Topping Centaurus R2R DAC Review

[MBO]

It's practically impossible to make a 24bit resistor ladder, because the precision needed of the resistors is so insane... That's why Delta/sigma 1bit DACs exist, and then oversample and noise shape for the rest ... 16 pure bits is still superior... Less than 0.000006% tolerance is needed for resistors in 24bit ladder. For 16bit 0.0015% (15 ppm), which is doable. That is why a 20bit R2R chip like the TI used in the Schiit MIB is quite incredible, using two per channel they can achieve 21 pure bits and also error correction... There also aren't any R2R ADC's, as this cannot be done - the equivalent is a SAR ADC, which is never used in recording studios (like LTC2380-24).

So, the best way is a discrete FIR DAC? It seems to combine all the best aspects of R2R and 1- Bit DACs??

 

[AdamFrandsen]

So, the best way is a discrete FIR DAC? It seems to combine all the best aspects of R2R and 1- Bit DACs??

FIR (Finite impulse response) is a filter that can be used pre DAC to alter the digital signal (high or low pass filter, oversample) it can also be used to downsample the signal, but it has nothing to do with a specific type of DAC technology, and discrete can be beneficial in some cases others not necessarily - as I just mentioned the TI R2R chip is by definition not discrete, but incredible engineering and superior to what can be achieved with discrete designs.

You may think of Chords FPGA approach, where the FPGA becomes the chip itself, this may be a great approach, but I am not sure of all the details - it uses pulse array as the conversion stage and has a WTA FIR filter, which is designed to maximise transient accuracy - exactly how they do this, I am not sure…But it is still a noise-shaped system, not a single on/off pulse array, maybe a handful of comparable raw bits, but likely not R2R level…

 

[NTTY]

This is the supposed performance of the Schiit using the newest TI R2R chip, again I would love to see it tested here as well! It uses 2 chips per channel, which with meticulous implementation could achieve 21bits of pure R2R... To me the Centaurus looks like it uses 16bit ladders from pictures of the board, but I am no expert in spotting this... Does anyone know if the additional bit depth up to 32bit is achieved by noise shaping and oversampling in the topping or simply truncating the additional bits?View attachment 416918

Now that I went for the Cosmos suite of devices for high precision measurements, I’d like to test it, if it wasn’t so expensive.

 

[KSTR]

FIR (Finite impulse response) is a filter that can be used pre DAC to alter the digital signal (high or low pass filter, oversample) it can also be used to downsample the signal, but it has nothing to do with a specific type of DAC technology, [...]

As a side note, FIR is a general technique and can be used in many contexts and one of them is DAC hardware, see for example this already famous thread about a discrete FIR DAC for DSD input: https://www.diyaudio.com/community/threads/return-to-zero-shift-register-firdac.379406/
The FIR filter *is* the output stage in this case, providing a sliding average (a simple form of a FIR filter).
It sure is debatable if we want to call such a DSD FIR DAC already a specific type of DAC technlology,

 

[AdamFrandsen]

Now that I went for the Cosmos suite of devices for high precision measurements, I’d like to test it, if it wasn’t so expensive.

It is indeed super expensive, but not really that surprising - the dac chips alone cost around 400 bucks

 

[lewdish]

I just wanna know if Topping can make a tube camp that measures well

 

[BeeKay]

I just wanna know if Topping can make a tube camp that measures well

They probably can. Ideally a class-D design with a tube-DSP option and fake valves (LED illuminated) sticking through the top.
Nah, would not expect them to come up with that. Too many competitive Chinese brands around doing exactly this and not much to win?

 

[Anonymous_Theos]

Wouldn't it be better to just go with a TOPPING D70 pro saber compared to this?

 

[oleg87]

Wouldn't it be better to just go with a TOPPING D70 pro saber compared to this?

Don't think that one has PEQ, does it? $700 is a pretty steep price for a pretty basic DAC.

 

[Anonymous_Theos]

Don't think that one has PEQ, does it? $700 is a pretty steep price for a pretty basic DAC.

I’m not aware if it have peq but apps sells certified refurbished D70 saber’s for under 500 that’s why I’m curious if I should consider one over the other i don’t mind using my eq software on pc

 

[unpluggged]

I’m not aware if it have peq but apps sells certified refurbished D70

"Certified" by whom?

 

[Anonymous_Theos]

"Certified" by whom?

Apos

 

[unpluggged]

Apos

What do they "certify" exactly? Do Apos have an approved certification lab? And it's funny considering even the new Toppings out of the factory are not certified by anyone (which aligns pretty well with their real-world reliability and safety record).

 

[Chagall]

What do they "certify" exactly? Do Apos have an approved certification lab? And it's funny considering even the new Toppings out of the factory are not certified by anyone (which aligns pretty well with their real-world reliability and safety record).

Instead of badgering a new member you can go on Apos and find out. Also, contact Topping support as to why they don't certify (according to you) their products.

 

[unpluggged]

Instead of badgering a new member you can go on Apos and find out. Also, contact Topping support as to why they don't certify (according to you) their products.

Oh, yeah. What else will you tell me to do?

And at least I know why the Guangzhou guys don't certify their gear: because nobody enforces them to do so. Good for them.

For the rest of us, it's not so good:

 

[Chagall]

Oh, yeah. What else will you tell me to do?

And at least I know why the Guangzhou guys don't certify their gear: because nobody enforces them to do so. Good for them.

For the rest of us, it's not so good:

View attachment 420311

Wishing you all the best, hoping it's just a teenage phase.

 

[Postlan]

Why do some people say that NOS sounds better? Can it be explained with any measurable data?

 

[Purité Audio]

I think it’s mainly older listeners who may have grown up with vinyl and prefer the distortion added by NOS/valve designs.
Keith

 

[Koeitje]

I think it’s mainly older listeners who may have grown up with vinyl and prefer the distortion added by NOS/valve designs.
Keith

Its mostly just placebo.

 

[MaxwellsEq]

Why do some people say that NOS sounds better?

This has been covered at length in this thread and throughout ASR.

To be clear there 4 top-level states
1) non-oversampling DAC with proper reconstruction filter (measures OK, sounds correct)
2) non-oversampling DAC with non-proper reconstruction filter (measures bad, sounds wrong, but some people prefer it)
3) oversampling DAC with proper reconstruction filter (measures OK, sounds correct)
4) oversampling DAC with pseudo-non-proper reconstruction filter (measures bad, sounds wrong, but some people prefer it)

1 and 3 measure well and sound correct. So you can have good measuring and sounding oversampling and good non-oversampling DACs. To be good, they must have an effective reconstruction filter.

2 and 4 measure wrong and sound wrong, but some people prefer them. The DACs are wrong because they are missing the essential reconstruction filter. Why do people like them? There's research that suggests some people have a preference for noisy distorted sound.