DIY DC power cables

[BigGuy]

On 5/4/2017 at 8:26 PM, JohnSwenson said:

OK here it is: cable shielding, how to make it work and how almost all cables have it wrong.

 

The first important question is, what is electrical shielding and how does it really work?

 

Lets look at the infamous "Faraday cage". For this discussion it is a metal box, with all sides well electrically connected (no gaps). Shielding means that electrical fields outside cannot be sensed inside. Conversely electrical fields inside cannot be sensed outside.

 

So how does this work? When an electric field from the outside world impinges on the box it causes the electrons in the metal to move, this rearrangement of the electrons creates an electric field inside the metal which exactly counteracts the external field, thus the field is essentially stopped at the periphery of the box.

 

The important aspect here is the part "electrons in the metal move". What are moving electrons called? Current. In order for current to flow there has to be a loop. Current will not flow unless there is a loop. In the case of the Faraday cage, the loop is the whole cage. Lets say an electric field impinges on a side of the box, this causes a current flow that goes all the way around the box back to where the field impinges on the box. If the sides are not electrically connected to each other the current cannot flow, thus the electrons cannot move, and the canceling electric field does not get setup, thus no shielding.

 

It turns out that for AC electric fields it is a little different. Some current can flow due to capacitances between conductors. The electrons can move a little bit one way, then a little bit back. If the frequency is high enough the back and forth movement, which is charging and discharging the capacitance is sufficient for shielding. For a given capacitance the lower the frequency the less effective the shielding. The electrons start moving which charges the capacitance, then stop moving when fully charged, they don't do the full movement necessary to produce the canceling field.

 

So what about shielded cables? I hope is now obvious that for shielding to be effective there needs to be a conductive path from one end of the shield to the other. If there is not such a path the only shielding that is going to happen is for high frequencies due to cpacitances involved with the shield.

 

The best way for the shielding to work properly is a separate wire connected to each end of the shield. This is sufficient for shielding from DC to very high frequencies. Note the shield does NOT have to be connected an earth ground, the "ground" of the circuit at either end, or any thing else for that matter. A cable with a shield the is not connected to anything else except itself (ie a separate wire from one end to the other of the shield) will be highly effective in shielding what is inside.

 

Where does this wire need to go? It can be either inside or outside the shield, but if it is inside it can couple to the signal wires inside, so it is usually best to have it outside the shield. Note it has to be insulated from the shield except for the ends where it connects to the shield. It should intersect as little of the external field as possible so it should NOT be tightly spiraled around the cable. Just running along side the shield is best, although a very loose spiral (say one turn per foot) is almost as good.

 

So some ramifications of this: The traditional "connect the shield to one end and let the other end float" is not good, it does not allow a loop so shielding does not happen very well. If you add the external wire connected to the shield at both ends, then you CAN connect one or both sides of the shield to the signal ground or some other ground, but you don't NEED to for effective shielding. You will find that in many cases leaving the shield completely disconnected from the rest of the circuit is the best way to go, you get the benefit of properly working shielding without any interaction of the shield with your system. You may wind wind up with static charges on the shielding so a resistance from the shield to ground may be useful in some cases in order to dissipate static charges.

 

So how come nobody does this? I don't know. My only guess is that cable shielding has been going on long before the actual mechanism for shielding was worked out, thus by the time it was understood, cable shielding was "standard" and nobody ever even thought about analyzing it based on an understanding of how shielding actually works.

 

But shouldn't the big companies know about this? It seems they don't. I have read several app notes from Belden that state that shielding is only effective at high frequencies, at audio frequencies and power supply frequencies (60Hz etc) it is totally ineffective.

 

Audio people are the only ones that seems to at least empirically know about this. Remember phono cartridges and preamps, there is a little green wire that goes from the "ground jack" on the preamp to the tonearm. Everybody assumes that this is to "ground the cartridge" but what it really does is provide a loop from one end of the interconnect shield to the other, it has nothing to do with whether it is "grounded" or not. So if you have (or had) a turntable you were actually taking advantage of this without realizing it.

 

So there you have it, shielding DOES work, but only if you provide a path from one end of the shield to the other. This is effective even if you don't connect the shield to anything else.

