TOPPING D90 III Sabre DAC Review

[zhgutov]

How good D90 III is at 192/384/768 kHz?

 

[Bleib]

How good D90 III is at 192/384/768 kHz?

Even at 16/44.1 better than human hearing or any loudspeakers. Most recordings are not HQ either

 

[Billy Budapest]

I see your D70 Pro and I raise you the SMSL SU-9 Pro. Same DAC chip, also balanced with 5.2v RMS output, MQA, bluetooth, $500. Some input differences though.

MQA is seen as a liability rather than a feature by many people here.

Given the dearth of MQA material, I see it more as superfluous and imposing licensing fees that I would rather not pay.

 

[theREALdotnet]

If two DACs sound the same, what is the value of buying the one that measures "best"?

You shouldn’t. You should just make sure that both DACs have state of the art performance, then select one based on how well it fits your requirements with regard to connectivity and functionality. You should research how mature and reliable its software/firmware is, whether it has unwelcome quirks (like clicking on sample rate changes), whether it messes with digital volume control settings without your input, whether it’s prone to fail within a few months, what the support from its maker is like, etc.

 

[BrooklynNick]

You shouldn’t. You should just make sure that both DACs have state of the art performance, then select one based on how well it fits your requirments with regard to connectivity and functionality. You should research how mature and reliable its software/firmware is, whether it has unwelcome quirks (like clicking on sample rate changes), whether it messes with digital volume control settings without your input, whether it’s prone to fail within a few months, what the support from its maker is like, etc.

It was a rhetorical question that you are removing from its context, but I basically agree with your response.

 

[Rhamnetin]

MQA is seen as a liability rather than a feature by many people here.

Given the dearth of MQA material, I see it more as superfluous and imposing licensing fees that I would rather not pay.

Yeah I despise MQA myself. I just mentioned it because it is a feature difference between them. SMSL is just insane for having 3 of the same DACs: SU-9 Pro, DO400 (this one has a headphone amp at least so it's justified), and D400ES (seems to cost more than the DO400 for some reason).

 

[MacCali]

So with 5 volts the performance will increase or decrease?

 

[zhgutov]

Even at 16/44.1 better than human hearing or any loudspeakers. Most recordings are not HQ either

This does not mean it will work in 768K mode at least as good as in 16/44.1 mode, maybe even much worse. That's why I'm asking.

 

[staticV3]

So with 5 volts the performance will increase or decrease?

It'll increase by a few dB.

Both because the DAC's SNR will rise by 2dB due to the higher voltage, and because Amir's APx555B is better at measuring 5V instead of 4V:

 

[Billy Budapest]

Yeah I despise MQA myself. I just mentioned it because it is a feature difference between them. SMSL is just insane for having 3 of the same DACs: SU-9 Pro, DO400 (this one has a headphone amp at least so it's justified), and D400ES (seems to cost more than the DO400 for some reason).

I almost see DSD support in a similar way. It is inferior to high-bitrate PCM by almost every metric, and there isn’t a whole lot of DSD media available (other than SACD). However, I am not aware of any licensing fees associated with including DSD conversion in a DAC so I am not opposed to it, even if it is something I will never use.

 

[welwynnick]

I almost see DSD support in a similar way. It is inferior to high-bitrate PCM by almost every metric, and there isn’t a whole lot of DSD media available (other than SACD). However, I am not aware of any licensing fees associated with including DSD conversion in a DAC so I am not opposed to it, even if it is something I will never use.

I'm a bit surprised to hear that. This is why.
In my experience, when any audio device digitises an analogue input, the SQ takes a big hit compared with analogue pass through.
I have two integrated amps that are almost identical - the older one has a DSD ADC on the analogue inputs, and the newer one has a PCM ADC.
They both perform identically with digital inputs, but with analogue, the newer one sounds significantly worse.
The DSD input sounds almost like its not digitised at all, it's that much better.

 

[zhgutov]

I'm a bit surprised to hear that. This is why.
In my experience, when any audio device digitises an analogue input, the SQ takes a big hit compared with analogue pass through.
I have two integrated amps that are almost identical - the older one has a DSD ADC on the analogue inputs, and the newer one has a PCM ADC.
They both perform identically with digital inputs, but with analogue, the newer one sounds significantly worse.
The DSD input sounds almost like its not digitised at all, it's that much better.

Maybe you DSD ADC is just better than your PCM ADC?
I cannot hear any difference when I pass some highres signal through my PCM ADC.

 

[Hifihedgehog]

I have been mulling over this DAC for a couple weeks, and I do not think it is definitively better than the Topping D70 Pro Sabre even from a purely objective sense.

My argument is simple:

While the D90 III's SINAD on its own is a smidge better, 123.8 dB versus 123.4 dB....

....the D70 Pro Sabre beats it in several other important areas including:

1. The D70 Pro Sabre offers an impressive 23 bits of distortion free range with a 192 Khz 32-tone multitone signal versus the 21.7 bits of the D90 III.

