Grounding

[Folsom]

jkenny, I think Dave mentioned SCIN. The article linked doesn't say which one is better for audio (braid/foil). It would depend on what range of RF you have the most of in your location; and some other small factors.

To know all about proper grounding I'd prefer to refer you to a book, without more specific questions.

 

[jkeny]

-About 95% that had the boxes and liked them had measurable grounding problems in the design of their equipment.
-extremely difficult - I meant how many of your total customers, not just ones with the grounding boxes?
-ground loop problems would be one way. Getting zapped might be another - I find there are often far more insidious issues with grounding problems which are not always audibly evident until something better is heard
-any manufacturer in business today has access to proper grounding technique through the internet. Whether they choose to take advantage of the improved performance thus available is an entirely different issue. But they have the responsibility to make sure their equipment works and is safe to operate. -I don't believe it is as

We did not test any of the so-called grounding devices! What we did is have the owners of the equipment do some simple measurements to find out how the rest of the equipment in their system was grounded- if it was done correctly or not. For all the 'grounding box' owners we surveyed, 100% also owned amps or preamps or other gear that had bugs in their grounding schemes.

Ok, now your recipe for simple measurements would be good for those here to test their own equipment - care to share?

 

[amirm]

I find your comments don't auger well for your objectivity in any testing you intend to perform on the device. Why not save everybody the time, bypass testing & just declare the device no better than a rock from the garden - why the need for the measurement pretence?

 

[jkeny]

jkenny, I think Dave mentioned SCIN. The article linked doesn't say which one is better for audio (braid/foil). It would depend on what range of RF you have the most of in your location; and some other small factors.

To know all about proper grounding I'd prefer to refer you to a book, without more specific questions.

Ok, it was Dave.
The article clearly states that foil + braid is the best solution for wideband shielding. It also says to avoid cables with a drain wire.
I have read quite a bit about grounding but don't pretend to be an anything close to an expert. I know that proper grounding in slower speed devices was a far simpler & intuitive proposition than the high speed devices that we use in our systems now. One of the biggest offenders in this scenario are computers & their high speed connections to our audio systems.

 

[Purite Audio]

95% of equipment measured poorly, there are more charlatans in this hobby than even I had imagined.
Keith.

 

[spazmatron]

the charlatans can make good music...

 

[Folsom]

-About 95% that had the boxes and liked them had measurable grounding problems in the design of their equipment.
-extremely difficult
-ground loop problems would be one way. Getting zapped might be another
-any manufacturer in business today has access to proper grounding technique through the internet. Whether they choose to take advantage of the improved performance thus available is an entirely different issue. But they have the responsibility to make sure their equipment works and is safe to operate.


Your comment seemed to imply that the grounding problem could be heard but at the same time could not be detected unless the equipment was opened up. It could be interpreted other ways as well so I thought you might want to clarify.

Here's a link to an Elgar unit on Ebay:

http://www.ebay.com/itm/ELGAR-5000B...823890?hash=item5429d11fd2:g:TVkAAOSwXshWqoHU

This unit uses an enormous isolation transformer, with a feedback winding. A high precision low distortion oscillator is synchronously locked to the line frequency; a comparator looks at its output vs. that of the isolation transformer. Then a feedback amplifier applies the resulting feedback signal to the feedback winding. The result is line voltage regulation and an utter lack of spikes, hash, with no harmonics of the line frequency (FWIW the 5th being the most deleterious BTW). So it can produce a clean sine wave right up to full power, which is in this case enough to run any audiophile system made- the unit is enormous. Its also physically noisy and so should be placed near the breaker box and then feed the audio room. All 'high end' audio conditioners fall flat on their faces by comparison. Smaller versions were made.




We did not test any of the so-called grounding devices! What we did is have the owners of the equipment do some simple measurements to find out how the rest of the equipment in their system was grounded- if it was done correctly or not. For all the 'grounding box' owners we surveyed, 100% also owned amps or preamps or other gear that had bugs in their grounding schemes.

Its my experience that when equipment is improperly grounded, its more susceptible to noise and ground loop intermodulations (read that as 'distortion'). So its really not a surprise that a box that does something with wires and the ground might have some sort of audible effect. Whether its an improvement or not is another matter and not germane to my study. At the same time, it is certainly an improvement when the equipment is properly grounded- one effect immediately audible is lower background noise, which should come as no surprise. But you don't need a box to get that, you just need the equipment to be built correctly in the first place.

Please explain the procedure for measurements, for the people with the boxes.

The Entreq's can only enhance a stereo. They can't correct a problem. It's possible they enhance one stereo more than another, but it's never a fix.

You've lumped ground and AC noise together. I'm not sure how to reply.... To check grounding yes you'd need to look in the equipment as a consumer, generally. Could you hear it? you'd only know if it came across as an artifact like a 60hz hum. For AC noise that can be contributed in part by grounding issues, again you don't know it was there by ear until it's gone.

