ISO Regen performance Improvement Cheap!

[Bricki]

I have changed all Ethernet cables in my network to supra 8 and it is significantly better than the audioquest pearl that it replaced....the improvements are there when you first plug in the supra cat8 but it does seem to take a long time to burn them in... Almost 2 weeks until it settles down and there seems to be an increase in trebble sweetness, depth/layering and more of a naturally rounded sound towards the end of burn in.

[Bricki]

3 hours ago, lmitche said:

OK, so after days of break-in and testing over here the Gotham Gac 4/1 Ultra pro with ends either connected directly to the LT3045 cards or terminated with Oyaide 5.5mm 2.1 DC plugs, beats any network cable (IBRA, UGREEN, or Supra CAT8 with or without a POE adapter). 

 

I have also replaced all my BJC CAT 6a with Supra Cat 8 using the Cable Matters tool less male ethernet plugs at $6 each. Thanks to whoever recommended this cable.

 

Too bad wushuliu chose not to get along.

 

Have fun,

 

Larry

Thanks for this comparison and the recommendation for the gotham ?. It's times like these that I curse the hirose connector on the sps-500 ?... It makes trying new dc cables very difficult. I thought I was done after getting my 4s6 "y" cable made by Ghent ?

[Bricki]

Just to confirm @lmitche... I assume you have already tried the canare 4s6 as a dc cable and the gotham beats it also?? Just double checking.. Sorry if you have already posted this elsewhere 

[Bricki]

45 minutes ago, lmitche said:

Yes, many of my DC cables were Canare star quad from Ghentaudio with JSSG treatment. They are all Gotham Audio star quad cables now.

OK thanks for that...I wonder if I can find a local stockist in Australia ?... Probably not

[Bricki]

Does anyone know what the gauge of the wire in the UGREEN is? 

 

Instead of ordering the gotham straight away, I thought I might try using my old audioquest pearl Ethernet cables. They are Cat7 28awg individually shielded pairs... I got some cheap passive PoE adapters from my local electronics store on my way home and as I was plugging everything together I was thinking "this is stupid, I am adding length to my DC cables and a bunch of cheap connectors".... I DID NOT REMOVE MY EXISTING GHENT CABLE..... I plugged my Ghent canare 4s6 "y" cable that comes out of my sps-500 into the PoE adapters and one 1.5 meter audioquest pearl Ethernet went to power my sMS-200 and the other side of the y cable went to my other length of pearl (through another set of PoE adapters) to power my sonicTransporter i5.

 

This whole thing is completely counter intuitive, I did not remove anything from my DC cables but rather added to it.

 

I have got a very nice upgrade in sound quality... More holographic, more natural, less trebble harshness tighter bass... VERY worthwhile.

 

This is just a guess but the only way this makes sense is if the Ethernet cable is acting as a filter. The twists in the cable are attenuating crosstalk between the pairs and acting as a filter between the psu and the powered device... The filter is stopping noise from getting back into the psu and creating a feedback loop... This is just a guess... If I remember correctly the sotm sps-500 has a filter to filter the noise coming back from the powered device to the psu - however this use of Ethernet cable has still brought a very nice improvement... 

 

The cables I used were 1.5 meters long.... So maybe shorter lengths have less of an impact??? This is just a theory... I haven't tried it yet 

[Bricki]

1 hour ago, JohnSwenson said:

My current ideas as to what matters for a DC cable is inductance and dielectric absorption. The inductance is primarily determined by the geometry of the wires, and dielectric absorption is determined by the insulation on the wires.

 

Starquad is by far the lowest inductance geometry, much better than twisted pair, 1/3 the inductance.

 

PVC (a very common wire insulation) has very high dielectric absoorption, polypropylene and teflon are very low (100 times less than PVC).

 

I think what has been happening here is that different cables have been trading these two off with each other. With the multiple twisted pairs you get inductance higher than starquad, but things like shielding, how tight the twist etc. are causing some differences in the inductance, and different cables have different dielectric materials.

 

I really think we need to try true starquad with low dielectric absorption insulation. The Canare is true starquad but uses PVC.

 

VH Audio is now selling just such a cable V-Quad Cu21 and Cu24  (21 and 24 AWG conductors) with true starquad and very low dielectric absorption. I think this would be a very good cable to try. It's not cheap ($24 or $27 per foot) but it may just be a very good combination.

 

Another cable I have been using for interconnects is Belden 1804A, starquad with polypropylene, but 28 AWG. If you guys are finding that 28 AWG conductors are all right it might be a good candidate. It is a LOT cheaper than the above VH cable.

