Reducing hum, noise and ground loops?

[Newmarket]

It is not. From any electronically balanced output you can take the signal from pin 2 referenced to pin 1. So the audio signal exsist between pins 2 and 1 also. It is just usually 6dB lower.

Yes you can. But, assuming some sort of balanced input, "we" don't do that because it's not the correct way to optimise the interconnection.
Pin 1 should go straight chassis. There is an AES paper / standard on this since IIRC more than 20 years

 

[moamps]

Yes you can. But, assuming some sort of balanced input, "we" don't do that because it's not the correct way to optimise the interconnection.

Balanced input will remain preserved if you connect the cable from pin 1 and 2 of the source to pin 3 and 2 of the load.
I quoted a statement "audio signal exists only between pins 2 and 3" that is fundamentally incorrect.

I have been a member on AES for more than 25 years and I know the whole story about the famous pin "1"

 

[Newmarket]

Balanced input will remain preserved if you connect the cable from pin 1 and 2 of the source to pin 3 and 2 of the load.
I quoted a statement "audio signal exists only between pins 2 and 3" that is fundamentally incorrect.

I have been a member on AES for more than 25 years and I know the whole story about the famous pin "1"

In that case you should be able to understand the statement in the context of the topic. But if you want to be stupidly pedantic then feel free to carry on FFS

 

[ccaudle]

From any electronically balanced output you can take the signal from pin 2 referenced to pin 1.

But that converts common mode noise to fully normal mode noise. By saying that "[t]he audio signal exists only between pins [3] and 2" it means that by definition the audio at pin 2 references pin 3. There may or may not also be a similar relationship between pin 2 and pin 1, but that is not guaranteed, and even if such relationship exists, it explicitly includes chassis or power supply common node voltage swings that you do not want as part of your audio signal.

 

[moamps]

.. But if you want to be stupidly pedantic then feel free to carry on FFS

That's how it always ends when the arguments run out.

 

[ruffrecords]

It is not. From any electronically balanced output you can take the signal from pin 2 referenced to pin 1. So the audio signal exsist between pins 2 and 1 also. It is just usually 6dB lower.

No, because pin 1 should not be connected to audio ground, it goes to chassis and then safety earth. The diff amp output 0V does not actually need to be connected to anything. The problem with diff amp inputs is they cannot often withstand the dc difference between send and receive ends.

Cheers

Ian

 

[moamps]

But that converts common mode noise to fully normal mode noise. By saying that "[t]he audio signal exists only between pins [3] and 2" it means that by definition the audio at pin 2 references pin 3. There may or may not also be a similar relationship between pin 2 and pin 1, but that is not guaranteed, and even if such relationship exists, it explicitly includes chassis or power supply common node voltage swings that you do not want as part of your audio signal.

Thank you for your post, you're absolutely right, it's not optimal but it's possible and used. It's not the first time I've come across that statement I commented on. For example, with many of my students, such thinking led to completely wrong conclusions about how to design the correct cable for connecting different types of compressors to an unbalanced insert input console. I won't go into that here any more, as it's obvious that I bore some privileged members when I insist on accuracy.

 

[moamps]

No, because pin 1 should not be connected to audio ground, it goes to chassis and then safety earth. The diff amp output 0V does not actually need to be connected to anything.

I don't agree and we can leave it at that. Anyone can check with a simple instrument if somewhere pin 1 is connected to the audio ground or not.