 

[Folsom]

Ok, it was Dave.
The article clearly states that foil + braid is the best solution for wideband shielding. It also says to avoid cables with a drain wire.
I have read quite a bit about grounding but don't pretend to be an anything close to an expert. I know that proper grounding in slower speed devices was a far simpler & intuitive proposition than the high speed devices that we use in our systems now. One of the biggest offenders in this scenario are computers & their high speed connections to our audio systems.

Change the word speed to frequency and you'd be better off. The higher the frequency the more complicated grounding is due to inductance.

I know what the article says, but I'd prefer to judge said cables by ear. For example FM frequency is much better blocked with a foil and shield since it's way above 25mhz. Also wireless devices pollution is even higher. The question is which potential noise is worse for the music? That's where the ears let us know. Perhaps the braid that deals with more noise that a stereo's cable and appliances will radiate sounds best, perhaps it does not. My personal experience thus far is not in favor of coaxial construction because I've never heard one that sounded particularly good.

 

[DaveC]

Ok, lets define what good component grounding is and how to test for it. There are some differences of opinion... I have uploaded a Rane paper on the subject, in figure 3 it gives several possible grounding schemes between signal and chassis.

One of my favorite designers, John Broski, advises to directly connect signal and chassis and isolate this ground from safety/house/earth ground using diodes or a cap (.001-.1 uF) and resistor (10R) paralleled... a Schurter DENO (ground isolation choke) could also be placed in series with any of these schemes, signal > chassis or (signal/chassis) > safety ground.

Is there a consensus as far as what's best and why you'd use different forms of isolation?

As far as testing, if you believe signal and safety ground should be isolated, you can simply test the resistance between an input jack's ground and the ground pin on the IEC. That's a pretty simple test! And between chassis and safety as well to see if they are isolated. Doing this should provide pretty easy assessment of how the component is grounded.

 

[Folsom]

The connection from chassis to safety ground from IEC socket needs to be high enough current to deal with a dead short. In otherwords that 10R of the R||C needs to be big wattage (which raises it's inductance a lot). It's more common to connect circuit ground with chassis using the resistor and capacitor for that reason.

For legal obligations you need to have safety ground attached to chassis if the circuit ground is connected to chassis in any way. If it is not, you can forgo the safety ground all together (like your Sony server Dave).

There's compilations that come up all ways.

First off if chassis and circuit have no connection then parasitic capacitance can share varying degrees of noise; safety ground to chassis may help but not 100%. However connecting chassis to circuit has it's own complications to avoid parasitic qualities as well. Sometimes having the RCA's ground not isolated to chassis works for noise on the incoming cables, but then allows the chassis to act something like an antenna toward the circuit. That's why the R||C is common. However sometimes even the R||C leaves an issue. If a circuit's impedance is very low and any path is opened up you can inject RF through the parasitic means even more so than any other alternative! That is more specific to devices that do signal processing and are literally right next to the chassis/isolated by mica or such.

There are some better practices, but best choices are related to the equipment. A more interesting discussion is grounding on the PCB board, the paths, loops, etc... This is where more mistakes than anywhere occur.

 

[Speedskater]

For engineering type people, interested in grounding and shielding, $15 USD (including shipping) will get you this 85 page journal. I would think that it's available from the AES in other countries as well for a reasonable cost. And it includes the John Windt "Hummer Tester" paper that's not obtainable anyplace else.

SHIELDS AND GROUNDS: SAFETY, POWER MAINS, STUDIO, CABLE AND EQUIPMENT
The June 1995 issue of the Journal was a definitive and comprehensive collection of information on this important topic. The seven papers by Neil Muncy and other experts in the field have been reprinted into a convenient guide for designers and practitioners.

http://www.aes.org/publications/specialpubs/journal_issues.cfm

 

[Atmasphere]

Ok, lets define what good component grounding is and how to test for it. There are some differences of opinion... I have uploaded a Rane paper on the subject, in figure 3 it gives several possible grounding schemes between signal and chassis.

One of my favorite designers, John Broski, advises to directly connect signal and chassis and isolate this ground from safety/house/earth ground using diodes or a cap (.001-.1 uF) and resistor (10R) paralleled... a Schurter DENO (ground isolation choke) could also be placed in series with any of these schemes, signal > chassis or (signal/chassis) > safety ground.

Is there a consensus as far as what's best and why you'd use different forms of isolation?

As far as testing, if you believe signal and safety ground should be isolated, you can simply test the resistance between an input jack's ground and the ground pin on the IEC. That's a pretty simple test! And between chassis and safety as well to see if they are isolated. Doing this should provide pretty easy assessment of how the component is grounded.

I don't think there is a consensus, but if the technique is used, generally speaking it will work. You do also want it to hold up in case of adverse situations!

The one problem we run into here is that often single-ended equipment gets mixed with balanced. The two technologies are incompatible. The result can be a ground loop despite the fact that both types of equipment employ good grounding technique. Yet in high end audio both have to play well together. The result is usually that the balanced line standard (AES file 48) gets abridged somehow. Most manufacturers do this (knowingly but mostly not knowingly) by having the outputs of pin 2 or 3 of the XLR have that output with respect to pin 1 rather than its opposite (either pin 3 or pin 2). We do it by isolating the chassis and circuit ground (which is easily restored obviously by connecting chassis to pin 1; a much easier thing to fix than the prior technique more commonly used; in this way we retain cable immunity).

