I'm connecting a recv A/V (pre-amp) that uses a two prong power plug to
a amp that uses a 3-prong power amp. My RCA interconnects were put
on without any regard to the arrow (didn't know it was there). What are
the symptoms? Ground-loop? I was troubleshooting a problem, and I'm
hoping this was the reason. I can't try it now, because I don't have the
equipment at this time (another thread)

Got a hum?

Got a hum?

I did, but the system is down right now. Saw a thread that was discussing
the interconnects - and this arrow - and the cables being directional -
but in what direction, and can it cause a hum?

There is no electrically accepted directionality of RCA cables, unless there is a screen for noise or hum, which one recommends to connect to ground only at one end, and orient the cables so this connection is at the source end.

Having said that, some audiophiles claim to hear directionality, but there is no electrical problem with swapping it araound to see if you agree.

Try wrapping the rca cords tightly around the preamp line cord the entire length of the line cord.

John

Uhhh... what?

Uhhh... what?
Wrap the ic's tightly aroung the line cord of the pre.

This does several things:

1. It eliminates dB/dt pickup between the ic shields, eliminating induced shield current (which is not compensated for). It also reduces hot to hot induction (left to right difference) as a result of field intercept.

2. It lowers the loop area between the line cord helical dipole field and the ic's (contrary to the belief that moving the wires farther away is a good thing). The more dipole field that goes outside the ic's, the lower the induced voltage.

3. It averages out the dipole field of the line cord. Since it is an average caused by twists, it integrates out to zero.

4. It reduces sensitivity to externally generated dB/dt fields, such as that caused by the romex in the wall.

Faraday's law of induction...

(aren't ya sorry ya asked?)

I do this whenever I have a long run of unbalanced line level signals to do. So far, I've used this technique successfully for runs up to 125 feet, in an auditorium with ten kilowatts of scr dimmers, a rather robust HVAC system, a toro 700 leaf blower and a 250 kilovolt vandegraf on stage. Noise and hum free.. Without any DI boxes, without transformers, totally unbalanced feeds, no rf connections..no grounding precautions at all, just what's in the line cord..

Cheers, John

All the lyrics are in reverse. ;)

How can you tell with the Toro leaf blower going?

How can you tell with the Toro leaf blower going?


Commutation noise is the first thing heard. The brushes.. The leafblower is embedded in a big box as part of a bernoulli demonstration, so the blower noise is a bit reduced, and the speakers are on either side of the stage, about 40 feet total.

Same thing with the van de graf, but in that case, it's the high frequency impulses of the sparks. (yes, the sparking can be heard over the kids screams...)

Cheers, John

Wrap the ic's tightly aroung the line cord of the pre.

This does several things:

1. It eliminates dB/dt pickup between the ic shields, eliminating induced shield current (which is not compensated for). It also reduces hot to hot induction (left to right difference) as a result of field intercept.

2. It lowers the loop area between the line cord helical dipole field and the ic's (contrary to the belief that moving the wires farther away is a good thing). The more dipole field that goes outside the ic's, the lower the induced voltage.

3. It averages out the dipole field of the line cord. Since it is an average caused by twists, it integrates out to zero.

4. It reduces sensitivity to externally generated dB/dt fields, such as that caused by the romex in the wall.

Faraday's law of induction...

(aren't ya sorry ya asked?)

Cheers, John Do you have an English translation? http://www.hometheatershack.com/forums/images/smilies/dizzy.gif

Regards,
Wayne A. Pflughaupt

Wrap the ic's tightly aroung the line cord of the pre.

This does several things:

1. It eliminates dB/dt pickup between the ic shields, eliminating induced shield current (which is not compensated for). It also reduces hot to hot induction (left to right difference) as a result of field intercept.

2. It lowers the loop area between the line cord helical dipole field and the ic's (contrary to the belief that moving the wires farther away is a good thing). The more dipole field that goes outside the ic's, the lower the induced voltage.

3. It averages out the dipole field of the line cord. Since it is an average caused by twists, it integrates out to zero.

4. It reduces sensitivity to externally generated dB/dt fields, such as that caused by the romex in the wall.

Faraday's law of induction...

(aren't ya sorry ya asked?)

I do this whenever I have a long run of unbalanced line level signals to do. So far, I've used this technique successfully for runs up to 125 feet, in an auditorium with ten kilowatts of scr dimmers, a rather robust HVAC system, a toro 700 leaf blower and a 250 kilovolt vandegraf on stage. Noise and hum free.. Without any DI boxes, without transformers, totally unbalanced feeds, no rf connections..no grounding precautions at all, just what's in the line cord..

Cheers, John

While this technique may reduce the induced magnetic coupling between the power cable and the interconnect, I don't think it would prevent ground loops caused by common-impedance coupling ( the more common problem). It would also be hard to believe any increased effectivness over that provided by the cable shield with regards to capacitively coupled electrostatic noise such as commutation brushes. You would almost certainly be better off with isolated or balanced interfaces.

While this technique may reduce the induced magnetic coupling between the power cable and the interconnect, I don't think it would prevent ground loops caused by common-impedance coupling ( the more common problem).

Given the fact that he has only three connections between the equipment...a two conductor power cord and two ic's, there is not a whole heck of a lot of paths for hum to take. Either it is created by the pre, the amp, or the result of wiring loops.

I address hum caused by external field sources, and detail how to make the ic's impervious to hum induced by the line cord. A line cord which is not coaxial produces external field. That is inviolate. Typical two wire line cords are zip type, and may not have a twist to them, this is worse.

You may not think this technique prevents ground loop issues, however I have successfully used this technique in various venues for the last seven years, in hostile e/m environments, and it worked without failure. So I know from both experience and field theory, so have recommended it.

It would also be hard to believe any increased effectivness over that provided by the cable shield with regards to capacitively coupled electrostatic noise such as commutation brushes. You would almost certainly be better off with isolated or balanced interfaces.

That is what most professional sound engineers would state. And sadly, they fall flat in cases where it is a magnetic induction issue and not the result of the electric field.

It is not a problem in electrostatics, but one of magnetodynamics. I see this error a lot. Why did you believe commutation noise is electrostatic? It is not. It is a result of high rate of change current creating high rate of change magnetic fields.

I do not provide sound engineering for the facility here, that is the job of another. When he runs the venue with his equipment, he needs DI boxes, he needs balanced runs to and from the stage, he needs transformer isolators...and when he STILL has hum and noise issues, he has to run RF.

I maintain my own system, and use it in the exact same venue...unbalanced, no isolation, nada. Takes me 15 minutes to setup, and I have never had a noise or hum issue. The auditorium I do this in has huge ground loop issues due to HVAC, lights, and roughly 20 megawatts of very noisy power supplies and fridges in close proximity (for those size loads, "proximity" is anything within half a mile) . Any wire which is run from the stage to the balcony traps a huge noise and hum voltage to ground. I avoid that.

Cheers, John

ps...BTW, my equipment does indeed have balanced capability. It has not been required (if it ain't broken, don't fix it). If my technique had not worked, I most certainly would have run balanced. I chose to keep the weight down.

Given the fact that he has only three connections between the equipment...a two conductor power cord and two ic's, there is not a whole heck of a lot of paths for hum to take. Either it is created by the pre, the amp, or the result of wiring loops.

I address hum caused by external field sources, and detail how to make the ic's impervious to hum induced by the line cord. A line cord which is not coaxial produces external field. That is inviolate. Typical two wire line cords are zip type, and may not have a twist to them, this is worse.

If you are referring to the OP we really don't know that's all the connections. He didn't say. Any source device connected to the pre-amp may ground it. The circuit could easily be grounded at both ends resulting in unintentional return paths and common-impedance coupling (ground loop).

I agree with you that hum caused by the radiated external magnetic field of the power cord could be reduced by what you are saying but this is rarely the chief problem. Common impedance coupling is more often the issue and wrapping the IC's around the power cord cannot fix that.

Commutator or other electric device noise, radiated directly or from a power line to a receptor circuit, would have a transient high-voltage high-impedance nature. It's caused by the making and breaking of contacts and the asssociated high voltage inductive kick. The coupling, represented by a lumped element capacitor would be primarily by electric field. Cable shields on many unbalanced IC's are very effective against this type of interference.

Gee, I almost wish I had a problem so I could try it!

If you are referring to the OP we really don't know that's all the connections. He didn't say. Any source device connected to the pre-amp may ground it. The circuit could easily be grounded at both ends resulting in unintentional return paths and common-impedance coupling (ground loop).

Agreed. More info would be nice, it could easily have been the cable connection to the receiver, or tv, whatever. My "tip" relates only to the rca's and line cord. Many would believe that the rca shielding is sufficient to allow the ic's to be draped past a magfield source, which is incorrect. My technique allows that without regard to the source.

I agree with you that hum caused by the radiated external magnetic field of the power cord could be reduced by what you are saying but this is rarely the chief problem. Common impedance coupling is more often the issue and wrapping the IC's around the power cord cannot fix that.

That depends on the nature of the source, as well as what you define as "common impedance coupling" Again, I provided a simple tip which is quite effective for a range of problems I have encountered.

Commutator or other electric device noise, radiated directly or from a power line to a receptor circuit, would have a transient high-voltage high-impedance nature. It's caused by the making and breaking of contacts and the asssociated high voltage inductive kick.

Commutation noise we speak of, will not exceed 6 KV, this being flashover at the outlet.

The external electric field due to commutation falls off rather quickly as a result of both potentials being quite close, the instance we speak of is an inch or so. Therefore, capacitive coupling to a system tens of feet away will be quite low.

My typical inductive kick back worry usually involves megajoules of stored energy, or 12 phase scr based supplies. For a toro leafblower, microhenries and 10 amps makes for very quick rate of rise. This rate of rise is along the entire AC distribution system, so broadcasts throughout the auditorium. Trusting eddy currents to quash the mag fields is a lost cause at 60 hz because of skin depth.

The coupling, represented by a lumped element capacitor would be primarily by electric field. Cable shields on many unbalanced IC's are very effective against this type of interference.

We agree electric is coupled by capacitive, and that shields are quite effective against it. We never disagreed on this point.

My point is the magnetic coupling, and I point out some very effective methods to alleviate it, methods which do not seem intuitive to most, cost nothing, and can be quickly done to either affirm or deny the issue.

Cheers, John

Many would believe that the rca shielding is sufficient to allow the ic's to be draped past a magfield source, which is incorrect.

My point is the magnetic coupling, and I point out some very effective methods to alleviate it, methods which do not seem intuitive to most

Agreed.

It's why I have argued here many times against the effectivness of shielded power cords and RG6 subwoofer cables.

Gee, I almost wish I had a problem so I could try it!

Never, ever, ever...wish for a problem. Murphy has ears everywhere.. :eek:

It's why I have argued here many times against the effectivness of shielded power cords
Shielding a twisted pair is useless for prevention of magfield at 60hz due to skin depth.

If you really need magsheild at 60hz, you need one of these:






Cheers, John

I do not provide sound engineering for the facility here, that is the job of another. When he runs the venue with his equipment, he needs DI boxes, he needs balanced runs to and from the stage, he needs transformer isolators...and when he STILL has hum and noise issues, he has to run RF.

I maintain my own system, and use it in the exact same venue...unbalanced, no isolation, nada. Takes me 15 minutes to setup, and I have never had a noise or hum issue. The auditorium I do this in has huge ground loop issues due to HVAC, lights, and roughly 20 megawatts of very noisy power supplies and fridges in close proximity (for those size loads, "proximity" is anything within half a mile) . Any wire which is run from the stage to the balcony traps a huge noise and hum voltage to ground. I avoid that.

Cheers, John

ps...BTW, my equipment does indeed have balanced capability. It has not been required (if it ain't broken, don't fix it). If my technique had not worked, I most certainly would have run balanced. I chose to keep the weight down.

Then that angineer is an idiot because using RF would not solve the issue either because the gear still has to connect to a mixing desk or amps or whatever to get the signal out.

We use DIs all the time, I would never connect a bass rig or keyboard rig right into a mixing desk. We do huge venues and I never have to live with hums buzzes etc.

The venue you use has issues because the audio system has no iso power feed like most theatres and stadiums and arenas do. Its a simple fix. Your just lucky your stuff works, the other guy is an idiot and has something wrong with some of his gear with more than likely a loose pin 1 or some part that has a pin 1 problem. Google Whitlock and pin1 to see more of this issue.

How many electrical engineers does it take to change a light bulb?
Its amazing how this thread morphed into transmission line theory on the relationship between loop area and magnetic coupling (db/dt) among tightly coupled conductors. Which BTW is not the same issue as ground loop hum caused by gear with 3-prong plugs interconnected with ubalanced cables.

Then that angineer is an idiot because using RF would not solve the issue either because the gear still has to connect to a mixing desk or amps or whatever to get the signal out.
Nah, he's pretty good. The problem is the building has no iso. It was made in '48, I believe, and the power problems cannot be removed but lived with. He does what he has to, that is all.

