Probably a stupid question, but I'll ask...

[G._Hoffman]

I think I'm starting to get a handle on the whole grounding/layout thing, but I have a question.  Why is it you don't shield DC wires?  I'm assuming there is a reason, and while I've been sort of blindly following what I've seen others do, I don't know WHY, and I've never seen an explanation in the books I've read.  They only ever suggest shielded wire for AC signal wires, but there are DC wires which have signal too - e.g., the wire from the plate to the coupling cap.  Why not shield them?


Gabriel

[phsyconoodler]

Because they aren't noisy.AC is.Simple as that.And...the DC is filtered.The AC is not.
  Do you see any capacitors in a heater circuit?No.

 

[PRR]

Wires don't know if they are AC or DC.

You generally shield high-impedance signal wires. The wiper from a 1Meg pot may be 250K impedance, pretty high. The signal from a AX7 plate is more like 40K, lower, less likely to accept external influence.

You shield small signals. Pickup. Also the signal from tone and volume controls, where signal has been cut down. After an amplier boosts it up, external influence has less influence.

You might shield BIG signal wires. When 240V AC power comes into a test-lab, they like it to be in metal pipe, which is a form of shielding. However you usually have few BIG "signals", and can usually lay-out to keep them away from small-signal points.

[Leevi]

I have used shielded wire in B+ since it goes quite often through the chassis.
If I remember correctly I had one case where an unshielded B+ wire was a reason for a parasitic oscillation.
/Leevi

[tubesornothing]

I think this is a good question.  We default to shielding the sensitive wires, but why not shield the source?  KoC recommended this for heater wires - I found it was overkill and I did not need it.  Currently I don't have any hi gain amps, so I am easily able to stay away from shielding anything.  For a couple of hi-gain prototypes I have built, it was just easier to shield one or two signal wires, rather than shield a whole bunch of B+ wires or other strong signal wires.

[RicharD]

I have a question about shielding signal wires.  Typically one only grounds 1 end of the shielded cable to prevent ground loops.  I usually ground the source end.  Does it matter which end you ground?

On a side note. I just overhauled an amp and I used all 5 pin terminal strips where the center terminal is grounded.  At each mounting point I ground off the chassis finish, bolted, and soldered the terminal strip to the chassis.  I clustered the component grounds by DC supply (ie. output tube cathodes to 1 ground point, PI grounds to another point, preamp tubes to another point, etc.)  This is only 1 ground wire in the whole amp and it runs from a truly grounded cathode to a terminal lug.  The chassis is my ground bus and it's dead quiet.

[tubesornothing]

I typically ground the source end too.  No clue where I picked that up, no clue why either.  It works though.

It is amazing the number of different grounding schemes that work.  Doug's practical scheme and Merlin's theory based scheme both work with no problems.  Yet both are quite different.  Mine is somewhere in between and works great too.  From repairs of various boutique amps and vintage amps I have seen a lot of other schemes that totally contradict each other, but are nice a quiet.

There ya go...

[RicharD]

>I typically ground the source end too.  No clue where I picked that up, no clue why either.  It works though.

My reasoning is based on studio experience.  Mic pre's etc. always have the ground lift on the output and this is because your source must be grounded if phantom power is required, also if your source is single ended, it's gotta be grounded. Typically your source (mic or geetar) is also a dead end so shielding is a huge benefit.  None of this really seems to apply when considering a shielded jumper within a shielded chassis.

>It is amazing the number of different grounding schemes that work.
It seems like a common denominator with grounding schemes is to keep your wall cord, power tubes, and preamp tubes at 3 different points.

[PRR]

> Does it matter which end you ground?

It often does.

You want a rule? "It Depends."

Amps are too complicated. Take something simple, a house.

A house electric system has a dirt rod.

Where?

Yeah, NEC tells us "at the entrance". But in many odd cases, there are better places.

If I build a 50 foot Tesla Coil in my back yard, it might be best to divert those kicks direct to the dirt under the Coil. In a serious case, that would have to be done, the main entrance ground broken with an isolation transformer.

The power company has trouble on the pole. This should probably be dumped to dirt at the street, not shot into all the houses. Indeed utility networks have hefty ground conductors and multiple dirt-rods for this reason.

If I run a sensitive short-wave radio set, I may need a short path to dirt, opposite my antenna, rather than the long run back to "entrance". This happens to be easy: an isolation tranny to pass radio is just a few turns. The radio chassis sits at entrance potential, but senses tower-dirt potential.

