[rb2013]

   
I wouldn't disagree with that.  That's why I will continue to use a single cell LiFe4Po battery on my Regen with a smps power supply trickle charging it.  All for less than $30 and it provides galvanic isolation on the Regen.  Of course you need to modify the Regen with a simple DIY and your DAC (preferable) does not need a 5V vbus feed.
I think folks overrate the effects of smps feeding noise back into the mains and how that effects SQ. 
 
Far better places for the LPS-1, such as my DAC/Hugo (coming soon) and the PPA USB card.

I agree on using the LPS-1 in other places - the max current it can supply does limit that though.
 
I can tell when using a SMPS in my system - even the iFi iPower.  Sold all mine.  Even the SMPS in the Mutec MC-3+ USB and AOIP REDNET 3 where detremental.  Those were sold as well.  I won't use anything by a LPS now.
 
I have not had much success with LiPo batteries  - they seem to dampen or limit dynamics.  Depending on which battery - they can have very high noise levels.  Higher then a well designed SMPS.  And their ability to supply current on demand is quite limited - judging by their slow rise times.

 

[Mediahound]

Question- if a DAC already has galvanic isolation on the USB does the Regen still help?

 

[elviscaprice]

Question- if a DAC already has galvanic isolation on the USB does the Regen still help?

​Depends on the galvanic isolation implementation.  If it did, it probably would be minimal as some have reported on the 2Qute and others.  I would try the Intona.  If you hear a difference then you have your answer on how good your DAC implementation is.
 
rb2013, I think it prudent to power any smps on a circuit of their own together.  That being said, the bigger problem with smps supplies is the AC leakage.  That being said, most smps supplies are poor and I agree with you, should be avoided but for a different reason.
 
As far as Lipo's, some are better than others.  All depends on where they are implemented.  I agree about speed, so at the DAC, definitely not.  Thus my DIY soon on the Hugo.

 

[rb2013]

 
​Depends on the galvanic isolation implementation.  If it did, it probably would be minimal as some have reported on the 2Qute and others.  I would try the Intona.  If you hear a difference then you have your answer on how good your DAC implementation is.
 
rb2013, I think it prudent to power any smps on a circuit of their own together.  That being said, the bigger problem with smps supplies is the AC leakage.  That being said, most smps supplies are poor and I agree with you, should be avoided but for a different reason.
 
As far as Lipo's, some are better than others.  All depends on where they are implemented.  I agree about speed, so at the DAC, definitely not.  Thus my DIY soon on the Hugo.

The Regen does not provide USB galvanic isolation.  It's main benefit is improving the USB signal integrity by closely matching inpedances (mismatched inpedences causes backwave reflections on the USB data lines) to the USB input - to min the activation of USB AGC circuits to try and overcome varying SI.  The ACG (Automatic Gain Control) on the USB processor increases 8K USB packet noise that can infect the power supply and ground plane.  This noise feeds into the audio clocks and creates jitter.  The second purpose of the Regen is to the feed a cleaner +5VDC power to the USB power leads on the DDC or DAC.  The Regen has excellent TI LDO regulator to filter the noise on the power output.
 
I have used the Startech GB LAN Ethernet USB extender for excellent GI.   Better then the Intona - but also more expensive.  The Startech (made by ICRON) has high power Xilinx SPARTAN 6 FPGA's at each end.  They translate the USB packets to IP packets, transmit the data stream over UTP CAT6 Ethernet (which has full GI built in) , then back again.

 
 
The Regen and the W4S Recovery, (which I prefer and I have had both) act as a great way to power a USB DDC or DAC with external  power.  The better SQ I heard from the Recovery is probably due to it using the ultra low phase noise Crystek CCHD-957 clock for reclocking, versus a XO in the Regen.  Although the TI regulator on the Regen is better with 4uv of noise vs the Recovery's 14uv.
 