We use DIs all the time, I would never connect a bass rig or keyboard rig right into a mixing desk. We do huge venues and I never have to live with hums buzzes etc.
He also uses DI's for the same reason.

The venue you use has issues because the audio system has no iso power feed like most theatres and stadiums and arenas do. Its a simple fix.
The building is older than I am, sheesh..I agree it's a simple fix, problem is we can't fix it..

Your just lucky your stuff works, the other guy is an idiot and has something wrong with some of his gear with more than likely a loose pin 1 or some part that has a pin 1 problem. Google Whitlock and pin1 to see more of this issue.

My stuff works because I found the problem and addressed it correctly. The other guy is working with what has been given him. The fact that he has inherited huge building power issues does not make him an idiot, nor does it mean his equipment is broken. I am suprised that you would call another an idiot without the benefit of knowledge about them or their circumstances.

Oh, btw...I forgot one very big thing...my source rack in the balcony is sufficiently small that I can power it from the stage via an extension cord. That cord is what I've wrapped with the unbalanced feed to the stage in addition to one balanced mike cord. I was quite suprised the mike worked without hum and noise, it's been seven years now..

The sound guy here can't feed his console and effects stuff from one measly power cord from the stage, he has to use a different circuit..

One conductor grounded at the stage and carried to the balcony, raises over 1.5 volts above ground because of induction. Some equipment has difficulty with that level of induction. I got around it by usin the old noggin.
Cheers, John

How many electrical engineers does it take to change a light bulb?
Its amazing how this thread morphed into transmission line theory on the relationship between loop area and magnetic coupling (db/dt) among tightly coupled conductors. Which BTW is not the same issue as ground loop hum caused by gear with 3-prong plugs interconnected with ubalanced cables.

The problem is not the coupling among tightly coupled conductors, but rather, conductors which are NOT tightly coupled. And that is indeed what causes ground loop hum.

Transmission line theory is much more fun to discuss..but it is not what has been discussed here.

Cheers, John

The problem is not the coupling among tightly coupled conductors, but rather, conductors which are NOT tightly coupled. And that is indeed what causes ground loop hum.

Transmission line theory is much more fun to discuss..but it is not what has been discussed here.

Cheers, John
Electrical and magnetic coupling between signal and power cables is not the cause of audible hum in home av/installations
The imaginary components of the resistancte of a cable does not fall under transmission line theory?

Electrical and magnetic coupling between signal and power cables is not the cause of audible hum in home av/installations
My experience contradicts that. Both at home, and whenever I do a gig.

Doesn't happen a lot, some never are "lucky". But it happens.

Ever hear a sub pop when a compressor motor turns on?.. Are you sure it's because of insufficient PSRR?

Ever come across someone who pops the ground off a 3 pronger to get rid of the hum?

Ever have somebody use an isolator on the 75 ohm feed to the receiver?

Induction. Magnetic induction.


The imaginary components of the resistancte of a cable does not fall under transmission line theory?
Imaginary component of "resistance"?.. You mean reactance?

Lumped element analysis of RLC circuits does not require transmission line theory. And the use of Faraday's law of induction also does not require transmission line theory.

Cheers, John

ps...and the answer to your initial question is ......42. :p

My experience contradicts that. Both at home, and whenever I do a gig.

Doesn't happen a lot, some never are "lucky". But it happens.

Ever hear a sub pop when a compressor motor turns on?.. Are you sure it's because of insufficient PSRR?

No, Its likely due to a sustained (several cycles) voltage drop caused by the initinal draw ( inrush) of the motor when it cycles on.

Ever come across someone who pops the ground off a 3 pronger to get rid of the hum?

Ever have somebody use an isolator on the 75 ohm feed to the receiver?

Induction. Magnetic induction.

No, its the equalization of ground pontentials over the shield/return of the unbalanced cables connecting two componennts with unequal ground potentials.


Imaginary component of "resistance"?.. You mean reactance?

No, I meant Impedance.


Lumped element analysis of RLC circuits does not require transmission line theory. And the use of Faraday's law of induction also does not require transmission line theory.

Cheers, John

ps...and the answer to your initial question is ......42. :p

So in short you're conceeding I'm correct and that analysis of the imaginary components of impedance (reactance) of cable, does fall under transmission line theory ?

No, Its likely due to a sustained (several cycles) voltage drop caused by the initinal draw ( inrush) of the motor when it cycles on.

That was also my initial thought when my system exhibited this pop. However, the voltage drop that ensued during startup did not show across the supply capacitors. A pop which occurs when the music is silent does not result in a droop across the supply caps. The pop I encountered was a result of a stimulus at the input of the sub cabinet.

By re-arranging the line cord of the sub with the ic to the sub, I was able to reduce the pop to below background. This was done without alteration to any other system...the compressor was untouched, the romex in the wall was untouched, I did not use any kind of line conditioner or any additional filtering. All I did was wrap the sub ic around the sub line cord all the way to the receiver.

No, its the equalization of ground pontentials over the shield/return of the unbalanced cables connecting two componennts with unequal ground potentials.

hmmm..I believe you must review faraday's law of induction. When a time varying magnetic is trapped by a conductive loop, current will flow within that loop. The vast majority of audio equipment that exists today is unable to adequately deal with shield currents at their inputs.

"Equilization of ground potentials", as you state, requires two things. Low resistance to reduce any IR drop due to current flow in the ground, and (this is the important part) reduction of flux trapping between the ground point and any loops of conductor.....ic sheilds form loops.

It would be interesting to see how you propose "equilization of ground potentials" for the cable coming into the house.. In my neck of the woods, we have a single bushing distribution system. This means, there are definitely ground currents to deal with and ground potentials to be dealt with. And on top of that, any neutral currents to the water pipe will form a rather large induction loop to the cable feed (remember, that current does not cancel out at the house feed), thereby increasing hum..

You probably think that a star ground is all that you need. Sigh..

No, I meant Impedance.

Ah....then you should have stated impedance.

However, I have not discussed the characteristic impedance of transmission lines. So do not understand why you have included this.

So in short you're conceeding I'm correct and that analysis of the imaginary components of impedance (reactance) of cable, does fall under transmission line theory ?

Again, I have not discussed any "imaginary components of impedance", so am still at a loss to understand why you are talking about it?

Faraday's law of induction is a simple law, simply applied to physically seperated wires. This is not rocket science, but simple stuff. There is no need for you to attempt to bring into the discussion things which are not relevant.

Cheers, John

No, its the equalization of ground pontentials over the shield/return of the unbalanced cables connecting two componennts with unequal ground potentials.

Apparently it take several engineers. :)

I agree with you in this regard. As I stated before, IMO the bigger problem is equipment with unequal ground potentials being connected and the result is common-impedance noise coupling. The classic ground loop we are all plagued with to some degree. Isolation is the cure is almost every case. That's not to say magnetic and electric field coupling don't play a role in generating EMI in audio equipment however. They certainly can.

The reason we are not discussing transmission lines here is we are dealing with electrically short lines. The circuit elements are lumped vs distributed as the wavelengths are relatively long. You can have imaginary reactances even on electrically short lines.

If you really need magsheild at 60hz, you need one of these:

Looks like copper. Non-magnetic shields are near useless at most practical thcknesses at preventing 60 Hz ingress.

We use DIs all the time

What's a DI?

Looks like copper. Non-magnetic shields are near useless at most practical thcknesses at preventing 60 Hz ingress.


He he...

Nobody said that cord was practical...sheesh.. ;) twas a joke son, a joke I tell ya..



Cheers, John

That was also my initial thought when my system exhibited this pop. However, the voltage drop that ensued during startup did not show across the supply capacitors. A pop which occurs when the music is silent does not result in a droop across the supply caps. The pop I encountered was a result of a stimulus at the input of the sub cabinet.

By re-arranging the line cord of the sub with the ic to the sub, I was able to reduce the pop to below background. This was done without alteration to any other system...the compressor was untouched, the romex in the wall was untouched, I did not use any kind of line conditioner or any additional filtering. All I did was wrap the sub ic around the sub line cord all the way to the receiver.

I had the same issue, and had complete success moving my a/v gear to another dedicated ac circuit. The voltage drop on its own may not be enough to cause the audible pop depending on the stiffness of your ac supply and the robustness of your gear’s power supply. But in combination with the transients that follow (high energy, high frequency, multi-phase events) it is likely to cause an audible pop.


hmmm..I believe you must review faraday's law of induction. When a time varying magnetic is trapped by a conductive loop, current will flow within that loop. The vast majority of audio equipment that exists today is unable to adequately deal with shield currents at their inputs.

Yes I understand Faraday’s law, my point being that its not applicable to run of the mill ground loops.



"Equilization of ground potentials", as you state, requires two things. Low resistance to reduce any IR drop due to current flow in the ground, and (this is the important part) reduction of flux trapping between the ground point and any loops of conductor.....ic sheilds form loops.

It would be interesting to see how you propose "equilization of ground potentials" for the cable coming into the house.. In my neck of the woods, we have a single bushing distribution system. This means, there are definitely ground currents to deal with and ground potentials to be dealt with. And on top of that, any neutral currents to the water pipe will form a rather large induction loop to the cable feed (remember, that current does not cancel out at the house feed), thereby increasing hum..

You probably think that a star ground is all that you need. Sigh..

This is done all the time at the main service panel/meter and on certain surge suppressors with signal grounding blocks.


Ah....then you should have stated impedance.

However, I have not discussed the characteristic impedance of transmission lines. So do not understand why you have included this.



Again, I have not discussed any "imaginary components of impedance", so am still at a loss to understand why you are talking about it?


Faraday's law of induction is a simple law, simply applied to physically seperated wires. This is not rocket science, but simple stuff. There is no need for you to attempt to bring into the discussion things which are not relevant.

Cheers, John
Your using Faraday’s law to justify wrapping interconnects around electrical cords, when its more aptly applied to the twisted pairs of cat 5e to increase bandwith and lower crosstalk.

What's a DI?


http://www.radialeng.com/di-j48.htm

Although sponsered by Jensen these papers are considered to be the best on this info. Sorry John but I have never myself nor have I ever met a real audio pro who does what you do with unbalanced lines. We dont need to go to such extremes it seems. I guess we dont tour in old enough buildings.

http://www.jensen-transformers.com/apps_wp.html

I had the same issue, and had complete success moving my a/v gear to another dedicated ac circuit. The voltage drop on its own may not be enough to cause the audible pop depending on the stiffness of your ac supply and the robustness of your gear’s power supply. But in combination with the transients that follow (high energy, high frequency, multi-phase events) it is likely to cause an audible pop.

Yah, sometimes just moving to a dedicated circuit is the best. I first moved the sub AC line, and that didn't resolve the pop. It turned out that the AC compressor run was standard romex, and it cruised right through the wall my ht system was against.. So the locked rotor current at startup was creating the magnetic field rather close to the sub loop. Wrapping the ic around the sub linecord eliminated that field pickup.


Yes I understand Faraday’s law, my point being that its not applicable to run of the mill ground loops.
That is incorrect. Ground loops cause trouble by two methods, IR drop, and trapped flux induction. To neglect the flux induction is ignoring potential problems..in my case, I've used an understanding of flux trapping to solve quite a few difficult problems.
This is done all the time at the main service panel/meter and on certain surge suppressors with signal grounding blocks.
Yup. And many times, simply grounding the tv cable feedin at that point does nothing to solve ht hum issues, an isolator is needed at times.


Your using Faraday’s law to justify wrapping interconnects around electrical cords, .
I am using Faraday's law of induction to explain why trapping of magnetic flux that is time varying can be a problem. It is a trivially EASY method of removing this source of hum and noise pickup, it costs nothing to try, and if it works, completely explains where the signals are coming from. If it is tried and it works, the only decision is, do you keep the cords wrapped therefore resolving the problem, or do you unwrap them and persue haphazard guesses in an attempt at solving it using random tweaks.
when its more aptly applied to the twisted pairs of cat 5e to increase bandwith and lower crosstalk

The need to use different twist pitches within a cat5e cable for pair isolation is well known and trivial.

If you place a balanced run of mike cable next to an extension cord which has the same twist pitch or an integer multiple of it, do you expect isolation? I don't...and I've experienced this. I found it and fixed it because I considered magnetic field and induction.

Electrostatic thinking of your type, is only half the science. I am not hampered in that regard..

Cheers, John

Sorry John but I have never myself nor have I ever met a real audio pro who does what you do with unbalanced lines. We dont need to go to such extremes it seems.


There is no need to be sorry. I also have not met a real audio pro who does what I do with unbalanced lines, nor do I recommend that a real audio pro use unbalanced runs for any of their equipment. That would be rather scary.