 

[amirm]

For engineering type people, interested in grounding and shielding, $15 USD (including shipping) will get you this 85 page journal. I would think that it's available from the AES in other countries as well for a reasonable cost. And it includes the John Windt "Hummer Tester" paper that's not obtainable anyplace else.

SHIELDS AND GROUNDS: SAFETY, POWER MAINS, STUDIO, CABLE AND EQUIPMENT
The June 1995 issue of the Journal was a definitive and comprehensive collection of information on this important topic. The seven papers by Neil Muncy and other experts in the field have been reprinted into a convenient guide for designers and practitioners.

http://www.aes.org/publications/specialpubs/journal_issues.cfm

Thanks. I just bought it for $10 since I am a member. Didn't realize they were papers I could download separately for free though . Regardless, seems like a very nice collection so I have plenty of bedtime reading to do.

 

[Speedskater]

OK, lets define what good component grounding is and how to test for it. There are some differences of opinion...
Is there a consensus as far as what's best and why you'd use different forms of isolation?

This question comes up from time to time in different threads over at the DIY forum. There are strong disagreements as to what is good or safe or legal.
Unfortunately it would take a person with great searching skills to find the threads or the posts in the forever long threads.

I managed to find one thread.
http://www.diyaudio.com/forums/soli...nagement-power-amplifiers-what-royal-way.html

 

[Speedskater]

Thanks. I just bought it for $10 since I am a member. Didn't realize they were papers I could download separately for free though . Regardless, seems like a very nice collection so I have plenty of bedtime reading to do.

It's the June 1995 AES Journal. All the original convention papers were revised for the journal (don't know how much).

 

[Atmasphere]

OK, lets define what good component grounding is and how to test for it. There are some differences of opinion...
Is there a consensus as far as what's best and why you'd use different forms of isolation?

This question comes up from time to time in different threads over at the DIY forum.l

The means that I laid out earlier is correct. The exact values might vary a bit which is fine. So yes, there is a consensus- amongst those that are interested in getting the ground to work right.

It comes up from time to time on DIYaudio because people don't know.

 

[amirm]

It's the June 1995 AES Journal. All the original convention papers were revised for the journal (don't know how much).

As a member I can download journal papers for free just the same. The download I got is an optical scan and slow to view but again, the information is good so don't mind the $10 I spent .

 

[Speedskater]

They really are grounding boxes and they really work.
Grounding = Bonding = keeping various metal parts at the same potential. Reducing the impedance between parts reduces SCIN, shield current induced noise, and results in less noise. This is indisputable and the fact grounding boxes do this is also indisputable.......................................

Jim Brown the go-to expert on noise and interference, just wrote this in a Ham radio group.

You are confused about the difference between "grounding" and "bonding." I suggest that you study this tutorial on the subject. A year or two ago, N0AX used it as the jumping off point for a series of QST columns. http://k9yc.com/GroundingAndAudio.pdf

In short, that wire you're using to "ground" your computer is acting as an antenna to radiate the RF trash that's on it's chassis. What you need is bonding. A connection to the earth is NOT part of a solution to hum, buzz, or RFI, and, as you have seen, can often be a cause of problems. It is needed only for lightning safety. Jim Brown
73, Jim K9YC

 

[DaveC]

Jim Brown the go-to expert on noise and interference, just wrote this in a Ham radio group.

You are confused about the difference between "grounding" and "bonding." I suggest that you study this tutorial on the subject. A year or two ago, N0AX used it as the jumping off point for a series of QST columns. http://k9yc.com/GroundingAndAudio.pdf

In short, that wire you're using to "ground" your computer is acting as an antenna to radiate the RF trash that's on it's chassis. What you need is bonding. A connection to the earth is NOT part of a solution to hum, buzz, or RFI, and, as you have seen, can often be a cause of problems. It is needed only for lightning safety. Jim Brown
73, Jim K9YC

Thanks, I'll check out the article... Bonding and grounding are often used interchangeably, I used bonding because that gets across the point of keeping component ground planes at the same potential while grounding seems to be a confusing term and leads people to believe the term in the literal sense, which I think we are both in agreement is not the case, as demonstrated by the last paragraph of your post.

Just looking at the first few slides, bonding is what I meant... but the terms are often used interchangeably. Grounding in his presentation means safety ground. I think this is a better way of describing the terms but people are going to use grounding when they mean bonding frequently.

EDIT: Further, the ground boxes accomplish step 2 of the hum/buzz reduction in the pdf. Step 1 is feeding the system from a source using the same ground, which many people do not do when they run multiple lines.

 

[Folsom]

I've found PG and SG to be useful terms. Power and signal ground. One reason is that where these two meet may have a huge influence on the sound. They're also distinct as part of the equipment not associated with enclosure/chassis.