The true experimental tweak thread

[lmitche]

I look forward to hearing the results of this test.

[lmitche]

Cornan great to hear it worked out well!

 

Wait until you hear what a LPS-1 or two and the ISO Regen bring!

 

Larry

[lmitche]

11 minutes ago, Cornan said:

 

I wish those electronic repair shops would be still around in Sweden. It is really hard to find good ones these days. If you find them they do not know zip about floating power supplies for sure. I use to have a guy that did all my audiophile soldering jobs a couple of years ago. I used him atleast 2-3 times a year. Now he is gone and I have'nt found anyone to replace him! ?

Vinyl repair shops are still quite easy to find, but I am pretty comfortable in that perticular area! ? I can solder and have reasonable good gears, but it is definately not a part of my CV! ?

I don't know if he is local, but this fellow may be able to help you.

 

http://sjostromaudio.net/shop/

[lmitche]

11 minutes ago, austinpop said:

Wait - aren't we the grounding nuts? 

I thought you were a time-nut!

 

http://www.leapsecond.com/time-nuts.htm

[lmitche]

1 hour ago, agladstone said:

I just saved you about 1 month of agony and decision making and testing back and forth over and over, hahaha (well at least that's been my past month! ) 

anyhow, if you're only using just the ISO Regen and no other USB devices like I am, I would say you have only two choices:

ootion A - USPCB out of computer / server into ISO R and then a USB Cable out into Brooklyn DAC 

Option B - the exact opposite:) 

- The Cardas Highspeed 2.0 USB cable is outstanding (especially with the Brooklyn) and I think I've tried them all ! 

However, it's NOT the cheapest out there! (If I recall, I think I paid $375 USD for a 700CM cable)? 

 

You may want to try the Lush cable from Phasure. For 200 euros maybe the best USB cable ever. Fantastic when used between the ISO Regen and DAC.  

[lmitche]

19 minutes ago, austinpop said:

 

Very interesting. That is the same DAC that Larry @lmitche swears by. Although, he has posted - can't remember in which thread - that it really comes into its own using HQPlayer to upsample to DSD512.

 

If you ever get your hands on it again, and want to try DSD512 upsampling, it would be interesting to get your take on it. You can always get an eval copy of HQPlayer.

Hi all,

 

Yep, the IFI microIdsd black label with Hqplayer DSD512 upsampling is what I use here.  Once heard, I sold my Brooklyn DAC, which I owned for about 4 months. I haven't looked back.

 

Mozes, if your system has the juice, you should give dsd512 a try with HQPlayer and Windows 10.  Let me know if you want some help on setup. There are 3 to 4 tricks that make a big difference (and 1000s that make a little difference), LOL.

 

Larry

[lmitche]

23 minutes ago, Cornan said:

 

All of this talk gives me flash-backs. Flash-backs of my Gopherts actually. They will improve in lower voltage range. If used in a higher voltage range the noise will increase quite dramatically towards the end of the voltage range. I wonder if this is not the same scenario with the iFi micro DSD being more immune to noises at higher upsampling than the Brooklyn DaC? 

The microIDSD uses two parallel Burr-Brown DAC chips called a DSD1793.  The designer of this chip reached out to me and shared that the internal DSD processing is separate from the PCM side, and uses a simple internal shift register for DSD playback rather like the T&A passive DSD DAC section. The DAC specsheet suggests these DACS are limited to DSD256, but with two of these chips in the microIDSD IFI have enabled DSD512. I have had a T&A DAC here several times for several months, and the IFI is damn close in SQ.  The biggest difference in SQ I attribute to the analog side with single ended output vs. the balanced output design of the T&A.

 

The microIDSD BL is meant to be used as a mobile dac powering headphones, but used as a stationary DAC still sounds best on it's internal battery which creates another level of isolation.

 

I am hoping that someday IFI will ship a balanced output version of this DAC as they have been threatening for many years.

[lmitche]

4 minutes ago, mozes said:

Thanks Larry, my Nimitra has a low powered cpu, it won't do dsd512 upsampling and I no longer have the IFI dac. I used the internal battery and if I powered it from an LPS-1 it would be better I guess as my Brooklyn is powered by a VR Mini

Yes, if I had a Brooklyn that is the PSU I'd use.

 

My Brooklyn was powered by a 12 volt sigma 11 lpsu that delivered a nice increase in SQ.

[lmitche]

5 hours ago, Cornan said:

More great news! One of my absolute favorite bands have released a new album.

 

"A Fever Dream" by Everything Everything

 

 

This is definately a must listen to. A great album! ?

Cornan, music for orgies?

[lmitche]

15 minutes ago, TubeMan said:

and what level is that,,,,,,

I think everyone here is looking for better sound ,,

Ludricrous level sound!

[lmitche]

1 hour ago, Cornan said:

 

Here is an interesting claim by Per-Anders Sjöström. Any support for his claims? ?

 

http://sjostromaudio.com/pages/index.php/hifi-projects/142-ssr01-sjoestroem-super-regulator?showall=&start=1

Is "Super" too much?