I put any SMPS (when I had them) on a Art Audio PB4X4Pro AC line isolator and filter (40dB of common and differential mode rejection).  I still prefer them out.  I have a separate PB4X4Pro for the DAC and PC as well.
 
I just did an experiment on the Startech - which has four USB ports.  One is used to fed my USB DDC, by way of the Recovery.  The other three I have filled with 256GB USB 3 flash drives.  I have transfer a load of audio files to them and the SQ is really better.  So the Startech can do double duty - GI and SI improvement to the USB DDC and GI (from the PC) isolated solid state storage.
 
PART 4 of my power supply thread series will include batteries, in addition to Ultra and Super Caps.  The last PART 5 will e dedicated to DIY power supplies - LPS and Battery.  So I would be interested in your DIY project.  I totally agree on not powering a DAC by battery.

 

[elviscaprice]

Very good information and guidance there rb2013.  Agree, Ethernet is another avenue in achieving galvanic isolation.  I think the choice between Ethernet or Intona USB for galvanic isolation still comes down to implementation as far as SQ achieved.  Hard to say which is better. 
That being said though, I think with Ethernet one is just passing the issues further down the stream but not nearly to the same degree as poor USB implementation.  Sooner or later you have to determine how to enter the DAC.  Also software issues and possible lack of versatility in audio formats/media player can become an issue with Ethernet.  Thus I think it's much easier to remain with USB thru the down stream. 
 
Think you misread my reply to the other poster, I never claimed the Regen had galvanic isolation.  But I like your detail on what the Regen does, spot on.   Only problem is getting clean power to it.  Thus the suggestion to the poster to just concentrate on adding an additional galvanic isolator first, such as the Intona, or as you suggest Ethernet.   Either way is good.
 
I'll definitely  update on how my DIY goes with the Chord Hugo and it's replacement of the Lipo batteries with the LPS-1.

 

[rb2013]

  Very good information and guidance there rb2013.  Agree, Ethernet is another avenue in achieving galvanic isolation.  I think the choice between Ethernet or Intona USB for galvanic isolation still comes down to implementation as far as SQ achieved.  Hard to say which is better. 
That being said though, I think with Ethernet one is just passing the issues further down the stream but not nearly to the same degree as poor USB implementation.  Sooner or later you have to determine how to enter the DAC.  Also software issues and possible lack of versatility in audio formats/media player can become an issue with Ethernet.  Thus I think it's much easier to remain with USB thru the down stream. 
 
Think you misread my reply to the other poster, I never claimed the Regen had galvanic isolation.  But I like your detail on what the Regen does, spot on.   Only problem is getting clean power to it.  Thus the suggestion to the poster to just concentrate on adding an additional galvanic isolator first, such as the Intona, or as you suggest Ethernet.   Either way is good.
 
I'll definitely  update on how my DIY goes with the Chord Hugo and it's replacement of the Lipo batteries with the LPS-1.

Well some knowledgeable folks have pointed out the issues with the Intona (dirty regulators on the output side of the GI bridge), and I have pointed out many times the very low cost and very high phase noise of the SiTime CMEMS clocks in the Intona.  These low end clocks - have to reclock the USB data stream.  They were chosen due to one of the design parameters of the Intona was use in 'industrial' applications - so the need to with stand +2G lateral accelerations.  And Crystal based clocks would fail under such conditions.  If you look at the SiTime clock lineup - their CMEMS rank at the bottom for performance.  So the Intona while addressing GI it creates other issues.
 
As Intona states:
 

Then the quality of the crystal oscillator is important too. Are you using, say, Crystek oscillators? It is a SiLabs MEMS oscillator, with measured Jitter around 2ps. This is, because 1. we have some customers that need mechanicly rugged hardware because they use our isolators at repeated >2g acceleration 24h/7d - crystals will fail here - and 2. power consumption of MEMS is 1/10 of crystals. As all voltage regulators are linear in the isolator, we had to pay attention to overall efficiency.

