10 metres audio cable going into PC = too long?

At this point, changing voltages would be prohibitively expensive. We do have 240-volt outlets for major appliances.

It's a trade-off. 120 V is safer from the electric shock standpoint, but

240 V greatly reduces the risk of starting a fire when a connection develops high resistance, since there will be only half as much current and thus 1/4 as much heat output. British ring wiring further reduces that risk. I wonder if ring wiring is legal in the US... if so, I might request it the next time I have something wired.

The output impedance of any modern IC-based audio equipment is probably going to be quite low (under 100 ohms) because the output impedance of an op-amp is inherently low. This includes line-level outputs that are intended to drive high-impedance inputs.

Differing impedances is harmless. The problem is *common* impedances.

My appologies, I didn't mean to go that far over your head.

On Thu, 20 Apr 2006 07:54:57 GMT, snipped-for-privacy@pearce.uk.com (Don Pearce) Gave us:

You just proved that you are even more of an idiot than he proved you are.

His diagrams didn't kill the thread, and no, you do NOT quote everything over and over again, dolt boy.

I quote what I want to quote. I produced a simple diagram that showed the essentials of a ground loop - two ground connections and a single signal connection. This was the opposite of what he claimed. He then went on to produce a slew of poor ascii diagrams, none of which demonstrated his contention. It was clear at that point that the thread was going nowhere.

So, are you saying he is right, and I am wrong? Do you believe that for a ground loop you need two signal connections and one ground connection?

Go ahead - make my day.

d

You produced a _block_ diagram, not a circuit diagram. You did not demonstrate that it even produced a ground loop, and in fact what you showed does *not* necessarily constitute a ground loop.

You still can't get the right quote. Two signals, one common path. There need not be a ground, and there need not be a loop.

And you can trust that everyone who actually does understand what happens is indeed saying you are wrong.

The diagram demonstrated exactly what it needed to demonstrate - no more and no less. To get a ground loop, you need to make a loop from the ground. To make a loop, you need the ground from one piece of kit to another by two separate paths. This can be a problem in two ways. Either, if the loop is physically large it can intercept magnetic hum fields and generate a casement flowing around the loop, or if other equipment is using the circuit, those currents flowing through the main ground path will do the job for you. The mechanism that turns these currents into emfs that actually couple into the input was irrelevant to the point I was making. I was showing that for a ground loop, you actually need to make a loop from the ground.

You diagrams, on the other hand were actually wrong. You were putting Earth symbols at locations where there is no Earth connection. In a house there is only one Earth connection, and that is on the company side of the consumer unit. All ground connections within the house are as I showed - simple wire connections between pieces of kit. In fact, the Earth connection is irrelevant to the phenomenon - ground loops would still happen if the whole lot were floating.

Maybe I'm wrong of course - in your house does each mains socket have an individual ground which is actually a stake driven into the Earth?

No? Didn't think so.

Wrong - as wrong as you can get. Let me tell you this again. You can have ground loops producing hum in a mono system. That is just one signal in case counting is a problem. What you need is the two ground connections. (Two is one more than one).

I leave them to judge.

d

Since I am reading (and replying) on an 'audio' newsgroup, my assumption is that people may be concerned up to around 20kHz, and want to be aware of any changes that might be considered worthy of note in the context of using good audio equipment in a domestic situation.

If I assume a cable capacitance of around 100pF/m, then a 5kOhm source would, I think, give a roll-off of the order of 1dB in the 15-20kHz region. (Assuming I managed to push the right buttons on my guess-box. :-) )

On this newsgroup I'd normally expect people to regard that as being large enough to be worth at least being aware of it.

You are probably correct. Alas, consumer equipment sometimes dissapoints out expectations - particularly when the makers have failed to provide the relevant data for the users. ;->

Once one becomes aware of the problem, and knows what adjustment to make to correct for it appropriately. Hence my initial comment. :-)

Slainte,

Jim

I would agree that modern circuits that use IC amps with feedback should be able to provide nominally low o/p impedances.

