How can ground loops not be a problem in single-ended connections?

[jdav]

I understand how ground loops can be a problem and where they come from. What I'm surprised of is in fact the opposite: why are ground loops often not a problem?

The specific context I'm thinking of is a CD reader connected via RCA cables to an amplifier, while both devices are earth grounded. There is an obvious ground loop, and yet this is the setup I've seen in most average home audio systems, most of the time without any noticeable hum (at the time I wasn't really paying attention so maybe there was an audible hum even though not very loud).

What is the reason? What measures can be taken in that case? The common ones I'm thinking about but I don't think are used are the following:
- Ground lifting : obviously not used in those situations.
- Signal isolation through audio transformer: if it is used sometimes I don't think it is systematic so it can't be the answer.
- Single-ended to differential: not used (I don't even know if it is useful with RCAs, but it was obviously not used in the situation I'm thinking about)
- The ground is only connected on one side: I cannot see how this a solution, since there is no more ground loop but the ground reference might very well by different on both side.

Is it because the shield is conductive enough that hums are usually low enough to not bother the average listener, but is nevertheless always present in those configurations?

Thanks

 

[Speedskater]

a] for RCA analog interconnects, use a coax cable with a very heavy braided shield.
b] plug everything into the same AC wall outlet (or better yet the same distribution strip).
c] use short AC cords.
d] chose components that do NOT have plastic insulator rings on their RCA chassis connectors. (this is the same as the XLR "Pin 1 Problem" )
if all else fails:
run a heavy wire from metal chassis to chassis.
* * * * * * * * * *
the idea is to try to get all the chassis's to the same potential.

 

[jdav]

You gave me very good advice on how to avoid ground loops. But my question is, knowing that the average home audio user didn't apply any of them, how it is possible that you don't here big hums everywhere.

 

[jdav]

d] chose components that do NOT have plastic insulator rings on their RCA chassis connectors. (this is the same as the XLR "Pin 1 Problem" )

By the way, could you explain why? The usual goal is to use separate analog and power ground, doesn't this go against it?

 

[Philbo King]

If there is little or no magnetic or electrostatically coupled noise to pick up, it is a non-problem.

 

[MaxwellsEq]

I understand how ground loops can be a problem and where they come from. What I'm surprised of is in fact the opposite: why are ground loops often not a problem?

The specific context I'm thinking of is a CD reader connected via RCA cables to an amplifier, while both devices are earth grounded. There is an obvious ground loop, and yet this is the setup I've seen in most average home audio systems, most of the time without any noticeable hum (at the time I wasn't really paying attention so maybe there was an audible hum even though not very loud).

What is the reason? What measures can be taken in that case? The common ones I'm thinking about but I don't think are used are the following:
- Ground lifting : obviously not used in those situations.
- Signal isolation through audio transformer: if it is used sometimes I don't think it is systematic so it can't be the answer.
- Single-ended to differential: not used (I don't even know if it is useful with RCAs, but it was obviously not used in the situation I'm thinking about)
- The ground is only connected on one side: I cannot see how this a solution, since there is no more ground loop but the ground reference might very well by different on both side.

Is it because the shield is conductive enough that hums are usually low enough to not bother the average listener, but is nevertheless always present in those configurations?

Thanks

Designers try their best to optimise how noise like this is minimised. However, as you suspect, it's always a risk. I suspect that if you used a top measurement device, you would detect mains noise in many single ended system, but it's probably more than 80dB down, so not noticeable.

 

[DVDdoug]

You gave me very good advice on how to avoid ground loops. But my question is, knowing that the average home audio user didn't apply any of them, how it is possible that you don't here big hums everywhere.

You only get hum when there is an AC voltage difference between the two (or more) grounds and there has to be enough voltage to hear it. With short connections in the same room you don't normally get audible ground loop problems.

It's also possible to build equipment where the audio ground has no connection (or no direct connection) to earth/power ground.

