(Unofficial) Topping D50 III DAC review and measurements

[staticV3]

My my question regarding PEQ is: How do I configure it in a Linux system such as Moode or Volumio? If it is a set and forget process I can do either via the remote/font panel or connecting once to a Windows machine, then I can get by. If not the PEQ is useless.

The PEQ can only be configured using "Topping Tune" for Windows.

It's set and forget. The PEQ settings are stored on device:

Also, read the user manual before asking!

 

[beefkabob]

Fantastic job!

 

[Phorize]

Nice work LDAC performance is impressive. Pity the balanced output level isn't higher for better fit with typical ncore/purifi power amps. Still alot of DAC for the money though.

 

[Rja4000]

Pity the balanced output level isn't higher for better fit with typical ncore/purifi power amps.

With 5V output, you just need 19dB gain on the power amp to reach 500W under 4 ohm.

 

[ArturoKiwi]

My my question regarding PEQ is: How do I configure it in a Linux system such as Moode or Volumio? If it is a set and forget process I can do either via the remote/font panel or connecting once to a Windows machine, then I can get by. If not the PEQ is useless.

In Moode you could use the internal brute fir.

 

[welwynnick]

Great DAC, great test, many thanks!

Makes it difficult to justify spending a lot of money.

Audio Precision or not, isn't that the highest SINAD ever measured on ASR?

 

[daniboun]

Good work nice dac )

 

[marcom22]

You did it an excellent work. Thanks to share it with us.

 

[Rja4000]

Audio Precision or not, isn't that the highest SINAD ever measured on ASR?

As I wrote above, it's not possible/wise to compare figures directly.
This is, by the way, why Audio Precision is so widely used: since it's a benchmark, to allow comparison.

 

[yanm]

Here is a review and measurements of the recent
Topping D50 III DAC

View attachment 363412

The D50 III is a small DAC.
From above, it looks like a square of approx 13x13cm .
I purchased it new from Audiophonics for around 230€ (pre-release discount included).
This is one of the cheapest DACs with balanced and unbalanced outputs, LDAC Bluetooth, SPDIF inputs, level control and remote.

It comes with usual connectors: USB, Optical and coaxial (RCA) SPDIF inputs, and of course a pair of unbalanced analog RCA outputs.
On top of that, it adds a pair of TRS balanced outputs.

View attachment 363413
In the box, there is also a (bluetooth + IR) BC-15A remote, but NO power supply.
This for sure helped to lower the price.
The DAC may be powered from any USB source, so just a USB power cable is provided.
It worked for me from any USB power source I had at hand, including a power bank.

Interestingly, it also has a 12V Trigger output, to remotely power on an external power amplifier, as an example.
This is a very welcome addition for me.

I didn't find any "fixed level output mode". So everything here is done at 0dB level unless specified.

The display is quite small, as you may see, and hardly readable from a distance.

The remote allows to set the volume, select the digital input, the analog output, the filter, chose between 2 presets, mute and display light level.

Topping says it's built around a pair of ES9039EQ2M DAC chips.

Important notice:
I measured the D50 III with firmware 1.21.
Previous versions of firmware have a very serious issue. (Read more here)
If you receive your D50 III, check the firmware version and upgrade to the last version before doing anything else.


Measurements

Measures are performed with following equipment:
RME ADI-2/4 Pro SE, as the main ADC
For very low levels, I used the RME 12Mic-D Mic preamp/interface or the RME UFX III interface.
For this review, I used a trial version of Audio Precision APx 500 Flex 8.1.2, where possible
(What's better than actual measurements to test and understand a software ?)
All other measurements where done with my ususal software, Virtins Multi Instrument (now version 3.9.9.6).


There are some limitations and differences compared to Amir's testing with an Audio Precision APx555.
That's more than enough to get a good idea of the device performances though.
I'll post comments where such differences are most important.



SINAD measurement

NB: I don't own an Audio Precision analyzer.
Conditions are therefore not exactly the same, so please don't compare this directly to Amir's figures.

TRS Balanced outputs 4V Dashboard

View attachment 363415

For this measurement, I used the APx 500 Flex software, with the RME ADI-2/4 Pro SE and an external E1DA APU notch filter.
I can only measure one channeml at a time,
This is left channel. Right channel is 0.14dB lower.


TRS Balanced outputs 5V Dashboard

At 5V, we get higher SINAD

View attachment 363416


RCA Unbalanced outputs 4V Dashboard

View attachment 363417


Bluetooth LDAC to TRS outputs 4V Dashboard

While at it, I also measured with Bluetooth LDAC

View attachment 363462

Limitations we see here are typical of the LDAC Codec.