As it is, I carefully inspect every ground connection at the beginning of each summer...this is very important simply because ground failure in my case is very loud and will probably toast the tweets..(the woofs are capable of rms dissipations 20% over the amp's clipping output, so that the fool who owns them cannot kill them (me being the "fool".)

As I said, I use unbalanced because the spool my snake is on is stuffed to capacity (I added a rg6 run to it last year for a television on stage), and if I didn't need to run balanced, I could use what I had.

My point in all this is: while my application is extreme, it is possible because I considered the magnetic field induction paths which wreak havoc with most equipment. For others such as you, it may be necessary to consider the mag effects I speak of someday...imagine the surprise (I finally spelled that right..) on somebody's face when you wrap the signal line around the line cord and the hum goes away??? It's so weird and contrary to intuition.

Consideration of magnetic field induced effects is simply a tool....something to have in your grab bag of knowledge in case it is required. It certainly does not substitute for engineered solutions, but rather, augments them.

I guess we dont tour in old enough buildings.

Some people have all the luck..

Cheers, John

Although sponsered by Jensen these papers are considered to be the best on this info. Sorry John but I have never myself nor have I ever met a real audio pro who does what you do with unbalanced lines. We dont need to go to such extremes it seems. I guess we dont tour in old enough buildings.

http://www.jensen-transformers.com/apps_wp.html


Holy mackeral... I just read that paper "Hum and buzz in unbalanced interconnect systems". WHAT A RUDIMENTARY TREATMENT OF THE TOPIC!!!

No wonder you don't understand what I'm talking about.. sheesh.

Are there any writeups which are thorough? Or is this it??

My goodness...I really did not think pro audio was so far behind in this area.

Cheers, John

ps...I apologise for any offence taken with regard to my remarks, but I am quite astonished by the level of that paper. :eek:

[QUOTE=jneutron]Holy mackeral... I just read that paper "Hum and buzz in unbalanced interconnect systems". WHAT A RUDIMENTARY TREATMENT OF THE TOPIC!!!

No wonder you don't understand what I'm talking about.. sheesh.

Are there any writeups which are thorough? Or is this it??

My goodness...I really did not think pro audio was so far behind in this area.



QUOTE]

I understand what your doing its just not needed there are easier ways to do it.

Bill Whitlock is an easy guy to get ahold of I would suggest contacting him and telling him he is so off base.
Rudimentry but true. And notice that deals with what folks will see in most home systems, not on my Meyer rig.The Pin 1 issue is something we run into.
http://www.audiodesignline.com/GLOBAL/electronics/designline/shared/article/showArticle.jhtml?articleId=187201359&pgno=4

These are more up to date.

http://www.rane.com/note110.html

http://www.rane.com/note151.html

I understand what your doing its just not needed there are easier ways to do it.

Bill Whitlock is an easy guy to get ahold of I would suggest contacting him and telling him he is so off base.
Rudimentry but true. And notice that deals with what folks will see in most home systems, not on my Meyer rig.The Pin 1 issue is something we run into.
http://www.audiodesignline.com/GLOBAL/electronics/designline/shared/article/showArticle.jhtml?articleId=187201359&pgno=4

These are more up to date.

http://www.rane.com/note110.html

http://www.rane.com/note151.html

Oh, thank goodness..the first one I looked at, sheesh...nuttin but capacitance???

The first of these new links is good, from Bill... it identifies the loop I speak of, and offers a method of looking for it. It does not, however, address the reason for the currents so does not offer ways to eliminate them at the source. My methodology eliminates the currents at the source..

Instead, he talks about designing immunity to the currents. Not bad, but it only goes part of the way with respect to design issues, I note he only talks about IR drop. There are also inductive coupling issues with internal chassis ground currents, which he does not address. Care must be taken with the chassis design, to assure that ground currents cannot couple to the internals, not just watching for IR drop.

In my line of work, this issue is not something that can be ignored..we gotta find microvolt and millivolt signal levels buried in 10 to 30 kiloamp widgits runnin 5 to 10 tesla internal fields and magnetic gradients in space and time that defy imagination....then we have half a second or so to dump a coupla hundred kilojoules outta the widgits..(I play wit big toys..) Needless to say, I gotta pay a lot closer attention to this magnetic goop than you need to..cause the AC line feeds to the power supplies radiates holy "heck" at 60 hz up (scr based supplies).

The Ranes look ok at first glance, but I'll have to read them when I return on the 23rd, I'm kinda busy preppin for vaca..

If you or Bill are interested, toss me a line, I'd be happy to elaborate..


Cheers, John

Ya
h, sometimes just moving to a dedicated circuit is the best. I first moved the sub AC line, and that didn't resolve the pop. It turned out that the AC compressor run was standard romex, and it cruised right through the wall my ht system was against.. So the locked rotor current at startup was creating the magnetic field rather close to the sub loop. Wrapping the ic around the sub linecord eliminated that field pickup. My experience is different; I simply ran a dedicated line to my HT and did not move my gear or any pre-existing house wiring. If your theory is correct, then I would still experience the pop, which I don’t. So the only conclusion I can draw is that the audible pop exhibited by my gear, a direct consequence of the electrical events initiated by a motor startup, was NOT inducted but conducted into my system.
That is incorrect. Ground loops cause trouble by two methods, IR drop, and trapped flux induction. To neglect the flux induction is ignoring potential problems..in my case, I've used an understanding of flux trapping to solve quite a few difficult problems.
Ok, I get what you’re saying, that the house wiring acts like a very low frequency antenna: an antenna’s resonant frequency has an inverse relationship with its loop area and the number of turns within that loop. Runs of Romex through a wall are not coiled, and the loop area of a ground loop can be very large, ideal conditions for a very low frequency antenna. BUT an antenna has only one resonant frequency, and the resonant frequency of a typical length interconnect is too high to act as a receiver to what your Romex wiring could conceivable transmit.



Yup. And many times, simply grounding the tv cable feedin at that point does nothing to solve ht hum issues, an isolator is needed at times. That’s because no 2 circuits are ever exactly alike, minute differences in impedance will always exist and some ground loop current will always flow if they were coupled.

I also wanted to point out by coiling an interconnect you increase its characteristic impedance (not good for your pre/pro output op amps) and increase its resonant frequency making it more susceptible to EMI.

In my line of work, this issue is not something that can be ignored..


Which us what? Rocket science?? :D

My experience is different; I simply ran a dedicated line to my HT and did not move my gear or any pre-existing house wiring. If your theory is correct, then I would still experience the pop, which I don’t. So the only conclusion I can draw is that the audible pop exhibited by my gear, a direct consequence of the electrical events initiated by a motor startup, was NOT inducted but conducted into my system.
My theory was correct for my situation, which was coupling from the AC line to the sub ic/ac loop. I am not saying it is the only way, just one of several.

Ok, I get what you’re saying, that the house wiring acts like a very low frequency antenna: an antenna’s resonant frequency has an inverse relationship with its loop area and the number of turns within that loop. Runs of Romex through a wall are not coiled, and the loop area of a ground loop can be very large, ideal conditions for a very low frequency antenna. BUT an antenna has only one resonant frequency, and the resonant frequency of a typical length interconnect is too high to act as a receiver to what your Romex wiring could conceivable transmit.
Antenna efficiency is certainly based on physical size, as you point out. However, the local coupling we are discussing is not based on antennae theory, but rather, transformers. We are not establishing a propagating wave in order to couple, just trapping the local flux.
I also wanted to point out by coiling an interconnect you increase its characteristic impedance (not good for your pre/pro output op amps) and increase its resonant frequency making it more susceptible to EMI.

Coiling an interconnect doesn't alter it's characteristic impedance. However, creating more loops does indeed make it more susceptible to emi as you say.
Which us what? Rocket science??

Nah, it ain't rocket science.. :D

I use superconductors.

Cheers, John

Superconductors are cool.

Sorry, couldn't resist

Antenna efficiency is certainly based on physical size, as you point out. However, the local coupling we are discussing is not based on antennae theory, but rather, transformers. We are not establishing a propagating wave in order to couple, just trapping the local flux.

Antennas and transformers are subclasses of Faraday’s law of induction (the superclass). If, as you claim, your theory is an instantiation of the transformer class, what are the values of the attributes of your object?
Coiling an interconnect doesn't alter it's characteristic impedance. However, creating more loops does indeed make it more susceptible to emi as you say.

You're correct , the characteristic impedance of the cable is unaltered but the impedance of the circuit formed by the interconnect and its transmitter and receiver will increase if the interconnect is coiled because it then acts as an air core inductor.

if the interconnect is coiled because it then acts as an air core inductor.


It acts as two inductors, each with opposing current flow, which largely cancels out the effect.

It acts as two inductors, each with opposing current flow, which largely cancels out the effect.
Hello Targus, welcome back.
Because of ground loop currents, a net-current field exists.

Antennas and transformers are subclasses of Faraday’s law of induction (the superclass). If, as you claim, your theory is an instantiation of the transformer class, what are the values of the attributes of your object?.

Actually, it is not a theory, but rather, practice and experience. And as you can see, the link you provided from Bill shows methods of detecting sensitivity to this transformer class of induced hum, but does not discuss methods of eliminating that hum induction.

My initial statement to the OP concerned methods of removing the induction, as I assume he runs unbalanced.

"values of the attributes?" sheesh, that was technical sounding... :eek:

Seriously, think about how much control you have over the ac distribution geometry, and the variation from site to site...The "remove the sensitivity to the current" approach is pretty much the only way people like you can control hum in your systems. My approach goes a bit beyond that, to "remove the current".

They are both good tools to use, but I stress (again), making the equipment insensitive is the first thing to do.

Values...ok, assume two 100 foot runs of romex to a common panel, one feeds amp, other pre. Assume they remain parallel (fairly decent assumption)..

Romex, 250 nH per foot, 25 uH total inductance to panel. Poweramp pulls 10 amperes sine at 60 hz. Neglect inductive comp at amp powerin, assume the supply has adequate psrr..

Half the dipole field of the romex is on either side of the romex ground feeding the amp. So, if another wire (the pre ground) traps half that flux, it will produce half the inductive reaction of the poweramp run.. This means, the ground to ground potential at the equipment will be half the inductive voltage of the poweramp romex.

So, V = L di/dt..

V = 25 times 10e-6 times 10 amps * 2 * PI * 60hz (cos 2PI 60 t)

V = 25 * 10e-6 * 3770 cos (wt) Edit: first pass I accidentally multiplied in another 25, making the numbers 25 times too large...sorry bout that.

V = 94250 * 10e-6 cos(wt)

V = 90 millivolts...that is the poweramp stray due to the romex inductance.

The maximum flux on one side of the romex is half, this means the amp to pre ground potential can be no more than 45 millivolts ..at 60 hz.

Poweramps pull haversines..second component is 180 hz, and the di/dt rate is triple...

These calcs show the induced voltage, you have to include mutual coupling and loop impedance to calculate the actual ground current.. Keep in mind, however, the signal conductor will not have such a low impedance, the unbalanced input will see the full 45 millivolts. balanced should cancel that out..


Antenna practice is generally far field, with planar waves...I've not seen any derivations with a radius of curvature assumption built into the TEM wavefront.

You're correct , the characteristic impedance of the cable is unaltered but the impedance of the circuit formed by the interconnect and its transmitter and receiver will increase if the interconnect is coiled because it then acts as an air core inductor.

It's unfortunate you choose "impedance" as terminology for several entities, it gets confusing because we do not share a common terminology...

Looping will indeed increase flux trapping effects..

Cheers, John

Actually, it is not a theory, but rather, practice and experience. And as you can see, the link you provided from Bill shows methods of detecting sensitivity to this transformer class of induced hum, but does not discuss methods of eliminating that hum induction.
I provided no such link.
My initial statement to the OP concerned methods of removing the induction, as I assume he runs unbalanced.
...by making the interconnect more susceptible to EMI and flux induction from net-currents?

Values...ok, assume two 100 foot runs of romex to a common panel, one feeds amp, other pre. Assume they remain parallel (fairly decent assumption)..

Ok, I think I finally understand, and NO, this is not a "fairly decent assumption" as ac circuits are not run parallel unless they're split phase. Also the field strength is inversely proportional to the square of the distance from the conductor, so simply putting some distance between your two circuits would mitigate the problem.

I provided no such link.

Go back and read the second round of links you provided me. In it you will find Bill using a transformer and a pair of antiparallel diodes to find sensitivity to ground currents in balanced equipment. He unfortunately only considers resistive drop signal and neglects inductive coupling within the chassis, but the article is still good.

Didn't you read the link you provided me??

...by making the interconnect more susceptible to EMI and flux induction from net-currents?

You continue to use incorrect theory. It's not antenna theory, stop thinking of it this way.

Ok, I think I finally understand, and NO, this is not a "fairly decent assumption" as ac circuits are not run parallel unless they're split phase.