Let's not think about radio transmitters dumping kilowatts toward dirt using Godzilla-grade dirt electrodes.

While power-ground is at "entrance", any lightning-rod dirt-ground must NOT go to service entrance, but be well separated.

So even houses are not simple. Amps are even less simple.

BTW: my house now has three almost-good ground-rods. Cut/slanted because ledge-rock is only 4 foot down here, but there is steady ground-water over the ledgerock so it's not too bad. Now I should probably get some wire and run it to the fusebox.....

[DummyLoad]

fusebox.....   ???

you're still running on a fusebox? eeek! 

[Fresh_Start]

Butter - your approach with the terminal strips is interesting.  Steel or aluminum chassis?

A related hijack:  I read somewhere that it's "bad practice" to use the shield as a signal ground.  However, doing just that solved a noise problem in a BF Super Reverb build.  It allowed me to ground the Reverb pot at the same point as the reverb circuit & trannie.  What's bad about this?  The only thing I can think of is some future tech removing the shielded wire and then wondering wtf happened?

BTW I'm going to try shielded heaters for the preamp tubes in my next build.  If it's possible with the layout, I'd like to try it without shielding first and then just to V1, etc.  It's a single-ended amp with relatively high gain so the extra effort may be worth it.

Cheers,

Chip

[RicharD]

>your approach with the terminal strips is interesting.  Steel or aluminum chassis?

Steel.  there is a picture in my OTB - Doric Amplifier thread.
http://www.el34world.com/Forum/index.php?topic=8882.0
Far from my prettiest work, but given the undocumented inferno I started with, I happy enough.

I have seen this grounding method in in many an old chassis.  The key is to make all the ground connections to the chassis secure.  Bare steel, bolted and soldered.

EDIT:
Attached updated (final) schematic.  Ground symbol indicates tie point chassis.

[jjasilli]

I think I'm starting to get a handle on the whole grounding/layout thing, but I have a question.  Why is it you don't shield DC wires?  I'm assuming there is a reason, and while I've been sort of blindly following what I've seen others do, I don't know WHY, and I've never seen an explanation in the books I've read.  They only ever suggest shielded wire for AC signal wires, but there are DC wires which have signal too - e.g., the wire from the plate to the coupling cap.  Why not shield them?
Gabriel

Also: note that shielding is not benign.  It causes parasitic capacitance which attenuates hi frequencies.  Also leads are often short and can be laid-out within the amp so as not to get affected by noise through induction.

[Ritchie200]

fusebox.....   ???

you're still running on a fusebox? eeek! 

I'd rather have fuses than breakers any day!  I've seen the data, I've watched the testing...

Besides, you need good reaction time in case the 50ft Tesla Coil has a problem!

Jim

[jjasilli]

@Freshstart:  "A related hijack:  I read somewhere that it's "bad practice" to use the shield as a signal ground.  However, doing just that solved a noise problem in a BF Super Reverb build.  It allowed me to ground the Reverb pot at the same point as the reverb circuit & trannie.  What's bad about this?  The only thing I can think of is some future tech removing the shielded wire and then wondering wtf happened?"

Guitar chords use the shield as 1/2 of the signal circuit.  It works.  The bad news is that ambient noise can get picked up by the shield which is carrying signal.  So you can go to mic cable which has 2 signal wires, both protected by an outer shield.  But this requires TRS plugs & jacks or XLR; it's unconventional and complicated.  Blah blah.

Also the noise in the shield is not part of the signal circuit.  The grounding scheme can cause the noise to drain to ground, while the signal current is pulled away form the noise to the 1st gain stage.  See Ch 15 of Merlin's Preamp Book (2d Edition) which is posted online.  (It's Ch 12 of the 1st Ed. hardcopy)

[rake]

More "dumb" but related questions.       

1. Doesn't the twisting of the heater wires help "shield" them?

2. Has anyone ever tried a three wire twist with one of the wires
       being a chassis ground? (just wondering if it would make any difference?

[jjasilli]

@rake.  1. Doesn't the twisting of the heater wires help "shield" them?
No.  Heater wires are a source of noise.  The AC current in the heater wires generates a "field" around the wire.  The field can induce noise into the signal circuit.  By twisting the heater supply wires tightly together, the opposing fields cancel each other out (hopefully) . This is a "humbucking" method.  It is not shielding. 

