USB Nervosa Thread Decrapifiers, pro interfaces, and bears oh my

Yes, and it's reference multiple places in this thread, including on this page (about 9 posts up).

1. It already uses a linear power supply (it's built into the unit, standard IEC socket input).
2. Replacing that LPS requires removing the existing supply and installing an all-new one (or doing component level upgrades to the PSU board).
3. I think Kitsune already offer a "KTE" version of the SU-1 that has a 1/50th the PSU noise for the stock PSU.

I'm highly inclined to doubt that it'd beat the Aries simply on the basis that, no matter how low-noise the PSU as the always-on USB implementation on the SU-1 is XMOS based, and that chipset tends to impart more noise than a decent LPS (or even a properly filtered SMPS). You'd have to try it in a specific setup to know for sure.

The SU-1 did not improve for me when fed with a JCAT/Intona isolator.

I think the KTE version of the SU-1, which has a lower noise LPS than the stock SU-1, as well as upgraded custom firmware, is ~$600 ... I can't see why you'd spend more than that just on a LPS that is going to require major internal surgery, or external interfacing, just to make work in the first place.

--

I would only use an SU-1, of any flavor, in two situations:

1. I needed a standard, common, interface to multiple DACs, possibly requiring multiple parallel output formats, and wanted to avoid running unique drivers for each (which is really only a Windows problem in the first place).
2. wanted a cost-effective I2S interface for a DAC that works best with that input.

Otherwise, I'd buy something else.

 

I don't know where this stuff comes from ...

I DO NOT play music directly from a PC in my rigs.

The SU-1 I have exists to give my Spring DAC an I2S feed and that SU-1 is NOT connected to a PC.

My pair of Yggdrasil's are each connected to an Auralic Aries (full-fat version) via AES.

My Gungnir is currently fed via it's BNC S/PDIF input using an NAI/AOIP box of my own design.

--

If I was going to run a direct connection from a PC, and assuming it was a Windows box with a suitable spare PCIe slot, I'd use a Lynx card with, or without a Mutec MC-3+ (depending on the re-clocking implementation of my DAC).

 

Until I am ready to buy a AOIP box of some type, I will continue to use my computer to play music through my Yggdrasil using USB. My current audio chain is fairly simple:

2016 MacBook Pro (A+ 3.x) --USB--> Singxer F-1 --SPDIF coax--> Yggdrasil --XLR--> Mjolnir 2 --SE--> Powered Nearfield Monitors

The DAC, amp, and monitors are all plugged into a TrippLite “PS” power strip. The power strip is plugged into a 2.5kVa Topaz isolation transformer.

I am still willing to see what I can do to improve the sound. I know the USB signal getting to the Singxer F-1 is not great so that signal could be improved. But, I was not convinced that any of the available USB "decrapifiers" were worth the cost or the effort. That was partly due to what I have learned from the people here. I have been intrigued with the what I had been reading about the UpTone Audio ISO Regen unit and the USPCB USB 2.0 A to B adaptor. The ISO Regen provides full power, ground, and signal isolation. The USPCB preserves signal integrity and impedance match between the ISO Regen and the USB device it is plugged into. I had an UpTone Audio LPS on hand to use as a power source for the ISO Regen.

My new audio chain looks like this:

2016 MacBook Pro (A+ 3.x) --USB--> ISO Regen --USPCB--> Singxer F-1 --SPDIF coax--> Yggdrasil --XLR--> Mjolnir 2 --SE--> Powered Nearfield Monitors

The following comments outline the changes I perceive with the ISO Regen and USPCB in place and powered on for more 15 hours:

The music and voices emanate from a quieter background.

The timbre of instruments and voices are a bit more defined. This is even true with bass sounds which tend to muddle together more.

The bass is more full yet has more detail and clarity. The mids and highs are a bit sharper without being harsher. The overall clarity and detail are a little bit better.

The soundstage isn’t any wider but I get a sense of more depth and height. Instruments and voices have much better isolation and their locations are more defined. The is the single biggest difference I hear in the music. I could hear this before but not with the laser like focus I get now. Instead of the sound coming from a general direction on the stage, I get a sense of the sound coming from a very specific location on the stage. The effect is subtle but it gives me a much better sense of being at a live performance.

The best analogy I can come up with is that with the ISO Regen, LPS-1, and USPCB in place versus what I started with is like comparing two photos of the same scene. One slightly out of focus with a color filter in place while the other is perfectly in focus with no filter. You can tell that both are of the same image. But, with the slightly out of focus image, you can’t quite see all the detail. Parts of the image are blurred together and some of the colors aren’t quite right. It’s pleasing to the eye until you see the perfectly focused image. You immediately realize what you were missing. You see all the details and the colors look better and more natural. It’s like a veil has been lifted from your eyes.