My goodness, you are very wrong. Haven't you ever had reno work done to a house? I had a 200 amp panel put in when an addition was added. The panel is on the front wall, the addition the back. EVERY single romex run is stapled to a floor joist in the basement, and there's about 10 runs that are parallel, absolutely parallel, and looks really nice and clean..

How many mechanics twist conductors prior to a conduit pull??? NONE that I know of.

I would venture to guess that you were thinking of electrically parallel?? I am not discussing that. Please read my posts more thouroughly..


Also the field strength is inversely proportional to the square of the distance from the conductor, so simply putting some distance between your two circuits would mitigate the problem.

The field strength of a single conductor falls off as 1/R. For a pair which create a dipole field, it is 1/R². For a quad set, it falls as 1/R3. In general, the external field strength falls as 1/Rn, and the field in the center of the wireset falls off as 1/Rn-1.

However, for a loop integral where one conductor is at the source (the amp ground wire), the integral loop capture asymptotically approaches a non-zero value, that value dependent on the actual total external flux being trapped. I provided worst case numbers for a 100 foot romex run.

For the problem I am discussing, your decision to "put some distance" worsens the problem. It is because of errors in thinking of this type that the best approach is to make the equipment insensitive to the ground current, as people in the field (such as yourself) do not have to worry about how the AC system is built. E/M goop is not always as simple as you are trying to make it.

Again, you are trying to use antenna theory in a problem which is not that. We are in actuality, discussing a small aspect of what I do for a living. If you choose to ignore what I have been saying, so be it...I present tools for you, you do not have to use them..

Cheers, John

ps..bummer, I forgot this site doesn't support equations. :(

John I provided the link to Bill. Regardless I gave up on this thread long ago. You seem to be the only one in the audio biz that wraps cables around each other etc. to prevent noise. I and many others dont need to go to such extremes to simply break a ground loop or avoid RF problems.

John I provided the link to Bill.

Ahhh, ok..

Sprung...my apologies...I thought you provided the link.


Regardless I gave up on this thread long ago.
You seem to be the only one in the audio biz that wraps cables around each other etc. to prevent noise. I and many others dont need to go to such extremes to simply break a ground loop or avoid RF problems.

You're probably right. I also know of nobody in their right mind who does what I did.

The primary reason I did such was cost. Since my involvement with the audio here is entirely voluntary and outside the scope of my normal work, I did not wish to incur any of my own expense to solve the problems facing me. This facility uses somewhere between 25 and 35 megawatts of power, with some very messy conversions, and the e/m enviro is quite hostile..

Nonetheless, the techniques I outline have served me well for everything work related, as well as audio apps from mike and phono up to the poweramps.

Cheers, John

Try wrapping the rca cords tightly around the preamp line cord the entire length of the line cord.

John

When you say "line cord" are you referring to the power cord?

When you say "line cord" are you referring to the power cord?
Yes.

The purpose is to reduce the loop area between the power cords and the IC. This works best if the sub powercord goes to the receiver also. If it goes to a separate wall outlet, it is most likely useless to wrap the two.

Oops, wrong thread....sub?? It'll work for anything else also..

But it still requires plugging the source into the receiver.


Cheers, John

Actually, it is not a theory, but rather, practice and experience. And as you can see, the link you provided from Bill shows methods of detecting sensitivity to this transformer class of induced hum, but does not discuss methods of eliminating that hum induction.

My initial statement to the OP concerned methods of removing the induction, as I assume he runs unbalanced.

"values of the attributes?" sheesh, that was technical sounding... :eek:

Seriously, think about how much control you have over the ac distribution geometry, and the variation from site to site...The "remove the sensitivity to the current" approach is pretty much the only way people like you can control hum in your systems. My approach goes a bit beyond that, to "remove the current".

They are both good tools to use, but I stress (again), making the equipment insensitive is the first thing to do.

Values...ok, assume two 100 foot runs of romex to a common panel, one feeds amp, other pre. Assume they remain parallel (fairly decent assumption)..

Romex, 250 nH per foot, 25 uH total inductance to panel. Poweramp pulls 10 amperes sine at 60 hz. Neglect inductive comp at amp powerin, assume the supply has adequate psrr..

Half the dipole field of the romex is on either side of the romex ground feeding the amp. So, if another wire (the pre ground) traps half that flux, it will produce half the inductive reaction of the poweramp run.. This means, the ground to ground potential at the equipment will be half the inductive voltage of the poweramp romex.

So, V = L di/dt..

V = 25 times 10e-6 times 10 amps * 2 * PI * 60hz (cos 2PI 60 t)

V = 25 * 10e-6 * 3770 cos (wt) Edit: first pass I accidentally multiplied in another 25, making the numbers 25 times too large...sorry bout that.

V = 94250 * 10e-6 cos(wt)

V = 90 millivolts...that is the poweramp stray due to the romex inductance.

The maximum flux on one side of the romex is half, this means the amp to pre ground potential can be no more than 45 millivolts ..at 60 hz.

Poweramps pull haversines..second component is 180 hz, and the di/dt rate is triple...

These calcs show the induced voltage, you have to include mutual coupling and loop impedance to calculate the actual ground current.. Keep in mind, however, the signal conductor will not have such a low impedance, the unbalanced input will see the full 45 millivolts. balanced should cancel that out..


Antenna practice is generally far field, with planar waves...I've not seen any derivations with a radius of curvature assumption built into the TEM wavefront.



It's unfortunate you choose "impedance" as terminology for several entities, it gets confusing because we do not share a common terminology...

Looping will indeed increase flux trapping effects..

Cheers, John
So To re-cap: the induced time varying voltage v(t) = [inductance of the inductor L (itself a function of the permeability of the inductor core and the number of n-turns of the conductor aka windings)] * [derivative of the time varying current di/dt ]. To be accurate this is a method of the inductor class. The transformer class is a grandchild of the inductor class, inheriting all its methods and fields adding only the n-turns of the secondary windings and a method to determine the final voltage at the secondary (a function of the n-turns of the windings of the secondary and the primary). The reluctance of the inductor core, and n-turns of the windings being paramount to the entire process. You use the series inductance of the conductor for L and the amp draw as the current in di/dt. This I understand, but do not agree with because, as Targus so eloquently stated, the inductance of the amp draw is mostly cancelled by the return current on the neutral.

My goodness, you are very wrong. Haven't you ever had reno work done to a house? I had a 200 amp panel put in when an addition was added. The panel is on the front wall, the addition the back. EVERY single romex run is stapled to a floor joist in the basement, and there's about 10 runs that are parallel, absolutely parallel, and looks really nice and clean..

How many mechanics twist conductors prior to a conduit pull??? NONE that I know of.

I would venture to guess that you were thinking of electrically parallel?? I am not discussing that. Please read my posts more thouroughly..
Paralell in my book means conductors in the same nonmetallic sheathing (Romex) or in the same flexible metallic conduit (BX). Split phase neutral circuits either feed 120 volt (2 circuits) or 240 volt (single circuit) loads on the same branch circuit, off the same breaker, and therefore truly run paralell the entire length of the the home run; from the overcurrent device (the panel/subpanel) to the outlets (the NEC's definition of a Branch circuit btw). What you describe is multiple homeruns of Romex for luminiaries and outlets in different rooms, they do not run paralell.
The field strength of a single conductor falls off as 1/R. For a pair which create a dipole field, it is 1/R². For a quad set, it falls as 1/R3. In general, the external field strength falls as 1/Rn, and the field in the center of the wireset falls off as 1/Rn-1.

However, for a loop integral where one conductor is at the source (the amp ground wire), the integral loop capture asymptotically approaches a non-zero value, that value dependent on the actual total external flux being trapped. I provided worst case numbers for a 100 foot romex run.

For the problem I am discussing, your decision to "put some distance" worsens the problem. It is because of errors in thinking of this type that the best approach is to make the equipment insensitive to the ground current, as people in the field (such as yourself) do not have to worry about how the AC system is built. E/M goop is not always as simple as you are trying to make it.

Again, you are trying to use antenna theory in a problem which is not that. We are in actuality, discussing a small aspect of what I do for a living. If you choose to ignore what I have been saying, so be it...I present tools for you, you do not have to use them..

Cheers, John

ps..bummer, I forgot this site doesn't support equations. At least we agree that the field strength falls. But the ground conductor of the Romex cannot act as the flux conductor or secondary winding of your transformer object. Any current on the ground conductor is due to common impedance coupling of both circuits over the unblanced interconnects between the amp and pre in your model. Given that you no longer have a stimulus current, it seems to me your model falls apart, I won't even pretend I understand what is the "integral loop capture"

So To re-cap: the induced time varying voltage v(t) = [inductance of the inductor L (itself a function of the permeability of the inductor core and the number of n-turns of the conductor aka windings)] * [derivative of the time varying current di/dt ]. To be accurate this is a method of the inductor class. The transformer class is a grandchild of the inductor class, inheriting all its methods and fields adding only the n-turns of the secondary windings and a method to determine the final voltage at the secondary (a function of the n-turns of the windings of the secondary and the primary). The reluctance of the inductor core, and n-turns of the windings being paramount to the entire process. You use the series inductance of the conductor for L and the amp draw as the current in di/dt. This I understand, but do not agree with because, as Targus so eloquently stated, the inductance of the amp draw is mostly cancelled by the return current on the neutral.

To be accurate, inductors and transformers are the same entity. The voltage that is produced in a loop of conductor is proportional to the rate of change of flux in that loop. It is immaterial whether that conductor is part of the current path, or simply picking up the flux from the current path.

The inductance of the amp draw is mostly cancelled???? Boy, have you got that mixed up. As I stated, romex runs about 250 nH per foot. There is nothing you can do to reduce that, without disassembling the romex and putting the hot and neutral together. It is geometry controlled.

If you wish to be accurate, the magnetic field caused by the amp draw is mostly cancelled, but that cancellation results in the 250 nH per foot number. And that inductance creates the field I'm talking about.

Paralell in my book means conductors in the same nonmetallic sheathing (Romex) or in the same flexible metallic conduit (BX). Split phase neutral circuits either feed 120 volt (2 circuits) or 240 volt (single circuit) loads on the same branch circuit, off the same breaker, and therefore truly run paralell the entire length of the the home run; from the overcurrent device (the panel/subpanel) to the outlets (the NEC's definition of a Branch circuit btw). What you describe is multiple homeruns of Romex for luminiaries and outlets in different rooms, they do not run paralell.


The definition of parallel I am using is the wires running side by side with the same seperation along the length. "they do not run parallel"???? You are kidding, right? Had I known you'd make a statement like that, I'da taken a simple picture...at the staples too, to show the clean work the mechanic did..

At least we agree that the field strength falls. But the ground conductor of the Romex cannot act as the flux conductor or secondary winding of your transformer object. Any current on the ground conductor is due to common impedance coupling of both circuits over the unblanced interconnects between the amp and pre in your model. Given that you no longer have a stimulus current, it seems to me your model falls apart, I won't even pretend I understand what is the "integral loop capture"


I won't even pretend I understand what is the "integral loop capture"

That is why you haven't understood a single word I've said here. You should have stated a while back that you do not understand the terms, that would have saved a lot of writing..all this time, you've been having a discussion on a subject that you are not understanding??

Faraday's law of induction defines the voltage generated around a loop as a result of the rate of change of the flux trapped by the loop. It is a simple integration of the total area enclosed by the loop. Ground loop currents are generated because the return leg of the loop is not the same wire as the start leg. As the return leg gets farther from the field generator, it is enclosing more and more flux from the field generator. Since the field is falling off, moving the return leg farther away asymptotically traps all the flux available, for a dipole generator such as romex, this end limit is 1/2 of the total flux of the romex.

You need to understand this field theory if you wish to discuss it reasonably. You are trying to apply incorrect concepts to the discussion.. If you wish, go to the hyperphysics site, they have some decent pages on the e/m stuff. If you do not understand how it applies to the discussion at hand, pm me with questions, I'll be happy to help you understand..

You stating a ground loop cannot be formed by the two romex grounds is entirely inconsistent with reality. You stating it cannot be.. does not change that reality...

Do yourself a favor. Measure the voltage between the grounds of two outlets on opposite sides of a room. It doesn't matter whether or not the outlets are being used, or in fact, if that circuit is being used. You will find a voltage between the grounds. WITHOUT any current being drawn.

You are measuring the voltage generated by the flux trapped within the loop you formed. It's not IR drop, there's no current through the #14 (minimum) to the panel..

Cheers, John

To be accurate, inductors and transformers are the same entity. The voltage that is produced in a loop of conductor is proportional to the rate of change of flux in that loop. It is immaterial whether that conductor is part of the current path, or simply picking up the flux from the current path.

The inductance of the amp draw is mostly cancelled???? Boy, have you got that mixed up. As I stated, romex runs about 250 nH per foot. There is nothing you can do to reduce that, without disassembling the romex and putting the hot and neutral together. It is geometry controlled.

