How to use screened wire

[jim68000]

(I'm probably using up my beginner points too quickly here, but here goes)

The first fuzz box I built suffered from incredible interference - I could even pick up what sounded like Filipino radio on it. I'm going to have a go at a Fuzz Face and I'd like to cut down the noise where possible, so I've ordered some screened wire from Musik Ding but I'm a bit puzzled what the best way to use the core and screen is. Do I connect both the core and the screen to the jacks like a guitar cable, so hot is core and shield connects through to the guitar cable, or is it one wire per connection and then solder the shield to a ground point? Looking at the pictures of the insides of DAM pedals it looks like they do the former, but I've also read the latter is the thing.

Which is better?

[thom]

AFAIK it's core as a normal wire (both ends) and shield at ground at one end only.

You might want to wait for more knowledgeable folks to confirm it, though  

[GibsonGM]

Yes, ground your shield at ONE end only, or you'll just make a ground loop.  With one end grounded, you pretty much make the shield another part of the chassis.   Neat trick, and VERY effective!

[peterg]

A small cap (0.00022uf/220@pf) from the input to ground will help reduce radio interference.

[Kesh]

sometimes radio frequencies won't be from wiring inside, but brought in from guitar or whatever. the circuit then detects it (ie changes it from radio frequencies to audio frequencies). so like peterg says, a cap chosen to kill everything above audio frequencies at the input to ground can sometimes help.

[deadastronaut]

add a switch for the sheild and have a radio too..

[jim68000]

Ha, it only worked if I pointed my strat in a particular direction, and had the tone control just so. But it was the best shortwave radio I've heard in years.

[pinkjimiphoton]

a small resistor in series with the input....say, 47-220r will also help.

you want to "telescope" the shields.

connect the shield to input jack ground, center conductor to hot. run that to your switch,  but make sure there is no connection with the shield on that end... only the center conductor should connect to your switch.

output jack should be the opposite... run the shield to ground on the switch itself, (or where the ground for your led is) and the center conductor to the pin on the switch...  at the output jack, only the center should be connected to the tip connection.

if need be, you can run shielded cable from the effect's circuit board back to the switch as well.. in that case, connect the shield of both wires to the circuit ground, and connect the center conductor to the switch. ground should only be connected to the circuit board itself.

that will help with noise quite a bit. you can probably get away with only shielding the input/output jacks to the switch.

there's other tricks, too... but for shielding, try this. always try to remember you want it like branches from a tree, and all grounds connected to one point, usually the input jack ground lug.

hope this helps ya bro

[GibsonGM]

Good advice, and also: be sure to trim the shield away from the 'cold' end, where you're attaching the center conductor to a switch or whatever.  You do NOT want that grounded braid touching your nice, hot circuits!!  Keep it out of the way.  I heat-shrink the end of my runs like this to prevent shorts.

It's ok to cut 1/2" to 3/4" of the plastic insulation, pull the center conductor thru the braid - or separate the 'hairs', if you have that type - and twist the loose braid.   So you have two 1/2" long wires, so to speak.    The last little amount of run is not shielded, is what I am trying to say....no practical way to do so, and it's not necessary anyway.   

Most of your noise/RFI will be eliminated doing this plus the cap to ground, with an R in series if necessary.  Only do what is necessary, and no more!  Keep the R low value to not change the FF input impedance much, tho, or it will sound different!    My money is on the shield/cap being your magic bullet. 

[Radcon]

I tried such wire on a recent (crowded) build.

I had some trouble with the grounded shields inadvertently contacting and grounding out various other things. Care with routing is obviously the preferred remedy. I had to insulate with a little hot glue in a few places. 

Results were good-relatively low noise on a high-gain circuit (Madbean Kokbox). 

[merlinb]

It depends how many conductors your cable has.

If it only has one hot wire and a shield (like a regular ol' guitar cable) then obviously you have to connect the shield at both ends since the shield is the return path for the current. If you connect it at one end then you've broken your signal path, unless you've made your ground connections elsewhere (in which case you might think about improving the whole way you're doing things, like using insulating jack sockets).

