AC Polarity and the reduction of noise and hum

[YashN]

Electrical codes should be respected, but they changed along the years. In the late 80's, most audio equipment coming from the US come with a two wire mains cables - today it will be immediately arrested at the customs!

They also vary depending on countries.

 

[DonH50]

Steve, I don't doubt it, and I am sure there are many that have no trouble including the majority of other systems I have set up for myself and others. Prior to this (in a different house) I ran a couple of (very!) cheap 20' (6 m) cables to my power amp and had no problems. Like I said, most don't need it, including me, most of the time. I think this is the second time in decades when I have had to go balanced in my system to eliminate hum. I could have used an isolator, natch, but I have XLR I/O at my pre/pro and subs and cables so went ahead and used them. My run is two 50' (~15 m) cables to the subs and they are on a different AC circuit. SE they hum; XLR, no hum. I am not really worried about signal loss or even coupled noise in this install, but it was an easy way to break the ground loop without purchasing anything extra.

There are technical advantages to balanced circuits but, like so many things, in practice and at audio frequencies the theory does not make it audible. There are disadvantages, too, mainly higher cost and power as well as the aforementioned space considerations. I suspect some of the negative sonic comments are related to either a poor balanced implementation (they are not all the same, and "quasi" circuits can be problematic in some cases), or the normally higher balanced output means gain gets turned down and noise is a bigger issue (the noise advantage is not if you have to turn the gain way down). The latter I have seen more when consumers use pro gear, which often expects much higher (maximum) signal levels.

FWIWFM - Don

 

[YashN]

I would guess the main reasons balanced connections are not used more widely are due to the cost, the lack of readily-available (and thus cheap) differential drivers, and panel area (XLRs are bigger than RCAs). Also the chicken and egg bit; nobody asks and other components don't have them so why should I?

Balanced, assuming fully differential, circuits offer lower distortion (cancellation of even order terms), higher SNR (3 dB for each doubling of signal level), and greater noise etc. rejection than comparable single-ended circuits, but XLRs cost more and more importantly the circuitry inside is more complex and costly. The vast majority of op-amps have single-ended outputs, though many DAC chips (the actual device inside, not the box) have differential outputs to reduce glitches, lower distortion, and all that jazz.

For me having a balanced output means I can run long lines to a couple of rear subs on a different AC circuit without ground-loop and noise problems but by and large balanced is not needed in consumer installation. That is not to say some would not benefit, of course, and a lot of folk struggling with ground loops likely wish their components had balanced I/O -- if they knew to ask for it. XLRs also offer positive capture and retention which may or may not matter to most. I would guess single-ended is fine for most people so not enough demand.

Great succinct explanation, Don.

Just as reducing Leakage Current with the method of swapping plug orientation, Balanced vs SE can have a big difference in SQ, and you don't even need a long run to notice it. It is a night-and-day difference, so it must be tried.

With my studio monitors, which are active and bi-amped each, I had the opportunity of comparing SE to Balanced on initial setup, and there was such a big difference that I never used the SE in my studio since.

In this respect, some of the high-end gear lags behind the studio gear.

It's a must-try, especially a true, properly Balanced implementation (i.e. without Pin 1 issue - some gear using XLR have this issue so give a distorted perspective of what Balanced should be, which was why Neil Muncy - RIP - was traveling around to show manufacturers and others how to resolve this).

Another thing one could try is Balanced Power, there are caveats to it one needs to be aware of regarding safety.

Another way to reduce Leakage Currents is to use a "disconnected" PSU like the Uptone Audio UltraCap LPS-1 or the Vinnie Rossi Pure-DC-4ever PSU.

There is another more complex form of Leakage Current which is less known but which can only be characterised by the interaction between at least 2 PSUs and their circuits in a chain and aided by the interconnects.

Clean Power and clean grounding is a complex matter, there is a fair chance that people doing default connections aren't getting to hear their gear properly.

Some SQ effects are also rather insidious as some may get used to them after a while. However, a rudimentary AB definitely shows the difference when changing configurations.