If you wish to be accurate, the magnetic field caused by the amp draw is mostly cancelled, but that cancellation results in the 250 nH per foot number. And that inductance creates the field I'm talking about.

If hot and neutral currents are identical and the ground is positioned exactly midway between them, no voltage would be induced in the ground wire because the magnetic fields ( a function of the series inductance of the conductor in this instance) are of equal strength but opposing polarity and would cancel. Induced voltages are even less likely to occur in Romex because of the uniform conductor geometry maintained by the NM sheathing. You'd have a better case with BX or Armored Cable.


The definition of parallel I am using is the wires running side by side with the same seperation along the length. "they do not run parallel"???? You are kidding, right? Had I known you'd make a statement like that, I'da taken a simple picture...at the staples too, to show the clean work the mechanic did..

I'm not kidding, I define parallel as the entire length of the branch circuit. Given that each of your home runs power luminaries and outlets not only in different areas of a room, but rooms in different parts of the house, how can you possibly define the runs as parallel?


I won't even pretend I understand what is the "integral loop capture"

That is why you haven't understood a single word I've said here. You should have stated a while back that you do not understand the terms, that would have saved a lot of writing..all this time, you've been having a discussion on a subject that you are not understanding??

Faraday's law of induction defines the voltage generated around a loop as a result of the rate of change of the flux trapped by the loop. It is a simple integration of the total area enclosed by the loop. Ground loop currents are generated because the return leg of the loop is not the same wire as the start leg. As the return leg gets farther from the field generator, it is enclosing more and more flux from the field generator. Since the field is falling off, moving the return leg farther away asymptotically traps all the flux available, for a dipole generator such as romex, this end limit is 1/2 of the total flux of the romex.

You need to understand this field theory if you wish to discuss it reasonably. You are trying to apply incorrect concepts to the discussion.. If you wish, go to the hyperphysics site, they have some decent pages on the e/m stuff. If you do not understand how it applies to the discussion at hand, pm me with questions, I'll be happy to help you understand..

You stating a ground loop cannot be formed by the two romex grounds is entirely inconsistent with reality. You stating it cannot be.. does not change that reality...

Do yourself a favor. Measure the voltage between the grounds of two outlets on opposite sides of a room. It doesn't matter whether or not the outlets are being used, or in fact, if that circuit is being used. You will find a voltage between the grounds. WITHOUT any current being drawn.

You are measuring the voltage generated by the flux trapped within the loop you formed. It's not IR drop, there's no current through the #14 (minimum) to the panel..

Cheers, John Hyperphysics, give me a break, I think you're over analyzing the situation. You're mired in negligible inductances that have no noticeable effect on the system. And ignoring the common impedance coupling which we all know, by experience, to be the cause of most noise in a system. For all I know you could of been referring to the integral of the time derivative. You make so many frequent references to calculus, for effect I suppose, its hard to know what your meaning is.

If hot and neutral currents are identical and the ground is positioned exactly midway between them, no voltage would be induced in the ground wire because the magnetic fields ( a function of the series inductance of the conductor in this instance) are of equal strength but opposing polarity and would cancel.
Correct.

Induced voltages are even less likely to occur in Romex because of the uniform conductor geometry maintained by the NM sheathing. You'd have a better case with BX or Armored Cable.
Again, correct.

Now, take a second wire, run it parallel to the romex under power but 2 inches away, connect the ground at the service panel...and measure the voltage that the powered romex induces between the ground wires.

The powered romex ground forms a loop with the external wire, it traps some of the flux coming from the powered romex.

You didn't do the simple measurement I asked you to, did you..?? It's there even if you refuse to measure it...

.... I define parallel as the entire length of the branch circuit. Given that each of your home runs power luminaries and outlets not only in different areas of a room, but rooms in different parts of the house, how can you possibly define the runs as parallel?

Hmmm, lets see. From the panel to a point 45 feet away from the panel, I have about 10 romex runs, all stapled to the same floor joist, side by side, with about 1/4 inch between the romex's. As I said, they are parallel. Physically parallel. And each of them broadcasts a dipole field when current flows through them.

Why are you having such a problem with the word "parallel"?

.... Hyperphysics, give me a break, I think you're over analyzing the situation. You're mired in negligible inductances that have no noticeable effect on the system.
As I detailed, while an inductance of 25 uH is trivially small for the device which is loading the line, there are consequences to other wires in the area. I guess you consider 45 millivolts at an unbalanced line input as being "negligible"?

In the auditorium where I run my 100 foot unbalanced cord, the induced voltage between the stage ground and the balcony ground is over 1.5 volts 60 hz, with hash spikes exceeding 10 to 20 volts. Negligible??

.... And ignoring the common impedance coupling which we all know, by experience, to be the cause of most noise in a system.

As I said earlier, you use the word impedance in different ways than I. I integrate the time varying flux, you say "common impedance". What I use can be integrated and normalized to get to what you refer to, but that is not necessary. My methods remove the culprit of noise induction...the flux trapping.

My method is useable in all environments, you with "common impedance", completely falls apart when multiple, multi-phase sources are involved..mine works regardless of the e/m enviro.

.... For all I know you could of been referring to the integral of the time derivative. You make so many frequent references to calculus, for effect I suppose, its hard to know what your meaning is.

Quite honestly, I've kept the analysis and verbage as low level as I can, so that you can follow. That has not worked, you are still stuck in antenna's and such.

The word "integral" for example, if you prefer I dumb it down, I guess I could say "an enclosed area where the total amount of flux lines going through the area is added up and the rate of change of all that flux through that area is derived, to provide the voltage that will appear around that area". But I suspect you still will not understand..perhaps I'm wrong??

PM me if you wish to learn what I'm talking about. Since you haven't, I guess you're not interested in understanding.


And do yourself a favor...try the test I recommended, then figure out why the voltage is there..it's simple to eliminate once you understand the cause.

Cheers, John

You can tell when you install your cables backwards when the music comes out in reverse.

You can tell when you install your cables backwards when the music comes out in reverse.

Thats's ok, I'll just use a mirror to reflect the sound, reversing the reverse music.

I'm lost by this thread by the way.

I was told by the dealer, that there was a person who actually listened to the
cables, and where ever the music sounds best - the arror points the way. :eek:

I was only asking, because I was troubleshooting a ground loop, and on one
side of the interconnects I have my 2-prong power cord pre-amp, and on
the otherside was my 3-prong power cord amp.

To be accurate, inductors and transformers are the same entity. The voltage that is produced in a loop of conductor is proportional to the rate of change of flux in that loop. It is immaterial whether that conductor is part of the current path, or simply picking up the flux from the current path.

The inductance of the amp draw is mostly cancelled???? Boy, have you got that mixed up. As I stated, romex runs about 250 nH per foot. There is nothing you can do to reduce that, without disassembling the romex and putting the hot and neutral together. It is geometry controlled.

If you wish to be accurate, the magnetic field caused by the amp draw is mostly cancelled, but that cancellation results in the 250 nH per foot number. And that inductance creates the field I'm talking about.


The definition of parallel I am using is the wires running side by side with the same seperation along the length. "they do not run parallel"???? You are kidding, right? Had I known you'd make a statement like that, I'da taken a simple picture...at the staples too, to show the clean work the mechanic did..




I won't even pretend I understand what is the "integral loop capture"

That is why you haven't understood a single word I've said here. You should have stated a while back that you do not understand the terms, that would have saved a lot of writing..all this time, you've been having a discussion on a subject that you are not understanding??

Faraday's law of induction defines the voltage generated around a loop as a result of the rate of change of the flux trapped by the loop. It is a simple integration of the total area enclosed by the loop. Ground loop currents are generated because the return leg of the loop is not the same wire as the start leg. As the return leg gets farther from the field generator, it is enclosing more and more flux from the field generator. Since the field is falling off, moving the return leg farther away asymptotically traps all the flux available, for a dipole generator such as romex, this end limit is 1/2 of the total flux of the romex.

You need to understand this field theory if you wish to discuss it reasonably. You are trying to apply incorrect concepts to the discussion.. If you wish, go to the hyperphysics site, they have some decent pages on the e/m stuff. If you do not understand how it applies to the discussion at hand, pm me with questions, I'll be happy to help you understand..

You stating a ground loop cannot be formed by the two romex grounds is entirely inconsistent with reality. You stating it cannot be.. does not change that reality...

Do yourself a favor. Measure the voltage between the grounds of two outlets on opposite sides of a room. It doesn't matter whether or not the outlets are being used, or in fact, if that circuit is being used. You will find a voltage between the grounds. WITHOUT any current being drawn.

You are measuring the voltage generated by the flux trapped within the loop you formed. It's not IR drop, there's no current through the #14 (minimum) to the panel..

Cheers, John You misunderstand me, I meant your use of the term "integral loop capture" besides not being standard terminology for flux density, is incorrect because I had already established that in your model no magnetic field exists to induce any current in in the romex ground. I was expressing my perplexity at your statement given what I had previously established.

Of course there is current on my circuit grounds, single ended telco lines like catv and sat couple with it, leaky power supplies couple with it, surge suppressors and filters that shunt to ground couple with it, the list goes on, flux trapping of the kind that you describe is very low on my list - and rightly so.

I’ll gladly PM you to continue this discussion but unless you concede that you are in err with respect the draw of the amp, we can't continue. You do not have a valid source of current for your model.

My apologies to the OP for taking this thread a bit off topic.

You misunderstand me, I meant your use of the term "integral loop capture" besides not being standard terminology for flux density.


Perhaps in your line of work..however, in my line of work, it is one of the most basic entities I work with.

"Integral loop capture" is a concept I MUST use to do my job. At this instant, I am using it to chase down 1 uVolt error signals on a 75 foot long cable which is five feet away from a 5 kiloamp 440 volt 12 phase power supply and a 2 tesla magnet. Given the complexity of the environment, it is absolutely impossible to calculate any kind of "common impedance", the term you are fond of.. I have to find and eliminate the loops which are seeing the flux within the environment (a hostile one).

Key to removing the inductive pickup is "visualizing" the loops and the flux, I don't recommend trying to use the equations, that is a lost cause.

...., is incorrect because I had already established that in your model no magnetic field exists to induce any current in in the romex ground. I was expressing my perplexity at your statement given what I had previously established.

No, you have only repeated what I have said all along, and misapplied it. I have never said that the ground of the powered romex develops a voltage as a result of flux from the hot and neutral. I said the romex NEXT to it forms a ground loop with the powered romex ground, and that loop is trapping the flux.

Go back and read my posts...you glossed by the important parts so continue to argue against something I have never said..

....,
I’ll gladly PM you to continue this discussion but unless you concede that you are in err with respect the draw of the amp, we can't continue. You do not have a valid source of current for your model.

You continue to miss the point. It's not the powered romex that develops the ground voltage, it's the ground loop that uses the powered romex ground as the return path.

Cheers, John

I'm lost by this thread by the way.

Sorry bout that. (pay not attention to the man behind that curtain...) :D

I was only asking, because I was troubleshooting a ground loop, and on one
side of the interconnects I have my 2-prong power cord pre-amp, and on
the otherside was my 3-prong power cord amp.

I figured that. Did you correct it, or is it still there?

ps...2 prong power cord? If it's zip cord, twist it so that it does a 360 every inch or two, that may help..and plug it into the same outlet the amp is using, or if your amp has it, the aux outlet of the amp.
John

Of course there is current on my circuit grounds, single ended telco lines like catv and sat couple with it, leaky power supplies couple with it, surge suppressors and filters that shunt to ground couple with it, the list goes on, flux trapping of the kind that you describe is very low on my list - and rightly so.

Hmmm...so lets see... I measure 1.5 volts ACrms between two points that are grounded to the panel with #12 wire....at 1.71 ohms per kilofoot..If I assume 100 feet of total wire, that's 170 milliohms.

1.5 volts/.17 ohms, 8.82 amps rms running through the ground.. Hmmm.

And where is this major current coming from? Leaks?


So when I do this same test on two outlets fed from the same breaker, daisied and on opposite sides of the room, with NOTHING plugged into either outlet, and get 3/4 volt rms...this is, leakage?? To a #12 wire, causing IR drop?? It's 25 feet outlet to outlet...42 milliohms..17.8 amps??? And nothing in the outlets?

Perhaps you need to modify your list.

Cheers, John

My apologies to the OP for taking this thread a bit off topic.

That's Ok. I moved on.

Sorry bout that. (pay not attention to the man behind that curtain...) :D



I figured that. Did you correct it, or is it still there?

ps...2 prong power cord? If it's zip cord, twist it so that it does a 360 every inch or two, that may help..and plug it into the same outlet the amp is using, or if your amp has it, the aux outlet of the amp.
John

The amp was the main cause of the hum. I tried a hum-x plug, it's like
a cheater plug, but doesn't remove the safety ground.

The main objective is to get rid of most of the noise, so that when I'm
listening to movies, I don't hear the hum from the surround speakers
when there's no audio.