Does the design deserve the epithet "super"? Yes, indeed! Can anyone come up with any regulator with less noise and lower output impedance? I don't think so. I would be happy if someone can prove me wrong. This design is extreme in the true sense when it comes to these two parameters. I believe that a couple of more parameters also are state of the art.

 

Quite a lot more expensive than other choices, but very appealing to me I must say! ?

 

http://sjostromaudio.net/shop/super-regulators/18-ssr02-sjostrom-super-regulator-built.html

 

And here is the technical data:

 

Operating voltage:	Max 50 V, less current at high voltage in. Min 4 volts more than output voltage.
Output voltage:	5-40 V, down to 3 volts is possible.
Hum and noise at full output current:	Down to 0.9 µV or below depending of component choices
Max current:	Approx 1 A peak
Max continuous current:	Approx 1 A depending of ambient temperature and mounting (cooling).
Dimensions:	66,0 (2.6") x 38,1 (1.5") mm

Isn't this fellow one of your neighbors?  Have you met him?

[lmitche]

1 hour ago, Cornan said:

I actually wondering if not the voltage regulator in the Regen Amber that is isolating the leakage loop? I have decided to buy three LT3045s and have one in front of the Aqvox, one in front of the IR and one in front of the Aries Mini (15v) for starters. It will be interesting to see what happens if I then remove the Regen Amber and the 5v injection. If it sounds good with ISO on I will be onto something really interesting. At least for me personally! ?

 

I'll guess that when I used the Regen Amber as a voltage regulator for the Aqvox switch that loop went right through the IR (ISO off) and Brooklyn. No wonder it did'nt improve things.?

Thanks for the USPCB Amber Regen suggestion.  I am listening to one now plugged into a new $14 USB card PCIE card, with both powered by a 5 volt lt3045. The USB card has 5 volt injection on a sata plug at the back of the card. A y-cable connects them together. A lush cable follows and last an ISO regen/USBPCB powered with a 6 volt lt3045.

 

Wow!

[lmitche]

1 hour ago, octaviars said:

As I get it it is all about how much voltage the regulator needs to drop the voltage as it will turn into heat. With LT3045 I would try to supply it as low voltage as possible but enough to still be able to regulate.

If you look at the scematics with 4 parallel LT3045 it says 300mV dropout which means that if you need 5v out you will need 5,3v in but that is perhaps on the edge so 5,5v may be on the safe side. You could always put a pre regulator board with a LT1083 or something like that with adjustable DC output to set it at right level before entering the LT3045.

 

 

I have two 5 volt 1 amp ldovr lt3045 cards running with 7 volt input from my lps-1s for the past week with no heatsink and no problems.

[lmitche]

5 minutes ago, octaviars said:

 

That sounds good but do you know how much power the stuff you have connected draws?

I'm powering a USB NIC, Adnaco USB box and HDD in an enclosure on one, Microidsd DAC and ISO REGEN on another,  last an Amber Regen. The LPS-1 on the HDD is the hottest. I don't have measurements of the current draw.

[lmitche]

5 hours ago, Forehaven said:

Is the pin on the right side the ground pin?

The three pins are symmetric.  Just solder the three into the board.  You can then read the markings in the card to learn where to move the connector to power up the card.

[lmitche]

4 hours ago, Cornan said:

 

I know Moussa! ? I am all for less, but to optimize things you need to make it overly complicated to understand what really matters. I use the voltage regulators as isolation for the DC path. Not intirely sure yet if they will work as intended, but as soon as I have got the last voltage regulator for the Aries Mini I'll give it another go. Hopefully that does it. Touch wood! ?

 

I have actually plans in a couple of weeks to rip out everything in my A setup and start all over just to find out what really matters. There is just too many things that have happened along the way, so I'll need to get a little perspective. I also want to try different AC mains power routes.

When I read about your experiences, I grow concerned that you are just moving the leakage current around.  I expect you would be well served with some good linear power supplies instead of the gophers. That should let you simplify things.

[lmitche]

19 minutes ago, Cornan said:

 

It is two LT3045 voltage regulators in series with 1v drop-down. This is a tip from @tapatrick based on RB2013/Tubelover2's posts. The LT3045 are ultra low noise voltage regulators that reduce the noise and probably also reduce leakage in the DC path.

They are very simple to make and will work great on most type of power supplies. Not only Gopherts at all.

 

Here is pictures where you can see the construction better.

 

 

 

I have made two versions with LT3045s with following output voltage:

 

1. For Aqvox switch. 6v & 5v. Powered with 7v.

2. For ISO Regen. 8v & 7v. Powered with 9v.

 

Cables used are from Canare 4S6 cables (just the double wires).

 

I hope that helps! ?

 

 

I've got three more lt3045s to use in series with the 3 already in use.  I hope to get the soldering done tomorrow. These little things really have a spectacular impact on SQ.  Glad to hear you were successful.