 

Just some thoughts.

 

John S.

Thanks so much for this very insightful post about dielectric absorption.

 

Because I know very little about this I apologise if my question sounds ill informed. 

 

Is it possible that the lower dielectric absorption of the Ethernet cable I added after my canare 4s6 cable has contributed to the sound quality improvement? I assume this was implied in your original post... But I thought I would ask because the improvement I heard was with both cables still in place... Clearly if the v-quad cu 21 cables can combine the 2 characteristics of low inductance and low dielectric absorption into one cable then it should be an improvement again... 

 

Thanks again for your insights 

[Bricki]

3 hours ago, totoxio said:

Thank you @bit01

 

This is some interesting information: If I understand correctly, ethernet cables could maybe act as a common mode noise filter, in part because each pair of conductors is twisted at a different rate. 

 

Could this explain the magic?

 

https://www.techrepublic.com/blog/data-center/some-interesting-twists-about-ethernet-cabling/

Yeah I'm sure we are hearing the combined benefits of the shielding in cat 7 cables and the twists acting as a filter.... I cut up one of my audioquest pearl Ethernet cables yesterday and made 3 half meter lengths to use with screw connectors... I plugged all in and nearly cried because it was so abysmal ?... I was expecting an increase in sound quality and got a decrease... I put my other 3 cables plus PoE adapters back in and this time I put two 1.5 meter lengths between my canare 4s6 cable out of my sps-500 and my sMS-200...these two lengths sound better than just one... I have already cut up one cable so I am reluctant to try cutting another one to get rid of the PoE adapters... I don't need to worry too much about voltage drop because my sps-500 is at 12v and the sMS-200 can run fine on 7-12v.

 

From my testing these Ethernet cables seem to be acting like a filter and the longer the better (although there must be a sweet spot)... A half meter length terminated with screw adapters didn't sound particularly good and 3 meters sounded better than 1.5 meters.... Obviously this is just what I heard in my system, others may have a different experience. ?

[Bricki]

Is anyone else finding there are further improvements from longer lengths?? 

 

I am now running 3 × 1.5 meter lengths of audioquest pearl chained together (with PoE adapters in between) and 3 lengths sounds better than 2...worth a try if you can afford the voltage drop.

 

I'd be interested to know what others find in their systems... But so far for me - the more Ethernet cable the better - so it's looking like common mode noise rejection on twisted pair is the mechanism for the improvement... 

[Bricki]

31 minutes ago, Quadman said:

 

 

Well I just put my cat 8 in today replacing 3-1M runs of uGreen cat 7 (round).  The cat 8 I cut to the just needed length which in every case but one was shorter that the cat 7.  So I hope you are wrong otherwise I may be stepping back.  First sonic impressions I will report on tomorrow.

Have you tried chaining your 3 × 1 meter lengths together to power your most critical device (endpoint or Iso regen)?? 

 

I'm interested to know what others find.. 

[Bricki]

Just now, mansr said:

I think you may have misunderstood how twisted pair wiring works and the meaning of common-mode noise rejection.

? I may very well have misunderstood ... I'm not going to pretend to know what I'm talking about and I'm sure that you have infinitely more knowledge than I do... I'm just looking for a reason for why longer runs of Ethernet cables sound better than shorter ones when used as dc cables in my system. ?

[Bricki]

26 minutes ago, totoxio said:

@Bricki I will try 2m uGreen round CAT7 tonight. I have been using 1m round and flat uGreen. I will report my first impressions tomorrow. The song I use the most for testing is Babylon sisters, by Steely Dan. It has a great dynamic range, bass, vocals, treble and also shows any sibilances very clearly.

 

 

Thanks, I will try this track when I get home. Looking forward to your impressions too ?

[Bricki]

Yesterday after reading about @Cornan's success with some coax, I picked up some belden 75ohm coax on my way home for AUD$1.80 per meter. This cable has 4 layers of shielding. 2 foil layers and 2 wire mesh layers. I twisted the shields together at each end and taped it together because I wasn't sure if I was going to use it for longer than 5 minutes ?. I was only able to listen for a few hours while it was burning in and my initial impressions are that it's better than the supra cat8, which is what I wanted - so I can put it back as an Ethernet cable. I hope to listen some more after 24hrs burn in and then possibly finish the cable off with some mini JSSG 360's at each end (I don't have enough braid to go the full length). 

