Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › DC umbilical length question.
New Posts  All Forums:Forum Nav:

DC umbilical length question.

post #1 of 17
Thread Starter 

How much does DC cable length matter?  I have some gear incoming that has a separate box for the power supply, and I wanted to keep them separated by a few feet (amp on the desk, power supply on a lower shelf) if possible.  I have some 18ga wire that I was going to twist up and shield to run between the two boxes.  It is a fairly low powered system if it matters (12V, 800ma).

post #2 of 17
Quote:
Originally Posted by FraGGleR View Post
 

How much does DC cable length matter?  I have some gear incoming that has a separate box for the power supply, and I wanted to keep them separated by a few feet (amp on the desk, power supply on a lower shelf) if possible.  I have some 18ga wire that I was going to twist up and shield to run between the two boxes.  It is a fairly low powered system if it matters (12V, 800ma).


It matters.  However, if you're worried about voltage drop, that can be easily calculated.  Interestingly enough, I am referring to the opposite.  Many of the linear-regulated power supplies that we are used to around here depend on the few fractions or more of ohms in the connecting cable to maintain stability.  Put that same power supply on the same PCB as the amplifier circuit, separated by only a small trace in the ground plane, and you may find that the power supply design is not stable - at all.

post #3 of 17
Thread Starter 

Thanks, Tomb.  Would you mind pointing me to resources that could relatively explain the principles so that I could calculate the drop or tell if the impedance of the cable would actually help/be necessary?  At this point it is a practice in theory since neither the amp nor the LPS designs have been finalized yet.  The two were designed to be separate though, and are going to come with a cable of at least a foot already.  In this case, do you think a couple extra feet would really matter?  I want to learn, but I am also itching to build a cable :)

post #4 of 17

You just asked a VERY broad question. 

 

The answer is "it depends."

 

Sorry, I did not realize I was helping to design/build a commercial product. 


Edited by nikongod - 4/6/14 at 7:48pm
post #5 of 17

12V, 800mA.

 

This is a respectable-sized wallwart, but far from the biggest. Get a 1A one. Take the last foot or so of the cable and pass it repeatedly through a ferrite toroid or other large bead, preferably lossy, with as low frequency as you can find. Stop only when just sufficient cable remains to make the connection.

 

DC cable length matters. The intent is to put distance between the AC parts of the circuit and the DC-using part. This is absolutely the cheapest way to reduce PSU noise (hum) in the output. Anything intended to operate from a remote supply should be designed and tested to be able to live with that.

 

What we have to be aware of is the nature of the interference we are are attempting to reduce.

 

A PSU, conventional or switch-mode, contains a transformer and some switching circuitry and some filtration circuitry. So we are talking about, on the one hand, magnetic fields, and on the other, switching noise. 

 

Magnetic interference occurs when fields from the transformer induce currents in nearby conductors. The only way to block magnetic fields is by using a magnetic material such as iron or steel, copper or aluminium will not work. You can move the transformer further away or change its orientation and attenuate the field that way. Transformers have field distributions with irregular shapes.

 

Switching noise is caused by high speed edges on the currents in a circuit. Fast rise-times and falls. If the rise time is short enough the frequency content will be high. The higher the frequency the greater likelihood of radiation. RF radiation behaves in non-intuitive ways. It likes to get on the outside of things and crawl around. An unregulated conventional linear PSU does not produce much switching noise however.

 

Where DC cable length really can matter is in terms of noise received or noise transmitted. If the circuit generates radio noise internally it can escape onto the power cable. If the PSU generates radio noise it can escape onto the power cable. Once it escapes onto the power cable, the chance of it finding an accommodating transmitting antenna is improved. Very often the receiving antenna will be an input or output cable.

 

Then there are waves of many dimensions wandering around in the aether, looking for an antenna of suitable dimensions to home in on, if you believe in reciprocity. We'd really prefer that any alighting on our power cord don't enter the device.

 

So there are conflicting reasons for deciding the cable length, other than the stray (uncontrolled, unmeasured) capacitance, inductance and resistance, and their effect on the PSU and amplifier performance, especially considering that those effects are in all probability inaudible, considering the number of wallwart-driven DIY amps working OK with little more consideration given to their design than buying a wallwart of appropriate voltage and current. Or the next size up.

