90SE comes in the same clothes as current class of Topping products which is just fine with me:

View attachment 135755

We have the usual inputs and outputs including Bluetooth:

View attachment 135756

The architecture here is based on ESS ES9038PRO DAC instead of AKM silicon. Hopefully this means better availability of parts and DACs.

Topping D90SE DAC Measurements
Out of box output voltage is our standard 4 volts but there is also a setting to adjust that to 5 volt (see test of that later). Let's start with USB input and balanced XLR output:

View attachment 135757

Oh wow! We have a record breaker here when it comes to SINAD:
View attachment 135759

View attachment 135760

The Audio Precision analyzer puts severe strain on getting such high numbers due to its own self-noise so I was surprised to see 123 dB limit broken. Distortion is down to an incredibly low -140 dB!

Switching to RCA output, we naturally lose some performance:

View attachment 135761

Noise level is kept to absolute minimum producing some of the best dynamic range numbers I have seen:

View attachment 135763

Temporarily switching to 5 volt output we get:

View attachment 135764

So not a major change but if you have an amp that can accept it, you can push your noise floor down a bit more. Remember, while distortions are hard to hear, hiss from speakers or headphones is not.

Intermodulation distortion (really the noise aspect) sets a new benchmark:

View attachment 135765

There was a time when DX3 Pro was the benchmark. See how much the D90SE beats it by. It is almost left in the dust!

Courtesy of resampling in ESS DAC, jitter is a forgotten word with either input:

View attachment 135766

Combined with the ultra noise floor, we can clearly resolve the low order bits of the 250Hz square wave at 24 bits! I had to remove the ASR logo to reveal them!

Linearity locked in and ran so perfectly, I had to scratch my eyes to believe it:

View attachment 135767

You have a choice of 7 filters and I was happy to see thee of them producing the correct filtering for 44.1 kHz sampling:

View attachment 135768

Attenuation was so good that I had to lower the bottom of the scale to show where they bottom out.

The only puzzler was THD+N versus frequency:

View attachment 135769

Seems like there is some kind of ESS signature going on with some filter setting. Getting the best performance required selecting filter #7 which has a bit of roll off in frequency response.

Multitone is nailed so hard that you could almost hear it nail screaming from being hit so hard:
View attachment 135770


Conclusions
It is 3:00am but I had to stay up and post this review. It is not every day that every barrier we have seen in DACs gets shattered to new levels. We are seeing extreme attention to every aspect of this design to squeeze this much performance out of this ESS DAC. Just when we though the race was over, here comes one more over achiever! Those of you who have made it a hobby to upgrade the latest and greatest, here is another one for you!

At the risk of stating the very obvious, it is my absolute pleasure to recommend the Topping D90SE.

Note: this is a new product. If you are not an early adopter type of person, you may want others to go before you. My testing is focused on performance only, not reliability or functionality.

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

Appreciate any donations using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/

Topping is just showing off now lol.

Does anybody get the impression that they didn’t like being second or third in the ASR DAC race?

It appears that company can do anything they want, and reasonably. I just hope it lasts I have two of their DACs now.

Thanks for the review, hats off to Topping. Even if we can’t hear it… It’s nice having a V-8 engine.

 

[firedog]

HQ Player

audiophilestyle.com

Jussi, of HQPlayer fame, found anomalies in the playback of the D90SE, especially with DSD. Put the DAC back in the box.
And note: he doesn't have uber expensive DACs, and uses some relatively inexpensive ones some of the time.

 

[Deleted member 43441]

HQ Player

audiophilestyle.com

Jussi, of HQPlayer fame, found anomalies in the playback of the D90SE, especially with DSD. Put the DAC back in the box.
And note: he doesn't have uber expensive DACs, and uses some relatively inexpensive ones some of the time.

I have no idea who that person is, but maybe he got a bad unit, or maybe I don’t know the history of the whole conversation, but I didn’t care to read it. If that was me and I found a problem, I would get ahold of Tech Support.

 

[Belker]

I have no idea who that person is, but maybe he got a bad unit, or maybe I don’t know the history of the whole conversation, but I didn’t care to read it. If that was me and I found a problem, I would’ve got a hold of Tech Support.

Anyway I just bought a Fryea + because I didn’t like the desktop look of the pre-90, and didn’t like paying for the extension. Plus the Freya + has an XLR and two RCA’s out, one more than the pre90. But I doubt the pre90 is broken, and we can always find someone online that’s had a problem.

You don’t care to read, but still feel the urge to comment? And doing so, bringing two totally unrelated products into the mix.

 

[Belker]

HQ Player

audiophilestyle.com

Jussi, of HQPlayer fame, found anomalies in the playback of the D90SE, especially with DSD. Put the DAC back in the box.
And note: he doesn't have uber expensive DACs, and uses some relatively inexpensive ones some of the time.