So far it sounds very dynamic with quite a rounded bass now. I'll see how I like it over the next couple of days. Thanks again everyone for sharing your experiments and ideas! ?

 

[Bricki]

Yeah it was jaycar. The one I got was this one. 

https://www.jaycar.com.au/belden-brand-quad-shield-rg-6-75-ohm-coax-sold-per-metre/p/WB2009

 

It has 4 layers of shielding. I think it's slightly different to the one @sandyk linked above. 

 

I'm not promising that it will sound amazing. But it's a cheap experiment ??

[Bricki]

6 minutes ago, sandyk said:

This close to 7.6mm diameter coax is very inflexible.  ( I have some)

 For DC use it is highly unlikely to outperform the cheaper Belden cable at 6mm OD after modification either.

Yeah it is very annoying and inflexible. I chose it because of the shielding and it was a cheap experiment. There may well be much better options at jaycar ?. 

[Bricki]

Thanks @Cornan and @charlesphoto for your on going experiments. I have been trying to follow along and have had some failures with some cables sounding worse than the coax one I made last week. However, I have moved my lps1 to sit on top of my sMS-200 separated by a wooden block and isolation discs and made the shortest possible positive wire with the largest gauge cut down to fit in the screw terminals. Then the same gauge negative wire somewhat longer to keep more distance between it and the positive. I have had it burning in for 24 hours now and it makes the lps1 like a new psu. Everything is a lot cleaner but not in a washed out clinical way - cleaner in a natural and relaxed way with natural tone and texture. Annoying screach on brass instruments and some vocal pieces has been hugely reduced, restoring a more natural and far less irritating sound. There is also more depth ?. Now I'm gonna have to order some of those kemets. Where's the best place to purchase? Mouser wants to charge $24 to ship to Australia ?. 

[Bricki]

@Cornan

 

Are these the same ones you are using? 

 

https://au.mouser.com/ProductDetail/KEMET/A750MS567M1EAAE015?qs=M6jHmRuQorWa3alz81gcwg%3d%3d

[Bricki]

8 minutes ago, Boomboy said:

@Bricki i had some sent over to oz from this seller on ebay .. took a couple of weeks . Im trying then this fri/weekend .

?Excellent! Thanks for that, that looks like a much more reasonable postage price ?

[Bricki]

11 minutes ago, Cornan said:

 

It is the correct Kemet A750 you´ve got there. I buy mine from RS-Components https://se.rs-online.com/web/p/aluminium-capacitors/1248271/. Their freight charges are just slightly better than Mouser´s though, so I think the suggestion from @Boomboy is a good option if you order just the Kemet´s!  

OK thanks ?

[Bricki]

Would someone be able to recommend the best way to implement some LT3045 boards between an sps-500 and sMS-200? 

 

I am an LT3045 newbie so I have a number of questions.... 

 

I have my sps-500 set at 12v because I have a "y" cable that is powering my sonicTransporter i5 that requires 12v. 

 

If I recall correctly, @Cornan said that it's best to drop the voltage from input to output by 3v if using 1 board and with boards in series it's best with 1v per board.

 

Given that my input voltage is 12v then I assume my best result would be to order a 9v board and a 8v board?? Is this correct? And I assume my sMS-200 will be fine at 8v because it's fine at 7v??

 

Also, do the boards come with the screw terminals already soldered on? 

 

Sorry for the dumb questions ??

[Bricki]

57 minutes ago, sandyk said:

 The short answer is that you can't use a single LT3045 .5A PCB because of the current draw of your sMS-200.

It's specifications are : Power requirement
  Input Voltage : +6.5 ~ 12Vdc
  Power : Max 15W

 This means that with an 8V supply you would need to be able to supply almost 2 Amps.

No worries, thanks Sandyk. 

 

I was looking at these boards which I think are 1 amp... Are these the ones others are using? 

 

http://www.ldovr.com/product-p/lt3045-a.htm

 

Also, the sMS-200 is fine at 1 amp 7v as long as you don't intend to use the functionality for an external usb drive or plug in a dac that requires the 5v for it's own power. So in my particular case it's fine at 7v 1 amp ??. Or at least it's fine at 7v 1.1 amp because I have also powered it from my lps1 ?

[Bricki]

28 minutes ago, jaaptina said:

I use two 1A LT3045 boards in series between LPS-1 and ultraRendu. No problems. No heat. But voltages are much closer to each other. 

What output voltages do you have on your boards? What's the input voltage range of the ultrarendu?