However, although in general you are probably correct, the snag is that life may not always be that simple. :-) For example, some outputs may have a low nominal output impedance, but have a relatively low current limit. Hence they may act like a low impedance source for low currents, but show slew-rate limiting when asked to drive a high capacitance load (long cable). For this reason the designer may even have included o/p series resistors to protect against this and avoid the IC being overloaded.

As I indicated, all we have so far is that the source is a 'stereo or TV' with no details of their actual outputs. Hence it will probably be OK, but as I pointed out, we can't really be sure in the absence of the relevant data.

IIRC the old IHFA-707 specs required audio signal sources to be tested with loads of 1000pF//10kOhms as the 'worst case'. I don't know what standards (if any!) makers of domestic TVs routinely adopt these days, but 10 meters of cable does seem to me to bring us into the area where the 1000pF limit might be an issue.

Slainte,

Jim

I'm a complete newbie when it comes to this stuff, so I'm jumping in with a question and hoping I'm in the right place. We have always hooked up our computer (in one room) to our stereo (in another) so we can play audio from the computer through the stereo speakers. We recently moved and in this new house we're getting a hum or buzz on the line when we play music from the computer. The cable from computer to stereo is 30 feet and there was no hum in the audio when we did this in our previous house. What can we do to get rid of the hum? We've tried physically moving the cable but that hasn't changed anything.

Thanks, Stacia

Use Toslink optical cable to link the two, assuming that the two have suitable ports.

Virtually *every* outside plant telephone cable is wired up exactly like that. There is a ground at both ends of each and every section (3000 or 6000 feet), and the shields from each coupled section are bonded to the other and to ground.

A three mile long section of cable might look just like this:

6000' 6000' 6000'

signal pair +==//==+ +==//==+ +==//==+ shield | | | | | | | +-+-+ +-+-+ | | | | | ----- ----- ----- ----- --- --- --- --- - - - -

Yet there is no ground loop, ever. Your description says that would virtually *always* cause a ground loop.

The problem is that you description does *NOT* describe the essence of a "ground loop". You have not shown where the coupling takes place, nor have you describe why it happens.

You apparently were unable to read the equivalent electrical diagram that I posted showing exactly what the mechanism is. Here is is again, though this time I've rearranged it slightly to meet your requirement that the load resister be to the right of the input. I hope the irony of the difference is not lost on you...

o---------+ | +-------+ | Rload | +-------+ Signal | Source +-------> connection =======//======= The mechanism that turns these currents into emfs that actually couple

You do *not* need a loop, and the "mechanism ... that actually couple" them *is* the only point that actually matters. If there is no interference, then there is no "ground loop". When there is interference, that mechanism is what causes it.

The mechanism is having a single common path for two different signals. (It often involves a ground system and that might well have a loop, but in fact is there is no requirement for it to be a ground system and no loop is required other than the normal closed circuit required by *every* signal path.)

There of course *is* an "Earth connection" at the points where I showed them, *by definition*.

There is not necessarily only a single earth connection point in any given house, but the circuit was not necessarily showing a circuit within a single house anyway. It could just as easily be between two houses a mile apart. Nothing in either scenario would change how it works as opposed to a single connection within one building.

You've never lived in a house with 1) natural gas lines, 2) water pipes connected to a well, 3) any kind of a ground system for a transmitting antenna, or probably some other odd circumstance that I haven't thought of? (I've lived and worked in buildings with all of that, and I've worked in buildings with more than one electrical distribution system too. Your statement is absurd.)

You cannot have "hum" and your desired "signal" without having two, count them, signals. Whether it is a "mono system", or something like RS-485 data, a telephone dial loop, a T1, video, or any of a number of other types of circuits and signals, the basics are exactly the same.

Two signals does *not* define a stereo audio system! (Though your comment pretty much does indicate the extremely limited concept of a "signal" that you have... and that probably explains your limited understanding of ground loops too.)

Has anyone suggested yet that you are even close to correct in any aspect of what you've said? Do you wonder why?

Yep, well thats balanced operation which as you say will go for miles over telephone copper lines without humm..