Shielding the heater wires is mentioned in this thread.  I suppose the intent is either to contain the noise in the wires of an AC heater supply; or to keep outside noise from getting induced into a DC heater supply.


2. "Has anyone ever tried a three wire twist with one of the wires being a chassis ground? (just wondering if it would make any difference?"
This would not accomplish anything.  A shield would have to surround a wire, not be merely tangled with it.

[HotBluePlates]

2. "Has anyone ever tried a three wire twist with one of the wires being a chassis ground? (just wondering if it would make any difference?"
This would not accomplish anything.  A shield would have to surround a wire, not be merely tangled with it.

Sorry for the small image.

Towards the upper left of the picture, right above the black fiberboard, there is a white wire wrapped around a bundle of other wires.  Take my word Fender soldered both ends of that white wire to the chassis to "shield" the wires in the bundle.  Tell-tale silverface wiring...

[bnwitt]

Yep.  I've seen a lot of silverface Twins with that.  Also on some master volume wires.  Komet amps shield the three wires on the Preamp load gain switch with a coiled outer shield wire connected to the chassis ground as well.

[jjasilli]

In vintage amps I've seen the ground lead "laced" around a bundle of other wires.  I thought that was done for neatness.  Today that's usually done with zip ties.  Are you saying the lacing is a form of shielding?

[HotBluePlates]

In vintage amps I've seen the ground lead "laced" around a bundle of other wires.  I thought that was done for neatness.  Today that's usually done with zip ties.  Are you saying the lacing is a form of shielding?

Not me.

I've never seen what you're talking about.  I've only seen plain fiber lacing cord using for lacing.

Separately, you'd need the "shield" formed by the wire to have very-complete coverage of whatever it's shielding.  That's why the Fender coiled white wire looks the way it does in those silverface amps.

...  Also on some master volume wires.  Komet amps shield the three wires on the Preamp load gain switch with a coiled outer shield wire connected to the chassis ground as well.

You'd think someone making amps for a living could invest in some multi-conductor shielded wire, so they didn't have to jury-rig a solution.  I wonder if they tried it & have a reason for not using it.

[jjasilli]

Here's an example of old fashioned "cable lacing" using chord.  I've also seen it done with with the ground lead wire.


It's hard for me to believe that the loose Fender wrapping actually provides a shield; but I'll take your word for it.

[bnwitt]

Theoretically, a wire with DC voltage has a static magnetic field around it so there are no changing flux lines there to cut through adjacent wires and impart unwanted noise.  In the real world there is usually some ripple present which would create a varying flux field but with adequate filtering it is minimal.  AC wires on the other hand have a continuously changing magnetic flux field which can affect adjacent wires and circuits.  This is the reason twisting heater wires is suggested which may or may not really help that much.  Having the two AC wires close together is supposed to encourage phase cancellation by the two out of phase AC currents in the wires.  Shielded input wires have the shield to catch (and drain to ground) any spurious magnetic fields from surrounding AC currents and keep it from the signal wire.  At least that is my understanding.  Based on that, I try to run most of my wires down along the chassis rather than fly them in order to use the chassis as an electron catcher so to speak.

[bnwitt]

Here's an example of old fashioned "cable lacing" using chord.  I've also seen it done with with the ground lead wire.


It's hard for me to believe that the loose Fender wrapping actually provides a shield; but I'll take your word for it.

Well it's hard to believe that the low mass control grid inside a vacuum tube (which has more open area than surface area) can control the electron flow but it does.  Electrons are invisible mysterious things indeed.

[HotBluePlates]

... It's hard for me to believe that the loose Fender wrapping actually provides a shield; but I'll take your word for it.

It's reasonably-complete coverage, for what it is.  Now, I didn't say I endorse doing it, I just said that's what Fender was trying to do with that wrapped wire.

As for the lacing, using a grounded wire would provide even less coverage than Fender wire wrapping, no?

[jjasilli]

Well, that's my point - lacing ≠ shielding; questioning the effectiveness of the Fender method as shielding (which is also a rather ugly lacing job    ).  OTOH, taking another look at the Fender wrapping, it does look closely wound enough to form a Faraday cage.


But, the question asked in the post was about the effectiveness of a "3 wire" twist of the heater supply, the 3rd wire being a ground wire.  That would be ineffective.  A wrap like the Fender method might be effective.