 

ICU say 'I am itching to build a cable.' Don't. There are a lot of poo-traps to shielding. If you go with a wallwart, the wire and bullet connector that it comes with will be fine.

 

So you could listen to the output, move the device closer to the transformer and see (hear) if you can hear (see) any difference? If being up close makes little appreciable difference, then you could use a short cable, and all this would be moot anyway.

 

w

post #6 of 17
Thread Starter 
Quote:
Originally Posted by wakibaki View Post
 

12V, 800mA.

 

This is a respectable-sized wallwart, but far from the biggest. Get a 1A one. Take the last foot or so of the cable and pass it repeatedly through a ferrite toroid or other large bead, preferably lossy, with as low frequency as you can find. Stop only when just sufficient cable remains to make the connection.

 

DC cable length matters. The intent is to put distance between the AC parts of the circuit and the DC-using part. This is absolutely the cheapest way to reduce PSU noise (hum) in the output. Anything intended to operate from a remote supply should be designed and tested to be able to live with that.

 

What we have to be aware of is the nature of the interference we are are attempting to reduce.

 

A PSU, conventional or switch-mode, contains a transformer and some switching circuitry and some filtration circuitry. So we are talking about, on the one hand, magnetic fields, and on the other, switching noise. 

 

Magnetic interference occurs when fields from the transformer induce currents in nearby conductors. The only way to block magnetic fields is by using a magnetic material such as iron or steel, copper or aluminium will not work. You can move the transformer further away or change its orientation and attenuate the field that way. Transformers have field distributions with irregular shapes.

 

Switching noise is caused by high speed edges on the currents in a circuit. Fast rise-times and falls. If the rise time is short enough the frequency content will be high. The higher the frequency the greater likelihood of radiation. RF radiation behaves in non-intuitive ways. It likes to get on the outside of things and crawl around. An unregulated conventional linear PSU does not produce much switching noise however.

 

Where DC cable length really can matter is in terms of noise received or noise transmitted. If the circuit generates radio noise internally it can escape onto the power cable. If the PSU generates radio noise it can escape onto the power cable. Once it escapes onto the power cable, the chance of it finding an accommodating transmitting antenna is improved. Very often the receiving antenna will be an input or output cable.

 

Then there are waves of many dimensions wandering around in the aether, looking for an antenna of suitable dimensions to home in on, if you believe in reciprocity. We'd really prefer that any alighting on our power cord don't enter the device.

 

So there are conflicting reasons for deciding the cable length, other than the stray (uncontrolled, unmeasured) capacitance, inductance and resistance, and their effect on the PSU and amplifier performance, especially considering that those effects are in all probability inaudible, considering the number of wallwart-driven DIY amps working OK with little more consideration given to their design than buying a wallwart of appropriate voltage and current. Or the next size up.

 

ICU say 'I am itching to build a cable.' Don't. There are a lot of poo-traps to shielding. If you go with a wallwart, the wire and bullet connector that it comes with will be fine.

 

So you could listen to the output, move the device closer to the transformer and see (hear) if you can hear (see) any difference? If being up close makes little appreciable difference, then you could use a short cable, and all this would be moot anyway.

 

w

  Thanks for the input!  The DAC/amp is the LHLabs Pulse X which is currently being designed with delivery expected closer to end of summer.  It will ship with a switch mode power supply (so I assume it would be able to handle some length like a "normal" component that has a wallwart), but LHLabs is offering up a linear power supply which I also signed up for.  I am assuming that it is meant to be stacked or sit next the main unit, but I had hoped to have the main unit on my desk and then keep the LPS on a lower shelf no more than a meter away.  They will be supplying an umbilical, but as of yet the final details have not been released.  I will likely have to wait until the designs are finalized before I can safely build the cable.  I was hoping it would be simpler than this so I could build a cable with some pretty DC connectors I picked up during a visit to Japan.  Looks like I was wrong :)

post #7 of 17
Quote:
Originally Posted by FraGGleR View Post
 

How much does DC cable length matter?  I have some gear incoming that has a separate box for the power supply, and I wanted to keep them separated by a few feet (amp on the desk, power supply on a lower shelf) if possible. It is a fairly low powered system if it matters (12V, 800ma).

