Acuhorn R2R DACs

[The Computer Audiophile]

Anyone heard any of the Acuhorn DACs?

 

Acuhorn R2R XT

Acuhorn R2R T

Acuhorn R2R TS

 

 

Balanced DAC & DSD Converter with Class-A Big DHT Tube Stage

 

The Acuhorn R2R XT four original triode for balanced gain in class A. Why is the best sound now. The music is revealed from an unknown space. The best solution directly into the grid of tube from discrete resistors of ladder R2R Spartan 6. Playback main digital formats. Supports audio devices via USB audio or SPDIF. Balanced independent modern Vicor module power supply

features

R2R ladder 28 bit resolution 384Khz dac by Soekris. Loaded FPGA Spartan 6 technology oversampling up to 3.072Mhz. Frequency range +0.1dB -1.0dB 20hz - 20Khz. Total harmonic distortion + noise <0.008%. Clock jitter RMS 0.8 pS typical. Resistors precision 28 bit 0.02% resistors. Signal to noise ratio 20 Khz bandwith 129 dB unweighted. Input SPDIF up to 24 bit 192 Khz. Input I2S PCM up to 24 bit 384 Khz. Input I2S DSD up to DoP128 and DSD256.

Autoselect between the USB audio device and SPDIF inputs with priority SPDIF. Manual switch of input selection SPDIF coax or optic. RCA single ended tube output: 2.8V RMS Zout 600 ohm.

Digital input USB Type B audio adapter Amanero combo384 compatible with Mac OSX®, Linux and Microsoft Windows®. Digital input optical Toslink. Analog outputs XLR balanced and RCA.

Pure class-A single ended tube stage. Type 300B, PX25, PX4, 45, 2A3 power triode. Caddock High Performance Film Resistors. Separate dual mono power supply.

Power supply is the best Vicorpower modules. DC-DC converters provide high power density and low noise High frequency Mhz switching. Advanced power processing in a small size. Autoranging 90-264Vac network voltage, for worldwide use.

technical specifications
input voltage mains select 90-132 or 180-264Vac
maximum input power 50W
discrete R-2R sign magnitude DAC type
4 pin base tube socket U4A
directly heated triode type 5 Volt, 4 Volt or 2.5Volt
3 signal input: USB audio and SPDIF coax and Toslink
2 audio output XLR balanced and RCA
internal cable OCC copper
terminal rhodium
weight 2,2kg
dimensions W 320 x D 320 x H 28mm

[AudioDoctor]

Mmmmmm tubes...

[audioking87]

Says i2s but there's no input anywhere for it.

[tedwoods]

I had their speakers.

Very well built, balanced and a bargain for the asking price.

The DAC uses the Soekris module, also used by Lampizator in the older version of the Atlantic DAC, but the power supply uses DC to DC converters, not linear power.

At the DAC's asking price though, there is some serious competition out there...

 

 

[numlog]

On Soekris board there are switching regulator used for digital only, to power FPGA at least, possibly the clock too but maybe not, FPGA is lower on the list of places you need linear supply, and when designed for high end audio in mind the SMPS can be less harmful.... linear reg used for analogue.

 

That is in the defense of the humble Soekris R2R board...  all power to the R2R board and tubes comes from switching power modules anyway, this DAC is actually using cheaper OEM soekris board that uses higher jitter clock and lower tolerance resistor (0.02% vs 0.01%) and sells for around 250 euro... dont know enough about tubes to say but that asking price seems high

 

The soekris DAC have natural voltage output from the resistor ladder, so the only thing a tube output stage can really provide in the DAC is gain (at the cost of transparency).

[Account Closed]

One thing that is not possible is 28 bit resolution. That would imply a S/N ratio that is not possible. They say the S/N is 129 db which is a little more than 21 bit resolution which is the most of any DAC that I have ever seen measurements for. Most of even the highest of the high-end DACs top out at 20 bit resolution. More affordable DACs are in the 18-20 bit range.

 

Probably what they mean is that the R2R chip is capable of 28 bit resolution but that is meaningless as there is always the inherent noise created when moving electrons around (especially when tubes are involved). And besides there is no recorded music above 24 bit. 

 

I find that misleading.

 

 

[numlog]

56 minutes ago, bobflood said:

One thing that is not possible is 28 bit resolution. That would imply a S/N ratio that is not possible. They say the S/N is 129 db which is a little more than 21 bit resolution which is the most of any DAC that I have ever seen measurements for. Most of even the highest of the high-end DACs top out at 20 bit resolution. More affordable DACs are in the 18-20 bit range.

 

Probably what they mean is that the R2R chip is capable of 28 bit resolution but that is meaningless as there is always the inherent noise created when moving electrons around (especially when tubes are involved). And besides there is no recorded music above 24 bit. 

 

I find that misleading.

 

 

From my limited understanding it is a 28 bit R2R ladder,  it can process 28 bits of data, this is important to avoid digital processing error with upsampling and digital volume control. The DAC actually converting all those bit to analogue accurately is a different story

 

However ASR showed soekris having ''perfect'' 20bit resolution which is top of the line performance even for ''32 bit'' delta sigma

 

However #2 that SNR figure is probably true but actual resolution on the DAC output is determined by Signal to noise + THD, according to asr thats around 90dB which is not even 16 bit resolution... what the ''20 bit'' figure they measured above actually means in practice I have no idea.