Oddly enough in the UK they don't as a rule use screened cable, the twisted balanced pair has very good rejection.

I seemed to think we were talking about domestic unbalanced lines here?.....

My circuit showed what kind of poor installation causes ground-loop induced hum. You get a ground loop in a domestic Hi Fi (and yes, that is what we are talking about here -forget your telephone cables).when you join to bits of kit together with coax cables, and also have them both connected to mains ground.

No better. You need to understand that the load is on the far end of the cable, not the same end as the source.

We are trying to work out how to prevent a domestic system suffering hum from a ground loop. So no, the exact mechanism by which the coupling takes place doesn't matter. What matters is how you connect your kit up.

Wrong. If you connect two bits of kit with a single coax cable, and there are no other connections, you will not have a loop that causes hum.

No there isn't, there is a connection to a bit of wire. Here in the UK that is a piece of thinnish bare copper contained within the twin-and-earth 2.5 squ mm cable that connects all domestic sockets to each other. At some distant point where the power enters the house there is an earth connection.

We are talking about a 10 metre cable connecting a PC to a Hi Fi.

Good for you! And the connection with the current question is?

Learn some terminology. In electronic engineering generally as well as audio, "signal" refers only to the wanted stuff. Hum, noise and distortion are *not* signal.

Yes they do. You will find that there is a left signal and a right signal. That, I'm afraid, is your lot.

I'm always open to learn something new of course. If you have other signals in your stereo do tell me all about them. Tell us all, in fact

- I'm sure we would all be glad of the new knowledge.

I'm still perfectly happy to let my words stand for judgment.

d

Ground loop isolator.

greg

Regular telephone cable does not have a shield on each pair, but does have a shield around the entire bundle of pairs. The above diagram shows the reason!

The effects of shielding is almost useless at 50-60 Hz AC power frequencies, which means that noise immunity would be only the common mode rejection ratio if there was no shield or if a shield is grounded at only one end. Instead the shield is grounded at both ends, which allows any induction to not only induce current into the cable pairs, but also into the shield. The shield has is a larger conductor than the pairs, hence has less resistance and therefore significantly more current flows. That current flow in the shield causes an opposing current to be induced into the signal pairs! And that reduces the amount of noise in the signal pair significantly below what it would be if common mode rejection was the only noise reduction mechanism.

The ground loop part is exactly the same in either case. The example above is just a very convenient way to demonstrate positively that cables *are* grounded at both ends, and that it not only does not necessarily cause ground loop noise, but actually is a way to reduce noise in the signal wires.

So please do comment on the above. If what you say is true, nobody would install telephone cables like that, because it would cause *huge* ground loop currents every time! It fits

*all* of your criteria perfectly. But it not only doesn't cause a ground loop, it is done specifically to *reduce* noise in the signal wires.

Incidentally, the exact same technique would work if the above were applied to a coax cable with one single ended audio channel on it. For example, a 10 meter connection between stereo components!

You "circuit" was identical to the above. It clearly does *not* necessarily cause ground loop induced hum.

You get hum in domestic HiFi for *exactly* the reasons I've stated.

If you ran separate ground cables all the way to the earth ground (or to some point which represents a low impedance connection to an earth ground), you would *not* have the common path that I've repeated demonstrated is the culprit. But instead people connect all of the equipment grounds to the ground wire in a power outlet, and because that is a very long

*common* path, *all* signals on it are induced into each circuit that is connect to it. That includes 60 Hz power and any number of audio (e.g., mono or stereo), video, RF, or whatever signals.

In your case, it is a stereo hifi, but the same applies to any other signal too.

You apparently have never worked with equivalent circuits, eh? The load is *not* at the far end of the cable. All that cable is doing is providing a connection to *two*, count them, signal sources.

The point in the above diagram is to emphasize the voltage divider made up of the two impedances, Rload and Rgrnd. The cable is shown

*only* because it was convenient to draw it that way and label it so that readers would understand what it was.

