Tír Na HiFi

Looks like I'm going to be the odd man out again. I tried putting a pre-charged Ultracap pair at the output of Ian's LiFePO4 supply's 3.3V rails to power my Ian GB FiFoPi/DM DAC stack a couple of months ago. AND after his brilliant idea, I tried it again last week.

Both times I preferred the LiFePO4 supply alone over the batt/UC combo. The latter lost some of the specificity and directness of just the battery. EVEN better and my favorite is just the 350F pair of UCs alone, but with roughly 1/100th of the Watt/Hours of an 26650 2500 mAh/8.25Wh cell (by my calculations, which could be wrong!), they don't run long. Even a 3000F UC pair is only roughly 1/10th the WH of the battery cell.

Very close (an almost imperceptable difference to my ears and in my setups) was the 350F UC pair float charged by an Uptone Audio LPS-1.2. BUT at $425 per LPS-1.2, that's not a solution for everyone.

It MIGHT be a difference in LiFePO4 cells and UCs used. I used real A123 cells from BatterySpace.com:

https://www.batteryspace.com/a123-lifep ... isted.aspx

AND I used Maxwell UCs PN 'BCAP0350 E20 T11 10' (Digikey PN 1182-101-ND). Ian has started using and likes Maxwell UCs PN 'BCAP0325 P270 S17' (Mouser PN 723-BCAP0325P270S17) better. I have some of the latter one here to wire up and try and will report back here and to Ian. My setup has good low-resistance high-current connectors to add the UCs, so I can test quickly... and also connect/disconnect them while music is playing to hear immediate results.

Originally I used NESSCAP parts from before they were purchased by Maxwell. My experience with them was that they had an upper-mid/lower-high emphasis... I heard this on both UC supplies and from LPS-1.2s using NESSCAP parts. I had put UCs aside until I tried the above Maxwell parts which sound pretty neutral to me. Luckily most of the LPS-1.2s use Tecate parts that seem as neutral as Maxwells to my ears.

Then onto the question of the primacy of an SD card player over a streamed source... I got a new Uptone Audio EtherRegen a couple of weeks ago. It is a switch with special reclocking circuits designed to totally eliminate both the noise and clock signatures from upstream sources. I've been running it almost constantly since and it has been improving with break-in. It has made a stunning improvement to my setups' SQ. I now need to go back and compare my SDTrans/Soekris setup to my Soekris that I connect to an RPi (will be an Allo USBBridge Sig now) and see how much diff I hear. Using an RPi 2B modified with linear regs to replace the onboard DC-DC converters put the RPi source feeding a Soekris into the same realm as my SDTrans/Soekris combo powered by a good, but not great supply, though an SDTrans-setup powered by my best supplies still was above it. I know using the USBBridge Sig will be another jump up from the modified RPi based on what I'm hearing using both a Katana and an Ian GB FiFoPi / DM DAC, with each in one of my setups. BUT the jump from the ER was larger than that from replacing modified RPis withthe USBBridge Sig and to my hearing, the 2nd largest single improvement I've made, only beaten by installing some tweaked audiophile-quality dedicated AC lines in a house I owned from 1989-1994.

The USBBridge Sig / ER addition (and the removal of the Isolator before the Katana... see below) has pushed my setups up from the level of:

'what should I do next to make my setups sound better'

to the levels of:

'what should I listen to next'... and:

'how do I get more of my library on that music server'

The latter prompted me to purchase 1Tb SD cards for music storage for those servers replacing the 25Gb ones there now. That should get me to having 3/4 of my library available. YAY!

While one of the Chinese SD card player setups with NDK SDAs and synchronous is a lot less expensive than a USBBridge Sig AND an EtherRegen combo, being able to use any reasonable music server or streaming source with little or no SQ difference adds a level of ease of use, utility, and source variety that no SD Card player can match IMHO. SO very attractive IMHO.