(Yes, connecting the shield at both ends may well create a ground loop via the metal enclosure, depending on how you've done things, but pedals are so small that the loop area will be pretty tiny, so it shouldn't be a problem. Ground loops of this type are not what cause problems for people in pedals. It is ground loops via the power supply that are the problem.)

But if your cable has two conductors plus a shield then you can use the internal conductors for the hot and ground connections, and then the shield protects BOTH. The shield should then be connected at one end only; it isjust a tubular extension of the enclosure, not a part of the working circuit.

[GibsonGM]

Merlin makes some very good points!  I was assuming you just wanted to carry your signal and had a single shielded conductor; the method would change if you have 2 conductors.  And of course pedals are running on lower voltages, which would not cause appreciable noise to be induced in short loops.  That condition would change if one moved up to tube amps.

[jim68000]

So it's just the same as the guitar cable connection on the other side of the socket? Core to tip and shield to ground? Or am I missing something?

[pinkjimiphoton]

yes, pretty much.

the diff is with a good guitar cable, you'll have two center conductors AND a shield. in that case, the shield is connected on one end only (which is what ya want) and the two center conductors become +/-

going from input jack to switch, you do NOT want to OR NEED to connect the shield at both ends. it defeats the entire purpose of star grounding or telescoping the shield.

you only need one connection to ground from the circuit. you want the shield connected to the ground lug of the jack , the center conductor connected to the tip of the jack. if you connect the shield to both ends, you create for all intents, an antennae. not what you want.

if using multi conductor shielded cable, IF you have a "send/return" kind of situation, then yes, you would connect both sides of the center conductors, but STILL you only want the shield connected on ONE SIDE ONLY. then, and only then, does it become an extension of "the box". connect at both ends, and suddenly you will have ground loops, and they will cause some kind of issue that may or may not become noticeable imho.

you should run the shield on the input jack to switch. MAYBE from the switch to the board. in that case, you would connect the shield to the jack side only, and have the center conductor only go to the switch from the jack. then you want the same thing going to the board if you feel you need more shielding... in that case, center conductor to switch and circuit board in, and shield connected to ground ON THE CIRCUIT BOARD.   you can do the same on the output of the circuit.

but the circuit ground itself should never be thru the "shielding". it should have a dedicated connection to whatever star ground you choose, and THAT becomes THE ONLY ground point as far as the circuit is concerned. that will make a huge diff in RFI and noise, and zero ground loop potential.

trust me... i use cheap ass sheet metal aluminum boxes from radio shack, and they SUCK at rfi rejection. but, if you wire the circuit as i suggest, it no longer matters. you could probably use a plastic box and be fine....(in fact, i have on a few things)

if you connect the shield to more than one end of each cable run, you no longer have a shielded cable, you have a capacitive antennae

jmo... your mileage may vary some, and true, in a metal box it should matter less...depending on the gain of the circuit....but if you're adding loops, you're overcomplicating stuff and probably making your pedal more susceptible to noise, interference, and other schmutz.

hope i don't come off like a douchebag here,  i know i am far from knowledgeable as many folks here, but having messed with high voltage tube stuff for a long time, and learning how to deal with interference and bizzarre parasitic oscillations, i think i am right.

it's quite possible i'm not, but i think that is kinda improbable on this one thing.

with a guitar cord, you want two conductor plus shield for best noise rejection.

one center conductor connects tip to both ends
one center conductor connects ground to both ends
shield connected to one side only, generally whatever is closest to actual ground potential (usually the amp side, as YOU don't want to be the ground for anything!!)

peace

[tubegeek]

shield connected to one side only, generally whatever is closest to actual ground potential (usually the amp side, as YOU don't want to be the ground for anything!!)

