Gustard R26 new R2R DAC

[davide256]

On 12/3/2022 at 9:59 AM, Qstik said:

Yes, I ordered directly from ShenzhenAudio at the Black Friday price + free shipping on Nov. 28.  I received it 4 days later via DHL in the Detroit area !!!  Not a single mark on the package.

 

It's burning in right now, but so far, it already sounds fabulous over I2S compared to my Metrum Onyx with updated DAC3 modules.  Settings are:  PCM Filter "Mid" , PCM NOS "ON" , DSD Direct "ON"

Thanks for your post! Still have a Metrum Octave that a Denafrips Ares II couldn't unseat so if you prefer it to the later model Metrum Onyx, that says quite a lot.

[GUTB]

I was interested in this DAC -- until I saw that it uses opamps on the output. I believe that all serious DACs have discrete output stages.

[rafa]

52 minutes ago, GUTB said:

I believe that

This what we believe often leads as to wrong conclusions ;) There is a lot of opamps in the hiend gear on the marked. I mean real hiend gear :)

Presents of opamps doesn't tell as anything because a lot depends on implementation/ whole design of the circuit pcb.

[gstew]

@GUTB curious where you got that info.

 

Gustard's product blurb on the output stage...

 

"LPF is specifically designed for the D/A converter using discrete devices. Final parameters of the circuit are adjusted with subjective listening. Different from IC s, discrete circuits provide the possibilities to control every detail in sound and performance. But the parasitic effects introduced huge challenges to development experience and device' s quality."

 

TIA!

 

Greg in Mississippi 

[GUTB]

On 12/17/2022 at 2:16 PM, gstew said:

@GUTB curious where you got that info.

 

Gustard's product blurb on the output stage...

 

"LPF is specifically designed for the D/A converter using discrete devices. Final parameters of the circuit are adjusted with subjective listening. Different from IC s, discrete circuits provide the possibilities to control every detail in sound and performance. But the parasitic effects introduced huge challenges to development experience and device' s quality."

 

TIA!

 

Greg in Mississippi 

 

I learned about these topics over the years from various high-quality sources (including Mishka and others on this forum). 

 

LPF (Low Pass Filter) may be done by a discrete circuit, but the signal amping is (probably, see: "hybrid") done with opamps. Why do they use opamps when discrete sounds better? In my opinion it's because these companies don't want to spend the engineering budget to put together a high-quality amp circuit. Additionally, companies who serve the budget end of the hobby / lifestyle know their customers aren't audiophiles to begin with and like to consume marketing based on measurements. A good opamp should measure better than a discrete circuit. 

[seeteeyou]

A New Point Of Reference - Gustard R26 Discrete R-2R Ladder DAC Review
https://soundnews.net/sources/dacs/a-new-point-of-reference-gustard-r26-discrete-r-2r-ladder-dac-review/

Quote

I wanted to be sure that at such prices, you could still find linear transformers (plural), overkill filtering and regulation stages, while swapping op-amps with an all-discrete output stage (hence, its name), an impressive digital board and clocking system.

 

Quote

Last, but certainly not least, you can spot a fully discrete output (LPF) stage consisting of hundreds of discrete devices that are making me warmer inside, knowing their unlimited potential.

 

---

  

On 11/26/2022 at 11:16 AM, gstew said:

Just ordered one via Amazon to take advantage of the $250 off Black Friday price.

 

Just wondering if you're having any experiences with the NOS mode of R26 yet? Quite a few PGGB P128 / P256 clips could be found below

 

https://audiowise-canada.myshopify.com/collections/pggb

https://audiowise-canada.myshopify.com/blogs/news/pggb-adds-high-precision-modes

[GUTB]

Actually I double-checked and yes the R26 uses discrete output circuits. I seemed to have confused it with the X16 which does use a hybrid output.

[gstew]

On 12/19/2022 at 11:48 PM, seeteeyou said:

<SNIP>

 

Just wondering if you're having any experiences with the NOS mode of R26 yet? Quite a few PGGB P128 / P256 clips could be found below

 

https://audiowise-canada.myshopify.com/collections/pggb

https://audiowise-canada.myshopify.com/blogs/news/pggb-adds-high-precision-modes

@seeteeyou,

 

No I do not. BUT MANY THANKS for this link, I have been wondering what next on the R26 besides letting it continue its break-in. Trying well-upsampled materials (like from PGGB) has been on my list! I'll be getting some of my reference tracks upsampled to 384 (the max the SDTrans will handle).