If I weren't dealing with someone like you, it would have looked more like this:

Signal Sources

Desired Cable Induction Ground Differential o o o | | | +--------+ +--------+ +--------+ | Rload1 | | Rload2 | | Rload3 | +--------+ +--------+ +--------+ | | | o---------------------o---------------------o | +-------+ | Rgrnd | your kit up.

You have adequately demonstrated that understanding the mechanism is a requirement to *fixing* the problem! You've incorrectly described what happens, why it happens, and what can be done to correct it.

Knowing how to avoid ground loops is what matters in how you "connect your kit up".

You also can't turn it on, as there is no AC connection.

The point is that people *do* want to turn it on, the do want various other components connected, and they *do* want to have various components grounded... both to avoid AC hum and to avoid electrical shock.

Hence there *are* going to be other connections to active components.

And regardless of that... an easy proof that what you say is not correct is known to almost anyone who has ever tried cabling microphone systems. Just hook up a microphone through several feet (10m would do!) of shielded cable and connect the shield at both ends... No other connections. No ground. No AC either. No loop. But just see what a nice ground loop that makes as soon as someone picks up the microphone! Bzzzzzzzzzzzzzz.

And of course *that* is what is shown in both the diagram I drew, and equally well in the one you drew!

The points at which I place the ground symbol are the points where the system is connected, via whatever means, to a ground. The impedance of that connection is represented in the equivalent circuit by Rgrnd.

You seem to have the mistaken idea that there has to be some sort of "kit", or equipment, connected on both sides of that cable to cause ground loop hum! Wrong!

You do realize that in your diagram there were *two* ground loops, right? Not one, two. See if you can find them both.

It works exactly the same, whether 10 meters or 10 miles.

You brought it up, what *was* the point? (None. You didn't make a valid point.)

Signal in one circuit is noise in the other. Your audio of course does not bother the AC power lines, but regardless it is nothing but noise when coupled into the power line. The AC power is the desired signal when on the power line, but noise when on the audio line.

The problem, once again, is having those two signals on a

*common* *path* (where *both* are the desired signal). In that way each of them becomes a noise for the other's circuit, even though they are in fact both the *desired* signal on the common path.

Thanks for the laughs. That is hilarious.

You are confusing cause and effect. A stereo has two signals, but two signals does not make a stereo. There are *many* things that have two signals that are never described as "stereo". The most obvious at this point is a single channel of audio and a single channel of AC power. Two signals... no stereo.

They are!

OK I will comment on the above. There are ground loops, and they induce hum. The hum doesn't reach the signal because it is common mode, while the signal is differential - the hum is thus cancelled.

Yes it does.

Try and understand. There is only one signal - that is the wanted signal. The rest is interference.

You now appear to have three separate load resistors. This just gets more and more bizarre.

I have not at any point addressed what should be done to correct it, but I will now. One must ensure that there is but a single ground connection between two pieces of equipment. In the UK these days that is simple. Amplifiers have mains plugs in which the ground is connected. All ancillary kit - tuners, CD players etc are safety-protected by double insulation, and have a mains plug with no ground connection. The only ground connection they have is via the coax cable, thus making sure that ground loops don't happen, and there is no hum.

There are these things called batteries, you see....

See my paragraph above.

Connect the shield to what at both ends? Try and write a little more clearly, please. And yes, I work regularly with microphones, and guess what? The shield is connected to things at both ends.

And now you can explain how somebody picking a microphone up could create a ground loop. Or is that simply your name for anything that goes buzz?

I suggest you look at your diagrams again. They show no such thing.

No - you have placed final earth connections there - not the same thing at all.

No, the kit is there to reveal it. The hum will be generated whatever.

There is one ground loop in my diagram. It comprises the coax screen in one direction, and the return via the mains ground cable. I challenge you to find a second.

No, I pointed out the errors in your diagrams. You turned it into this nonsense in a desperate attempt to deflect the criticism.

In audio the signal is the music. All else is interference.

AC power is not a signal - signal is information, not power.

What two signals are you talking about now?

I'm glad you are amused.

I know a little girl with an iPod. I suspect even should would call you an idiot.

They are? What is that supposed to mean? What are what?

Get a grip, please.

d