ESS i general, and specifically Topping, seems to be a bad match with HQPlayer/DSD.

 

[firedog]

ESS i general, and specifically Topping, seems to be a bad match with HQPlayer/DSD.

Don't really think that's true. Jussi doesn't make negative comments about ESS in general. He's specifically relating to his testing with this model.
He upsamples everything to high rates - either DSD or PCM with his proprietary system/filtering - so he also measures performance that way.
When doing so, he found anomalies.

 

[Belker]

Don't really think that's true. Jussi doesn't make negative comments about ESS in general. He's specifically relating to his testing with this model.
He upsamples everything to high rates - either DSD or PCM with his proprietary system/filtering - so he also measures performance that way.
When doing so, he found anomalies.

Yes, but I was referring to the fact that ESS don’t allow direct DSD bypass. AKM do. Also, Topping seems reluctant to support 48k base DSD. On top of that we have the anomalies shown here.

 

[Deleted member 43441]

You don’t care to read, but still feel the urge to comment? And doing so, bringing two totally unrelated products into the mix.

In my experience if we listen to every Tom Dick and Harry on the Internet… we’d never buy anything.

As stated above somebody will always have a bad opinion about something that other people regard as either good enough, very good, or stellar. That’s been my experience, so I don’t go searching the Internet looking for people that have said bad things about something that I know tested instrument grade on ASR.

Edit;

I changed part of this, and if there’s know issues I would like to know about it.

 

[srkbear]

Well, yes, thank you, but no. I already have 8 different devices plugged to the power behind my audio rack and I definetely want to avoid adding another one.

I know this is an audio forum, but after over 100 pages of wows over this dac there are very little info regarding the crucial failure of optical and coaxial inputs!

I really want to warn everyone who wants to use this dac with optical input, especially with tv as a source. Those dropouts disqualify this dac to me completely.

Now, after 3 months of use, I am not only annoyed but angry as hell.
It is impossible to watch tv with this dac. I am embarassed when I invite some guests to watch live sports on tv and droputs are like every 20-40-50 seconds. And those questions "why/what is that, what's wrong?"
Maybe JohnYang1997 would tell ME what should I do now with my dac now? Somehow he buried his head in the sand after first info about dropouts here.

There’s quite a bit posted in this thread about this limitation in the Topping D90se; I know this because I participated extensively in these discussions. We don’t know for certain what the root cause is, but Topping has attributed this to compatibility issues with their cs8416 receiver module. Unfortunately I haven’t seen John Yang dropping by this forum for quite some time—he seemed to disappear around the time that folks took a break from high-fiving him for winning the SINAD contest to call him on the carpet around this particular “bug”.

I discovered the issue when trying to use my Oppo or Sony SACD player as a transport via the Topping’s S/PDIF inputs, and noted audio dropouts every 3-4 seconds. I tried the same with a Gustard x26 pro and didn’t find the same issue, although I ultimately went with the Topping because I preferred its overall sound signature and higher voltage output.

I did manage to find a workaround for this via a Kanex Pro HAECOAX HDMI audio extractor that I got from Amazon for $70 bucks (link is below). I’ve detailed my results with this method several times on this thread, but basically when I connected the HAECOAX via HDMI to the audio output on my Sony UBP-x1000ES, and then ran the signal from the extractor via coax to my Topping, the audio quality was outstanding and there were no dropouts at all.

An unexpected bonus I noted with the HAECOAX and the Sony player specifically was that when the Sony’s DSEE Extreme engine was turned on, and the HAECOAX’s 5.1 audio toggle was engaged, my Redbook CDs were decoded by the Topping upscaled at 88.2 kHz, and SACDs were decoded at 176.4 kHz—a benefit not available by running coaxial from the Sony direct to the Topping. It’s quite a delight to have my regular old CDs sounding so exceptional; I’m not certain what is going on with the upscaling, but the differences in soundstage and bass slam are fairly stark, at least to my ears.

I’m delighted that I found this solution, because overall I remain convinced that every dollar I spent on the D90se was worthwhile, and with this successful and minor workaround I couldn’t possibly justify shelling out an extra $600 for the x26 Pro—for what I consider a bulky, inferior sounding and performing DAC in comparison, despite its fancy innards.

I would be interested to hear whether anyone has discovered similar audio dropout issues with the new Gustard x18, which does not utilize the CS8416 receiver and is fairly matched with the D90se (albeit the latter still slightly in the lead in terms of measurements)…

(KANEXPRO HAECOAX / HDMI Audio Deembedder https://www.amazon.com/dp/B00F5W3S2U/ref=cm_sw_r_cp_api_glt_i_8FAKR4EVPFBPQ5BJRSS2

 

[Deleted member 43441]

There’s quite a bit posted in this thread about this limitation in the Topping D90se; I know this because I participated extensively in these discussions. We don’t know for certain what the root cause is, but Topping has attributed this to compatibility issues with their cs8416 receiver module. Unfortunately I haven’t seen John Yang dropping by this forum for quite some time—he seemed to disappear around the time that folks took a break from high-fiving him for winning the SINAD contest to call him on the carpet around this particular “bug”.