 

I have moved from my stock "spare" SDTrans384 to the one I modified with LT3045 regulator boards from LDOVR.com replacing the older stock regulators. That was a nice upgrade when it was feeding my modified Soekris DAM 1021 and was here too, feeding the R26 from its I2S-over-HDMI output. At some point I'll have to try it with IanCanada's new-ish HDMIPi Pro, where he uses discrete LVDS circuitry for the lowest possible jitter.

 

Ian believes his latest HDMIPi Pro may be the lowest jitter source available of this type. I'll be sending the R26 to him to try and also compare his HDMIPiPro against my stock SDTrans384 (and the R26 against the Holo Spring 3 he recently purchased to try an R2R DAC).

 

R26 impressions... with about 2 1/2 weeks break-in, it is getting good and promising. I don't have any comparable commercial DACs for comparison, BUT my DIY'd setup using an Allo USBBridge Signature, Ian's FiFoPi Q7 (recently upgraded to this), his ES9038Q2M Dual Mono DAC, and his recently released OPA861 I/V stage with seriously over-the-top power supplies is ahead of it in all areas, including detailing, resolution, spatial presentation, and ability to keep instruments and voices separate especially when the music gets complex. 

 

I'll give the R26 at least 4 weeks of constant run-time before I form any lasting impressions. It has continually and significantly improved significantly since new and cold. AND while I haven't fired it up to compare yet, from memory it is well ahead of my modified Soekris DAM 1021 setup.

 

I'll also compare the I2S-over-HDMI osurce to USB USB from a Sonore UltraRendu I have here.

 

Greg in Mississippi

[gstew]

@GUTB,

 

Thanks for this response. I generally agree with what you wrote, though I have heard DACs using discrete opamps (like the licensed Sparkos Lab circuit used in the Allo.com Katana) that compete with non-opamp discrete and/or the best transformer output stages.

 

My take is that Gustard has incrementally improved their DAC systems' designs and each successive model builds on the next, so where they had opamp-only output stages at first they now include discrete building blocks.

 

OTOH, you may be right on their use of opamps in the output stage of the R26. I took closeup pix of much of the interior before I put it in one of my setups for break-in and audition to assess the potential for DIY improvements. I saw a couple of opamps in the output stage area, OPA1642s (IMHO a fairly good sounding one).

 

How they affect the final result will of course be dependent on how they are used... I am GUESSING that the R2R ladder output is voltage similar to that of the Soekris circuits, so no I/V conversion would be needed (which is where opamps have particular challenges though again, with proper design and implementation AND with the use of discrete opamps, I have heard some opamp I/V stages sound reasonably good). Perhaps they are used as a buffers. Perhaps as DC servos. IDK. I doubt that Gustard will release the circuit (very interested in seeing it if they do), but their effect may be large or not.

 

I posted my current assessment of the unit in a previous post. While I don't consider the R26 to be fully broken-in at this point, it was a fairly good assessment, though my super-over-the top DIY setup was IMHO better in pretty much all parameters. AND in all fairness, my DIY setup, with the music server, networking gear, and the over-the-top power supplies (12 in toto), would likely be a $10k USD commercial unit.

 

Greg in Mississippi

 

P.S. DIY improvement potential assessment of the R26... not much to do nor much that can be done. The unit appears to be well though-out, designed, and implemented and likely a bargain for its price point. The overall integration is VERY tight and because of that, there are not a lot of opportunities for improvement without a significant chance of degrading the sonics by disrupting their system design (or out and out breaking the unit). To contrast, OTOH, I found the Sony HAP Z1-ES a very fertile unit for DIY improvements, many of which I documented on a thread on that unit at DIYAudio.

[Dandou]

I received my R26 about three weeks ago, and it is still breaking-in. 

I'm surprised by the sound quality when I use the LAN input of its built-in streamer. The result is superior to a much more sophisticated streaming setup, combined to an Audio-GD DI-20HE, that I've been using, and that I meticulously optimized over time. 

 

Full details about this streaming setup on my profile: 

 

https://audiophilestyle.com/profile/45620-dandou/?tab=field_core_pfield_3

 

[botrytis]

Anyone try ver 1.4 out there.

[winged creature]

2 hours ago, botrytis said:

Anyone try ver 1.4 out there.