Ahh its a ISO transformer it seems. And correct me if I am wrong but it does have all three prongs as well? Yes this should be safe.

Ahh its a ISO transformer it seems. And correct me if I am wrong but it does have all three prongs as well? Yes this should be safe.


I found it here:

http://www.ebtechaudio.com/humxdes.html

I don't think it's an iso transformer, it looks waay too small.

I was thinking it is perhaps an antiparallel set of diodes in the ground path. Perhaps even a bridge rectifier with the ground passing from AC to AC terminal, and Pos to Neg Shorted, giving two diode drops in either direction. Use a big enough bridge, and it should be capable of handling the fault currents in a typical 15 amp service..

I've had to use this type of ground break at work for low level stuff.

I can't tell if it's UL or CE approved, though..But it's a neat little device..

Cheers, John

I found it here:

http://www.ebtechaudio.com/humxdes.html

I don't think it's an iso transformer, it looks waay too small.

I was thinking it is perhaps an antiparallel set of diodes in the ground path. Perhaps even a bridge rectifier with the ground passing from AC to AC terminal, and Pos to Neg Shorted, giving two diode drops in either direction. Use a big enough bridge, and it should be capable of handling the fault currents in a typical 15 amp service..

I've had to use this type of ground break at work for low level stuff.

I can't tell if it's UL or CE approved, though..But it's a neat little device..

Cheers, JohnIts probably appropriately rated power diodes wired to only allow the ground to conduct under severe fault current conditions. Or possibly wired as rectifiers, as dc ground loop will not cause safety ground-referencing signals to modulate as ac ground loop would. Diodes can fail open though, possibly defeating the safety ground. Not an issue behind an isolation transformer of a power supply where they normally are located, but a majaor issue in the safety ground path in the humx implementation. Therefore this is not the safest of options. I'd leave the safety ground path unaltered. It’s reckless and probably against code to do otherwise.
I re-read your posts carefully, I still don't agree that any conductor of the nearby pre-amp branch circuit is asymmetrically exposed to the magnetic fields around the amp circuit conductors. The distances between the hot and neutral conductors in romex are too minute for any relatively nearby wire to see their respective fields as asymmetric. The distances necessary for what you hypothesize, only occur if the romex runs are bundled, which is not done because code requires you significantly de-rate such circuits.

Diodes can fail open though, possibly defeating the safety ground.

That depends on how the diodes are manufactured. The TO-247 style or TO-220 packages are typically dice soldered to the copper backplate, with wirebonds from the die top to the leads. Even the largest aluminum wirebond I've ultrasonically welded (20 mil) will probably fail open prior to clearing a 15 amp breaker. Some manu's use the copper leadframe formed to bring the side leads to the top of the die, and solder it. Even those, a fault can blow the silicon and "de-seal" the encapsulation epoxy, again...open. So yes, the bulk of diodes are not acceptable safetywise.

Diodes which are brazed directly to the moly or tungsten are not as bad, they will tend to blow short....however, that is also not a guarantee of safety, as the glass passivation around the edge of the die is not very substantial nowadays with the passivated/molded devices.

Bridges which are made with brazed construction axial lead diodes or brazed leadless slugs are the best candidates for groundloop elimination. When they are made into a bridge, they are encapsulated within a potting compound. If the compound is something like emerson and cuming 2851 KT, the compressive strength of the epoxy runs in the 17 kilo-psi range at room temp, so the unit will be more likely to retain structural integrity during a fault. It'll sure blow the die to failure, but more than likely, a short. This applies mainly to the aluminum case style of bridges, however.

I agree with your concerns, as I do not believe there are any specifications covering the failure mechanism or end result of a regular diode in general, so there can not be a guarantee of fail to short always..

I re-read your posts carefully, I still don't agree that any conductor of the nearby pre-amp branch circuit is asymmetrically exposed to the magnetic fields around the amp circuit conductors. The distances between the hot and neutral conductors in romex are too minute for any relatively nearby wire to see their respective fields as asymmetric. The distances necessary for what you hypothesize, only occur if the romex runs are bundled, which is not done because code requires you significantly de-rate such circuits.

If you try to measure the field around a romex under power with a multi-turn loop, for example, you can easily see the dipole nature of the field, the fact that it falls off with distance, and the general directivity of the field. Everybody agrees on these points. The point I am making, which is being ignored or misunderstood, is that two romex runs which go to a common panel is the condition which is guaranteed to trap the most flux in the ground loop.

I concur on your bundling point, that's never been a contention. However, it is common practice to use one staple to hold two runs of romex to a joist, as it is stapling them side by side. It is that parallel config that is the worst condition.

Remember, I'm not talking about a bowl of spagetti in a conduit, this is absolutely non-random conductor placement. Imagine a mike cord where the signal pair doesn't twist.

Cheers, John

That depends on how the diodes are manufactured. The TO-247 style or TO-220 packages are typically dice soldered to the copper backplate, with wirebonds from the die top to the leads. Even the largest aluminum wirebond I've ultrasonically welded (20 mil) will probably fail open prior to clearing a 15 amp breaker. Some manu's use the copper leadframe formed to bring the side leads to the top of the die, and solder it. Even those, a fault can blow the silicon and "de-seal" the encapsulation epoxy, again...open. So yes, the bulk of diodes are not acceptable safetywise.

Diodes which are brazed directly to the moly or tungsten are not as bad, they will tend to blow short....however, that is also not a guarantee of safety, as the glass passivation around the edge of the die is not very substantial nowadays with the passivated/molded devices.

Bridges which are made with brazed construction axial lead diodes or brazed leadless slugs are the best candidates for groundloop elimination. When they are made into a bridge, they are encapsulated within a potting compound. If the compound is something like emerson and cuming 2851 KT, the compressive strength of the epoxy runs in the 17 kilo-psi range at room temp, so the unit will be more likely to retain structural integrity during a fault. It'll sure blow the die to failure, but more than likely, a short. This applies mainly to the aluminum case style of bridges, however.

I agree with your concerns, as I do not believe there are any specifications covering the failure mechanism or end result of a regular diode in general, so there can not be a guarantee of fail to short always..
showing off again. Hopefully this'll hold up to scrutiny


If you try to measure the field around a romex under power with a multi-turn loop, for example, you can easily see the dipole nature of the field, the fact that it falls off with distance, and the general directivity of the field. Everybody agrees on these points. The point I am making, which is being ignored or misunderstood, is that two romex runs which go to a common panel is the condition which is guaranteed to trap the most flux in the ground loop.

I concur on your bundling point, that's never been a contention. However, it is common practice to use one staple to hold two runs of romex to a joist, as it is stapling them side by side. It is that parallel config that is the worst condition.

Remember, I'm not talking about a bowl of spagetti in a conduit, this is absolutely non-random conductor placement. Imagine a mike cord where the signal pair doesn't twist.

Cheers, John Field symmetry of the romex under load is exactly what is under contention, I disagree that it is possible for the field to be asymmetric with constant polarity for any given length with respect to any parallel run as you have defined it. This assumes a uniformity among romex runs that can never really occur at distances not normal for the reasons I previously stated. Randomness is the norm, the net result of which is little or no coupling between branch circuits. Mic cable is twisted to maximize CMR for low voltage runs, not necessary when you’re drawing tens of amps at hundreds volts such as it is in the world of ac mains, as if mu volts of differential emi noise affects power supplies.
The common ground at the panel is not sufficient in it self to induce ground loop.

showing off again. Hopefully this'll hold up to scrutiny .

:p . Actually, it's from a past life..so it's not like I'm smart or anything, but rather just what I remember..

Try as I could, I was unable to find a link to the super-rectifier passivation scheme that General Instrument patented, so could not provide a detailed construction picture of it nor how weak it was. The patent expired back in '93 I believe, so I guess it's not something they hawk anymore. I could scan my old databook if anybody were interested..

I just remembered the name of the slug diodes, they called em "cells". The company I worked for didn't use cells, but regular axial diodes for the mil bridge product..

The 25 amp single phase mil bridges had 4 regular 3 amp glass diodes internally. With adequate internal heat spreaders of copper, it is easy to get four 3 amp diodes to work up to 25 amps full wave, as the 3 amp limit is a thermal one anyway. Since the diodes were oriented horizontally, the only way for the silicon to blow up and push the moly's away (open circuit) would be to deform the aluminum case outward. That won't happen. Cells are oriented so that when they blow, it is towards the epoxy surface.

Field symmetry of the romex under load is exactly what is under contention, I disagree that it is possible for the field to be asymmetric with constant polarity for any given length with respect to any parallel run as you have defined it..

No, field symmetry is not in contention...we both agree on that.

The assymetry which traps flux is the other wire's fault. The primary wire (the powered one) creates the field. The only thing required is a pickup loop which is assymetrical to the primary's magnetic field. Asymmetry causes a net non zero flux capture. A second romex that is not stapled on top of the first, introduces that asymmetry.

Mic cable is twisted to maximize CMR for low voltage runs.

I know..

, not necessary when you’re drawing tens of amps at hundreds volts such as it is in the world of ac mains, as if mu volts of differential emi noise affects power supplies..

Nobody said it affects power supplies...to whom are you arguing on this point?


The common ground at the panel is not sufficient in it self to induce ground loop.

Nobody has said that either. Specially me...

The fact is, a wire travelling through ANY time varying magnetic field whatsoever, never, ever, has a voltage potential from one end to the other...(note: this is as long as you reference the ends though the wire itself, for example, a coaxial construction.)

It's when you try to measure it to confirm it, that you form a loop. Then, Faraday's law kicks in. (note: for the coax example, the inner conductor does not trap any flux with respect to the outer shield, so no voltage.. if you form a loop/gap combo with the coax, then a potential will exist across the gap, as that potential path completes the loop.)

Same here. Any one ground wire connecting to the panel, does NOT form a voltage, nor a loop. It's the SECOND one that does it. That, and the interconnect between two "appliances" at the far end of two romex runs. Then, you live at the whim of the trapped flux..

A ground loop is called a "loop" for a very specific reason. You are still trying to explain how a romex can't cause voltage on it's own ground wire (yet nobody has said such), and completely ignoring what I've been talking about all along....the LOOP.

Cheers, John

ps...the diode stuff will certainly hold up to scrutiny..if you really wanna call me on it, I can give you names of the experts in the industry to ask. I'm sure some of them are still alive and working..well, least I think so... :o

Modesty (sure you’re not smart) and magnanimity (I appreciated the dumbing down) all displayed within your last post. I prefer the show off. I've learned quite a lot actually. I'm not ignoring the loop, what I'm disputing is the source of your stray flux, it cannot be the amp draw. I think you were checkmated a while back, now in your obstinacy, you’re adding new squares to the board by changing the lay out of your romex runs to explain conductor exposure to asymmetric (dipole) fields which due to the standard distances and randomness of inter romex spacing and the geometry of romex conductors itself, is highly unlikely.
If I may humbly suggest that the source of your ground loop current simply be stray return current from unbalanced signals coupling with circuits with a far greater ground reference.

Modesty (sure you’re not smart) and magnanimity (I appreciated the dumbing down) all displayed within your last post.

Neither is by design. More like a drunkards walk.. ;)

I'm not ignoring the loop, what I'm disputing is the source of your stray flux, it cannot be the amp draw. I think you were checkmated a while back, now in your obstinacy, you’re adding new squares to the board by changing the lay out of your romex runs to explain conductor exposure to asymmetric (dipole) fields which due to the standard distances and randomness of inter romex conductor spacing and the geometry of romex conductors itself, is highly unlikely.

Umm, where to start...

first, obstinance...again, it's a drunkards walk...

All romex has an external AC field. The question is, what configuration can trap some of it and produce a voltage.

I'm not adding new squares. My explanation has been consistent throughout. You just have not understood.

I believe the fault most likely falls to both of us. I am not explaining well, and you are unable to follow what I haven't stated clearly.

If I may humbly suggest that the source of your ground loop current simply be stray return current from unbalanced signals coupling with circuits with a far greater ground reference.

What is a far greater ground reference?

Mental exercise...

On the stage, there is an outlet. In the balcony, another one.

I run an unbalanced line from the balcony to the stage, connecting amp to mixer. It hums. I check with a voltmeter, and find 2.5 volts ac between the ic and the mixer ground.. Sheesh, no wonder...so, I leave the IC connected at the stage, check from the ic shield to the outlet ground in the balcony. again, 2.5 vac. Where is this coming from??? Easy. The magnetic environment is causing an induced voltage to appear around the loop formed by the IC and the building ground.

Normal solution: run balanced lines, let the cmmr of the balanced set fix the problem..

Problem? Yup....I ain't normal. :p (actually, I didn't have enough balanced cable at the time, I had about 10 minutes till showtime on a sunday morning, and live 30 minutes away...) :eek:

So, How to get around this magnetically induced voltage? Simple, get rid of the loop.