Then there's the question of to use Isolators or not. I did a bit of a dive into I2S isolator setups and posted some thoughts on a thread on DIYAudio. I need to update that with this summary:

"Isolators work well when followed by a good reclocking scheme such as the synchronous one most use here OR Ian's FiFoPi-style/Allo Kali asynchronous reclocking. Isolators DON"T work well when there is no following reclocking (such as in an Allo Boss or Katana) or with only so-so reclocking (such as the FPGA reclocking on a Soekris)."

I've been a little cautious about removing the isolators on my Soerkris setups. BUT realizing the above, I'll do that soon to those setups.

AND I've got the parts to setup the synchronous reclocking most use here on a couple of the latest inexpensive Chinese SD Card players. BUT I also have some Acko S03 boards. These are synchronous reclocking boards designed to work with BeagleBone Blacks, but could also be used on one of these SD Card players with some care in implementation. I have one of these out to try too as I'm curious if the SD Card players are sufficiently quiet as to not benefit from the isolators built onto the S03 boards. Thoughts?

Greg in Mississippi

Hi Greg

Thanks for the comprehensive report and interesting that you prefer without the UCs. I have used the Mouser PN 723-BCAP0325P270S17s and genuine A123s. Definitely works for me although we all have different setups and desires. I am also keeping an eye on Andreas Mori next group buy for his clocks.

I have read your glowing reports on ER which I have been keeping an eye on to see if it is the way to go for streaming.

The reason I find the SD card powered by Lifepo4s so attractive as it gives what I am looking for ie more pleasurable listening, more emotion. I have all the detail etc I need. I want rid of any traces off digital sound in voices especially as I listen to a lot of vocalists. A tall order I know!

Can you describe what you find so striking about the ER? If it adds any hardness or brightness no matter what other benefits then I'm reluctant to go there..

Patrick,

The impact of the ER is exactly the opposite of what you don't want. No sense of glare, hardness, and the overall sonic palette is definitely not bright, yet still very detailed and delicate yet dynamically impactful when appropriate in the highs. It is what I hear with any lowering of both noise and jitter on the I2S signal. AND I suspect it brings out the best results from both the tweak (some would say whack) power supplies I use and the USBBridge Sig.

Bottom line is I hear a lessening of digital-related artifacts and the sonics moved SIGNIFICANTLY towards that of my SDTrans setup. All the emotion, PRAT, instrument separation / color / texture, micro and macro dynamics and such, none of the hardness, glare, thinness, and brightness.

I need to go back to compare my SDTrans->Soekris versus my RPi (now USBBridge Sig)->Soekris setups to get a good feeling for how close my network player source with USBBridge Sig and ER is to my SDTrans setup. In the year or so since I've run the SDTran->Soekris, I've upgraded my default powering schemes significantly. When I recently tried the Maxwell float-charged UC pairs on the 2 5V inputs on my somewhat modified Allo Katana, I was unprepared for the magnitude of the improvement. My listening to that configuration about a year earlier did not include my mods to the Katana which mostly improved the power feed connections. AND I also was using the NESSCAP parts which added an upper-mid/lower-treble emphasis.

Adding a UC pair on the RPi 5V feed was another unexpectedly significant improvement. In both of these cases there are regulators OR DC-DC converters between the UCs and the circuits, so I expected a smaller delta. The replacement of the stock RPi with a USBBridge Sig was also a significant step up, as you know. In my Ian GB setup, I had moved to one of the RPi 2Bs I had modified with linear regulators replacing the DC-DC converters. That was a nice bump, but the USBBridge Sig improved on it quite a bit too.

SO while my impression is that what I'm hearing now beats what I heard the last time I played the SDTrans->Soerkis, it was not at the same level of power sophistication as what I use now. So some updates and a good head-to-head trial is needed.

One interesting side-effect is that I have been noticing and to some level bothered by a additional hardness and edge once I removed the Allo Isolator under the Katana. The Katana is the worse-case for use of an isolator... no reclocking after the isolator, so the added jitter is not corrected. When using an RPi as the player endpoint, the noise reduction of the isolator made the tradeoff of the added jitter worthwhile to my ears. On the MUCH lower noise USBBridge Sig, the increase in detailing and instrument separation was preferable. BUT there was some added hardness and edge from being directly connected to a computing device. I tried some things with some limited success in curing it, with a list of additional things to try though they were all more invasive and more work.