 

Yep I'm using 1.4, works great.  Apparently it gets rid of DSD pop sounds.  I find that internal streamer is better than the external options i fed it.

[botrytis]

I  at the 'ain't broke, don't mess with it' frame of mind

 

[mikeypas]

On 12/21/2022 at 7:52 PM, Dandou said:

I received my R26 about three weeks ago, and it is still breaking-in. 

I'm surprised by the sound quality when I use the LAN input of its built-in streamer. The result is superior to a much more sophisticated streaming setup, combined to an Audio-GD DI-20HE, that I've been using, and that I meticulously optimized over time. 

 

Full details about this streaming setup on my profile: 

 

https://audiophilestyle.com/profile/45620-dandou/?tab=field_core_pfield_3

 

 

I'm still going back and forth between U18-I2S-R26 vs  LAN-R26 (ordered the FMCs, haven't received yet)

I find the U-18 + R26 -> somewhat lower, less pronounced, bass less punchy, but feels more detailed, greater separation, more refined

The LAN connection -> feels louder, more punchy, somewhat less detailed but more "fun/musical", slightly more muddled compared to U18. 

 

It's a tough comparison for me. Switching back and forth causes my A3.5 to crash occasionally, so hard to go back and forth rapidly. I'm still waiting for the FMCs , so maybe that will turn the tide. 

 

 

[ChrisG]

I'm going to sell my R26 and will post it in the next couple of days - it's broken in, but I never registered it, so it comes with an unused warranty card. I am selling because I prefer my Rockna Wavedream Signature XLR. 

 

If anyone is looking for one, DM me. Thanks. 

 

Chris 

[gstew]

Update on my R26 experience so far...

 

First, it is good. AND my assessment so far is that it does not beat my DIY setup in resolution, detailing, black backgrounds, and soundstage depth, width, and height. BUT that setup is wildly over-the-top, especially around the power supplies used, and if made into a retail product would cost several times that of the R26, so not surprising.

 

I'm listening to 44.1 ripped CDs via either I2S-over-HDMI from a SDTrans384 SD Card Player or USB via an original-version (Not V3) Sonore UltraRendu, with both powered by an Uptone Audio LPS-1/1.2s. 

 

My modified SDTrans384 (LDOVR.com LT3045 regulator boards replacing the on-board stock regulators) beat a stock one. BUT I was surprised that the UltraRendu on limited listening seems to beat them both. I think this is both a testament to how good the UR is as a source AND that the SDTrans384's I2S-over-HDMI output uses an LVDS chip (TI DS90LV047A) that is over 10-years-old (as @superdad has posted, there are recent chips that provide measurably lower-jitter on the outputs than most others).

 

I need to spend more time comparing these two inputs and against my DIY setup. 

 

In the meantime, I tried downloading the PGGB upsampled files from Audiowise as @seeteeyou mentioned. No dice so far, they are higher rate than the SDTrans's max 384. I'm checking with Audiowise on getting 384 versions and also will try the sample files via the UltraRendu. If they sound good, I'll get some of my tracks/files converted for further comparisons.

 

This all brings me to a couple questions...

 

First, my standard digital file playback is via the Logitech Squeezebox scheme only playing back files stored locally on my LMS server computer. I use the UltraRendu in this mode, but as far as I've read so far, the R26 built-in streamer does handle this. Am I wrong? OR if not, what recommendations do people have for a good sounding, easy to setup streaming SW I can use into the R26's Ethernet input? I would only be playing local files, as above.

 

Second, I took awhile to understand how to get the UltraRendu connect via USB consistently. Most of that was my learning curve, the biggest breakthrough was when I found a comment in one of the UR support threads saying that the USB receiver had to be up and connected when the UltraRendu boots. Power-cycling the R26 just before power-cycling the UR works every time. BUT I get inconsistent results selecting the I2S input, then back to USB. Most of the time the USB input quickly cycles from indicating PCM 44.1k to DSD. SOMETIMES it will cycle back after a few seconds, but most of the time it does not. AND if I don't power-cycle the R26 before rebooting the UR, it often comes up as DSD on the USB input too.

 

Before I start trying to troubleshoot this via both Sonore and Gustard support, has anyone else experienced similar via the R26 USB input? OR have any ideas here? I admit I haven't updated to the 1.4v of the firmware yet, I need to do so and see if it impacts this issue.

 

Thanks in advance for any answers!