 

[fpitas]

Designers try their best to optimise how noise like this is minimised. However, as you suspect, it's always a risk. I suspect that if you used a top measurement device, you would detect mains noise in many single ended system, but it's probably more than 80dB down, so not noticeable.

This is the real answer. Single-ended is not the best way to transmit low-level signals, but it's good enough for most domestic uses as long as equipment is all on the same circuit.

 

[jdav]

If there is little or no magnetic or electrostatically coupled noose to pick up, it is a non-problem.

In what situation could it possible be little 50 Hz noise in a home studio? Unless you put your kitchen oven beside your audio system, shouldn't be the 50 Hz noise comparable in all homes?

Designers try their best to optimise how noise like this is minimised.

How could you design to minimize this noise? In single-ended connection, once it's there you cannot do anything to not reject it without rejecting all other 50 Hz signal (maybe you could probe mains phase and do something clever but it would be more costly than doing it differential), or can you?

It's also possible to build equipment where the audio ground has no connection (or no direct connection) to earth/power ground.

What are the cons of doing this? Is it common? I think I've almost always seen audio ground tied to earth, but my experience is quite low.

 

[Zapper]

Expanding on some previous replies: the voltage between two points in a ground loop depends on the strength and frequency of the AC magnetic field, its orientation with respect to the ground loop, the area enclosed by the loop, and the resistance between the two points in the loop. In the case of the RCA cable the resistance is the shield resistance of a single cable, divided by how many cables connect the two instruments. Ground loops aren't a big problem in domestic environments because the amplitude of the magnetic fields are very small, their frequency is low (50/60 Hz), the loop areas are usually small, and the resistance between two instruments through the cable shields of short RCA cables is small. Of course, that depends on what you define as "big". If you want a signal clean to the LSB of a 24bit DAC, then unbalanced connections simply will not do.

 

[Geert]

What are the cons of doing this? Is it common? I think I've almost always seen audio ground tied to earth, but my experience is quite low.

You make it an IEC Class II or Class III device, so protection doesn't rely on earthing https://www.xppower.com/resources/blog/iec-protection-classes-for-power-supplies

 

[Philbo King]

In what situation could it possible be little 50 Hz noise in a home studio? Unless you put your kitchen oven beside your audio system, shouldn't be the 50 Hz noise comparable in all homes?


How could you design to minimize this noise? In single-ended connection, once it's there you cannot do anything to not reject it without rejecting all other 50 Hz signal (maybe you could probe mains phase and do something clever but it would be more costly than doing it differential), or can you?


What are the cons of doing this? Is it common? I think I've almost always seen audio ground tied to earth, but my experience is quite low.

It isn't just the amount of interference in the environment, it is also the the gears susceptibility to it. It is not as simple as it seems.
The are various categories of EMI susceptibility as discussed here:

Electromagnetic compatibility - Wikipedia

en.m.wikipedia.org

I don't understand why you are looking for problems that don't exist in a given hypothetical setup. It's more productive to solve problems that do exist in a real system...

 

[jdav]

You make it an IEC Class II or Class III device, so protection doesn't rely on earthing https://www.xppower.com/resources/blog/iec-protection-classes-for-power-supplies

You could still earth the chassis but not connect the audio ground to earth. Would this be a problem? EDIT: I guess leaving the audio ground floating means leaving all the circuit floating which would not qualify for Class I.

I don't understand why you are looking for problems that don't exist in a given hypothetical setup. It's more productive to solve problems that do exist in a real system...

I'm not looking for problem, I'm learning. My learning method is based on critical thinking to make sure I understand things properly. The question I asked was precisely testing my understanding.

 

[Zapper]

In what situation could it possible be little 50 Hz noise in a home studio? Unless you put your kitchen oven beside your audio system, shouldn't be the 50 Hz noise comparable in all homes?

Yes, it's usually low.

How could you design to minimize this noise? In single-ended connection, once it's there you cannot do anything to not reject it without rejecting all other 50 Hz signal (maybe you could probe mains phase and do something clever but it would be more costly than doing it differential), or can you?


What are the cons of doing this? Is it common? I think I've almost always seen audio ground tied to earth, but my experience is quite low.

