The issue of ground loops comes up fairly often among system installers, as it is a frequent problem with a simple cause that is not widely understood. The simple cause is as follows: the connections carrying audio and video signals between most equipment are of the unbalanced type. In an unbalanced connection, the signal is conveyed as a relatively low AC voltage on a single shielded conductor. Of course, voltage is by definition a difference in potential between two given points in an electrical circuit. At the unbalanced input to an audio or video device, the AC voltage carrying the signal is measured between the shielded center conductor of the cable (or pin of RCA, BNC, or F connector) and the outer shield (or connector shell).

This in and of itself does not pose a problem. However, on nearly every piece of audio/video equipment, the input/output connector shells are electrically connected the equipment chassis. In equipment with grounded power cords, the chassis is in turn connected to the building earth ground via the mains wiring between the wall outlet into which the equipment is plugged and the breaker panel. Now, if there is only one piece of grounded equipment in the entire system, there is still no problem. However, as soon as you add another piece of grounded equipment (and keep in mind that properly earthed antenna and/or cable TV feeds also present a ground connection to equipment not otherwise grounded), the system as a whole now has multiple connections to ground - hence a ground loop.

However, the presence of a ground loop still does not guaranty a noise problem. To understand how a ground loop can cause noise problems, one must understand that – while "ground" should ideally be a uniform zero-voltage reference throughout a building's electrical system (with no difference in voltage between any two ground points, such as outlet ground pins, for example) – this is never really the case at all. For the fact is, the ground conductors in the house wiring have a resistance, just like any other conductor. And for various reasons there is always a small current being carried by these ground conductors back to the breaker panel. Those familiar with basic Ohm's law will understand that each of the building's ground wires will have a voltage drop along its length in proportion to its resistance and the current passing through it. Therefore, each receptacle will have a measurable voltage between its ground pin and the ground buss in the breaker panel, and – by extension – to the ground pins of nearly every other outlet in the house. (It's worth noting that this difference in ground potential will tend to be greatest between outlets on different phases, for the voltage on an outlet's ground pin is more or less in phase with that circuit's hot line, and as we know, voltage between two opposite phases is equal to the sum of the two with respect to earth.)

So, finally we’re able to get to the meat of the problem. We take two pieces of equipment in difference areas of a home, each plugged in to an adjacent outlet with a grounded power cord. Let’s say that one is a high-end A/V receiver in an equipment closet in the hall, and the other a powered subwoofer in the nearby home theater. Now, we plug a shielded RCA cable into the subwoofer output on the back of the receiver and run that cable to the subwoofer in the home theater, but don’t yet connect it to the sub. Were we to kneel down behind the subwoofer and place an AC voltmeter between the connector shell of the RCA cable coming from the receiver and the connector shell of the input jack on the back of the sub, we would note a significant voltage, likely a couple volts or more (of 60Hz AC, here in North America). Why? Going back, remember that the shields of the RCA connectors of both the receiver and subwoofer are connected to their respective chassis, which are then connected to the ground pins of the respective wall outlets via the grounded power cables, which are in turn at different potentials thanks to Ohm’s law and current leakage into the ground wiring between each of those outlets and the breaker panel in the basement.

Now, here’s the fun part: plug the RCA cable into the subwoofer and turn the sub on. HUM! Why? Now the chassis of the sub and receiver (which had been lifted to different ground potentials as just established) are electrically connected to one another via the shorter path provided by the shield in the RCA cable. As we know from basic electrical theory, the voltage difference between the two has resulted in a current in the shield of the RCA cable. Ohm’s law once again comes into play here, for the resistance of the RCA cable shield assures us that it – while serving to reduce the voltage difference between the chassis of the two devices – has not eliminated it altogether, for there is now a voltage drop down the length of the RCA cable shield. Remember that the audio signal being reproduced by the sub consists of the AC voltage as measured between the center pin of its RCA input jack and that jack’s outer shell. Trouble is, the signal at the sub input is no longer the same as that at the receiver output, for receiver output signal is the voltage between the center pin of its RCA output jack and that jack’s outer shell, and the voltage drop down the shield of the RCA cable (which consists almost entirely of 60Hz AC) ends up being added to the AC waveform at the sub input as a result.