 

John S.

 

 

 

 

 

 

Thank you, John.  This has got to be one of the best DIY improvement posts and one I will be trying.

 

A couple of questions...

 

For John:  Given the thread topic, the JSSG has great applicability to DC power cables.  Does it also offer value for other cables, e.g., USB, interconnects, AC power cables, etc., and, if so, how would they rank?

Which would be better...tinned or copper braid?

 

For anyone:  Can anyone provide sources of braided shield here in USA that people have done business with?

I bought some copper braid shield from a motorcycle shop some time ago intended not only for protecting but blinging-up gas line tubing.

[BigGuy]

12 hours ago, JohnSwenson said:

The JSSG is primarily effective at lower frequencies. As mentioned in my post there is enough parasitic capacitance in most shielded cables to provide shielding at high frequencies. Where traditional doesn't work well is low frequencies. SO audio cables are a great place for it. High frequency cables (such as USB etc) no so much. If done right it would be effective at shielding power line frequency interference  for high speed cables (which MAY be useful), BUT unless the cable was specifically designed for it adding it to an existing cable will more than likely mess up the high frequency performance of the cable.

 

John S.

Thank you, John, for the additional explanation and the caution re applying JSSG to high frequency cable, e.g., USB.  The few DC cables in my system are less than 12 inches in length so guessing it might not be worth the bother to do these.  Any thoughts about the length where the shielding effect would become significant.

 

lmitche, I did see offerings of tinned copper braid on Amazon.  Given what I would expect to be relative ease of soldering to copper  sleeve, any special reason to pick tinned?  If not, then I guess it becomes issue of aesthetics.

[BigGuy]

8 conductors!  Thinking that would be star quad for D+ and D- plus shielded star for 5V and GND.  Are the data conductors also shielded?  Hard to imagine trying to get all those conductors soldered to the pins?!  I had a design in mind for a DIY cable but was not considering star quad.  This thread has likely changed my  mind.

 

Looking forward to your impressions.

 

FWIW, CAD has announced 2 new USB cables.  Gulp!

https://daviddenyerpr.files.wordpress.com/2017/09/press-release-cad-usb-cable-i-ii-2017-09.pdf

[BigGuy]

29 minutes ago, Speedskater said:

Fair Radio in Lima Ohio has a barrel full.

https://fairradio.com/

I think barrel is empty since I could not find any reference to shielding, etc. in their full product list.

 

While on the subject I have been reading that Mu-metal provides better shielding properties than copper or tin-plated  copper braid shielding.  Definitely not a shielding expert myself and, in this case, a little knowledge may be a dangerous thing!

 

Anyone have further info on this with potential application to cable shielding?

 

Just got in the parts to make up some grounding umbilicals for my LPSs as well as an outlet box for power supply and grounding plugs.  To someone's comment, I DID buy GREEN wire for the umbilicals and a few of these for fun... 

http://www.hardwareandtools.com/cooper-wiring-bp3867-4gn-high-visibility-plug-fluorescent-green-efda-7490.html

[BigGuy]

1 hour ago, Speedskater said:

Fair Radio is not an internet store. It's a dusty old building with stuff sitting on shelf's for decades.

The braid (in random lengths & sizes) is in a 55 gallon cardboard drum near the door.

Send them an e-mail or phone.

Ah!  That would explain why could not find "barrel" on their list! 

 

I found this supplier of copper braid...

 

<https://www.techflex.com/land_metal.asp>

 

Speaking of copper braid, got some clarification on use of Mu-metal.  Mu-metal is best used for EMI, which is low frequency.  Copper is the best for RFI, which is high frequency and likely the biggest issue with cables.

[BigGuy]

18 minutes ago, sima66 said:

I will use DIY Canare 4SE6 cable.

Do you think that 8 wires will fit into one hole......if I order two way only?

Does anyone know what is the connecting material inside of the terminal block?

I have used the Wago connectors which have metal to metal contact inside the connector and the wires are secured extremely well.

https://www.amazon.com/s/?ie=UTF8&keywords=wago+connector&tag=mh0b-20&index=aps&hvadid=78340207358489&hvqmt=e&hvbmt=be&hvdev=c&ref=pd_sl_2binredxhe_e

 

Number of wires per hole would be predicated on wire gage.