 

 

Quote:
Originally Posted by FraGGleR View Post
 

Thanks, Tomb.  Would you mind pointing me to resources that could relatively explain the principles so that I could calculate the drop or tell if the impedance of the cable would actually help/be necessary?  At this point it is a practice in theory since neither the amp nor the LPS designs have been finalized yet.  The two were designed to be separate though, and are going to come with a cable of at least a foot already.  In this case, do you think a couple extra feet would really matter?  I want to learn, but I am also itching to build a cable :)

 

Haha, as I was reading along those 3 things stuck out at me and I was thinking "I wonder if he could be asking about the Geek Pulse & LPS"? That's awesome that it turned out you were because I'm also signed up for those components and want to do a pretty similar thing as you. 

 

I opted for the LPS-4, so I am looking at powering a few components from the DC outputs. I asked about it and Larry confirmed they have 12V @ 1-1.2A on tap, so you've got some extra leeway for what you'd like to power if you got the LPS-4. I'm planning to power the Geek Pulse X-fi, an internet switch, and a router. The switch and the router will be close to one another, but the Geek Pulse X-fi will be in another room completely. 

 

Because of that I'm also wondering what what kind of length I can run the cables out to without loss. Unfortunately I'm looking at something more like 7m, not your easy 1m :-(

post #8 of 17
Quote:
Originally Posted by nikongod View Post

 

Sorry, I did not realize I was helping to design/build a commercial product. 

 

I'm pretty sure FraGGleR, just like me, is looking to make up his own DIY DC cable for a product that is going to supply one, but it'll be one doesn't meet his needs. I don't know him from Adam, but I think he's trying to fill in the gap like I am. 

 

I guess we've got to get more info on the actual output of the power supply before we can actually get some advice though, right? What do he/we/I need to find?

post #9 of 17
Quote:
Originally Posted by FraGGleR View Post
 

I will likely have to wait until the designs are finalized before I can safely build the cable.  I was hoping it would be simpler than this so I could build a cable with some pretty DC connectors I picked up during a visit to Japan.  Looks like I was wrong :)

 

These aren't the Oyaide DC-2.1 or 2.5Gs you got in the Akihabara shop, are they? If they are, I'm the other guy from that thread that said he'd be getting a set from his friend in Japan. They're now on their way as we speak :-D

post #10 of 17
Thread Starter 
Quote:
Originally Posted by Benny-x View Post
 

 

These aren't the Oyaide DC-2.1 or 2.5Gs you got in the Akihabara shop, are they? If they are, I'm the other guy from that thread that said he'd be getting a set from his friend in Japan. They're now on their way as we speak :-D

Haha, yes, that's me and those are the connectors.  And to Nikongod, I have zero affiliation with LHLabs other than I backed their Indiegogo campaign to get their new DAC/amp.  I have been making my own headphone cables and interconnects for years, but have never done any power cables, so I was looking for some help.  Not sure why you think I am designing a commercial product.  Anyways, I appreciate everyone's advice so far.

post #11 of 17

Haha, small world. I'm happy that what ever you manage to find here and what ever questions I ask in the mean time will 100% benefit both of us :-D

 

I've also given DIY a shot before. I've done regular AC power cables, but they're kinda the most basic of the basic. Didn't do anything special to them, so it was never worth asking about.

 

About these DC cables; because of the low voltage and current, I figured there'd have to be some drop out when you started to extend the length. Anyone have any ideas?

 

Like I said before, what else do I need to provide so that someone has what they need to work with to help us out here?

post #12 of 17
Thread Starter 
Quote:
Originally Posted by Benny-x View Post
 

Haha, small world. I'm happy that what ever you manage to find here and what ever questions I ask in the mean time will 100% benefit both of us :-D

 

I've also given DIY a shot before. I've done regular AC power cables, but they're kinda the most basic of the basic. Didn't do anything special to them, so it was never worth asking about.

 

About these DC cables; because of the low voltage and current, I figured there'd have to be some drop out when you started to extend the length. Anyone have any ideas?

 

Like I said before, what else do I need to provide so that someone has what they need to work with to help us out here?