The differences I hear are nothing more than subtle. It’s when you take them as a whole that you realize how much the listening experience is enhanced. The clarity and detail improvements along with the voice and instrument isolation on the soundstage make for a much more realistic and better sounding listening experience.

 

Thanks, I was hoping you would respond as you have some and have had all of my chain.

 

Replacing the internal power supply is easy. Just unscrew the IEC plate & unplug a cable & replace it with something like this
https://kitsunehifi.com/product/singxer-su-1-dc-power-conversion-kit-kitsunehifi-2-1mmx5-5mm/

With that you can easily use any 3rd party power supply on the market to power the SU-1, though choosing which is a completely different story.

I also think the upgraded KTE version still has the same power supply as the stock one.

 

Same topology, different parts, yielding 1/50th the supply noise.

It's a linear design in either case.

Spending $500-800 on a PSU for a $399 device seems a bit daft to me. Frankly, so does spending an extra $289. The USB chipset will still produce noise that is far more problematic than the SU-1's standard LPS, and no amount of faffing around with third-party supplies and fiddling-with-shit-for-the-sake-of-it is going to change that.

The cost of an SU-1 w/ some dubiously effective, unnecessarily expensive, after-market LPS is still more than something like the Aries and, unless you need that I2S output is unlikely to yield an audible benefit.

Or it's the difference between a Gungnir MB and an Yggdrasil.

Both of which will yield more audible changes than piddling around with a USB-fed DDC.

Seriously, if it's not people adding as many dodgy gadgets into the chain as possible, it's fiddling with the fundamental design choices for the stuff, mostly without understanding how/why such things either don't do anything useful, or if they do certainly not for the reasons claimed. And the commonly cited "drastic improvements" for such fannying around seem to be much more accurately described as "barely audible, if at all, on a really good day with a super-resolving chain" and, at least as commonly, "can't tell the difference at all the second they don't know what they're listening to".

 

Anyone tried this?
https://www.thomann.de/es/amphenol_amphe_dante_single_channel.htm

 

Outputs analogue and requires 48V POE.
http://www.amphenolaudio.com/products/dante/adapter/amphe-dante/

Does not do AES/EBU out.

 

You might as well just get the Focusrite AM2 then. Dante in, 2-channel balanced out plus a headphone jack (separate volume control on each).

 

This right here, this is what bugs me. Am I the only one around here who is reminded of mid-90s Mac publishing house ADB dongle and/or SCSI chain hell? Something is always going to be prone to acting up when you chain so much random shit.

Call me a killjoy if you will, but simplicity and elegance, using fewer, better-quality and more appropriate components in a chain tends to produce more pleasant (and reliable results)- at any pricepoint. Chaining random shit together endlessly just seems a bit like cargo cult behaviour that you'd expect to find on HF.

 

People who paired the SU-1 with a LPS said they noticed immediately benefits. While Spending $800 on a power supply does seem overreaching, but the Aries costs what? $2,000?

Speaking of which, does your Aries come with a external LPS?

I could send you the SU-1 DC conversion kit to test out on your SU-1 & Aries power supply to see if it does really elevate the SU-1 or not.

 

Lots of people say lots of things ... especially when they've spent unrecoverable funds on something that, if it makes a difference at all, is unlikely to be more than extremely subtle. Funny how these things always sound BETTER and never worse. Unless they had some means to meaningfully A/B it, they're at least as likely to be subject to expectation bias as anything else.

Spending $800 on a PSU for a $399 device, when that same sum as an upgrade would get you a better DAC (or amp or headphone) which is vastly more likely to yield a meaningful and audible benefit, is a bit silly. You might get improved results from the SU-1 ... but I'd be stunned if they were anything more than extremely subtle (and if they are, that suggests serious problems with the internal power-regulator and isolation claims made for the SU-1 as it is).

The top-of-the-line version of the Aries, which does include a low-noise LPS, retails at $1,599.

I don't think anyone actually pays this much ... the only new one I bought I paid $1,199 for. The others were bought used for between $750 and $900. But, regardless, it's a different type and class of device. The SU-1 is just a USB sourced DDC and doesn't offer any of the other functionality of the Aries, nor does it perform as well in the simple DDC/interface role.

Yes.

Standard part with the non-LE models.

If I wanted to replace the PSU in my SU-1 I could do it with parts on hand. And if I was going to do that, I'd start with one of myriad lab-grade LPS units I have around all of which will comfortably outperform your typical "upgrade" audiophile-targeted units.