[BigGuy]

6 minutes ago, sima66 said:

 

I ordered, just in case, this 5 way 10 pieces, shipping included for $2.39.

 

https://www.ebay.com/itm/2-3-5-Way-Reusable-Spring-Lever-Terminal-Block-Electric-Cable-Connector-Wire/142486098372?_trkparms=aid%3D777003%26algo%3DDISCL.MBE%26ao%3D2%26asc%3D41375%26meid%3Df79f397fdfe148ea9d402500424cd679%26pid%3D100012%26rk%3D3%26rkt%3D7%26sd%3D332401300319&_trksid=p2047675.c100012.m1985

 

I don't know what is the difference between this ones and the Wago from Amazon....beside the price!

Nice find.     Good price as well.   Probably the same difference as Coca Cola and Cola!  ;-)

[BigGuy]

9 minutes ago, Cornan said:

 

Alex is Aleksandar Tsankov at ATL Hi-Fi https://www.atlhifi.com who made many of my ac mains starquad cables, DC cable with JSSG, starquad & star-earth wired PSD, balanced IT with floating & grounded center tap and of course the USB starquad cable with JSSG.

 

The construction of the USB cable is clearly visable on this picture.

 

 

It is a starquad, so all wires are double. Totally 8  conductors. GND plus 5v and Data- plus Data+ wires are housed in separate shielding sleeves. The insulated JSSG loop is soldered to all four shield ends and are not in contact with the metal parts of the USB A or B plugs. That's basically it!

 

As I said in the other thread I am in contact with Aleksandar and we are working out a optimized version of the USB cable with Mundorf gold/silver wires, silver plated shield sleeves and better plugs. I am only doing this because I think it will be my final USB cable. I want it to be as good as it can be and not unimportant good looking as well. I want it to look something like this cable in the end.

 

 

Anyway, the details are still under discussions. I might change the GND wire to pure silver since I beleive that all 0v related wires usually sounds better with high silver content.

 

The price for the original one is around EUR30/pc which is very affordable, especially when you consider the hassle to solder 8 conductors to a small USB plug. I will let you know what the new one will be in the end, but based on the material it will of course be a lot more. Since Aleksandar is away for a week or so and are working on shunt regulators with starquad Kelvin cable and plugs for me as well (for DC cable path) it might take 2-3 weeks before I know how much it will be.

Your description of the custom cable suggest that all bases have been covered but have couple of questions...

 

1) Is it not better to have the 4 data conductors twisted and then shielded rather than D+ separate from D-?  (same with 5V and GND)

2) what was the length of the original 30Eu ($35USD) cable?

[BigGuy]

7 minutes ago, Cornan said:

 

1) They are separate! The GND & 5v is in one shield. The Data- and Data+ is in the other separate shield. The JSSG loop is soldered to all 4 shield ends and are not in contact with bare wires or plugs.

2) 50cm

 

?

Thanks.  Misunderstood "the four shield ends" as being 4 shield rather than 4 ends (of 2 shields) so design makes sense.

 20cm probably keeps it fairly flexible which is good thing.

 

Does gold/silver or even plain silver offer measured advantage for USB transmission vs copper?  I understand that silver is better than conductor than copper but, of the three, gold is the worst.  Or, is gold there to prevent silver corrosion>

[BigGuy]

8 minutes ago, Cornan said:

 

No problem! ? It is 50cm, not 20cm. It is not as flexible as normal USB cables due to the construction, but it is surely flexible enough (it is more flexible at a one direction and less flexible at the other direction).

 

I have used Elijah Audio Konvertible Lite before. It uses Mundorf gold/silver wires for Data and Audience wires for GND+5v. I like the sound of the Mundorf gold/silver wires. That's basically why I want them, plus that the Mundorf is silver cables with gold plating which makes the sound more full. Pure silver on positive wires can sound too thin in my own experience. Pure silver wires for negative is knowledge I have gained through grounding experiments during many years plus also via different cables using gold for positive and silver for negative wires that usually sounds great. In the end this is a cable that I pick the things that I beleive in to make it a optimized version of the basic one. I cannot say for sure if it will improve much, but I beleive it will! ?