 
The Startech/Icron/PS Audio GB LAN USB Ethernet extenders - offer a unique solution to USB GI - and enhance the USB SI.  This is not any ordinary USB extender - but one capable of very high throughput (the PS Audio LANRover is made by ICRON and uses a similar method - just with lower cost ASIC chips vs the SPARTAN 6 FPGA).  So can pass an improved USB SI to 384K and 256DSD data streams with ease.  The sender box employs high quality XO clocks for the recreated USB output.  And even better both units can be fed a LPS power source - this improves the SQ even further.  I have 24VDC LPS feeding the REX and a 5VDC LPS feeding the LEX.
 
This lead to finding an even better solution to computer audio - doing away with USB completely!  See my AOIP thread (Audio over IP).  The Rednet boxes which use the DANTE AOIP protocol have created waves in many audio circles.  Virtually everyone who has compared sota or near sota USB to AOIP agrees that AOIP is superior.
 
Here is my thread on that subject - really a seachange in computer audio.  But the Startech Ethernet solution does work wonders for USB.  I have tried virtually every USB gizmo made - and this was a major SQ improvement.
http://www.head-fi.org/t/806827/audio-over-ip-rednet-3-16-review-aes67-sets-a-new-standard-for-computer-audio

 

[motberg]

 I have tried vritually every USB gizmo made  

Have you tried the Intona in your system and compared it directly to the Startech/Icron/PS Audio GB LAN USB Ethernet extenders?

 

[rb2013]

  Have you tried the Intona in your system and compared it directly to the Startech/Icron/PS Audio GB LAN USB Ethernet extenders?

Yes and I much prefer the Startech - esp now that it performs so well as a solid state - GI'd storage solution.

 

[elviscaprice]

 
Well some knowledgeable folks have pointed out the issues with the Intona (dirty regulators on the output side of the GI bridge), and I have pointed out many times the very low cost and very high phase noise of the SiTime CMEMS clocks in the Intona.  These low end clocks - have to reclock the USB data stream.  They were chosen due to one of the design parameters of the Intona was use in 'industrial' applications - so the need to with stand +2G lateral accelerations.  And Crystal based clocks would fail under such conditions.  If you look at the SiTime clock lineup - their CMEMS rank at the bottom for performance.  So the Intona while addressing GI it creates other issues.
 
As Intona states:
 

 
The Startech/Icron/PS Audio GB LAN USB Ethernet extenders - offer a unique solution to USB GI - and enhance the USB SI.  This is not any ordinary USB extender - but one capable of very high throughput (the PS Audio LANRover is made by ICRON and uses a similar method - just with lower cost ASIC chips vs the SPARTAN 6 FPGA).  So can pass an improved USB SI to 384K and 256DSD data streams with ease.  The sender box employs high quality XO clocks for the recreated USB output.  And even better both units can be fed a LPS power source - this improves the SQ even further.  I have 24VDC LPS feeding the REX and a 5VDC LPS feeding the LEX.
 
This lead to finding an even better solution to computer audio - doing away with USB completely!  See my AOIP thread (Audio over IP).  The Rednet boxes which use the DANTE AOIP protocol have created waves in many audio circles.  Virtually everyone who has compared sota or near sota USB to AOIP agrees that AOIP is superior.
 
Here is my thread on that subject - really a seachange in computer audio.  But the Startech Ethernet solution does work wonders for USB.  I have tried virtually every USB gizmo made - and this was a major SQ improvement.
http://www.head-fi.org/t/806827/audio-over-ip-rednet-3-16-review-aes67-sets-a-new-standard-for-computer-audio

​Basically I am ending up in the same place with the Intona no 5V vbus/ Regan (jKenny mod) no 5V vbus as the Startech.  The XO clock being of similar nature between the Regen and Startech.  Without the need to improve power source for galvanic isolation, which you will have to do with the Startech.  Ethernet also has it's issues as some are finding out with galvanic isolation.
I still maintain it's better to remain in the USB domain.  Your just pushing the issues further away with Ethernet and causing more software limitations.  But it is a viable alternative if it fulfills your needs.
Best to have a DAC that doesn't require any 5V vbus.