It’s this last concept that is often hardest to grasp, so I’ll put it another way. Remember that with an unbalanced signal connection, the original signal is represented as the AC voltage between the center conductor and the outer shell of a device output jack. The shield of a cable connected to the output jack insures that the signal voltage on the center conductor remains largely unchanged from one end to the other. Were the shield itself to remain at the same potential along its length with respect to the shell of the output jack, then the signal would be faithfully transported, free of noise. However, in the presence of a ground loop, a substantial current is transported by the shield of the cable, resulting in a measurable 60Hz AC voltage on the connector shell at the far end of the cable when measured with respect to the connector shell at the output (or source) end of the cable. Because the audio signal being reproduced by the sub consists of the AC voltage as measured between the center pin of its RCA input jack and that jack’s outer shell, and the shell of the sub input jack now has a 60Hz AC voltage on it with respect to the outer shell of receiver’s output jack (from which the signal originated), this 60Hz AC voltage is now present in the subwoofer input and is audibly reproduced.

Now that we understand how ground loops are created, how do we avoid them? First I’ll tell you how not to fix them: ground lifters and cheater plugs are valuable diagnostic aids but should never be employed as a permanent measure. Design you systems properly from the get-go and you’ll seldom need such band-aids. Also, I’ve heard it suggested that cutting the shield connection at one end of an unbalanced signal would eliminate hum. In fact, it will make it much, much worse!

The first thing to keep in mind when designing your systems is that the building power distribution network used by your equipment is part of the system! Whether you’re doing new construction or retrofit, place your equipment on dedicated circuits, on the same phase. For more modest installations, a single 20A circuit may suffice.

In any event, when the system design calls for source equipment to be located apart from other audio/video equipment such as subwoofers, displays, and projectors, run the power wiring in a star configuration. That is, run the power from the panel directly to the central source equipment, and then branch from there to the outlying equipment. This not only helps keep the outlying equipment at the same ground potential as the source equipment, but it also permits you to do surge protection, filtration and/or power switching at your central source location.

When cable TV or antenna connections present ground loops, use RF isolation transformers (such as Xantech’s ground breaker). An isolation transformer works by severing the electrical connection between two pieces of equipment. Instead, the source is connected to one side of a 1:1 transformer, which inductively couples the signal to the other side, which in turn feeds the receiving equipment.

When your mains wiring can’t be run in accordance with the ideal layout, or ground noise has entered the system in spite of you best efforts at prevention, break the grounds in signal cables by employing audio or video isolation transformers such as those in the IsoMax line from Jensen Transformer. Or consider running balanced audio and/or video over twisted pair using passive baluns, active transceivers or equipment already equipped with the appropriate I/O. When using audio isolation transformers (particularly for subwoofer feeds), keep in mind that it takes a high quality (expensive) transformer to pass low frequencies without significant roll-off.

Good Luck!

I'm not sure I understand the signifigance of using the same phase. Aren't both phases referenced to a single ground point? And isn't that what we're after?

Yes, both (or all) phases use the same ground point. However, by the time the building wiring reaches the loads an AC voltage is present on the respective ground conductors. This AC waveform is of the same phase as the "hot" in its circuit. In a typical residence with split-phase service, each of the two "phases" is 180 degrees opposed to the other. So if you compare the waveform on the ground pin of an outlet on one phase with that of an outlet on the other, they will also be 180 degrees opposed to one another and therefore additive. For example, if the ground at the first outlet is showing 5 volts of ripple and that from an outlet on the other phase is showing 7 volts, there will be a total of 12 volts of ripple when measured between the two. This higher voltage differential will create a more severe ground hum in interconnected equipment plugged into these two circuits.

Conversely, with outlets on the same phase the opposite is true. The respective ground ripples are in phase and therefore have the effect of nullifying one another to some degree. From the previous example, the 5 volts on the ground of the first outlet will subtract from the 7 volts of the second, yielding a difference of only 2 volts.

All of the preceding assumes that the contamination of all building grounds consists only of the same 60Hz sinusoidal waveform. In fact, it's a bit more complicated than that, as switching and other non-linear loads dump all kinds of garbage into their respective points in the ground system. While these other noise components certainly play a part in signal corruption, it's still the 60Hz component that causes the most trouble. Besides, placing our equipment on dedicated circuits leaves us less susceptible to ground contamination from other loads.

Finally, someone who can explain this better than me! Great post M. Anyone taling asking any questions about ground loops and hum patterns should automatically be refered to this post. Great post again.