Yes, 50cm, not 20.

 

Interesting observations re copper, silver and gold.  Will be interested in getting the cost with cable configured this way. 

If "reasonable", may also get in the queue.

[BigGuy]

14 minutes ago, BigGuy said:

Yes, 50cm, not 20.

 

Interesting observations re copper, silver and gold.  Will be interested in getting the cost with cable configured this way. 

If "reasonable", may also get in the queue.

FWIW, I am using the Pangea 1/2m Ag USB cable whose construction looks to be similar to what is described with the custom cable other than it not being starquad...

• Designed specifically to deliver high-speed digital audio signals
• 24-AWG design optimized for fast, low-jitter signal delivery
• Pure, solid silver signal conductors
• Precision-twisted conductors for superior noise rejection
• Triple-shielded for optimized noise isolation
• USB High Speed 2.0 rating supports up to 480 Mbit/s data transfer rate
• 24K Gold-plated contacts for superior electrical contact
• Special two-shot molded connector bodies protect vital digital connections for stable signal transfer and long life

 

[BigGuy]

7 minutes ago, charlesphoto said:

 

What do you think the minimum gauge is that would work with the JSGT? Looking at silver wire prices they get up there, and I’d need a number of feet. 

+1 on a suggested gage.  My issue is cost/length as well.  May need to prioritize uses to those power supplies closest to DAC.

 

Anyone buy insulated silver wire from a particular vendor in US?

[BigGuy]

3 minutes ago, charlesphoto said:

So far this is about the best price I've found on bulk silver wires. Scroll to almost the bottom of the page for bulk pricing with or without a teflon sleeve. 

 

http://www.tempoelectric.com/cables.htm

Yes, the Tempo Electric ARE the best prices I have seen so far.  Nice find plus you can get "flavored" jackets!   Way better than VH Audio at $6.00/foot which at that price would be better suited to a necklace for my wife!  ;-)  Since I need on the order of several feet for grounds, the cost with VH starts approaching or exceeding the cost of the protected device!

 

Even at Tempo prices it will be minimal application.
 

[BigGuy]

6 minutes ago, Cornan said:

 

+1 Those new Oyiade DC plugs are much smaller than you might think (unless the good old Oyiade plugs is used that for example Uptone Audio uses for their DC cables. AFAIK they are not possible to buy anymore) @fayerichard. I have messed it up soldering Furutech Alpha-18 (18awg) before ordering Ghent DC cables. I am not a soldering pro though. It is very easy that the positive and negative poles come into contact using greater gauge than 18awg using new Oyiade IME.

A trick recommended to me for DIYing USB cables, but applicable here, is to separate conductors, etc., with Teflon plumbers tape to avoid unwanted contact points.

[BigGuy]

Speaking about cables...

 

Hope I am not going too far off-thread but thought someone might be able to help.

 

Anyone ever come across a rigid USB A-Male to B-male adapter to replace actual cable?

[BigGuy]

20 minutes ago, Cornan said:

 

I guess you haven't got a USB Regen or ISO Regen, but isn't this rigid enough? https://www.amazon.com/HDE®-USB-Type-Male-Adapter/dp/B002B9AR8I

Thanks.   Guess I just searched on the wrong description.  :-(

 

Been  thinking about the ISO.  Is it significantly better SQ than the original?  I guess at some point I need to deal with jitter, etc.

[BigGuy]

 

1 hour ago, charlesphoto said:

You mean like this: https://uptoneaudio.com/products/uspcb-a-b-adapter

 

sounds amazing. Can’t go wrong.

Not breaking the bank to try one but I was questioning the SQ of circuit board+switch vs a few wires.

 

Speaking of switches for the 5V, I had been using the SBooster VBus and later VBus2 between USB cable and my Manhattan I DAC.  I recently upgraded the DAC to the Manhattan II, which incidentally was a surprising jump in SQ, the VBusses did not work.  Apparently the new configuration needs to see the 5V even tho' it is self powered.  Does anyone now or perhaps posit an explanation as to why some USB implementations need to "see" 5V and others not?  I DO like the concept of eliminating the 5V as a source of pollution.