 

[RogerD]

What I have read is SE is good for up to 20 feet. Balanced is used generally for > than 20 feet. Btw one of my AG440 machines electronics has been converted to use SE . The engineer stated that he found the SE was quieter. I have used both and essentially The heavier grounding made the type of connection moot in my system.

 

[DaveC]

What I have read is SE is good for up to 20 feet. Balanced is used generally for > than 20 feet. Btw one of my AG440 machines electronics has been converted to use SE . The engineer stated that he found the SE was quieter. I have used both and essentially The heavier grounding made the type of connection moot in my system.

Like most things, implementation is most important!

 

[DonH50]

50 feet of pretty cruddy cable driven from a 100-ohm source (fairly typical SS preamp output) still has >1 MHz of bandwidth and negligible loss (<0.1 dB) into a high-impedance (10 k-ohm) amplifier input. I wouldn't run it that far in a live sound venue but in a home unless there is significant noise pick-up it should be fine. IME the problem with SE connections in homes tends to be coupled noise -- from other lines (power and signal), light dimmers, EMI from motors and other stuff in the house, etc. -- and ground loops. A good balanced connection can offer much higher noise rejection not only from the intrinsic common-mode rejection of the differential signal wires but also because the outer shield (ground) no longer carries signal current, better isolating the inner conductors from external noise. Current always forms a loop so, when the outer shield of an interconnect is modulated by noise, the receiver at the other end cannot tell noise from signal. Whether, or how much, any of this matters is a function of the myriad variables affecting your own installation. I have had 10 feet of SE cable cause problems in one system, and 100' work just fine in another.

 

[microstrip]

(...) Balanced, assuming fully differential, circuits offer lower distortion (cancellation of even order terms), higher SNR (3 dB for each doubling of signal level), and greater noise etc. rejection than comparable single-ended circuits, but XLRs cost more and more importantly the circuitry inside is more complex and costly. The vast majority of op-amps have single-ended outputs, though many DAC chips (the actual device inside, not the box) have differential outputs to reduce glitches, lower distortion, and all that jazz.(...)

In the audiophile world XLR plugs cost much less than RCA's - many cable manufacturers use expensive RCA's and inexpensive Switchcraft or Neutrik gold plated pin XLRs.

IMHO in the high-end what matters most is what I quoted in bold - these improvements carry a sound signature, and some designers prefer other sound signatures, so they choose single ended. Fortunately there was never a Carver test on single ended versus balanced, otherwise we would be also debating it!

I am astonished that Ralph Karsten of Atmasphere is not chiming - he would explain us why XLRs are used for balanced mode, but most equipment using them with positive and negative phase are not truly balanced differential.

 

[BruceD]

Maybe we all need one (or two :b)! of these?

Jarvis Isolation Transformer

BruceD

 

[DaveC]

In the audiophile world XLR plugs cost much less than RCA's - many cable manufacturers use expensive RCA's and inexpensive Switchcraft or Neutrik gold plated pin XLRs.

IMHO in the high-end what matters most is what I quoted in bold - these improvements carry a sound signature, and some designers prefer other sound signatures, so they choose single ended. Fortunately there was never a Carver test on single ended versus balanced, otherwise we would be also debating it!

I am astonished that Ralph Karsten of Atmasphere is not chiming - he would explain us why XLRs are used for balanced mode, but most equipment using them with positive and negative phase are not truly balanced differential.

IMO balanced needs to be implemented with low overall distortion, otherwise the cancellation of even order harmonic distortion while leaving the odd order unaffected makes the sound clinical and annoying. With low distortion components like PranaFidelity's pre and amp this works really well, it sounds great. We also haven't mentioned the fact that passive components in balanced circuits should be matched closely and tight tolerance parts cost more. Despite that there's still potential for the + and - legs to be treated differently. There is also a potential for the legs to get off due to uneven input impedance of the most basic differential inputs, if you have, say, a passive balanced attenuator going into a basic differential input it's not going to work out, the input would need to be buffered to insulate it from those effects.