I discovered the issue when trying to use my Oppo or Sony SACD player as a transport via the Topping’s S/PDIF inputs, and noted audio dropouts every 3-4 seconds. I tried the same with a Gustard x26 pro and didn’t find the same issue, although I ultimately went with the Topping because I preferred its overall sound signature and higher voltage output.

I did manage to find a workaround for this via a Kanex Pro HAECOAX HDMI audio extractor that I got from Amazon for $70 bucks (link is below). I’ve detailed my results with this method several times on this thread, but basically when I connected the HAECOAX via HDMI to the audio output on my Sony UBP-x1000ES, and then ran the signal from the extractor via coax to my Topping, the audio quality was outstanding and there were no dropouts at all.

An unexpected bonus I noted with the HAECOAX and the Sony player specifically was that when the Sony’s DSEE Extreme engine was turned on, and the HAECOAX’s 5.1 audio toggle was engaged, my Redbook CDs were decoded by the Topping upscaled at 88.2 kHz, and SACDs were decoded at 176.4 kHz—a benefit not available by running coaxial from the Sony direct to the Topping. It’s quite a delight to have my regular old CDs sounding so exceptional; I’m not certain what is going on with the upscaling, but the differences in soundstage and bass slam are fairly stark, at least to my ears.

I’m delighted that I found this solution, because overall I remain convinced that every dollar I spent on the D90se was worthwhile, and with this successful and minor workaround I couldn’t possibly justify shelling out an extra $600 for the x26 Pro—for what I consider a bulky, inferior sounding and performing DAC in comparison, despite its fancy innards.

I would be interested to hear whether anyone has discovered similar audio dropout issues with the new Gustard x18, which does not utilize the CS8416 receiver and is fairly matched with the D90se (albeit the latter still slightly in the lead in terms of measurements)…

(KANEXPRO HAECOAX / HDMI Audio Deembedder https://www.amazon.com/dp/B00F5W3S2U/ref=cm_sw_r_cp_api_glt_i_8FAKR4EVPFBPQ5BJRSS2

Looks like I need to read the entire thread, my fault for not keeping up with it. Thanks for the info.

 

[srkbear]

Yes, but I was referring to the fact that ESS don’t allow direct DSD bypass. AKM do. Also, Topping seems reluctant to support 48k base DSD. On top of that we have the anomalies shown here.

I’m not sure what you mean by “direct DSD bypass”, unless you’re referring to a rarified and vanishing market of AKM-based RME ADI-2 DACs that processed DSD without an extra delta-sigma step—those DAC’s days are numbered unless AKM refreshes the 4493 and maintains this capability in its future iterations, but either way, is there much evidence that this method offers any salient advantages?

I’m also aware of the theoretical quantization noise reductions in 48k-based DSD sampling, although I would definitely say that DACs that offer this DSD architecture are exceedingly hard to come by (blame the original folks at Sony who decided that 64 was the best base for SACDs over 30 years ago). But again, with upscaling to DSD 256 or 512 readily available on most DACs in the D90se category, with quantization noise being pushed so far above frequencies that even dogs or moths can hear, is this too a practical concern?

It sounds like what you’re asking of your DACs is pretty esoteric—again I wonder whether pursuing such a DAC offers audible vs didactic advantages…

 

[srkbear]

Looks like I need to read the entire thread, my fault for not keeping up with it. Thanks for the info.

Well it’s a fairly monumental thread, and I had to search back for it a few dozen pages to find the discussion when I discovered the problem myself. Just keeping a possible solution accessible for folks dropping by later in the story

 

[Deleted member 43441]

Well it’s a fairly monumental thread, and I had to search back for it a few dozen pages to find the discussion when I discovered the problem myself. Just keeping a possible solution accessible for folks dropping by later in the story

Do you think Siri could read it to me?

Just kidding, I will go back and look.

 

[Deleted member 43441]

You’ll have to forgive me, I ended up posting some of this in the wrong thread. So I deleted it and I’ll refrain from posting at 2:30 in the morning. I guess they call that sleep posting.

Surely this should be plug and play, not that I’m giving up my non-se version of the D90 but I would like to know what the solution is.

 

[linger63]

I discovered the issue when trying to use my Oppo or Sony SACD player as a transport via the Topping’s S/PDIF inputs, and noted audio dropouts every 3-4 seconds. I tried the same with a Gustard x26 pro and didn’t find the same issue, although I ultimately went with the Topping because I preferred its overall sound signature and higher voltage output.