 

[rb2013]

 
​Basically I am ending up in the same place with the Intona no 5V vbus/ Regan (jKenny mod) no 5V vbus as the Startech.  The XO clock being of similar nature between the Regen and Startech.  Without the need to improve power source for galvanic isolation, which you will have to do with the Startech.  Ethernet also has it's issues as some are finding out with galvanic isolation.
I still maintain it's better to remain in the USB domain.  Your just pushing the issues further away with Ethernet and causing more software limitations.  But it is a viable alternative if it fulfills your needs.
Best to have a DAC that doesn't require any 5V vbus.

I'll have to check out that Jkenny mod on the Regen - I had three of his Hifaces - back in the old days.  I agree clocks on the Regen and Startech are similar and much better then the Intona's.
 
The issues with Ethernet and GI - is the use of STP Cat6 and Cat 7 cables versus UTP.  I tried a FMC fiber between the Startech boxes and it really didn't help.  But I was using a BJC Cat6 UTP cable.  BTW you can use a Cat6 or Cat7 STP - you just have to disconnect one end of the shield from the plug.  On my AOIP thread we had a member who works in a server farm - massive amounts of disk drives and power in one building - a EMI soup and he never had an issue with using UTP cables.  But you could use a GISO or FMC in between if one is really paranoid about GI and Ethernet.
 
Let me say I have been down the USB only path - way down and adding the Startech has been a revelation.
 
Here is what I have owned and heard and my rankings:
REDNET 3/Cerious Graphene/Mutec 3+ USB (SPDIF)/Antelope OCX (RN wClock)           270
REDNET 3/Cerious/Mutec 3+ USB/Audience au24 se digital cable                               250
REDNET 3/Cerious/Mutec 3+ USB (SPDIF reclocker)/AS Sliver Statement dig cable        240
REDNET 3/Cerious Power Cord                                                                               220
Singxer F-1 DC30W/Cerious/Recovery/DCiPur/ iPur2/Startech GB LAN Iso USB               170
Mutec 3+ Smart Clock USB/Cerious Power Cord                                                        155
Singxer F-1 DC30W/Cerious/Recovery/DCiPur/ iPur2                                                   145
PUC2 Lite TeraDak DC30W/Cerious/Regen                                                               135
Singxer F-1 DC30W/Cerious                                                                                   135
DXIO Silver/TeraDak DC-30W/Cerious                                                                      130
Singxer X-1 DC30W/Cerious/Recovery/DCiPur/iPur2                                                    125
PUC2 Lite - USB power                                                                                          110
Singxer F-1  Stock feed                                                                                          110
Breeze/Cerious Graph/WBT RCA Nexgen                                                                   109
Breeze DU-U8 with Cerious Graphene                                                                      108
 Breeze DU-U8 (Talema version)                                                                              100
Breeze DU-U8 (BingZi version)                                                                                 95
Hydra Z with LPS                                                                                                    92
Melodious MX-U8 (upgraded caps)                                                                             85
Melodious MX-U8 (stock)                                                                                          81
Gustard U12 (upgraded caps)                                                                                    76
Gustard U12 stock                                                                                                   72
iDAC DAC2 (used as a DDC)                                                                                      65
Musiland USB3.0 US Dragon                                                                                      65
M2Tech EVO with LPS                                                                                              60
Audiophileo 2  USB Power                                                                                         50
M2Tech Hiface                                                                                                         40
 

 

[Audio Addict]

 
Yes and I much prefer the Startech - esp now that it performs so well as a solid state - GI'd storage solution.

 
Do you know where you can purchase it instead of the 599 PS Audio LanRover?