On plugs it's actually hard to find good sounding XLR plugs, the SC and Neutrik offerings aren't high end and sound like it. Furutech's 700 series are made by another major mfg, I forget who offhand but they are better as they don't add harshness, but they are too smooth, a better trade-off ime but still a major trade-off. They are inexpensive though... Furutech 600 and CF series are excellent but $$$, Xhadow are also very good, silverplated is not my preference for platings but they are nice. There are also some ridiculously expensive options that are ridiculous. It's easier to find decent sounding, reasonably priced RCA plugs. Once you've tried Furutech CF plugs it's hard to accept anything else but not everyone wants to spend that kind of cash.

 

[DaveC]

Maybe we all need one (or two :b)! of these?

Jarvis Isolation Transformer

BruceD

View attachment 30749

That looks nice, smaller iso trafos sound bad though, big ones are $ and if you're going to do it you might as well go with a balanced-power trafo with 20A capacity.

 

[microstrip]

(...) On plugs it's actually hard to find good sounding XLR plugs, the SC and Neutrik offerings aren't high end and sound like it. Furutech's 700 series are made by another major mfg, I forget who offhand but they are better as they don't add harshness, but they are too smooth, a better trade-off ime but still a major trade-off. They are inexpensive though... Furutech 600 and CF series are excellent but $$$, Xhadow are also very good, silverplated is not my preference for platings but they are nice. There are also some ridiculously expensive options that are ridiculous. It's easier to find decent sounding, reasonably priced RCA plugs. Once you've tried Furutech CF plugs it's hard to accept anything else but not everyone wants to spend that kind of cash.

Thanks. All the equipment I own has Neutrik gold plated connectors, so I can not see any advantage in using a different type of metal for the contact. I was told very nice things about the expensive Furutech XLRs (Crystal cables uses them the in their ultra-expensive top cables), but never assembled similar cables with different plugs to compare XLRs.

 

[DaveC]

Thanks. All the equipment I own has Neutrik gold plated connectors, so I can not see any advantage in using a different type of metal for the contact. I was told very nice things about the expensive Furutech XLRs (Crystal cables uses them the in their ultra-expensive top cables), but never assembled similar cables with different plugs to compare XLRs.

The gold plated Neutriks aren't bad, better than the silver plated ones, they can add some harshness and accentuate leading edges. The Furutech CF uses rhodium plated pure copper, definitely worth a try. The fancy housing actually makes a big difference, the 600 series are the same with the exception of a brass body, but the CF are noticeably better. The increase in resolution and better tone are worth it over the Neutriks.

 

[YashN]

What I have read is SE is good for up to 20 feet. Balanced is used generally for > than 20 feet

My studio monitors connection was far shorter than that since they were nearfield...

 

[YashN]

Maybe we all need one (or two :b)! of these?

Jarvis Isolation Transformer

A good isolation transformer is definitely a great way to reduce Leakage Currents.

This one is huge, but most setups can get back using smaller ones. If you do go this route though, it is important to find an isolation transformer that would contribute little to parasitics. That means looking for one where the stated inter-winding capacitance is really low.

 

[YashN]

Take seriously what the man wrote,
AC shock can be quite serious. Don't mess with the AC with cheater plugs and whatnots... We are talking about safety here: Life and deaths issues. Not trivial sonic pursuits.

Feel free to listen to him yourself if you believe his advice is good...

There isn't any cheater plug when the device is two-pronged as I mentioned.

This isn't about safety here, and nobody was saying to ignore safety in this thread but of course, some are hell-bent on making it about this when it isn't.

The same people have offered zero appropriate advice on actual reduction of Leakage Currents.

 

[mcduman]

earlier in the thread i posted the results of my measurements with normal and inverted polarities: http://www.whatsbestforum.com/showt...rity-and-the-reduction-of-noise-and-hum/page4

my pre (an Ayre Twenty series unit) has the following readings: Normal 39.2V. Inverted. 84.8V. So, has lower leakage in the "normal"position" which I have been using. But as per amir's post the question now is what are we trying to minimize: the leakage in individual units OR the difference between units. because if i swap the orientation of the plug for the Ayre i get a reading which is very close to my DAC (100V) and the difference is now 15V whereas for the "correct" orientations it was higher at 60V.