I’m delighted that I found this solution, because overall I remain convinced that every dollar I spent on the D90se was worthwhile, and with this successful and minor workaround I couldn’t possibly justify shelling out an extra $600 for the x26 Pro—for what I consider a bulky, inferior sounding and performing DAC in comparison, despite its fancy innards.

How did you manage to come to this conclusion?
Surely a couple of SINAD points at this high level would be totally inaudible.
As far as voltage goes the X26 Pro hits 5.1V whilst the D90 SE hits 5V so don't know where you were going there.
You also mention sound signatures???

 

[srkbear]

What a strange comment for this Forum..........
How did you manage to come to this conclusion?

How is it strange for this forum? I wasn’t attempting to be controversial. The Gustard x26 Pro is an outstanding DAC to be certain—and as you might know I almost owned one. But I have a headphone audio setup that is small in form factor (my amp is the iFi Pro iCAN Signature, which lines up perfectly with the Topping), and despite my admiration for the dual ES9038 Pros and toroidal transformers (the guts of the x26 Pro are a work of art), it’s hard for me to justify the extra $600 given the Topping’s superior measurements and my experience listening to it for several months.

As I’ve said elsewhere, I think the flagship price point for DACs has dropped over the past year. With recent offerings from Topping (and the new Gustard x18) leading the SINAD wars, I think it’s becoming increasingly difficult to justify shelling out more than a grand for a DAC—I think Amir has proven time and again that we’re perhaps reaching an audible technological limit, until perhaps AKM or ESS manages to produce a new chip that works miracles.

I’ve come to the conclusion that my money is better spent on headphones and amps, in that order. What are your thoughts on this?

 

[Belker]

I’m not sure what you mean by “direct DSD bypass”, unless you’re referring to a rarified and vanishing market of AKM-based RME ADI-2 DACs that processed DSD without an extra delta-sigma step—those DAC’s days are numbered unless AKM refreshes the 4493 and maintains this capability in its future iterations, but either way, is there much evidence that this method offers any salient advantages?

I’m also aware of the theoretical quantization noise reductions in 48k-based DSD sampling, although I would definitely say that DACs that offer this DSD architecture are exceedingly hard to come by (blame the original folks at Sony who decided that 64 was the best base for SACDs over 30 years ago). But again, with upscaling to DSD 256 or 512 readily available on most DACs in the D90se category, with quantization noise being pushed so far above frequencies that even dogs or moths can hear, is this too a practical concern?

It sounds like what you’re asking of your DACs is pretty esoteric—again I wonder whether pursuing such a DAC offers audible vs didactic advantages…

Yes, I’m referring to the AKM implementation. They are ramping up production now with successors, so let’s see what they can do. Whether these HQPlayer tricks are audible or not is of course an interesting question. I haven’t seen any controlled tests. But let’s not derail this thread with that…

 

[Miska]

I changed part of this, and if there’s know issues I would like to know about it.

Here's an example what I get on DAC output when I input -63 dBFS 1 kHz tone at 705.6k rate:

Note that this doesn't happen for example with -3 dBFS input level. But the same happens on 0 - 22.05 kHz linear sweep at -3 dBFS level once input reaches about 10 kHz frequency. This same behavior also happens with DSD input rates. Also 352.8k PCM input misbehaves depending on input signal causing wide band dirty noise floor.

While SMSL M500 mkII with same DAC chip doesn't have this behavior. But there seems to be difference that I suspect D90SE runs the DAC chip in asynchronous clock mode (ASRC on) while SMSL seems to run it synchronous clock mode (ASRC off).

I have now tested D90SE connected to two different computers (Intel and AMD) and with three different measurement devices. At home and at office...

I will later test it through I2S input too, to check if it is misbehavior of the USB interface (wouldn't be surprising for MQA-enabled XMOS interfaces).

 

[srkbear]

How did you manage to come to this conclusion?
Surely a couple of SINAD points at this high level would be totally inaudible.
As far as voltage goes the X26 Pro hits 5.1V whilst the D90 SE hits 5V so don't know where you were going there.
You also mention sound signatures???

Amir demonstrated that the x26 pro output lower practical voltages via XLR compared to the Topping—his results are posted elsewhere on here. Side by side the X26 pro definitely is the quieter DAC in an audible way. And by “sound signature” I just meant overall sound quality and detail retrieval—I shouldn’t have suggested tone.

Other than the differences in voltage output, I agree that the two DACs would be indistinguishable in terms of sound quality—which is my overriding point in terms of cost.

 

[srkbear]

Yes, I’m referring to the AKM implementation. They are ramping up production now with successors, so let’s see what they can do. Whether these HQPlayer tricks are audible or not is of course an interesting question. I haven’t seen any controlled tests. But let’s not derail this thread with that…

I certainly agree that that is a separate topic, as long as we don’t fault the D90se for lacking DSD Direct or 48-base DSD decoding on its own thread