The actual ground *potential* isn't the issue, it's LOWEST IMPEDANCE to *AC* ground. That way, any noise current on the shield will develop the smallest noise voltage into the circuit's input connection. So that could be at 50V instead of .5V (higher potential) but at lower AC impedance.

For the same reason, you might connect the outer foil of a capacitor at the higher-DC end, so long as that end was closer to being a short for AC. That's an example of a similar situation which I suspect you'd be familiar with inside a tube amp: the plate end of a coupling cap might be at a higher DC voltage but be lower impedance to the power supply B+ than the grid connection of the following stage, which would be low DC voltage but perhaps a high impedance to ground.

[DIMstompboxes]

-Using shielded wire-

Shielded wire can prevent noise and oscillation on signal carrying wires in effects and amps.
While it is not a cure for a bad layout or other problems, even very well thought out layouts will pick up some noise.
Circuits with high gain are most susceptible to noise etc.  

I use shielded wire from the input and output jacks to the switch and on gain and volume controls in amps.  
These are the main place to use shielded wire. What you are try to do is shield the high gain and high impedance
wire runs from radiated noise .  The longer the unshielded wires length the more chance noise has as getting in.
You only need to connect the shield at one end because the braided shield will work if both ends are not  connected and
if you do connect both ends you may have a problems with ground loops and signals traveling across the shielded connection.
I use Teflon wire because it will not melt or shrink with a soldering iron.  
If you use wire other than Teflon you must keep an eye out for melting the inner core's insulation,
which will short out the shielded braid to the inner conductor.


Start by striping off the outer insulation. I usually use a razor to do this. It takes only a small amount of pressure.  You don't want to cut or nick the braid strands.


Push the braid back with your fingers. This loosens the braid and sets you up for the next step.


Take a small sharp object and pick out the braid. A jewelers screwdriver of a multimeter probe tip works well. Start at the top and work toward the bottom near the insulation.

Once all the braid is picked out you can bend it back about 90 degrees and twist it up. Then bend the whole twist back parallel with the insulated wire.

Now you can solder on a short piece of wire to make your "ground wire". Cut off the excess braid shield . Cut a small piece of heat shrink tubing and shrink it over the connection. Now you have two insulated wires to connect to the jacks or switch. For the opposite end of the wire I strip about an inch of the insulation and pick out the braid, then cut off the shielding so you have a inch of inner conductor sticking,

out  of the shielded wire.. Add bit of shrink tubing over the bare cut end where the inner conductor and the braid and outer  insulation meet and you wont have any surprise shorts to the slight bit of braid that pokes out.
Make sure to only connect the shield at one end. I usually connect it at the jacks sleeve terminal. For the wiring going to the switch I connect it at the switch where the ground for the LED connects.

Heres what the pot looks like with the added Shielded wiring.
This is the gain pot.  Notice the add on ground wire length...nice!

Here's what the switch looks like. The blue shrink wrap is insulating the ground wire soldered to the shielding braid of on of the wires.


Here's the finished box . In this case a Dr Boogey.   

Thanks to you John Lyons

DIMstompboxes

[pinkjimiphoton]

shield connected to one side only, generally whatever is closest to actual ground potential (usually the amp side, as YOU don't want to be the ground for anything!!)

The actual ground *potential* isn't the issue, it's LOWEST IMPEDANCE to *AC* ground. That way, any noise current on the shield will develop the smallest noise voltage into the circuit's input connection. So that could be at 50V instead of .5V (higher potential) but at lower AC impedance.

For the same reason, you might connect the outer foil of a capacitor at the higher-DC end, so long as that end was closer to being a short for AC. That's an example of a similar situation which I suspect you'd be familiar with inside a tube amp: the plate end of a coupling cap might be at a higher DC voltage but be lower impedance to the power supply B+ than the grid connection of the following stage, which would be low DC voltage but perhaps a high impedance to ground.

what tubegeek said.

ground will still be connected via the drain wire in multi conductor cabling, so the shield's only real purpose is to telescope the grounded chassis along the length of the cable.