So I ran a 100 foot 3 prong extension cord from the amp rack to the mixer. Then I ran the unbalanced line feed with the cord, spiralling it around the extension cord.

Then, I measured the potential between the IC shield and the extension cord ground. ZERO volts. Then, ran a coupla amps through the cord (200 watts light bulbs), remeasured the ic ground to cord ground...still, a big fat, really sweet....ZERO volts. Major sweet...

Plugged mixer, ran the system...never a problem, 7 years now...

Reason it woiked? I prevented any magnetic field in the auditorium from getting between the IC and the extension cord ground.. Dead silencio....

Even with the 250Kilovolt vandegraf on stage, 5 feet from the amp.

I never have to worry about any AC currents or magnetic fields in the auditorium, because I do not trap them between the safety ground and the signal wires.

Here's a pic, not the best cause the coloration of the loops are not that clear.


arrrr...pic too wide..waittaminute... :eek:

ok...the orange loop area is defined by the two grounds of the two romex runs. The pink area is at the equipment. The total loop voltage is the sum of the voltage induced by the flux in the orange area plus that in the pink area. (sorry for the colors.).

Not only is the voltage caused by the flux of the amp being trapped, anything else in the vicinity that creates a magnetic field can induce a voltage in that loop. To reduce the induced voltage, either get rid of the magnetic fields, or get rid of the loop. My solution, by wrapping the ic around the line cord, gets rid of the loop. In this picture, envision that by pulling the IC by it's name, to the left until it hits the panel box. While that puts the IC closer to the ac lines, it reduces the orange loop area. Though counterintuitive, it works.

Cheers, John

Neither is by design. More like a drunkards walk.. ;)


What is a far greater ground reference?


My apologies Prof. Neutron for the use of incorrect terminology.
How about branch circuit ground references at different potentials.

Or perhaps you meant actual numbers (Ohms), of which I'm unable to supply you at this time as I don't have any resistance figures for typical fixtures and devices such as metal boxes and self grounding receptacles that are normally in series with the branch circuit ground..

My apologies Prof. Neutron for the use of incorrect terminology.
How about branch circuit ground references at different potentials...

Hey, a promotion...cool..does that mean I get the babes also?? :)

Ground references can only be at different potentials because of two reasons.

1. IR drop.

2. Hostile magnetic field environment.

I believe we both concur that IR drop is not the prime cause of 60 hz ac potential difference for normal lengths of #12 wires. The insulation would go first for the voltages typically measured.

Leaving the trapping of magnetic field flux. Which requires absolutely no current to develop a voltage potential.


Or perhaps you meant actual numbers (Ohms), of which I'm unable to supply you at this time as I don't have any resistance figures for typical fixtures and devices such as metal boxes and self grounding receptacles that are normally in series with the branch circuit ground..

No, resistance is not a factor here.

Cheers, John

I run an unbalanced line from the balcony to the stage, connecting amp to mixer. It hums. I check with a voltmeter, and find 2.5 volts ac between the ic and the mixer ground.. Sheesh, no wonder...so, I leave the IC connected at the stage, check from the ic shield to the outlet ground in the balcony. again, 2.5 vac. Where is this coming from??? Easy. The magnetic environment is causing an induced voltage to appear around the loop formed by the IC and the building ground.

Normal solution: run balanced lines, let the cmmr of the balanced set fix the problem..



All you did was make one path to ground by running it all off the same circuit. Its the right thing to do. You had a different potential to ground before on more than likely different circuits.

The normal solution is to do what you did with the power. And its a good guess that you wouldnt need to wrap the cable anymore. Again your the only person I have ever seen do something like this.It seems you are making a mountain out of a mole hill.

All you did was make one path to ground by running it all off the same circuit. Its the right thing to do. You had a different potential to ground before on more than likely different circuits.

It's not that easy.

See, the first thing I did was run the powercord and the IC's as a test. While the hum reduced, it was not gone. In addition, the van de graf made some major noise when it was on. (typical rise times of the sparks is in the sub microsecond range). There is a 50 foot or so area where I have to dangle the cord and IC, and I hadn't spiralled the two initially. As a result, there was still a rather large loop area between the two, and consequently, it picked up noise. (I was afraid to pull the IC as tight as the powercord, it's far more delicate.)

So, I dropped the cords down, twisted them together, re-hung them, and poof. no runs, no drips, no errors...case closed. After that day, I put a nylon braid tube over the whole kittencaboodle. Now, I can run this combo anywhere, even next to a 5000 amp/150 volt power supply...no hum..


The normal solution is to do what you did with the power. And its a good guess that you wouldnt need to wrap the cable anymore. Again your the only person I have ever seen do something like this.It seems you are making a mountain out of a mole hill.

As I said, the powercord was almost, but not quite, enough..I still had to wrap the IC around it to close the loop.

As for mountain/molehill..I solved a problem by figuring out the source of the noise, and eliminated it. simple, done forever.

Cheers, John

All you did was make one path to ground by running it all off the same circuit. Its the right thing to do. You had a different potential to ground before on more than likely different circuits.

The normal solution is to do what you did with the power. And its a good guess that you wouldnt need to wrap the cable anymore. Again your the only person I have ever seen do something like this.It seems you are making a mountain out of a mole hill.
My thoughts exactly.

My thoughts exactly.


Don't lose sight of the ball, guys. The premise that the hum is being caused by a difference in ground potential is only plausible when you think about a large venue where you can attribute the hum to some unknown, untestable entity. You can claim three phase imbalance causing neutral drops, light dimmers, capacitance, any host of things that may or may not be real.. I go for the real problem, loop pickup.

In my case, with only one IC set and one cord, with trivially small power requirements at the source rack, it's easy. What I did is perfect. For large venues, it is not possible to do.

When somebody at home plugs an amp into the top outlet of the duplex, and the pre in the bottom outlet of the same duplex, YOU two are claiming by default that there cannot be any hum, EVER.

Half the human race can tell you otherwise from experience.

I point out a method for eliminating even that hum. It's not rocket science..


Cheers, John

When somebody at home plugs an amp into the top outlet of the duplex, and the pre in the bottom outlet of the same duplex, YOU two are claiming by default that there cannot be any hum, EVER.



No that is not what I am claiming. I am claiming in this case it was all you needed to do.It has nothing to do with venue size in the least.

I could list all the causes for loops but these guys did it for me.

http://www.gbaudio.co.uk/data/ground.htm

No that is not what I am claiming. I am claiming in this case it was all you needed to do.It has nothing to do with venue size in the least.



Then you are making an inaccurate and glib claim which is not grounded in fact.

Inaccurate because I first tried the separate cord without regard to the loop formed between it and the unbalanced IC from the stage to the balcony. Most of the noise and hum disappeared, but not all.

Inaccurate because removal of that loop via a simple twisting together of the two wires solved the final portion of the hum and noise pickup. Zero hum and noise.

Glib because you have made a claim without knowing all the facts.

Lest you feel these statements are too harsh, go back a couple of posts and re-read what I said..

This: ""I believe the fault most likely falls to both of us. I am not explaining well, and you are unable to follow what I haven't stated clearly.""

Note I said you are unable to follow what I haven't stated clearly. I blame the bulk of our disagreement on the fact that I haven't been as clear as was needed for the discussion.

I could list all the causes for loops but these guys did it for me.

http://www.gbaudio.co.uk/data/ground.htm

When I get a chance, I'll read this..thank you.

Given what I've seen over the years in the professional audio arena, I have serious doubts it will be complete or entirely accurate...however, if it is at the level I hope it to be, I will certainly have no problem eating my "preconception" for lunch (on forum), as well as learning more myself (a bonus)..

Thanks.

Cheers, John

I found it here:

http://www.ebtechaudio.com/humxdes.html

I don't think it's an iso transformer, it looks waay too small.

I was thinking it is perhaps an antiparallel set of diodes in the ground path. Perhaps even a bridge rectifier with the ground passing from AC to AC terminal, and Pos to Neg Shorted, giving two diode drops in either direction. Use a big enough bridge, and it should be capable of handling the fault currents in a typical 15 amp service..

I've had to use this type of ground break at work for low level stuff.

I can't tell if it's UL or CE approved, though..But it's a neat little device..

Cheers, John
I use them and they work in most cases (remove the hum). I can't tell any downsides.

Ok, so I took a look at the link provided by speco2003

Not bad. Didn't go through it all, but it's pretty good.

Unfortunately, it only glosses on what I speak of, a casual reference, no meat, no potatoes..

First .

""If there is any current folowing in any wires, there is then some potentital difference which causes current to flow in other wires also which causes problems. The loop will also act as coil and pick current from the changing magnetic fields around it. Wire loop acts also like an antenna picking up radio signals. ""

First part, IR drop...consistent..been there, done that..

Second part (I bolded it) is the crux of what I've been detailing all along.. Unfortunately, not much is said about reducing it, it has to be implied by the reader...most will not do so. However, I have been saying it all along.

Second :

""Even a very small induced voltage can cause a very large current in a ground conductor loop, because the resistance (and inductance) are very low. These currents can indeed be tens of amps. Current induction can be caused for example by cables carrying high currents and from transformers. ""

Yup, been there also, and done that also. And provided an example of the level of voltages maximum one can expect from external loops near a romex cable 100 feet long carrying 10 amperes AC.

It's also why NEC requires all three phases go through the same tube at a service panel, as the resulting eddies at the penetration are capable of turning the panel metal orange (and it ain't paint).

It's also why a copier in one wing of my building died. A three phase 500mcm run feeding a medium size scr based 12 phase power supply impressed (by induction) 150 volt spikes on the neutral as well as the supporting cable tray's, that made it into the distribution system of the building.

It is specifically the induction component that I removed by twisting the IC around the line cord.

The next quote is also applicable. I note that all of us concur as to it's utility, however I find that it is insufficient to cover all I need. For youze guys, with balanced systems, it is all you need to do your job.

""If you have equipments connected to many power outlets, then try to connect them all to one grounded extension cord and then plug this cord to one grounded outlet. This solves usually many problems, because this creates star-like grounding scheme for those equipments. ""

As I noted, I used the line cord to immediately form a star ground in the system, but found that was insufficient.

Star grounds are pretty good, but they are only a stop-gap. Someday I'll detail where star grounds are weak and fail to meet my needs.

Cheers, John

Jneutron

I have no background, knowledge, etc. of what you are talking about but I do think I know "tone" when I see it, hear it, etc. and your posts to me (as in my opinion)come off as very competitive, as if you were at a science debate where one side wins. You also are very dimissive of others and their chosen profession. I do think someone asked your profession but I did not see the answer (but may have missed it). I am not trying to start a fight because I am sure with your probable IQ you would cream me. Just trying to maybe tone down the competitive remarks (which then brings out all the "competitiors")

Jneutron

I have no background, knowledge, etc. of what you are talking about but I do think I know "tone" when I see it, hear it, etc. and your posts to me (as in my opinion)come off as very competitive, as if you were at a science debate where one side wins. You also are very dimissive of others and their chosen profession. I do think someone asked your profession but I did not see the answer (but may have missed it). I am not trying to start a fight because I am sure with your probable IQ you would cream me. Just trying to maybe tone down the competitive remarks (which then brings out all the "competitiors")

1. "competitive"...hmm, maybe..

2. "one side wins"... no, incorrect. when accuracy prevails, everybody wins.

3. "dismissive of others".. absolutely not.

4. "and their chosen profession". again, absolutely not.

5. superconducting magnets.

6. "not trying to start a fight"...why even say that??? You are pointing out how my posts are being perceived by you (and perhaps others)...nothing more, nothing less. I thank you for your statements, and certainly consider your words in the spirit you present them.

7. "probable IQ"...yah, right. certainly fooled you... :p (note: I believe you are confusing "intelligence" with an understanding of a topic. I love this e/m garbage, it is so easy for me to understand...but an intelligent person wouldn't staple his finger, now would he??)

8. "maybe tone down the competitive remarks". competitive is not by design. Forums are very tough to convey attitude, as facial and body expression is so important to discussion. But yes, since you point out how you perceive my attitude, I will certainly reflect on it..

Thank you

Cheers, John

Someday I'll detail where star grounds are weak and fail to meet my needs.

Cheers, John

Please do. And present it at AES. The rest of the audio industry awaits as you show them that they have been doing it wrong for years. You always say us balanced guys. Sorry but I have to interface everything from the best balanced gear to the cheapest unbalanced everyday, as do many others and NONE of us have ever gone to your extremes to make them work.And if you read the complete article it deals with unbalanced gear.It is interesting that every article presented to you by real pros in the field is quickly dismissed out of hand by you. I am done with you and this thread.I tried to be nice and not get into my azzhole mode with you, but I feel it coming on.