I measured chassis to chassis voltage differential between the units for both orientations and got lower readings (.27mv) when the difference is minimized rather than individual units (.32mV). so, what do i do now? (cables are balanced)

 

[amirm]

That's the funniest thing I've read today. Keep it coming

That wasn't funny. This is funny:

Was at a recent audio show chatting with the designer who had just built the mother of all amplifiers (Boulder class but not Boulder). He was showing a prototype and said that he was planning on taking out the unbalanced connections and only ship with balanced. I asked him if he was not concerned with audiophile belief that unbalanced sounded better. He grinned his teeth in anger and said for that reason, he may have to leave unbalanced in there.

So go ahead and use this misguided interface designed some 90 years ago by RCA to hook up a "phonograph" to their radio. Me? I like a design where the chassis ground is not used for signal (unbalanced). Differential signaling is where it is at and in high-end systems costing tens of thousands of dollars, it better be standard issue or I walk. We wouldn't be having this topic if we had all moved to differential.

 

[amirm]

FWIW Don I run a 9 meter SE IC from my preamp to my amp and there is zero noise or hum. When I questioned my manufacturer he said he will never use balanced in his components. Further he said if I had any noise, hum or loss of signal he would refund my purchase price in total. I've never had a problem. To say that their consumers aren't smart enough to ask for balanced is ludicrous in consumer systems that are done right

I really like Vladimir and now have even more respect for him for using his sales acumen to get you past your objection. Thankfully we are all so deaf in low frequencies that even when hum is electrically there, we don't hear it until it gets to elevated levels.

Here is the deal though and there is no getting around it. Ground loops can occur when you change cables, equipment, distance between gear, etc. If you develop such a problem three years from now, you really think a dealer will give you your money back and have the salesman cough up the commision he made to go to Hawaii? I don't think so.

Advocate excellence and I will be with you Steve. Advocate good enough for you and you will have to go it alone.

 

[steve williams]

I really like Vladimir and now have even more respect for him for using his sales acumen to get you past your objection. Thankfully we are all so deaf in low frequencies that even when hum is electrically there, we don't hear it until it gets to elevated levels.

Here is the deal though and there is no getting around it. Ground loops can occur when you change cables, equipment, distance between gear, etc. If you develop such a problem three years from now, you really think a dealer will give you your money back and have the salesman cough up the commision he made to go to Hawaii? I don't think so.

Advocate excellence and I will be with you Steve. Advocate good enough for you and you will have to go it alone.

You're getting even funnier. Keep it coming

FWiW I choose never to go with you. Why don't you spew your so called acumen over in your little spot of the Internet. I'll take my system over yours any day. Oh but lest we forget no one knows what your system is let alone ever heard it. Keep on trucking. You've convinced a little over 200 people in your middle earth. I suggest you hang out there where the chosen few seek your so called knowledge. You add nothing here.

 

[microstrip]

earlier in the thread i posted the results of my measurements with normal and inverted polarities: http://www.whatsbestforum.com/showt...rity-and-the-reduction-of-noise-and-hum/page4

my pre (an Ayre Twenty series unit) has the following readings: Normal 39.2V. Inverted. 84.8V. So, has lower leakage in the "normal"position" which I have been using. But as per amir's post the question now is what are we trying to minimize: the leakage in individual units OR the difference between units. because if i swap the orientation of the plug for the Ayre i get a reading which is very close to my DAC (100V) and the difference is now 15V whereas for the "correct" orientations it was higher at 60V.

I measured chassis to chassis voltage differential between the units for both orientations and got lower readings (.27mv) when the difference is minimized rather than individual units (.32mV). so, what do i do now? (cables are balanced)

IMHO now you should follow Nelson Pass advice - decide for yourself. Listen and if possible then tell us about your findings - it is the purpose of audio forums. The success or not of these practices depends on your particular mains installation, mains RF contamination and equipment. Also, remember that most balanced equipment has poor rejection outside the audio band, or is implemented poorly. Considering your excellent equipment I suppose your are not looking for hum rejection, but subjective sound quality.