 

[Jen]

Lets add something else to this subject. The 'chassis' is a grounded faraday cage providing both shielding and shock protection for electronics held within. It is never supposed to be a conductor, (other than in radio transmitters). Audio common is contained within the grounded chassis and IS where the aidio circuits and power supply all connect. In some cases, that 'common' is connected to the chassis at one point, or brought out to a binding post near another binding post that IS connected to the grounded chassis, allowing selective connection of the ground to the 'common'. At various inputs and outputs, we sometimes see 2 or 3 pins of an XLR with capacitors to the chassis, not always a hard short of 1 to chassis... the caps are dumping RF to real ground and inhibit the egress of RF into the shielded enclosure. The outer shell of the XLR, pin 4 in essence, is always at chassis ground potential. Pin 1 is indeed the problem and even after decades of discussion still triggers lively conversations.. yeah its A problem because its function is not always as expected...., consider this:
In common 2 hot wire shielded cable, the shield carrys no signal at all, and does need to connect to the chassis as an extension of the faraday shield, that connection is more often done in the gear, rather than in the cable XLR, though it may be sometimes found in the cable. However, I have found with phantom power on mics (that runs return DC current through pin 1) it is better to use 3 inner conductors, not 2 as is usual practice , so that pin 1 can be routed to audio and power supply common inside the chassis, and reserve the shielding as pure shielding connected to shell/chassis. In this case the use of RF shunt cap on 1 to chassis is essential, along with a small ferrite just after the shunt. In such a case of cable with 3 internal wires and a shield with phantom power, pin one may or may-not be shorted to the faraday shield grounded chassis, if not, hard shorted, then it is at RF by the cap and ferrite network. Aint it grand, the complexities we endure, and the occasional rare exceptions to otherwise good rules.

 

[Chrisfromthepast]

https://centralindianaaes.files.wordpress.com/2012/09/indy-aes-2012-seminar-w-notes-v1-0.pdfI hope this link still works.
Im just a soundguy with a soldering iron, but this document usually solves my problems as it relates to this argument.
Page 40-72
Stop the unbalanced loops and you solve it all.

 

[Dennis F.]

Chris -
Thanks for the link. Very helpful.

 

[Newmarket]

Lets add something else to this subject. The 'chassis' is a grounded faraday cage providing both shielding and shock protection for electronics held within. It is never supposed to be a conductor, (other than in radio transmitters). Audio common is contained within the grounded chassis and IS where the aidio circuits and power supply all connect. In some cases, that 'common' is connected to the chassis at one point, or brought out to a binding post near another binding post that IS connected to the grounded chassis, allowing selective connection of the ground to the 'common'. At various inputs and outputs, we sometimes see 2 or 3 pins of an XLR with capacitors to the chassis, not always a hard short of 1 to chassis... the caps are dumping RF to real ground and inhibit the egress of RF into the shielded enclosure. The outer shell of the XLR, pin 4 in essence, is always at chassis ground potential. Pin 1 is indeed the problem and even after decades of discussion still triggers lively conversations.. yeah its A problem because its function is not always as expected...., consider this:
In common 2 hot wire shielded cable, the shield carrys no signal at all, and does need to connect to the chassis as an extension of the faraday shield, that connection is more often done in the gear, rather than in the cable XLR, though it may be sometimes found in the cable. However, I have found with phantom power on mics (that runs return DC current through pin 1) it is better to use 3 inner conductors, not 2 as is usual practice , so that pin 1 can be routed to audio and power supply common inside the chassis, and reserve the shielding as pure shielding connected to shell/chassis. In this case the use of RF shunt cap on 1 to chassis is essential, along with a small ferrite just after the shunt. In such a case of cable with 3 internal wires and a shield with phantom power, pin one may or may-not be shorted to the faraday shield grounded chassis, if not, hard shorted, then it is at RF by the cap and ferrite network. Aint it grand, the complexities we endure, and the occasional rare exceptions to otherwise good rules.

Do you have a diagram / sketch of that for clarity ?

 

[sahib]

I don't agree and we can leave it at that. Anyone can check with a simple instrument if somewhere pin 1 is connected to the audio ground or not.

Aside ccaudle's point, you are correct on this particular point.

Analogue/signal ground is connected to the mains safety earth at chassis stud point, ideally very near the IEC inlet.

Therefore through the mains lead of two (or more) equipment, their analogue/signal grounds are also connected together.