[BigGuy]

1 hour ago, Cornan said:

 

Without the metal housing on new Oyiade DC plugs it will fit. With the metal housing may good be with you! ?

If you take the advice from @BigGuy and wrap the ends in Teflon tape (very thin) you've got a slightly better chance to succeed. Since you are in the border of what is possible or not, all I can say is go for it and good speed if you try. I know Uptone Audio sells their huge gauge DC cables separately. If you really want to know how a thick DC cable sounds like I would warmly recommend that you buy it from Alex C instead of going into mission impossible.

Having recently constructed a DC cable using "new(er)" Oyaide barrel plug, I can attest to the fact that DIYing with large gauge wire would be a challenge.

 

This brings up a question from an "inquiring mind"...

 

I know that 12V auto jumper cables are typical quite large gauge but here we are talking about large current draw, not voltage.  If we are supplying 12V, etc., to devices that are drawing on the order of 1A or so, how low a gauge is electrically justifiable?

 

[BigGuy]

4 minutes ago, Cornan said:

 

I have the Brooklyn DAC and the change from my previous Pioneer U-05 was troublesome. U-05 didn't require 5v for handshake but Brooklyn did. In summary the jump from U-05 to Brooklyn wasn't as far as the cost. I was a bit dissapointed. I had to find another solution. IR I thought. Nope. I had to get a Teradak USB power splitter to get it to sound as good as the U-05. It was not until I got the BluWave USB to Spdif with a powerbank powering the GI output of the BluWave that I really thought that the Brooklyn DAC was worth the money. Now it is worth more than double price of U-05 for sure...but it took me roughly a year to end there. 

I too would recommend replacement of the OEM SMPS for the Brooklyn with an external power supply.  A friend's Brooklyn with Sbooster power supply was amazingly good...pretty close to Manhattan I, IMHO, which I owned.

 

Although I have heard/read good things about the BluWave, not sure I want to go there given I spent considerable money on the Adnaco Glass/USB setup.

[BigGuy]

4 minutes ago, Cornan said:

 

IMO it is due to distance. Keep it short and do not mind the gauge. If you need >50 cm think about greater gauge. DC does'nt require greater gauge than Canare above 50cm. JSSG will improve SQ even if it is a starquad configuration IME.

Guess I need to rethink the length of the DC cable that I grafted onto a "brick" laptop LPS to supply 5V to the Adnaco.   I am guessing it is over 6 feet long, a decision made thinking to improve its long term use and geographic location relative to the target device.

 

 

[BigGuy]

11 minutes ago, Cornan said:

 

Think like this. The less power the less lenght. The more power the more lenght. 5v equals roughly max 5-50cm. 120v equals roughly max 100-120 cm. 240v equals roughly max 200-240cm. Mind you that this is just rough lenghts based on own experiences...but makes it easy to remember.

 

Good crutch!  At my age, these can be important.  ;-)

[BigGuy]

13 hours ago, JohnSwenson said:

I've been doing measurements of this, they show that at least for digital circuits, the inductance of the cable is probably more important than the raw DC resistance. Many digital devices do not draw a continuous current, the load can fluctuate rapidly and by a fair amount. The impedance of the cable at the higher frequencies (above line frequency) can be significant. A piece of 14 AWG romex  has a much higher impedance than a 20AWG starquad.

 

When I started measuring all this I was rather stunned by how much impedance regular cables add the higher load frequencies. The results at the load are MANY MANY times worse than the regulator itself. Starquad has 9 times lower inductance than zipcord, that makes a big difference, even for short cables.

 

I haven't done much checking of different connector types so I'm not sure about that.

 

This is all happening now since I got my power supply impedance tester up and running last week, but I found out it wasn't even good enough. I now have to build a differential preamp for it in order to be able to properly measure the impedance of the regulator. At the end of a cable works fine, but when measuring the impedance of the regulator the voltage drop across across the output jack ground causes massive ground loops in the measurement system that totally destroys any accuracy. So yet again I need to build a high impedance low noise differential amp to measure, of all things, power supplies! I'm glad I can build these myself for less than $100, the commercial ones go for $8000, and are $1500 used on ebay.