Dynamic range

Due to the way I perform this measurement, I was not able to use Audio Precision software directly for this measurement.
However, I now use the same CCIR-2k weighting AP uses.
NB: I forgot to re-measure Voltage. So voltage shown is not accurate. Should be around 4mV for -60dB at 4V.

The level you see in the plot is as seen by the RME UFX III I use as ADC for this measurement. It is set for 60dB gain, 0dBFS here is around -41.5dBu.

View attachment 363444


View attachment 363447

Note that 132dB with A-Weighting is the figure Topping publishes in the D50 III Specs.


Linearity

View attachment 363450

Thanks to Amir, I was able to use his own project/template.
Values shoudl therefore be more in line with his results.


Multitone Measurements

View attachment 363451

Text-book perfect


Sine 50Hz under 600 Ohm load measurement
(Torture test)

I ran this new test Amir added in the recent DAC measurements.

I run this test with an home-made 600 Ohm load.

View attachment 363456


Output impedance

TRS (balanced) output

Open circuit Voltage	V1	3,9809​	Vrms
Load resistance	RL	614,3​	Ohm
Loaded circuit Voltage	V2	3,4292​	Vrms
Zo=RL*(V1/V2-1)	Zo	98,8​	Ohm

RCA (inbalanced) output

Open circuit Voltage	V1	2,0308​	Vrms
Load resistance	RL	614,2​	Ohm
Loaded circuit Voltage	V2	1,8753​	Vrms
Zo=RL*(V1/V2-1)	Zo	50,9​	Ohm

(Reference)

Measured at 997Hz only, with a calibrated Brymen BM869s


IMD measurement

View attachment 363460

Excellent results: Very low noise and just a hint of distortion close to 0dBFS

A few comments about this measurement:
This should be much closer to Amir's measurement with Audio Precision than what I was measuring before.
AP is using IMD+Noise, not just IMD.


Jitter

View attachment 363461

Jitter performance is very close to text-book perfect !
All plots perfectly superpose for a perfect picture, whatever the source.


Filters comparison

View attachment 363490

View attachment 363491

Default filter is Mode 3 (the red plot), which is the one I've used for all other measurements.

Filters Impulse responses

I recorded each Filter's impulse timeplot
From which we may deduct each filter's name:

Mode 1: Minimum phase
Mode 2: Linear phase apodizing fast roll-off
Mode 3: Linear phase fast roll-off
Mode 4: Linear phase fast roll-off low ripple
Mode 5: Linear phase slow roll-off
Mode 6: Minimum phase fast roll-off
Mode 7: Minimum phase slow roll-off
Mode 8: Minimum phase slow roll-off low dispersion:

View attachment 363636View attachment 363637View attachment 363638View attachment 363639

View attachment 363640View attachment 363641View attachment 363642View attachment 363643

Measured with RME ADI-2/4 Pro se @768kHz with Virtins MultiInstrument 3.9.9.6
See also the "ESS ES9039Q2M data sheet"


THD+N ratio vs Frequency (90kHz BW)

View attachment 363671
Here is what the noise at higher frequencies looks like:

View attachment 363678


Intersample overs

As I usually do now, I tested for intersample overs, up to +3dB overload, and distortion is showing only above +2.8dB or so.
This is excellent performance, and much better than the other Topping ESS DACs I have measured, almost on par with the RME performance.

I don't know if Intersample overshoots are actually a problem in real life, but not with this DAC anyway.

View attachment 363681


Parametric EQ functionality (USB input only)

The EQ requires an application called "Topping Tune"

View attachment 363683

It allows configuration of parametric EQ
Apparently, you may store various configurations locally (on your computer) or on the device.
You may have up to 10 bands in each Parametric EQ configuration.
And you may store up to 5 different configs in the device.

I initially had a short test of EQ, to check if that was working when using another source than USB - it does not.
I also checked if that was working when using USB from my Android smartphone - and it does.

Now, playing more with the App, I have to say it's quite buggy at that stage (v1.02).
It happens very often that you change a setting in the app on a device config but, even if it shows the proper curve on the app display, it's not actually activated in the device !
So it's probably better to configure your EQ locally and then export it to your device.

I've done a quick test to check how it matches RME EQ, as an example.

View attachment 363684

As you may see, it matches pretty well.
This is here a 3 band PEQ test, with maximum gain (+12dB) and attenuation (-12dB) and different Q for each frequency.
The main mismatch is that the 100Hz peak on Topping seems to be at 105Hz instead.


Conclusion

This DAC is really impressive !
It ticks most boxes in my list for a very reasonable price.
Including the trigger output and the Bluetooth remote (a clear improvement on IR remote).

The parametric EQ is also a very nice addition.
Too bad it's stricly limited to use with USB input.


What more could you want?
Add a fully functional preamp functionality: with control for Balance, Mono,...
and of course a parametric EQ for all inputs, and our quest for the perfect, affordable DAC would then be over.
Will that be for the version IV ?