Mike

I'm telling you I've read this "paper" a dozen times before. If I am mistaken someone should talk to M about writing for them.

Julie?

bruno...who are you?
thanks,
julie

Maybe Larry will chime in on this one.

It might just be our smallish market, but I get the most blank-faced stares when I ask the electrician to be sure the two or three outlets I am concerned with are on the same phase and nothing else in the house is on those circuits.

They are good guys. I like them. They generally do the electical part of the house well. They just don't understand. If they were just wire pulling helpers, I could understand. What I don't understand is a journeyman electrician that has no clue as to what being on the same phase means, and I've bumped into many of them.

I also don't understand why, when I stress... This outlet needs to be on a dedicated circuit all by itself, I find lights with dimmer switches on it after the house is done. If he needs permission from the contractor or home owner to add a circuit. Fine. Tell me. I'll talk with them and make sure he gets that permission. Just don't tell me you are going to do that and then don't.

These things DO matter.

What we need is a Larry Fine type heading up all the electrical companies.

IE Somebody that understands.

As far as phase mattering or not, I'd ten times rather fight a ground loop problem that was in phase... because there's a good chance it won't be there in the first place, and if it is there, it's easier to solve.

As a precaution I'd like to offer a suggestion. When all grounds are at the same potential, NO ground loop current can flow. The short version of that is, no ground loop problems. If non-signal current isn't flowing in the ground between two pieces of equipment, you will not have a ground loop problem. And you can take that to the bank. Think in terms of SIGNAL PATH ground. Not just power line ground. Signal path ground means the shield on your interconnect cables.

Power circuits may be at 120 volts (and thats the terms electricians think in, after all, that is their job. You are not going to start an air conditioner compressor with a few millivolts.)

Signal circuits (audio and video) are measured in millivolts. If there is a 60hz. 'power line' current of just a few millivolts flowing in the ground path, in unbalanced consumer equipment, it is flowing on the same wire that carries the return path for the signal. It becomes a part of the signal. Thus ground loop problems. It's really that simple.

If you want to be overly safe, any place you have equipment, make sure you have a ground wire going all the way back to the utility entrance ground rod. That is the defacto ground reference point for the house. 99% of the time, you won't need it.

Commercial broadcast installations are grounded using a 'star' grounding system. Star grounding is, simply put, like home running your grounds. They all go back to one central point and the utility entrance ground rod is a very good reference point. Generally, that is overkill for a residential environment if your circuits are all on the same phase and don't make any side trips.

You can daisy (lasy) chain grounds, just like you can telephones. That is part of what causes ground loops and unwanted current flow in the ground circuit. It may be minute, but that creates a difference of voltage potential between two pieces of equipment. And yes, you can use a voltmeter with the two leads connected to ground points on two pieces of equipment to actually measure that difference. If it's zero, you do not have a ground loop problem. If it's a few millivolts, you just might.

In consumer application, you rarely need a ground system that dramatic... as long as you can convince your electrician you really do need all your circuits on the same phase, with no side trips to pick up a lighting circuit, or a deep freezer, or an alarm system (lots of digital noise there).

Electricians are not bad guys. It's just that their industry needs to start educating them in areas where not volts, but millivolts matter. And on grounds, millivolts do matter.

There are only two way to combat ground loops.

Give the equipment a solid and independent ground back to the utility entrance.

Break the ground at some point in the system.

Commercial installations use balanced lines. That simply means there is a pair of wires that carry the signal, and a separate shield wire that connects to ground to drain off stray voltages.

Consumer installations generally use an unbalanced cable. A hot lead and a ground. In this case, the ground does double duty. It carrys the return path for the signal, and it dumps unwanted voltages to ground.

Properly installed balanced lines have their shields connected at just one end. Usually the end of the cable that connects to an input. ( In practice, you take noise measurements with the shield connected first to one side, then to the other, then you pick the connection that provides the least noise.)

The shield is left floating at the other end. Stray voltages picked up are drained off to ground at the end connected to ground, but since the other end is not connected, no ground current can flow.

Most consumer installations have a multitude of ground from all the RCA cables connected. Many times you can solve ground loop problems by applying the same prinicple.

If you are having ground loop problems, try breaking the shield of the RCA cable connecting the two pieces of equipment. You might think that would kill your signal path. It won't. You have other grounds that are carrying the same signal. If you didn't, you wouldn't have ground loop problems in the first place.

