I'm overseas waiting for some equipment that is weeks late and I've finally run out of things to do, so I'm looking up stuff. I found my -- probably our -- concept of ground loops is not quite right. This is from a T&B White Paper on grounding and bonding a STP system, [Link: ampnetconnect.com].

I have to start by pointing out what they say here about grounding and bonding.  Grounding is connection to the actual earth ground, or to the structural steel of a building.  Bonding is connecting together the chassis and shields of equipment so that there is a low resistance path to the ground.  That is, all of our cable shields, RCA plugs, F connectors, etc do not ground the cable but rather bond it to the bonding system that is connected to the ground.  So when they say "bonding" below, they mean what we usually mean when we say "grounding."


Guess what: it's not a ground loop if it is only a ground path that you can define as a loop because of what is connected to what; it is only a ground loop if there is a problem! A ground loop is not a wiring layout, it is the name given to a current under certain circumstances!


Now that we know that it's only called a ground loop if there is a problem, that totally makes sense.  [The article did not previously list a misconception with shielded cable.]


Now that's sloppy writing and perhaps sloppy thinking.  It should be "the voltage difference... IS more than 1 Volt."

This was illustrated at the Middle Atlantic grounding seminar of a couple years ago, but they used ground current as the indicator.  You can measure milliamps of current from location to location much more easily than you can measure small voltages between loactions.


Again, confirmed at that seminar.  They showed cables that had ground resistance, that is, poor bonding, created hum bars in a TV while cables with very low resistance did not.  The trick, too, was that the noisy cable was expensive while the quiet one was a piece of crap.


I love this one.  I once moved a ridiculously complicated set of two A/V systems for a guy.  He was in an apartment where they were on opposite sides of a wall.  He them moved to where they were thirty feet apart.  Halfway through the interconnection, I gave him TV to watch, and it had horrible hum bars.  He wanted me to fix it.  I demurred, offering to fix it when all the wiring was done.  When all the wiring was done, with no special effort on my part, the hum bars were gone, too.  The whole thing had to be in to be tested.


This is really hard to test.  AC current on the shield is good enough, and 10 mA is usually a problem.  The thing is, you'll need to spend $250 to get a current meter that can read that small of a current because you have to use a clip-on type.


I was shocked -- shocked, I say! -- to find out that if an actual loop circuit of ground wiring has no problem, then it is not a ground loop.  It makes sense, though -- we only notice a loop in the ground when we see a problem.

"Ran across some interesting things about ground loops" 


8 words that have never been assembled together in the English language until Ernie got a hold of them.

Ernie so if they bond really well do they all hold hands and sing Kumbaya? :) For some of us mere mortals this is the only way that we can be on the same level as you! So I hope you don't mind a little well intended poking and get your minds out of the gutter to everyone else not that kind! Most of what you post effects me the same way when my brother starts on gravity or quantum mechanics it makes my head hurt!

Ernie you are a treasure and thank you for sharing it with us.

Marc

Here, this can be helpful when trying to decipher Ernie's posts when he starts getting all technical

[Link: sites.google.com]

So,basically,the only solution for eliminating ground loop issues is the one that works?

Singing Kumbaya isn't necessary. Think of it as an "electronic group hug".

Another huge mis-conception in thinking that "star grounding" is using one screw to attach all grounds to the same point. It's not a physical topology, it means "at the same electrical potential". This is important- running long ground wires so they can attach at the same point is a good way to pick up noises that wouldn't have been picked up before.

For those who may have forgotten, "electrical potential" or "difference in potential" is voltage (voltage is potential energy) and 'equipotential' means there's no difference in potential between whatever is being referred to by this. If current can be measured between two devices, they aren't completely bonded.

If anyone has a garage/out-building with a sub-panel, they might know that code requires a bonding conductor to be run to and attached to the main breaker panel, with ground rods installed at the garage/out-building and also at the main panel. The ground rods assure that the panel is grounded but the bonding conductor assures that the two panels are equipotential. Rather than use the circuits to conduct the difference, the bonding conductor does this.

The 'loop' is created when a bond isn't adequate and the difference finds that the operating circuit is the easiest path to ground.

This is something CEDIA only touched on when I was there for certification, although I haven't been to the higher installer training- from what I have heard, they don't go too far into this, the way BICSI does.

You funny. These ideas are plenty interesting to have thought of when you suddenly have hum in a system. It's not nice to go into a battle of wits unarmed or only half armed.


Nicely put. As my friend Tom points out, you always find something in the last place you look (sometimes, such as on my desk, that is also the first place I look, too).

And your wording points out that we don't need to solve ground loops (as we normally call them), we need to solve ground loop issues; the guys who wrote that article say that if there is no issue, it isn't a ground loop. I wrote with their terminology. In a way it's silly to only call it a ground loop if there is a problem, but it is important for us to know that there are ground loops all over the place all the time that don't have problems. That's because adding one wire or removing one wire can cause hum, bit it's not necessarily that wire that's causing or solving the problem.  If you think it is, you can be totally baffled.

My favorite ground loop solution was in a school-type cassette recorder in the early 80s. These thing had three-wire plugs and hummed like mad. Idiot PC board designer ran the power amp ground all over the place, past a lot of low signal level grounds to the line out ground. Using this with any other three-pronged product caused hum. He solved the problem by adding a 3.3 ohm resistor between the PC board's ground and the chassis ground. When used by itself, that resistor hardly mattered. When a ground loop was formed, that resistor was high enough to force ground currents to go through the third prong power ground and not past the low signal level stages, so no hum!