 

Greg in Mississippi

[stefano_mbp]

3 hours ago, gstew said:

my standard digital file playback is via the Logitech Squeezebox scheme only playing back files stored locally on my LMS server computer. I use the UltraRendu in this mode, but as far as I've read so far, the R26 built-in streamer does handle this

You can do it using LMS UPNP/DLNA bridge plugin

[gstew]

Thanks, checking that out, a nice almost 100 page thread on the slimdevices forums. 

 

I run LMS on a Zotac ZBOX-Pi320-w3 with either W8 or W10. A tiny (smaller than 2 of your cellphones put face to face) package that I can easily run on a 5V linear supply. I have spares, so will load one up with LMS, DL the plugin, configure it, and give it a try.

 

Again, thanks!

 

Greg

[kennyb123]

19 hours ago, gstew said:

Second, I took awhile to understand how to get the UltraRendu connect via USB consistently. Most of that was my learning curve, the biggest breakthrough was when I found a comment in one of the UR support threads saying that the USB receiver had to be up and connected when the UltraRendu boots. Power-cycling the R26 just before power-cycling the UR works every time. BUT I get inconsistent results selecting the I2S input, then back to USB. Most of the time the USB input quickly cycles from indicating PCM 44.1k to DSD. SOMETIMES it will cycle back after a few seconds, but most of the time it does not. AND if I don't power-cycle the R26 before rebooting the UR, it often comes up as DSD on the USB input too.

Are you powering down these devices?  Best to keep the UR and Gustard powered up all the time.  What you describe isn't uncommon when using endpoints running Linux.  

 

My x26pro absolutely loves being fed PGGB'd tracks scaled to 24 bit / 16 FS and processed at 256 bit precision.  I assume that the R26 is like the x26pro in that it has a NOS mode or equivalent that does minimal harm to incoming scaled music.

[Miska]

16 hours ago, kennyb123 said:

I assume that the R26 is like the x26pro in that it has a NOS mode or equivalent that does minimal harm to incoming scaled music.

 

These two are very different animals.

 

X26 Pro is ESS based delta-sigma DAC which performs best when fed at DSD512 (through proper 512 fs upsampling). And if you send PCM there, 32-bit is best material for it's on-chip DSP. The "NOS" mode on such only bypasses the first FIR digital filter stage. But all the rest very simplistic oversampling is still there, of course in addition to the modulator. On ESS, there's no way to completely bypass on-chip DSP.

 

R26 is discrete R2R ladder DAC. Based on the measurement data I've seen, it wouldn't send more that 16-bit there. Sending too many bits just increases distortion. So far, Holo and Denafrips gap at 20-bit, and sending more bits just makes distortion pop up. Such true PCM DACs cannot have linearity much beyond 20-bit, which is already extremely good result, due to physical impossibility to create precise enough resistors. When we use such DAC at high sampling rate and only in it's linear region using a noise-shaper designed to correct such DAC, we can correct the linearity errors and the ladder's non-linearity distortion disappears. Sending "too few" bits to such doesn't really cause much harm, but sending one bit too much and the distortion pops up.

 

So these two DAC models have practically nothing in common when it comes to D/A conversion, they need very different handling/input data to optimize their performance.

 

[Dandou]

The R26 performs at its best with DSD512 upsampling. 

16-768 comes very close, and is also very good. 

 

The DAC being set at NOS "on", and DSD Direct "on". 

[Zauurx]

I currently have a Pegasus but am particularly interested in this R26 in order to gain (also) on the naa/dac link. Some declare it superior to the Aquarius, yet the measurements are not maddening.

I used the Pegasus with an excellent ZenStream an DSD256. 

[Dandou]

3 minutes ago, Zauurx said:

I currently have a Pegasus but am particularly interested in this R26 in order to gain (also) on the naa/dac link. Some declare it superior to the Aquarius, yet the measurements are not maddening.

I used the Pegasus with an excellent ZenStream an DSD256. 

The R26 replaced my Pegasus. It's a superior DAC. 

[Zauurx]

You can say more... même en français et en mp ;)

[Dandou]

16 minutes ago, Zauurx said:

You can say more... même en français et en mp ;)

The Pegasus is a good DAC, and I was happy with it. 

The R26 is more resolving. The sound is more dynamic (does not shy in comparison to a Delta-Sigmas). The sound stage is wider. The separation between the instruments is better. It has an excellent built-in streamer.