 

John S.

I am a bit confused, John.

You mention "...for digital circuits, the INDUCTANCE of the cable is probably more important than DC resistance."

Then you mention that 20AWG starquad is lower IMPEDANCE than 14AWG Romex.

The parameters IMPEDANCE and INDUCTANCE are both mentioned in the next paragraph.

Related parameters appear to be RESISTANCE and REACTANCE.

Doing a bit of reading which confused me a bit more, it appears that some of these terms are used for AC circuits and the others for DC.

Is there a simplified way for me, etal., to understand these terms?

Am I correct in the "take away" being that star quad would be the preferred way to make DC power cables?

Then, JSGT of course.

[BigGuy]

I would also be interested in knowing how short a starquad umbilical can be before it does not make sense.  In my case, I have an NOS brick style LPS which happens to have a 5V output.  I grafted on a barrel connector Digitek pigtail that is 6' long so overall DC cable length is about 8'.  Based on what I have read here, I would like to potentially improve the power quality to my target device by either (1) trimming the existing graft to 12-24 inches so I still have placement leeway or (2) replacing the graft with a similar length of starquad umbilical, perhaps with an Oyaide (or other recommendation?) barrel connector.

 

The other question in my mind is whether this umbilical treatment makes sense with an LPS as described?

 

Probably already discussed ad nauseam but does a JSSG make sense on these DC cables?

[BigGuy]

On 12/16/2017 at 4:09 PM, JohnSwenson said:

OK, really quick and dirty primer. (not terribly accurate, no formulas, but hopefully gets the idea across)

 

Resistance is a property of conductors that tends to impede the flow of electricity. Resistance is usually defined as a DC property. It is measured in ohms.

 

Impedance is a frequency dependent equivalent to resistance. It is not a single number, it is of the form: ohms at a specific frequency. You can either specify it at specific frequencies or have a graph that plots ohms VS frequency.

 

Reacatnce is much more complicated, it is what gives rise to impedance. There are two types of reactance: capacitive and inductive. Capacitive is created by electric fields and inductive is created by magnetic fields. They behave oppositely to each other, at a given frequency increasing inductance increases impedance, increasing capacitance decreases impedance. For a given value of inductance increasing frequency increases impedance, for a given value of capacitance increasing frequency decreases impedance.

 

Every real circuit has a combination of resistance, inductance and capacitance. It is the interplay between these that give rise to how circuits behave.

 

The reality is way more complicated involving phase as well, butI'm leaving all that out to cover the very basic properties.

 

John S.

 

Thanks for the explanation, John.  Enough detail for me!

[BigGuy]

12 hours ago, Cornan said:

 

If you have'nt tried JSSG on USB cables yet I can recommend it! ?Ethernet is great with JSSG too, but I guess you've done it already? So far my ac power cables are mostly starquad which is cumulative throughout in my setup. I have'nt tried JSSG on them yet so I would appreciate if you could post back your impressions when you've tried.

 

Just read contents of this thread which seems to have ended...

 

http://jplay.eu/forum/index.php?/topic/1128-can-the-guys-whove-used-emirfi-paper-sotm-3m-etc-post-photos-here-of-how-you-did-it/page-4

 

What does anyone think about putting an "air spacer", cotton sleeve or polyethylene foam tube <http://www.thefoamfactory.com/closedcellfoam/polyethylene.html>

(scroll down to near bottom) around the DC, USB, etc., cable before shielding? 

The tube might need to be split to avoid issues with terminations.  Certainly prices are reasonable.

 

Not sure where anyone is buying braid but did locate on Amazon...

 

https://www.amazon.com/Tubular-Copper-Bright-Diameter-Length/dp/B003R501TA/ref=sr_1_1?s=music&ie=UTF8&qid=1513612475&sr=8-1&keywords=tubular+copper+braid

 

Not sure whether tin plated is "better" than copper but I prefer the latter aesthetically and it is pretty easy to create a solder pad for attaching the ground wire...just put something that does not melt or burn inside the braid tube.