Besides, if you plugged the Aries LPS into the SU-1 with that conversion kit you would, very shortly there after, no longer have a functional SU-1.

You can't just plug random LPS units into random hardware and except them to work properly.

 

This Intona thing looks interesting for the galvanic isolation. I ended with the Yggdrasil A2 and MJ2 plugged into a Furman AC-215A (more for protection than noise), but that's plugged into a different wall outlet than the computer. So because of that I've switched to SPDIF over TOSLINK. No noise, no ground loop, and no obvious problems, so no rush. How long on average has the wait been once ordered for shipping to the USA?

 

I would reconsider using TOSLINK. It would seem to be the ideal solution but it has its own set of issues. Increased jitter, cable issues, and problems supporting higher sample rates are among them.

If you want isolation, instead of the Intona, I would look at the new UpTone Audio ISO Regen. It does what the Intona Industrial does plus more. I am testing one right now and it does work quite well.

 

Thanks winders. I'll take a look. Assume the Yggdrasil already does a good job re-clocking, but if the Regen helps, great. The house is wired funny and plugs in the same room are sometimes on different circuit breakers, and even when not, it's a fairly long ground wire run. Still I'd like to go back to USB. The TOSLINK is working, but agree it has issues and limitations (does avoid the ground loop though).

 

From my experience:

Top AES source (Auralic Aries) > good AES source (microRendu+LPS-1>SU-1) > good USB source (microRendu+LPS-1). Not by a lot, but noticeable even by my (very musical and sharp hearing) wife who did not know anything about the changes. IMHO, USB decrapifiers are the proverbial lipstick-on-a-pig in that they do not take out a main source of switching noise, the USB receiver inside the DAC.

 

Sure. I am waiting on a particular AOIP device that will have Ethernet in (obviously) and AES/UBU out. Until then, I will continue to use my MacBook Pro/Singxer F-1 setup to feed my Yggdrasil using SPDIF coax. I have to say that the ISO Regen, at least in my setup, offers a welcome improvement in overall sound quality. The biggest difference I noticed was in the soundstage. The width didn't change but the height and depth improved noticeably. The instruments and voices have much better isolation and their locations are more defined.

I can't wait to get rid of USB completely!!

 

Meh. My Raspberry Pi has an i2s to SPDIF board that gives it transformer isolation and solid clocks on the output. The box runs from a low noise PSU and is a completely independent streaming endpoint- independent of both USB bullshit and my computer. I can kick back and listen to music without the distraction and noise of the computer on at all... and best of all, the whole thing cost less than that Intona box. People spend far too much on polishing the turd of USB into their DACs.

Hell, it's cheap enough to be sane to use for cheaper DACs like Modi Multibit and Bifrost Multibit even, where they don't have fancy-pants VCXO reclocking- the difference is massive. My Gungnir Multibit is just fine with it too, though, over BNC. That said, people with a Yggdrasil should be considering a high quality AES source like an Aries, Lynx card or some Rednet kit. Once you've spent that much on a DAC, it should be fed nicely.

 

I've only had the Yggdrasil a week, so limited testing time. I did some very limited comparison of TOSLINK vs USB, and admit I could not immediately identify any difference, though had a slight impression that TOSLINK was a bit sweeter sounding, maybe. Yggdrasil can only switch sources so fast.

I am hopeful that Yggdrasil will take care of re-clocking but the back and forth testing was limited, and then there is that most of what I am listening to is 16/44.1 (either my ripped CDs (non compressed)), and Spotify. That leads to the questions of is it better to send at native rate, trust that Windows does no damage to the data if volume is set at 100% and all enhancements are disabled, and let Yggdrasil do all the up-sampling? or up-sample on the computer plus Yggdrasil may add it's own up-sampling after that?

Anyway I could hear a difference over USB after I moved the Yggdrasil/MJ2 to a different wall plug, but the USB connection creates a large ground loop in that case versus TOSLINK. Ground loops can sometimes go unnoticed, but they never help, hence the interest in isolating the USB connection.

That Lynx card looks interesting, maybe an upgrade for later this year.

 

I send all content to my Yggdrasil at its native bit depth and bitrate. Since Yggdrasil's max input bitrate is 192 kbps and it upsamples to 352.8 kbps or 384 kbps, it is going to upsample at least 2x no matter what. The only software I have tried upsampling with is Audirvana. I did not think it sounded better than Yggdrasil doing all the upsampling.

The majority of folks over at CA would tell you that software upsampling is superior. The only explanation they seem to have is that computers have more processing power and therefore can do more and better filtering. They also will tell you have more options in filter choices and can pick the one that sounds best to you. I guess that may be true. I sure like the sound I get from Mike Moffat's filter!