 

[Matt Syson]

There is a massive problem in that the term'balanced and the way 'balance' is implemented is VERY sloppy and when you consider that it needs to consider the frequency range that 'balance' is maintained it gets a hell of a lot worse. Life was so simple when it was mostly mains related 'interference' (Hum) and transformers with even mediocre 'balance' at frequencies above say 30KHz solved almost all problems. With switchmode supplies everywhere and people running around with transmitters (telephones) in their pockets it really shows up all the badly designed gear where 'balance' is NOT the only requirement to keep RF interference out (or indeed in). The RF engineers world has crash landed into 'audio land' and while 'hum bars' used to be a thing on video monitors (TV) widespread RF is now wreaking havoc on audio gear. pin 1 of an XLR should always be taken to chassis to maintain Faraday screening but in turn the inside of gear MUST allow for whatever MIGHT be connected to the outside and deal with it properly. Input circuits and output circuits have to be designed such that they do not demodulate 'RF imposed' on any connection, either 'balanced or unbalanced (differential AND common mode). an XLR is a reasonable 'balanced' connector (Audio to low RF) but any form of jack connector is hopeless at medium and high RF frequencies because their capacitance to 'shield is different (concentric conductor construction).

 

[thor.zmt]

I noticed that I was getting hum/noise from some of my outboard gear

As others have remarked, there are many possible sources for this him.

The suggestion of running a FFT (I use Wavegene and Wavespectra among other software) and then chasing the problem down to individual cables and devices by adding Device after device is spot on.

It is also useful to observe the spectrum of the noise, which can give hints regarding the source.

Hum dominated by 50/60 Hz is from mains (and usually earth loops), while 100/120Hz dominated hum usually indicates a power supply problem, either a fault from aging or a design flaw.

Some hum/noise came from my DIY outboard which I built with ground lift switches on them but they didn't reduce or stop any hum/noise when I toggled them.

This suggests the problem is not caused by an earthloop on this device.

I was reading online that having all my gear plugged into one power socket could be a solution

The solution depends on the problem. Using a single wall socket from which all equipment is powered via extension vmcables, long cables etc. certainly reduces some specific problems.

I am concerned that it may put a strain on one socket having all the equipment plugged that way?

Check the current ratings of the equipment and add up, or better measure it using one of these plug in power meters used on household appliances to find vampire loads and power hogs to bring down the electricity bill.

Add everything up. Usually main sockets are good for well over 10A which in the USA is over 1kW and in Europe and other 230V countries 2kW.

Be careful with cheap Chinese extension cables and distribution boards. Don't be a Wallmartian.

It may be worth to make a diy project of a mains distribution Unit (or more than one) using preferably neutrik powercon connectors (otherwise for USA and USA voltages use "hospital grade" outlets), extra heavy star wired Earth connections and Filter units for sockets equipment with switched mode power supplies is plugged into.

Make your own powercon mains cables and plan for a cascading set of distribution units. Mains cables should low contact resistance connectors (again, powercon scores big) and low resistance earth connections.

These can help get a few extra dB in noise reduction from ground/earth loops which are essentially unavoidable if equipment is used in a complex audio system while using equipment in compliance with electrical.safety regulations.

For extra credit, you can make an isolated balance power system, if all you are running is low powered gear, 500VA transformers with dual 55V secondary windings are commercially available. Only drawback, all gear must be able to run on 110V.

The benefit is that with the power supply now 55-0-55V you are operating everything at what is classed as "safe extra low voltage" in most applicable ordinances, laws et al.

This means you no longer need to comply with electrical safety regulations for mains powered equipment and you lift the safety earth connections (as one example) to your hearts content.

I read that using a UPS could be a possible fix so that I isolate the outboard gear and help stop the ground loop from happening.

Unlikely, a UPS must follow electrical safety rules, it could only help if the UPS is disconnected from the wall and supplying power from it's battery, which is usually accompanied by a lot of fan noise, is not very efficient and inconvenient. Never mind the risk of the unit having run flat in the middle of the perfect take.