There are physical laws that apply to ground loops. In practice, trial and error are usually needed to find a solution.

Educating the electricians you routinely work with is the best line of defense. If you can get them to get all your power outlets at the same potential (on the same phase of the AC line) you won't face very many ground loop problems.

That said, ground loops caused by CATV connections seem to be a separate breed. Always have a CATV ground break adaptor handy. They are inexpensive, work well and do not interfer with the CATV signal. They break the shield path by opening it physically, then bridging it with a capacitor. The capacitor looks like an open circuit to 60 hz. signals (where ground loops occur), but looks like a straight wire at RF frequencies where the true CATV signal lives.

Bravo, Mr. Bruno!

Alan

I appreciate your contributions, Warren, and the support of another vocal advocate for the application of sound engineering (no pun intended) to the vanquishing of system ground noise. However, I believe a few of your points may deserve some clarification. That said, I realize that this is not the forum for a lengthy debate of system grounding methods. Those interested in the finer points of grounding will find countless articles within the back issues if the Journal of the Audio Engineering Society.


"Overly safe" is right. If one were to adhere to the guidelines I outlined, the measure you propose would almost certainly be unnecessary. Also, I believe that there are code issues related to the running and use of independent grounds of the sort you describe. (The electricians on the board can provide a footnote here, but I would hope that this thread not stray too far off the subject.) Furthermore, I would advocate a slightly different ground topology (see next point).


The "star" grounding does generally involve running all grounds to a central point, as you say. However, placing that point at the utility entrance is not ideal for the suppression of ground noise. Rather, the hub of the star should be placed at the central nexus for the A/V system wiring. In a home theater, this would likely be the equipment rack containing the receiver/processor. In a professional recording studio, this is at the patch bay for the mixing console. From there, the ground is extended to the outlying equipment (within the mains wiring, per my earlier post), roughly paralleling the signal cables to each respective device. Keep in mind that the objective here is to minimize the difference in ground potential between each interconnected pair of A/V components. Placing the hub of the star at the main equipment rack typically assures us of the shortest average ground path (hence the lowest average impedance along each respective path, and the least ripple).



Again, I contend that the grounding technique I described is more effective and less cumbersome.


If you're having ground loop problems, by all means, try breaking the shield of the RCA cable connecting the two pieces of equipment. I guaranty that the problem will be much, much worse. For by doing so, you've now taken away the fairly direct signal grounding path provided by the shield and left in its place a lengthy run of ground wire through the building's electrical system, having a considerable 60Hz AC voltage differential already present along its length. if you want a simple demonstration of this, place a grounded sub on one circuit and a grounded receiver on another. Turn the sub on and adjust its input trim to a typical setting. Plug an RCA cable into the receiver and slowly plug the other end into the sub. If you have some experience connecting consumer audio equipment, you know what happens next: the center pin if the RCA makes contact before the RCA barrel, hence the shield is not yet connected. And should you dwell for a while with the connector partially mated in this manner, the resulting roar of clipped 60Hz will soon have your ears bleeding and/or your sub crying for mercy.

So, as I said before: don't cut the shield on an unbalanced A/V cable. It's a critical part of the circuit. Instead, use a isolation transformer as described previously.

Parts of my reply were referenced to RF systems. Not the best of answers for this forum. Other parts, well suited to this forum. Too late at night to go through it all.

I will reply tomorrow.

i think i have a ground problem but it is only affecting my video, a greenish tint is rolling on one of several 50" pannys we have in a bar.

time warner cable
mini rgb cable from pace cable box
thanks for help

They absoloutely do!

If it's only one, It might be a problem with the display itself

Actually M, we are not as far apart on some points as it seems at first glance. Just different ways of looking at it. I think the biggest point that should be made about ground loops is that most of the time they can be avoided if you understand what causes them. Yet there's always that odd time.


Depends on the application. For the general consumer one, I'll agree with you on that. That's why I said 99% of the time, you won't need it. Take care of the power source issues in the first place and you can probably change that to 99.9% if not more.

Star grounds always involve running all grounds to a central point. That is where they get their name.

The central part might be in an equipment rack, or at some other point. But always, each piece of equipment goes back to a central ground in a star pattern. NO daisy chaining grounds.