For the ORIGINAL poster and anyone else who cares here are 2 great refs for interconnects and grounding.The first one covers most interconnects you would ever come into.


http://rane.com/note110.html

http://rane.com/note151.html

Please do. And present it at AES. The rest of the audio industry awaits as you show them that they have been doing it wrong for years. You always say us balanced guys. Sorry but I have to interface everything from the best balanced gear to the cheapest unbalanced everyday, as do many others and NONE of us have ever gone to your extremes to make them work.And if you read the complete article it deals with unbalanced gear.It is interesting that every article presented to you by real pros in the field is quickly dismissed out of hand by you. I am done with you and this thread.I tried to be nice and not get into my azzhole mode with you, but I feel it coming on.

You tried to be nice???

Lets review...this is the very very first thing you said to me...

Then that angineer is an idiot because........

Then, in the same post, this gem.....

Your just lucky your stuff works, the other guy is an idiot .......

I have been quite tolerant of your piss poor attitude from the jump.

Pointing out your errors is not "dismissing" you, nor is it dismissing your experience.

Pointing out a weak area of an otherwise good webpage is also not dismissing.

While you glibly call my actions "dismissing" of the whole rest of the planet, it is really just your reaction to having what you believe questioned. The people who write those articles would not do what you are now doing. They would listen.

If you wish to learn, let me know. But keep your silly attitude and your glib remarks to yourself.

Cheers, John

From IEEE STD 1050-2004 , IEEE Guide for Instrumentation and Control Equipment Grounding in Generating Stations
http://ieeexplore.ieee.org/iel5/10160/32470/01515972.pdf?arnumber=1515972

Radial routing of I&C cable. Circuits should not be looped with a single conductor from one piece of equipment to another with the return conductor in another cable. Both supply and return conductors should be in a common cable to minimize loop area and avoid the large magnetic induction possible using separately routed single conductors. This means that both secondary leads of CT’s should be in the same cable, both positive and negative dc leads should be in the same cable, and all three phases and neutral of voltage transformer (VT) secondary leads should be in the same cable.
Any grounding leads associated with any circuit must also be tightly routed with the circuit conductors and also must be carried inside of the same shield, conduit, or other raceway with the associated circuit conductors. This is necessary to keep the impedance of these circuits as low as possible by permitting close interaction of the opposing magnetic fields occurring on them. If the supply and return signal lines are discrete wires (as opposed to being part of a factory produced multi-conductor cable), they should be laid as close to each other as possible within the same raceway to minimize the loop area and reduce susceptibility to interference from inductive coupling.
The greatest practical amount of symmetrical twisting per unit of length of all signal supply and return conductors along with any dedicated grounding conductor, is desirable. Twisting of conductors during installation must be commensurate with avoiding damage to the conductor’s insulation.
On power circuits, such tight twisting is not normally feasible, however, the most practical twisting should be applied. Do not twist a number of otherwise untwisted pairs of wires going to different circuits together as they can create unwanted crosstalk and a higher impedance for the DM currents in any related pair of signal conductors in such a bundle. However, gently twisting together already twisted signal pairs to form a larger cable is generally advisable if done carefully. There are established rules for doing this such as those promulgated by telecommunications companies, and these rules should be followed. Approximately 1 turn/meter is a generally useful amount of twist to use on multi-pair cables and on power wiring.

Not directly audio, but still the same loop characteristics of long wires encompassing magnetic fields. This also talks about the voltage differences in widely separated earth grounds.

I believe this is what JNeutron has reduced with his twisting. This is a very interesting conversation....

Not directly audio, but still the same loop characteristics of long wires encompassing magnetic fields. This also talks about the voltage differences in widely separated earth grounds.

I believe this is what JNeutron has reduced with his twisting. This is a very interesting conversation....

You are good. Wow, that is an excellent writeup... What is a CT?

ps..on second thought, I think I may just print the whole 101 pages out and put em in a notebook. Excellent link...thanks.

Cheers, John

I think it's interesting too and while I don't have any hum problems, I personally have found some of the suggestions and reasoning interesting. I hope that if someone posts that they have a problem that all who presently 'think' they're combatants take it upon themselves to help the poster out. You see, I think much of which appears to be highly confrontational comes from the medium (posting on a discussion site) where we're all given way too much time to think. We don't have the benefit of picking up on voice inflections, mannerisms, and the instantaneous back and forth that an actual conversation would permit. Hence, what may seem confrontational or a tone may be nothing more than one's style of writing. To my way of thinking anyways, having a more complete picture of things does not mean that one side is wrong. It's just incomplete. Maybe it's incomplete because the solutions offered work in say 90% of the cases, but it's nice to know the other 10% (arbitrary numbers on my part). Frankly I see speco, who I like (not in the biblical sense mind you!) and JNeutron on the same side of things both with practical and valid experiences in somewhat similar as well as quite divergent environments leading to solutions that have parallel and entirely unique solutions.
I personally would be loathe to to lose the experiences of all participants and suggest in the spirit of friendship and common goals that we realize what's important.

http://www.drsusanblock.com/eros/NudeBeer2.gif

1. "competitive"...hmm, maybe..

2. "one side wins"... no, incorrect. when accuracy prevails, everybody wins.

3. "dismissive of others".. absolutely not.

4. "and their chosen profession". again, absolutely not.

5. superconducting magnets.

6. "not trying to start a fight"...why even say that??? You are pointing out how my posts are being perceived by you (and perhaps others)...nothing more, nothing less. I thank you for your statements, and certainly consider your words in the spirit you present them.

7. "probable IQ"...yah, right. certainly fooled you... :p (note: I believe you are confusing "intelligence" with an understanding of a topic. I love this e/m garbage, it is so easy for me to understand...but an intelligent person wouldn't staple his finger, now would he??)

8. "maybe tone down the competitive remarks". competitive is not by design. Forums are very tough to convey attitude, as facial and body expression is so important to discussion. But yes, since you point out how you perceive my attitude, I will certainly reflect on it..

Thank you

Cheers, JohnProbably just my bad interpretation then, carry on.

Probably just my bad interpretation then, carry on.

That's not what I said..it can easily be my writing style..but not my intent.

You were nice enough to state how you interpreted things, that was admirable.

Speco was indeed a bad boy from the jump..

Chu: well stated, I kinda said the same, (my #8), but clearly you were far more elequent..

btw, the pic didn't show up...

Cheers, John

btw, the pic didn't show up...

Cheers, JohnIt didn't show up for me either, does that mean my sever at work is preventing it.....which probably means it is a good one. :D

It didn't show up for me either, does that mean my sever at work is preventing it.....which probably means it is a good one. :D

That was also my fear....darn it!! :mad:

Cheers, John

Really? How bizarre! I see it just fine. http://www.drsusanblock.com/eros/NudeBeer2.gif

Really? How bizarre! I see it just fine. http://www.drsusanblock.com/eros/NudeBeer2.gif
Well, I will have to go home and look at it tonight as my company did block it :(

Wasn't all that good really. I just know how difficult it is to communicate fully the importance of what one has to share.

http://www.krug.com.by/images/photo-brasil/brasil-beach.jpg

Wasn't all that good really. I just know how difficult it is to communicate fully the importance of what one has to share.

http://www.krug.com.by/images/photo-brasil/brasil-beach.jpg
Then again it wasn't all that bad either :D

Had a chance to study the IEEE paper....here's the relevant parts..

4.2

Both capacitive and inductive coupling refer primarily to near-field coupling. This is also often referred to as the reactive field where the energy is stored or as stray (parasitic) reactive coupling. A control circuit or cable is considered to be in the near field of an electromagnetic source when the source to circuit distance is
less than l/6 λ of the highest source frequency.

4.2.1

In the reactive or near field, where the energy is stored, the E-field and H-field can be considered independently where the wave impedance does not apply. E-fields are generated by and most easily interact with high-impedance, voltage-driven circuitry, such as a straight wire or dipole. E-fields are most readily coupled by capacitor action. H-fields are generated by and
most readily interact with low-impedance, current-driven circuitry, such as a wire loop. H-fields are most readily coupled by transformer action.


Note that section 4.2.2 refers to "common impedance"..this is not what I've been talking about.

Heres the meat...there is one writeup issue, I'll explain after the quote..

4.2.4 Inductive coupling
This type of coupling is also known as magnetic coupling. The various circuits of any system ultimately exist as closed loops. These loops have mutual inductances that are directly proportional to the area enclosed by the loops as shown in Figure 5. Interaction between the loops is essentially a transformer action between the aggressor interference source and the sensitive victim circuit. Even dc circuits produce a strongly changing magnetic field when their current is interrupted.

When a current change occurs in one of these circuits, a changing electromagnetic field through the area of its loop is produced. A voltage will be induced when some of this magnetic flux passes through a second circuit. The amplitude of the induced voltage is directly proportional to the area of the second circuit which
encloses the flux from the disturbing circuit. The induced voltage is determined from Equation (4):
E = M di/dt (4)
where:
E is the induced voltage in the second circuit,
M is the mutual inductance (amount of flux),
di/dt is the current change rate in the first circuit.
For magnetic coupling, the mutual inductance is a direct function of the coupled length of the conductors and an inverse function of the distance between conductors. Low-impedance circuits are more susceptible to
inductive coupling. Coupling varies according to the cube in the very near field, and then quickly follows classical square-law thereafter.
Both capacitive and inductive coupling are functions of the time derivative or rate of change of the source field (dφ/dt). Therefore, the interference coupling factor for a fixed coupling loop geometry increases with the higher frequency content of the transient current in the aggressor loop.

(I love that term...aggressor loop...)and victim loop... :p

They detail how the coupling is an inverse function of the distance....that is correct when the whole "victim" loop is moving away from source...

I am talking about the case where ONE of the "victim" loop wires is actually the center ground conductor of the agressor loop, that be the romex. In that case, moving the other victim loop wire away from the agressor loop makes the loop bigger. Wrong way..that is why I twisted the victim wire around the agressor, that lowers the loop.

My suggestion to the OP about twisting the agressor was to kill the net magnetic field.

Heres a writeup that I do have an issue with: it is in 4.3.3, shielding:

The use of a shield to attenuate H-field interference (a near-field phenomena) on victim conductors in a cable or similar circuit depends upon the shield operating in a fashion other than warping the magnetic flux away from the victim conductors or other component to be protected. In this second case, the magnetic flux lines are allowed to penetrate the enclosing shield and to reach into the victim conductors without much (if any) attenuation. This is achieved by grounding the shield at both ends, with the signal conductors contained inside the shield ground referenced to the same ground point as the shield at each end. The shield will then carry an induced current from the aggressor H-field, and the contained victim conductors will also carry an induced current from the aggressor H-field, with all of the induced currents being 180° out of phase with the aggressor current. Since the impedance of the shield’s closed loop (it is a larger mass of metal that is also
grounded at both ends) is much lower than that of the signal conductors contained within (they are smaller and feed into circuitry), this allows more induced current flow to occur in the shield than in the signal conductors
contained therein, so the shield’s current also develops its own magnetic field proportional to the current flowing in it, and this is also closely coupled (it is an H-field) to the victim signal conductors along with the originally induced current from the aggressor source. Therefore, the shield’s magnetic field now
induces its own current into the victim conductors with another 180° phase, thereby producing two current flows in the victim conductors that are out-of-phase with one another. These two induced currents cancel each other out in the victim conductors, and what is left is the greatly attenuated “noise” on the victim circuit, which cannot be removed without involving efforts with a diminishing rate of return for expended effort.



A cylindrical sheet of current, such as that which travels in a shield, does NOT create a field inside the cylinder. That is a consequence of maxwell's equations. So, the shield current can only cause induced voltage between the signal line and ground, meaning the geometry of the coax to ground factors in while that is not the same factor which generated the shield current or signal current. In fact, they consider the signal "current", but faraday's law of induction specifies the signal voltage, the current is dependent on the loop impedance of the signal line. That could be 10 kohms..

A cylindrical shield configuration is exactly what I used to shield 1 microvolt signals from some 1 Mhz pulses at 20 Khz rep rate ground currents...I made the currents flow through the external, cylindrical shield. Without the cylindrical shield to carry the error current, the induced voltage was over 500 millivolts. With it, the signal lines do not see any magnetic field caused by the current.

So, this writeup is not exactly fully consistent with e/m theory.

Here's two pics depicting the magnetic field intensity of a cylindrical shield carrying current.
Notes:
1. top pic.. Red is highest intensity clockwise field direction, blue lowest cw...yellow is highest ccw.green lowest ccw. Black is no field. Bot pic, red high cw, blue high ccw. Sorry I changed scaling...

2. You can see some bleed into the center near the wires. This is caused by the texture of the model, the small circles depict wires. Finer wires and finer shield reduces the texturing bleed. A pipe eliminates this entirely.

3. I didn't run the full area because what is there is several hours on a pentium-500.

4. The analytical solutions agree entirely with this computer model. But this site doesn't support equations, and quite honestly, I don't either... ;)

Cheers, John
ps...I like Chu's pic better...