You can float the shield of the RCA jack and apply the signal and the shield to a differential amplifier. This breaks the ground loop and makes the signal insensitive to the ground voltage. It works until the ground loop voltage exceeds the common mode of the diff amp. This configuration has a lot of liabilities for things like ESD (electrostatic discharge) and RF interference so normally isn't used. Those who care about ground loops use differential connections for analog signals.

Transformer coupling isn't often used in modern audio but it is another way to break ground loops in cables. A transformer between signal and shield acts as differential receiver to the cable.

 

[Philbo King]

You could still earth the chassis but not connect the audio ground to earth. Would this be a problem?

I'm not looking for problem, I'm learning. My learning method is based on critical thinking to make sure I understand things properly. The question I asked was precisely testing my understanding.

There are as many more possible solutions than there are causes. The T&EE (trial and error engineering) method is used on existing equipment. You try something, measure the result, and repeat, till you find the optimum cure for your setup in your environment.

This hints at the issue: Consumers do not have access to, nor expertise needed, speciallized test gear for magnetic loop, IR voltage between gear grounds. So you try things, till you find one approach that is "good enough" while not being outrageously expensive (such as redesigning and rebuilding all the electrical interfaces of a piece of gear).

Or you can sidestep the problem completely by using gear that supports optical fiber interfaces.

 

[Geert]

You could still earth the chassis but not connect the audio ground to earth. Would this be a problem?

No problem. Not that long ago the average CD player or cassette deck wasn't earthed. The power cables didn't even have a earth connection.

 

[Zapper]

No problem. Not that long ago the average CD player or cassette deck wasn't earthed. The power cables didn't even have a earth connection.

It will be a problem if "audio ground" means the cable shield. If you ground chassis to earth but float the cable shield, all the voltage induced around the ground loop gets applied directly to the preamp input. Try it! You WILL hear the 50/60 Hz. However I agree that the chassis does not have to be connected to earth ground, and few are. The earth ground connections are another source of ground loops.

 

[AnalogSteph]

Not that long ago the average CD player or cassette deck wasn't earthed. The power cables didn't even have a earth connection.

Indeed. In my mind a CD player that is not an IEC Class II device counts as an absolute rarity. I am aware of a line of TEAC CD/SACD players from the 2000s that were Class I (which promptly reflected in their measurements), but otherwise none of the big mass-market Japanese brands would have done that.

In these days of SMPS everything, you might actually come across Class II devices with a 3-prong mains lead. Then the mains filter leakage currents can be dumped straight into PE, while device ground might be connected to PE via 4.7-10 nF worth of Y2 capacitor or so at best. That's a much better option than having the mains filter leakage go down the audio cable, which I think was happening with e.g. my JBL 104s on the unbalanced input - dead quiet once plugged in, but the amount of hum if you had them unplugged and were touching the (3.5 mm) plug was legitimately scary.

(I have previously suggested retrofits along those lines for old Yamaha integrated amps with sort of a switch-mode regulated kind of power supply deal sporting corresponding mains filtering, i.e. A-760/960/1060 (II), the ones with the recall for the notorious RIFA caps blowing up. [Which was commendable for Yamaha to be doing after like 25 years, but these were nice amps and I'd rather invest a few bucks into new Y2 caps!] Unsurprisingly, in stock form these have a fair bit of mains hum on inputs left open if no other components provides an earth connection for them. Back in their day that would generally have been the radio-receiving fish via its rooftop antenna, which of course are not nearly as common now.)

 

[KSTR]

Unbalanced connections always cause measurable degradation, even when all devices are 2-prong (Class-II). Much depends on the power supply technology used.
However, in practice the degradation must be strong enough to be audible in the specific situation and setup.

 

[Sir Sanders Zingmore]

a] for RCA analog interconnects, use a coax cable with a very heavy braided shield.

Interesting. Do you have an example of these?
I have what I think is a ground loop issue. The hum varies with different interconnects, so I‘d like to try the type you suggest