THAT HARDNESS AND EDGE IS ENTIRELY GONE NOW WITH THE ADDITION OF THE ER!

Who knew that much of what we object to in computer audio was upstream noise and clock signatures that the ER removes? Well, who knew beside John Swenson? In conversations with him over the last several years, he's been working on this technique and reported pretty remarkable differences in his test setups. The key aspect his technique does that others don't is removing the lingering signatures of the clocks and other digital processing upstream of the ER. Since I had an upstream setup designed to be good-sounding and low-noise (all linear supplies with two critical components, the switch and final FMC powered by Uptone Audio LPS-1.2s), I was not expecting much of an improvement. What I'm hearing suggests to me that there is/are significant mechanisms causing 'digital sound' in networked / streaming player setups that we don't really know of or understand at this point.

AND I don't want to minimize the contribution of the USBBridge Sig. I doubt I'd hear what I am hearing with a stock or even linear-modified RPi.

AND it also matters that I'm powering the ER with an over-the-top combo of a pair of LPS-1.2's at 12V paralleled through a 10V MPAudio dual-section 3||LT3045 board. So about $1000 of power supply feeding that $640 ER! I've done some Q&D tests and the ER is still quite good with either the stock supply (SMPS... YUCK, but still good sounding) or some lesser linear supplies. BUT my preliminary assessment is that my power setup DOES add some magic to the impact of the ER. I was thinking of how to setup a float-charged UC set for 12V without a lot of hassle, but realized that I can power the ER at 10V with a couple of UC pairs set to 5V in series, so I just have to try some powering and balancing options between the 2 pairs. That might be even better!

All of this is why I realize I need to compare my ER'd player configuration to that of my SDTrans... and likely also the hot-rod TirNaHiFi SDCard player, which could be better than my modified SDTrans. The ER'd setup MAY make those a needlessly user-hostile way to play digital music.

Greg in Mississippi

P.S. Patrick, More thoughts on the SD Card players... the ONLY reason I use them is for better SQ. While my largely DIY'd system reeks of hair shirt chic, the tradeoff between significantly better SQ versus access to a larger percentage of my digital music library will always come down on the side of SQ.

When I first modified an RPi 2B with the LDOVR Mezzanine board to replace the on-board DC-DC converters, the SQ jump was quite nice. Those RPis feeding I2S to something like a Soekris brought the networked player SQ with my then-best supplies up to that of my SDTrans source with a good, but not top quality supply. So at that point I'd still keep the SD Card players around for the top SQ experience, sorta like my equally hair-shirt phono setups.

BUT the SQ I'm hearing now with the ER / USBBridge Sig / dual LPS-1.2->MPAudio board with the player and DAC mostly powered by UC pair buffers has me wondering if the SQ of that setup will come close to or equal that of an equivalently powered SDC Card player. IF that happens, that'll be the time to put the SD Card players away and focus on the networked players. Or at least until I can spring for a pair of DuCulon clocks to use for reclocking an SD Card player setup.

In the meantime, my 1Tb SD Cards came today, so I'll be loading them up with all of the selections from my digital library that I now want to hear on my newly invigorated networked player setup.

Since international shipping would cost me quite a bit, I'd like to ask Greg for a favor to give it a shot since it could be a relatively affordable option for ER. Are HDBaseT and REVConnect just hype or maybe they're the real deal?

https://cdn2.hubspot.net/hubfs/282305/O ... pdf#page=8


https://www.bicsi.org/docs/default-sour ... df#page=24


More information

https://www.belden.com/products/enterpr ... -uhd-media
https://www.belden.com/hubfs/resources/ ... lletin.pdf