How can 20 bits of resolution be measured when the DAC output is less than 16 bit?

 

[Account Closed]

1 hour ago, numlog said:

From my limited understanding it is a 28 bit R2R ladder,  it can process 28 bits of data, this is important to avoid digital processing error with upsampling and digital volume control. The DAC actually converting all those bit to analogue accurately is a different story

 

However ASR showed soekris having ''perfect'' 20bit resolution which is top of the line performance even for ''32 bit'' delta sigma

 

However #2 that SNR figure is probably true but actual resolution on the DAC output is determined by Signal to noise + THD, according to asr thats around 90dB which is not even 16 bit resolution... what the ''20 bit'' figure they measured above actually means in practice I have no idea.

How can 20 bits of resolution be measured when the DAC output is less than 16 bit?

 

I hadn't considered the use of digital volume control in the 28 bit number. That wasn't really clear. You are right about the SNR/THD number of course and whenever tubes are used in a DAC output stage the useful resolution measured in bits will be lower due to the inherent noise tubes create. Many people really like the sound of tubes (myself included) but there is a price to be paid. I have a tube pre after my all solid state DAC but I don't kid myself about the trade off

[Miska]

In most cases even if R2R is what ever bit, it makes sense to measure what real world word length gives best low level linearity and then use that instead.

 

For example with Holo Audio DACs, 20-bit TPDF dithered gives much better results than for example 24-bit data. And running it at 1.5 MHz 16-bit with 9th order noise shaping gives practically same SNR and linearity up to 100 kHz bandwidth as 20-bit TPDF at lower rates.

 

So it really depends and needs to be evaluated properly for each DAC array.

 

[Miska]

.

[numlog]

Have you found that with other R2R DACs? Holo spring uses a different technique to achieve good linearity, some kind of digital compensation, most DACs like Soekris seem to use some other method, based on PCM1704 DACs.

[The Computer Audiophile]

53 minutes ago, Miska said:

In most cases even if R2R is what ever bit, it makes sense to measure what real world word length gives best low level linearity and then use that instead.

 

For example with Holo Audio DACs, 20-bit TPDF dithered gives much better results than for example 24-bit data. And running it at 1.5 MHz 16-bit with 9th order noise shaping gives practically same SNR and linearity up to 100 kHz bandwidth as 20-bit TPDF at lower rates.

 

So it really depends and needs to be evaluated properly for each DAC array.

 

Excellent points. 
 

Looking at single numbers In a vacuum is always so misleading. 

[rickca]

6 minutes ago, The Computer Audiophile said:

Looking at single numbers In a vacuum is always so misleading. 

What, so I can't just pick the DAC with the highest SINAD? 😋

[tedwoods]

7 hours ago, numlog said:

On Soekris board there are switching regulator used for digital only, to power FPGA at least, possibly the clock too but maybe not, FPGA is lower on the list of places you need linear supply, and when designed for high end audio in mind the SMPS can be less harmful.... linear reg used for analogue.

 

That is in the defense of the humble Soekris R2R board...  all power to the R2R board and tubes comes from switching power modules anyway, this DAC is actually using cheaper OEM soekris board that uses higher jitter clock and lower tolerance resistor (0.02% vs 0.01%) and sells for around 250 euro... dont know enough about tubes to say but that asking price seems high

 

The soekris DAC have natural voltage output from the resistor ladder, so the only thing a tube output stage can really provide in the DAC is gain (at the cost of transparency).

I'm not knocking down the Soekris or smps as a rule.

It's the bang for the buck ratio that puzzles me and some of the literature.

For comparative purposes, the similarly priced full blown (balanced-ladder volume control, etc.) Lampizator Baltic looks like that :

 

[semente]

7 hours ago, rickca said:

What, so I can't just pick the DAC with the highest SINAD? 😋

 

Sinad? Isn't that a famous guy who almost died in a shipwreck?

[numlog]

22 hours ago, The Computer Audiophile said:

Excellent points. 
 

Looking at single numbers In a vacuum is always so misleading. 

what do you mean by this in relation to what he said?

 

Is looking at a single factor in a vacuum any different to a number? I believe dithering would impact another factor, SNR, for some reason I dont understand SNR or THD doesnt seem to directly relate to the linearity ''bits'' measurable on output but it would impact the dynamic range of the music.

 

If we are going to put a factor into a vacuum is linearity the right one? I have seen delta sigma DACs measured on asr that fall a few bits below 16 bit linearity and yet they were not deemed ''bad'', I dont* know enough about this measurement to understand why not.

 

And when it comes to actual sound quality Isnt putting these measurements as a whole in vacuum misleading since they are all mostly considered to be at inaudible levels?

There isnt single area R2R DACs can beat DS in measurements, so if measurements if a concern why bother with them at all

[Superdad]

On 11/30/2019 at 1:17 AM, semente said:

 

Sinad? Isn't that a famous guy who almost died in a shipwreck?

Sinad O’Connor almost died in a shipwreck?

Man, I’ve got to get better about keeping up with news of the world. Too much time spent on this audio forum.