 

[Hi Rez]

  To spend $400+ for a power supply to feed the Regen that feeds a USB DDC - is maybe overkill?
Not saying the LPS-1 is not top draw - would love to have one.

 
My thoughts exactly....
 
Then I tried the LPS-1 with a Regen -> Ayre QB-9 DSD.  It replaced an UpTone JS-2.
 
My Regen path started with the stock SMPS.  With the SMPS, the Regen added a naturalness to the music.  Wouldn't have called it a night and day difference, but I wouldn't have called it just a subtle difference either.
 
The stock SMPS was upgraded to the JS-2.  I was surprised by the magnitude of the change.   Upgrading to the JS-2 was a larger change than initially adding the Regen with the stock SMPS.
 
I had originally ordered the LPS-1 to use in my transportable rig, but just for laughs tried it powering the Regen in my main system.  I was not expecting much of a change as I was already using a top linear PS (JS-2)  to power the Regen - color me surprised.  Switching from the JS-2 to the LPS-1 was a larger improvement than going from the SMPS to the JS-2.  Not what I expected at all.
 
The choice between a $200 linear power supply and the LPS-1 might not be so easy after all.  My experience might have the LPS-1 being the better value for the money.
 
I was already sold on clean power with the JS-2.  The LPS-1 just make me more sold....
 
YMMV.

 

[Superdad]

 
I'll have to check out that Jkenny mod on the Regen - I had three of his Hifaces - back in the old days.  I agree clocks on the Regen and Startech are similar and much better then the Intona's.
 
 

 
Hi Rob:
 
Check the clock speed on the Starch/ICRON receive end and I think you will find that clock is for the Spartan FPGA.  It has been a while since I read ICRON's developer kit notes, but I recall that they synthesize the USB clock--either in the FPGA or in a USB PHY chip.  If you don't see more than one XO on the board than that is what is happening.  Maybe it does not matter too much since ICRON embeds an entire USB protocol engine into the FPGA--though I thought they still had some PHYs.
 
Signal integrity is not the Startech/ICRON's strong suit--galvanic isolation is (especially if you power the receive end in a way that does not again give a path for AC leakage currents). Hopefully PS Audio chose to add a dedicate USB clock to their LANRover version as that would seem obvious.  
 
We've been toying with ICRON's SO-DIMM form-factor version for a while but have not made a decision about making a REGEN-like baseboard (great clock, impedance match, ultra-low-noise power network--all the obvious stuff) and offering it to market.  ICRON's OEM prices are not low, so a product of that sort would end up priced similarly to the $600 LAN Rover.

 

[jermaink]

Can an iFi iPower 9V/1.5A be a suitable 'energising' supply for the UltraCap LPS-1? It is apparently a tad shy on the recommended current. I guess it also depends on the amount of current drawn downstream..

 

[Superdad]

  Can an iFi iPower 9V/1.5A be a suitable 'energising' supply for the UltraCap LPS-1? It is apparently a tad shy on the recommended current. I guess it also depends on the amount of current drawn downstream..

There are two different versions of the 9V iFi iPower:  The early ones were rated at 1.5A, and the latter, current ones are capable of 2.0A.  The latter work fine to "energize" the LPS-1 while the former do not.  (We have both versions of the iPower, so this is tested fact, not speculation.)
 
The strict requirements for an "energizing"/charging supply for the LPS-1 are:
7.5V/2.5A
9V/2.0A
12V/1.5A
 
Current capability is of course allowed to be greater (e.g. 9V/3A), but those are the amperage capabilities necessary across the acceptable 7.5-12V range.
Frankly, since the quality of the "energizing" supply makes ZERO difference to the output of the LPS-1, folks ought to just get the 7.5V/2.93A Mean Well that comes with the LPS-1 (as opposed to saving $15 by omitting it), and use their iPower SMPS warts elsewhere.