2183R by the foot

https://www.markertek.com/product/bl-21 ... k-per-foot

2183P by the foot

https://www.pacrad.com/belden-2183p-hdb ... 3-awg.html

REVConnect Core Pack

https://www.falcontech.com/REVConnect-C ... ew-b50.htm

Unshielded REVConnect plugs

https://www.falcontech.com/Modular-Plug ... ubk-s1.htm

Shielded REVConnect plugs

https://www.falcontech.com/Modular-Plug ... sme-s1.htm

REVConnect Termination Tool

https://www.amazon.com/Crimp-Belden-CAT ... B075XKV7WS
https://hermanproav.com/product/250902/ ... ation-Tool
https://www.falcontech.com/REVConnect-T ... vutt01.htm

They also have custom cable options here for roughly 65 bucks @ 6' (Data → CAT6A → Belden → BLD2183R-BLK-A1K → CABREVCONSHLD)

https://www.pacrad.com/custom-cable-builder

gstew wrote: ↑Wed Nov 27, 2019 9:13 pm The impact of the ER is exactly the opposite of what you don't want. No sense of glare, hardness, and the overall sonic palette is definitely not bright, yet still very detailed and delicate yet dynamically impactful when appropriate in the highs.

THAT HARDNESS AND EDGE IS ENTIRELY GONE NOW WITH THE ADDITION OF THE ER!

AND I don't want to minimize the contribution of the USBBridge Sig.

All of this is why I realize I need to compare my ER'd player configuration to that of my SDTrans... and likely also the hot-rod TirNaHiFi SDCard player ....The ER'd setup MAY make those a needlessly user-hostile way to play digital music.

Invaluable info Greg, appreciate as always your in depth reporting especially how crucial power supplies are.

This is what I didn't want to hear :) as I will have to add the ER to my purchase list now. I wasn't sure how it would fit being an ethernet device which is ideally placed before the DAC. I can see combined with a USBridge sig would be the way to go.. but a significant spend and will have to wait until next year.

VERY curious how your set up compares to the SD trans/SD card player route

Just wondering if Pearse were happy about his SD3.75 reference digital turntable, the same Chinese seller just released another flavor with dual Cirrus Logic CS43198 recently

https://item.taobao.com/item.htm?id=610424235375

He's also offering this 32GB eMMC in SD card form factor, not SLC by any means but the price is right

https://item.taobao.com/item.htm?id=606416340810

The card reader must have the Genesys Logic GL3224 chip (e.g. Rocketek CR11 and UGREEN CM104) in order to be able to recognize eMMC

https://github.com/dvdfreitag/USB3-eMMC
http://www.genesyslogic.com/en/product_view.php?show=53

Interesting stuff if we're able to get SLC eMMC somewhere with decent size and price

https://whycan.cn/t_2778.html
https://item.taobao.com/item.htm?id=568261155076

Unfortunately Seeteeyou I still have not received it yet, post from China has been really poos recently.
Great to see further developments in the SD Player segment...still surprised more have not found out about this exceptional playback solution and the terrific sound they produce.

In the middle of moving house and work at the moment so not getting time to do much in audio....roll on 2020 :-)

Thanks Pearse and Patrick for your updates, I'll keep an eye on this thread.

There's something else I'd like to know as well, someone in Taiwan actually spent almost 5500 bucks / 5000 Euros (600,000 yen) on some kinda Ian's FIFO DAC

http://www.my-hiend.com/vbb/showthread. ... -HI-FI電腦訊源（連載）&p=240486#post240486

Somehow they're assuming / believing that for whatever magical reasons, merely throwing Ian's FIFO into the mix would achieve that "Holy Grail" called source immunity. Did that stuff sound more like unicorn or something? Everyone should be looking for such creature in the wild but it's nowhere to be found. And then the funniest part should be paying VERY good money for a megabuck product while getting a pair of cheap TCXO in return, not even OCXO or anything like that. Dude, maybe they could very well have that "legendary" DuCULoN for burning such serious cash.

I still recall that Patrick should have tried quite a few clocks on Ian's FIFO before, it's mission critical to get the best clocks one could afford since they're playing such an important role. I just couldn't stop laughing when those clocks were put under the spotlight. If source immunity were actually a piece of cake to achieve, Patrick wouldn't need to try so many different sources over the years.