I don't, but just based on what folks here have said, it is going to depend on the LPS and the Pulse X specifically as to how sensitive they might be to any dropout.  I posted on LHLabs about it, but so far most of my power cable questions have been met with crickets.  Might have to wait until they get more product out and have time to answer.

post #13 of 17

Well, although we haven't found out any more here, I did just finish reading up on another thread over at JPlay where a few members said that they'd made up their own DC cables and added shielding to GREAT effect. There cables were inside a PC, a much noisier environment, but still.

 

According to one member, and I think so too from the DC cables I've opened up over the years, there's little if any shielding in them. Depending on where you sit on that argument, you might think that not important and you might think it bad. I'll be adding shielding to mine just to be safe, but who knows.

 

As for what we need to make, we else do we need to bring here so that we can get the ball rolling?

 

I'm really not sure what else we'll need to provide. The default length is going to <2m (I bet, I can't see it being any longer) and we've got the output at 12V @ ~1A. We know that our device needs to receive 12V @ 800mA, so what else is there to factor in? We're aiming to make up the cables ourselves, so can't someone just tell us what we need and we can curb the variables to get that outcome?

 

I'd like to be able to keep as near to 12V @ 1A over my 7m length as possible.

post #14 of 17
Thread Starter 
Quote:
Originally Posted by Benny-x View Post
 

Well, although we haven't found out any more here, I did just finish reading up on another thread over at JPlay where a few members said that they'd made up their own DC cables and added shielding to GREAT effect. There cables were inside a PC, a much noisier environment, but still.

 

According to one member, and I think so too from the DC cables I've opened up over the years, there's little if any shielding in them. Depending on where you sit on that argument, you might think that not important and you might think it bad. I'll be adding shielding to mine just to be safe, but who knows.

 

As for what we need to make, we else do we need to bring here so that we can get the ball rolling?

 

I'm really not sure what else we'll need to provide. The default length is going to <2m (I bet, I can't see it being any longer) and we've got the output at 12V @ ~1A. We know that our device needs to receive 12V @ 800mA, so what else is there to factor in? We're aiming to make up the cables ourselves, so can't someone just tell us what we need and we can curb the variables to get that outcome?

 

I'd like to be able to keep as near to 12V @ 1A over my 7m length as possible.

I had planned to shield since I have a bunch of shielding materials (aluminum mylar as well as copper braid).  Overkill, I am sure, but so as long as there isn't a downside, I am going to stick with it.

 

I am probably safe at the lengths I am looking at since I can't imagine them making something so sensitive that an extra foot or two will make a huge difference.  I am pretty curious as to what you find out the feasibility of a 7m DC cable.  I'm hoping once more stuff ships and the Wave campaign closes Larry might be able to give us some direct info. 

post #15 of 17
Quote:
Originally Posted by FraGGleR View Post
 

I had planned to shield since I have a bunch of shielding materials (aluminum mylar as well as copper braid).  Overkill, I am sure, but so as long as there isn't a downside, I am going to stick with it.

 

I am probably safe at the lengths I am looking at since I can't imagine them making something so sensitive that an extra foot or two will make a huge difference.  I am pretty curious as to what you find out the feasibility of a 7m DC cable.  I'm hoping once more stuff ships and the Wave campaign closes Larry might be able to give us some direct info. 

Yeah, I'd guess that at your normal ~1m length everything will be fine. And you've got extra stuff for shielding too! I guess you're set :-D

 

As for my 7m run, there's no craziness or nervosa going on there, it's just that I want to power my switch and router with the LPS-4, but it's going to be beside the Geek Pulse-Xfi, and the other two are in another room. So I just need 2 long runs to connect the 2, that's all. 

 

As for the new campaign, that's a bit annoying for me. Sure they can say "this won't affect the timeliness of our other campaigns ~AT ALL~", but that's ********. Of course it's going to eat up their time. I guess they're thinking they better strike while the iron is hot, but still, I don't think I'll be getting any more involved in any of the ongoing stuff. I'll just let it unfold and take what I get, I'm tired of asking questions and waiting and whatever. I'm sure the Wave will also have all kinds of crazy upgrades and turn out costing at least twice as much too...

 

I've been down the portable DAC/amp for my Android phone before. It's better, for sure. But it's not that impressive because of ALL the additional environmental noise. It drowns out too much of any gain(benefits) you get. Will you be picking one up?

New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › DC umbilical length question.