[lmitche]

17 minutes ago, elan120 said:

 

With these power supplies, I don't think using the LT3045 regulator would be a good idea.

Why would you say that?

[lmitche]

43 minutes ago, elan120 said:

 

That will limit the current output down to just 1A, and in addition, these are great quality power suppliers Ted has with very low output noise already, plus stepping down from 9.5~12VDC down to 5VDC is rather steep.

7 to 5 volts is working well here from an lps-1 feeding an USB card internal power  connector. There is a definite SQ benefit. @tapatrickwhat's the biggest voltage drop you've successfully used? I agree 9 to 5 is likely too far but then 9.5 to 7 to 5 should work fine, so it's a good use of serial lt3045s.

 

The lps-1 and Iso Regen both use linear technology LDOs internally.

 

I haven't tried grounding these LDOs and see no need to do so here as everything is lps powered except the PC motherboard which is galvanically isolated via all connections.

 

 

[lmitche]

41 minutes ago, elan120 said:

I should have more details from my earlier post.  The power supplies from Ted's list I don't think would be a good idea using the LT3045 are the two Hynes and the JS-2.  On the other hand, I have been using single LT3045 (not the parallel LT3045) to step LPS-1 down from 7VDC to 5VDC to 3.3VDC for my SU-1 with very good result.

I'd be very surprised if the lt3045s didn't improve the SQ of the devices powered by the JS-2 and PHynes supplies.

[lmitche]

15 minutes ago, Forehaven said:

Larry and tapatrick, if you've already tried, what the incremental improvement by going 3 over two boards in series.  Other than specs of .04 to .02uv is all I've read/seen so far.

 

I haven't tried this and so can't comment. I have no plans for more boards.

15 minutes ago, Forehaven said:

 

And I have a Teradak DW30 coming...how important is it to exchange the caps?  I have never done an exchange before.

Again, I have no idea.

[lmitche]

23 hours ago, lmitche said:

I've got three more lt3045s to use in series with the 3 already in use.  I hope to get the soldering done tomorrow. These little things really have a spectacular impact on SQ.  Glad to hear you were successful.

Spent the morning soldering, and early afternoon listening.  Thumbs up on the serial lt3045 solution.  More blackness, image stability and timber are the results here.

 

Recommended!

[lmitche]

8 minutes ago, sig8 said:

Larry: You are using 2 or 3 regulators? 

2 in serial, for the Adnaco and usb card. One 6 volt for Regen with split to 5 volt for DAC vbus injection.

 

Moved ISO Regen to behind PC feeding lush cable to DAC. Just an incredibly rich dense image with high detail and depth.

[lmitche]

20 minutes ago, tapatrick said:

I've been studying the data sheet for the LT3045s particularly the Applications p15-26 and doing some experiments.

 

I had some 0.1uF 630V Orange Drop Capacitors and tried them with the Elnas and this did help the treble as you suggested @sandyk, also has cured an occasional instability on startup.

 .."However, if a distantly located supply powers the LT3045-1, use a larger value input capacitor. Use a rough guideline of 1μF (in addition to the 4.7μF minimum) per 6" of wire length. The minimum input capacitance needed to stabilize the application also varies with the output capacitance as well as the load current. Placing additional capacitance on the LT3045-1’s output helps. However, this requires significantly more capacitance compared to additional input bypassing. Series resistance between the supply and the LT3045-1 input also helps stabilize the application; as little as 0.1Ω to 0.5Ω suffices. This impedance dampens the LC tank circuit at the expense of dropout voltage. A better alternative is to use a higher ESR tantalum or electrolytic capacitor at the LT3045-1 input in parallel with a 4.7μF ceramic capacitor."..

 

I also added some tiny ferrite beads on input, as mentioned in the data sheet in case of very high frequency spikes. Probably not needed but..

..."However, the very high frequency (hundreds of MHz) “spikes” — beyond the LT3045-1’s bandwidth — associated with the switcher’s power switch transition times will almost directly pass through the LT3045-1. While the output capacitor is partly intended to absorb these spikes, its ESL will limit its ability at these frequencies. A ferrite bead or even the inductance associated with a short (e.g. 0.5”) PCB trace between the switcher’s output and the LT3045-1’s input can serve as an LC-filter to suppress these very high frequency spikes.

 

http://cds.linear.com/docs/en/datasheet/3045fa.pdf

Tapatrick, are you powering an analog or digital device with your tweaked lt3045 board?

[lmitche]

28 minutes ago, tapatrick said:

Hi @lmitche I'm powering a chain: Aqvox switch, ISO Regen and Bluewave USB to spdif board, each has a dual in series pair of LT3045s, see signature for details. I'm engaged in lots of experiments but all sounding superb..

Thanks, got it.  I'm always surprised to hear analogish tweaks like audio capacitor quality impacting digital components. While I appreciate that a digital signal is transmitted in the analog domain, a one to one correspondence between a capacitors quality impact in the analog world to the digital world surprises me. I am sure what you are hearing is correct, so thanks for sharing.