It may still be worth adding a UPS if power outages and/or brownout are common.

Is there generally a rule on how you group your equipment to the-same power source to prevent hum/loop noise?

Yes. Starbearth for both safety and low noise. Minimum loop impedance for earth connections from equipment to outlet and between outlets.

But I suggest to first isolate and diagnose the problem before running around for solutions. The problem may very different from what you think.

When I read "tube ...." and especially if a little aged, then I'd first suspect Power supply issues. Modern standards and expectations for hum and noise differ notably from the 1960's.

Thor

 

[Newmarket]

It is also useful to observe the spectrum of the noise, which can give hints regarding the source.

Hum dominated by 50/60 Hz is from mains (and usually earth loops), while 100/120Hz dominated hum usually indicates a power supply problem, either a fault from aging or a design flaw.

+1

 

[scott2000]

Does the ESR of a capacitor have a different meaning in today's noisy environment? Meaning, does what is/was acceptable and considered "good" for a particular capacitance and voltage still hold true?

Are there things that can be addressed in older circuits to modify them to deal better with the wireless routers and such?

 

[Voyager10]

There is a massive problem in that the term'balanced and the way 'balance' is implemented is VERY sloppy

That's very true. Having 'pin 3' being the opposite of 'pin 2' has some useful advantages (e.g. greater output signal level for a given supply voltage) but it's nothing to do with interference rejection in a 3-wire system (which is: any interference affects both 2 and 3 equally, and the receiving input has good CMRR).

"Balanced" seems to be used for both ideas, which is perhaps why they end up confused.

 

[Matt Syson]

Are there things that can be addressed in older circuits to modify them to deal better with the wireless routers and such?
Not much really because the 'purist' nuts will tell you it is modified and therefore sounds 'wrong' so apart from careful attention to a bunch of small but significant issues things are what they are so you have to live with it. a 'Hum' at Abbey Road studios in London was 'cured' by fixing a Neutral to earth fault in a mains outlet in another room 3 floors higher in the building. Valve gear TENDS to be more resiliant against accidental demodulation of RF that has been picked up unlike semiconductor circuits that are all to happy to 'rectify' (demodulate) signals that are at a high level. even the OUTPUT of a transistor power amp CAN and will demodulate RF that come in via speaker cables so unplugging the INPUT actually achieves nothing. The negative feedback loop of the power amp can't handle externally applied RF. Good CMMR (common mode rejection) ONLY works properly if it extends from almost DC to GHz and includes the sending circuit, ALL cabling/connectors AND the receiving amplifier stage.
Bearing in mind the wavelength of modulated Microwave signals (mobile phones) it is a very tall order. Thus effective common AND differential mode filtering is necessary.

 

[thor.zmt]

Does the ESR of a capacitor have a different meaning in today's noisy environment?

Yes and no. The whole electrical behaviour must be considered. Generally, if you have an older piece of equipment, that has unacceptable power supply noise (Tube and transistors) replacing all electrolytic capacitors will usually lower noise, if the device is >> 10 Years old.

Modern electrolytic capacitors usually are smaller in physical size for the same value and voltage ratings and usually have lower ESR if you purchase a quality product. So don't be one Ali Baba's forty thieves.

If for a replacement Power supply capacitor you go with the same PHYSICAL size and make the value larger, you will reduce power supply noise substantially.

Only in devices with tube rectifiers it is necessary to be careful increasing the values of power supply capacitors that lack any significant resistance between cathode and Power supply capacitor. Solid state power supplies usually tolerate 10 or more times the capacitance without issues.

Meaning, does what is/was acceptable and considered "good" for a particular capacitance and voltage still hold true?

Up to a point yes, however often the state of the art has advanced and we have products that previously were not possible.

Are there things that can be addressed in older circuits to modify them to deal better with the wireless routers and such?

In order to preserve SNR here I split this off to here:

Are there things that can be addressed in older circuits to modify them to deal better with the wireless routers and such?

Thor