Anyway, as is, this DAC is a terrific value for the money !


For me,
This is just excellent !

Thanks for the review @Rja4000 !

For me, this is also excellent. I’ve finally a reason to upgrading my topping E50. Room correction was missing on my audio chain. It is pity that the PEQ software is only running on Windows but I would rather that Topping iron out the software bug before porting it to MacOS.

Have you measured power consumption? I assume that’s less than 2W as it seems to be able to be powered via a standard USB-A… I am asking because power consumption was one of the reasons I did not splurge for a RME DAC (the ADI-2 DAC FS is spec‘ed 7W idle and 18W full load) - the other, was I could not justify the cost for just adding room correction.

 

[Rja4000]

Have you measured power consumption?

No, I didn't.
But I can do it.

I did not splurge for a RME DAC (the ADI-2 DAC FS is spec‘ed 7W idle and 18W full load) - the other, was I could not justify the cost for just adding room correction.

The RME adds way more than that, IMO.
It remains a reference to beat for quality, flexibility and functionalities.
The price is different, for sure.

 

[yanm]

The RME adds way more than that, IMO.
It remains a reference to beat for quality, flexibility and functionalities.
The price is different, for sure.

True, it’s a professional device and the build quality and features look incredible… but an overkill just for my use case.

 

[yanm]

No, I didn't.
But I can do it.

Would be nice to see if the DAC can be powered by 500mA-limited computer USB port or by a phone charger. On the other hand, that should be stated in the spec (I didn’t found those figures but I’ve only cursorily glanced at Topping webpage).

 

[restorer-john]

Also, read the user manual before asking!

Heil, mein Führer!

Are you serious? He's just asking a question!

 

[Andreas007]

Heil, mein Führer!

Are you serious? He's just asking a question!

Are you serious?
I get the joke.
But as a German it has serious impact, still.

 

[Ahmonge]

Great job, @Rja4000! Would it be possible to run the 16 vs 24 bits at -90 DBFS with a higher fequency, let’s say 10,000 Hz?

 

[Palladium]

Nice work LDAC performance is impressive. Pity the balanced output level isn't higher for better fit with typical ncore/purifi power amps. Still alot of DAC for the money though.

What world are you in where 5V RMS isn't high enough?

My D50 III is currently shipping enroute to replace the Fiio BTA30 Pro, whee.

 

[Aperiodic]

Intersample overs

As I usually do now, I tested for intersample overs, up to +3dB overload, and distortion is showing only above +2.8dB or so.
This is excellent performance, and much better than the other Topping ESS DACs I have measured, almost on par with the RME performance.

I don't know if Intersample overshoots are actually a problem in real life, but not with this DAC anyway.

A link to a paper on the nature and significance of intersample overs, published by Benchmark:

Intersample Overs in CD Recordings

Intersample Overs are a Common Occurrence in CD Recordings We have frequently used Steely Dan's Gaslighting Abbie from Two Against Nature in our listening tests. This is a spectacular CD recording with lots of dynamics and a low noise floor. Nevertheless, in a little over 5 minutes, this track...

benchmarkmedia.com

Note that the content relates to the Redbook standard but findings may shed some light on the high res vs not debate. (According to Benchmark, the most severe effects are at 1/4 the sampling rate, which would be 11.025 kHz for a 44.1 source- well within the ‘standard’ definition of the human hearing range. 96k would peak at 24k (above ’audibility’) and so on.
The paper is from 2017 so a lot has doubtless changed as far as hardware but it would be interesting to see by how much. And I suspect that the statement that most listening is still Redbook or lower spec is probably still true.

I would like to see a random test of this parameter alone on a random sampling of DACs to see how much variation there is between older/newer, cheaper / costlier, this chip / that chip etc and how it relates (or not) to other performance characteristics. Maybe somebody with the gear, the skills, and the time would be interested in taking on such a project, or perhaps Amir would even consider adding it to the test regimen. I’d donate my “$9 phone dongle” which, being a ~6yo “el cheapo” implementation should not hang with today’s stuff on the bench even though it’s subjectIvey satisfactory.

 

[Rja4000]

I assume that’s less than 2W

2.19W

Would be nice to see if the DAC can be powered by 500mA-limited computer USB port

<0.45A

By the way, my RME ADI-2/4 Pro SE:
Idle: 0.22W
Loaded (Stereo 32 ohm Headphones delivering 2x1.9W): 14.49W

 

[Multicore]

It allows configuration of parametric EQ
Apparently, you may store various configurations locally (on your computer) or on the device.
You may have up to 10 bands in each Parametric EQ configuration.
And you may store up to 5 different configs in the device.

Dang! Talk about icing on a cake. And cute too: Topping Tune.