Unfortuantely that is exactly what unbalanced consumer interconnect cables do. Daisy chain with multiple grounds. With balanced gear, it's no big deal to lift the shield at one end to break a ground loop. Your signal path remains intact. In fact, that is routinely done on properly installed balanced circuits.

Lifting the shield on one end or the other will ALMOST always give you a lower noise figure. And hum ground loops and hum bars are noise factors.

Ground loops can be formed by just one ground wire, or can be caused by having multiple grounds involved. A piece of equipment that has a grounded power cord, or a CATV ground or a satellite dish ground adding extra grounds being examples. The additional grounds, at differing potentials can cause problems. Generally speaking, more than one ground is involved in all consumer applications.

In commercial broadcast applications with millivolt signal levels having to compete with high power RF transmitters, I'll opt to keep my options open. Grounds are much tricker there. I'll overkill the ground system so I have other options to try.

I was going back through memories of grounding nightmares from the past in dealing with RF. While I was writing my reply. For the purpose of this forum, I definitely should have stuck to 60 Hz. problems to avoid confusion.

That's my bad.

|

I'll agree with that to an extent. But I definitely wouldn't guarantee it. It certainly can make it worse. I have also seen it make a great improvement if not solve the issue outright. And there is a solid technical reason why. More on that later.



I wasn’t suggesting it is a cure all, just something else to keep in your bag of tricks in a tight spot. In your subwoofer example, let’s assume both the sub and the amplifier feeding it have grounded power cords.

If the outlets they are plugged into are properly grounded, the connection that the interconnect shield normally makes is already established before you ever plug the interconnect in. You might get a 'blip" when you slowly plug the interconnect cable in, but you won't get a sustained, ear bleeding hum.

I agree, It can happen, but you are not likely to get more hum by breaking the shield at one end of the cable if both units are using grounded power cords and the power circuit ground is solid. It can also happen they you will wind up with far less noise.

There can be a difference depending on which end of the cable you break the ground path in. Try it both ways and pick the one with the lowest noise.

Here’s why this can work, and work well in some situations. The signal path will still be complete because of the additional ground involved. So no matter what, the sub will still be thumping along.

Yes, the low side of the path will probably be longer and have the potential to cause more problems. But in the real world, that does not always happen.

Stray voltages the interconnect shield picks up can no longer flow between the two pieces of equipment via the shield, yet they will still be drained off to ground at the end where the shield is still intact.

The difference of voltage potential between the two pieces of equipment can no longer cause a current flow along the interconnect shield. And that shield makes a 100% perfectly parallel path with the hot side of the connection in your interconnect cable. VERY close, totally parallel proximity the full length of the run. And depending on supwoofer location, that can be a long distance in some installations.

Breaking the power ground will almost always cure the problem but that is not wise. Breaking the only other ground ground involved in this case will frequently do the same thing, and is legal. And in many cases works just fine. It’s worth trying before going to the expense of an isolation transformer (Whether signal path or power supply transformer).

Give it a shot. If it works properly, use it. If not, move on to the more expensive cure. I’m not disagreeing with the principle of isolation transformers at all. I prefer to try simple fixes before spending more money.

IF having the amplifier on one phase and the subwoofer on the alternate phase of the houses electrical system is the cause of the problem… that can be a different animal. An isolation transformer can certainly help there, but the best cure for that is to never let that happen in the first place. Install it right. Far cheaper, less parts to fail.

The sub should be powered from the same surge protector the rest of the equipment is plugged into... Even if it is located on a different wall.

In the latter case, the outlet it receives power from should not be wired back to the breaker panel. It should go to where the surge strip will land. That end can be terminated with an "inlet" connector so you can just use a power cable to jump from the surge strip to the inlet connector. (Projectors and TV sets should be handled the same way by the way.) Powering ALL equipment from the same location is ALMOST a guaranteed elimination of audio hum and video hum bar problems.

An isolation transformer on the power line that feeds the sub would also work in this case. I prefer the less parts method. Not much price difference between a good quality isolation transformer and hiring an electrician to reroute the subs power feed to the surge protector. And one less piece of equipment to malfunction.

Hampering a guess, I believe "M" and I would both agree on one thing.

And that is, wire it correctly in the beginning, and have less problems to deal with in the end.

"M" makes some good points. If you are struggling with ground loop problems, go back and reread his post. It is very good information.






This message was edited by Warren on 03/10/05 00:16 ET.