2. The D70 Pro Sabre offers lower distortion (flat across the audible spectrum) with its default filter, versus the D90 III which has a visible rise above 8 KHz.

Add to that that the D70 Pro Sabre has a spectrum analyzer and beautiful color display and interface. Subjective, yes, but consider this. Suppose you are presented with these two DACs, one with very, very, very slightly better SINAD while the other has a bit better multitone and filter distortion/noise performance. Then this other has also a modern user interface which is a huge plus over the blatantly 90s digital alarm clock design of the flagship. Suddenly, the D90 III really isn't that compelling and maybe even shouldn't have gotten the flagship pricing treatment.

Indeed, the D90 III has a technically world-class SINAD but Topping really needs to stop playing this one-trick-pony game of SINAD. Otherwise, they will lose sight of other key objective measures and other extremely important features in I/O and UI that users also value. And before they know it, they will get passed up by competitors like up-and-comer EverSolo who has truly upped the ante in integration, I/O, and UI.

As always, feel free to correct me. We are all here to learn and have fun together!

 

[Rhamnetin]

I'm inclined to agree with you. Those differences are particularly interesting because both use the same DAC. That rise with frequency appears in Topping's own measurements, but to a lesser degree.

The SMSL D400ES (D70 Pro Saber competitor) behaves more like the D70 here too in that it's flatter. SMSL sadly doesn't provide this measurement themselves - I'm curious how their competitor to this, the SU-X, performs here. I would have eventually offered to send mine to Amir to review, but it already kicked the bucket.

- Edit: This would follow the pattern going on with Topping's headphone amps, where the A70 Pro is king.

 

[Hifihedgehog]

I would have eventually offered to send mine to Amir to review, but it already kicked the bucket.

I saw that part in the SU-X thread, which is most regrettable because honestly lifespan is the most important part since longevity should be at least 10 years easily in modern electronics. A month lifespan? That signifies serious issues in the soldering or assembly process or even the design itself.

 

[theREALdotnet]

As always, feel free to correct me. We are all here to learn and have fun together!

It really is a no-brainer, isn’t it? Even though the D90 III has dual DAC chips, this added expense doesn’t seem to yield any results. Hence, unless you need IIS input (who needs that?) or fancy the Bluetooth remote control, go with the D70 Pro.

 

[amirm]

How good D90 III is at 192/384/768 kHz?

Multitone test is at 192 kHz.

 

[zhgutov]

Multitone test is at 192 kHz.

I recently found very strange behavior on D90SE when I play high frequency signal (10K+) in one channel.
In this case I see a huge rise of the noise floor at 768 kHz (up to -50 dBFS!).
Something very similar happens at 384 kHz, but it is much lower.
This happens sometimes as a spike at 192 kHz.
44.1/48/88.2/96 kHz are good.
I'm wondering if my unit is broken, or is it the common behavior of D90xxxx line?

UPD. I can actually reproduce this mess with any input even at 192 kHz.
But the effect is not as bad as for 768 kHz.
And D90SE seems to be perfectly fine at 44.1/48/88.2/96 kHz.

 

[welwynnick]

I have been mulling over this DAC for a couple weeks, and I do not think it is definitively better than the Topping D70 Pro Sabre even from a purely objective sense.
My argument is simple:
While the D90 III's SINAD on its own is a smidge better, 123.8 dB versus 123.4 dB....
....the D70 Pro Sabre beats it in several other important areas including:
1. The D70 Pro Sabre offers an impressive 23 bits of distortion free range with a 192 Khz 32-tone multitone signal versus the 21.7 bits of the D90 III.

Firstly, I don't think that last bit is fair to the D90.3 because you need to judge the performance from the chart and not from the number of bits.
The bits equivalent, I believe, is just Amir eye-balling the chart and doing some quick mental arithmetic (divide by 6.02).
When you compare the charts, they're to all intents and purposes completely identical.
Second, I agree about the distortion vs frequency.
Third, looking at the 1kHz FFT, the SINAD is very similar as the noise floor is very similar. However the D90.3 distortion harmonics are lower.
Finally, because of the limitations of the APx555, you probably need to look at Topping's own measurements to make the best comparison.
Yes, it's splitting hairs and it's entirely academic, but do I think the D90.3 is probably the best, at least until Amir can test the SU-X.

 

[JEntwistle]

It really is a no-brainer, isn’t it? Even though the D90 III has dual DAC chips, this added expense doesn’t seem to yield any results. Hence, unless you need IIS input (who needs that?) or fancy the Bluetooth remote control, go with the D70 Pro.

I agree. It seems to be all about whether you need those specific features. I also wonder why they did not upgrade to the fancy display that the D70 Pro Sabre has. Is the IIS input a big thing to implement from an engineering perspective and/or does it preclude being able to add the display?

I'm struggling to understand why the D90 does not really seem like the flagship model, which you would expect to have everything.