Thursday 24 April 2014

Power Struggles

My recent post regarding the PS Audio Jewel C7 Power Cord I got for my Mac Mini has prompted a number of comments, emails and PMs.  Let’s face it, anything you write on Power Cords is guaranteed to have this effect!  More than one of you have inquired as to the effects that I think may be responsible for the audible results, and so I though a post on that broader subject may be of interest.  Having said that, if you are looking for answers, this is the wrong place.  I don’t have them.

Many people object to the notion of the audibility of power cords almost as a fundamental point of principle.  After all, if a power cord is audible, so the argument goes, then so is your house’s wiring, and the power company’s wiring that delivers the power to your house, and so on.  But there’s the rub.  I have no doubt that these things ARE all audible.  It’s just that there are some things that you can easily upgrade, and some things that you can’t.  Your country’s Power Utility generally falls into the latter category.

In principle, any audio device which is electrically powered can be viewed as a device that draws an electrical signal from the mains, and shapes it to its own purpose.  Take a power amplifier.  It doesn’t really amplify the input from its preamplifier.  What it does is to take the mains power, and use it to fashion a bigger copy of the input signal.  The difference may be arguably semantic, but as someone who spent many years working with Optical Amplifiers I thought the distinction was interesting.  Therefore, it maybe helps to think that just about everything that comes out of an electronic device came into it in the first place via its power cord.

So it might be more rational to take the position that a power cord cannot possibly NOT affect the sound.  Of course you can then argue that any effects ought to be, for all practical purposes, inaudible.  That would be a rational philosophical starting point.  It at least forces the skeptics into the position where they have to stipulate the effect, and elaborate upon a rationale for supposing it inaudible.

What sonic effects can a power cord actually exhibit?  In principle, the mains voltage waveform is a 50Hz or 60Hz pure sinusoid.  However, the current that flows along a power cord is not so simple.  Inside the audio apparatus the power cord first connects to a Power Supply.  The power supply draws current from the mains, via the power cord.  This current does not flow in a nice sinusoid.  Rather, it flows as a series of spikes.  Imagine being 5 years old again, and blowing though a straw into a glass of water.  The air that you blow through the straw - into the water - flows at a more or less constant rate.  But once the air gets into the water it forms bubbles.  The bubbles rise to the surface where they hang around for a while and then burst.  The air is actually released back into the atmosphere in a series of sudden events, as each bubble bursts.  A Power Supply is a bit like that in reverse.  It consumes the current it needs in a series of very short spikes, like bubbles bursting.

You might choose to design a power cord based on allowing those current spikes to travel more freely along the power cord.  But I’m not sure about that, and I have two sets of reservations.  The first is that I’m not sure that an everyday Belden power cord (such as is shipped with most high-end audio equipment) is particularly deficient in that regard.  The second is that, based on how a power supply works, it is not clear to me how bandwidth limitations imposed upon the current spikes (which is one thing that a power cord might be doing) translates into audible effects in the equipment’s functionality.  Of course nothing I’ve said proves that it doesn’t.  It’s just that I don’t clearly see clear technical arguments for supposing that it does.

The other way to look at a power cord is to consider things that are transmitted along the power cord that you don’t want to see there.  And the biggest bugaboo is Radio-Frequency (RF) interference.  RF is pervasive.  Unlike ‘ordinary’ electrical signals, it is not constrained to traveling along the wires.  If it wants to, it is perfectly capable of jumping out of the wires and traveling through open air.  This, after all, is how our Cell Phones, Radios, Wi-Fi, GPS, and Microwave Ovens all works.  You wouldn’t believe how many ordinary household appliances generate RF interference.  Even those which are not designed to work in the RF.  Like your refrigerator, dimmer switches, air conditioning, even the street lights outside your house.  Add to these your computers and TV sets, and you can see how your household mains wiring can soon be seriously contaminated by RF interference.  If you live in a densely inhabited urban environment, it can be quite serious.

At BitPerfect, we took the concept of RF interference quite seriously.  Our Power Cords were designed to transmit all frequencies - including RF - down the cable as cleanly as possible, and to accept the reflected RF reflected from the power supply’s input RF filter, and transmit it effectively back down the power cord and back into the mains wiring where it will eventually dissipate.  That’s one design approach, and we don’t make any great claims for it, except that it seems to work rather well.

A lot of focus on power cords seems to be based on the cable construction.  But our experience is that the connectors at either end are very significant contributors.  The connector has two jobs to do.  First it has to connect securely to the internal cabling, and that connection itself is a major aspect of cable design.  Second, the connector has to mate satisfactorily with the receptacle into which it is plugged.  This latter is not something which is under the designer’s control.  And it begs a question that is applicable to all sorts of cabling applications - to what extent does a cable comprising a high-end connector require a matching mating connector to plug it into?  My experiments suggest that this is an important consideration.  (This is something I struggled with when designing BitPerfect’s own XLR interconnect cables.)

I want to also mention something that is seen regularly in specifications for cables, and most particularly power cords.  “Cryogenic Treatment” requires that the metal parts comprising the connectors, and often the cable itself, be exposed to “cryogenic” (i.e. seriously, seriously cold) temperatures for a period of several hours.  What this is supposed to do is unclear, but it is known that the crystalline structure of some metals can be altered by such cryogenic treatment.  But “can alter” is not the same as “does alter”, and I have yet to see any definitive reports of actual physical transmutation being observed on cryo-treated commercially-sold audio components.  I have purchased both cryogenically treated and untreated power connectors, and I’m sure I can’t hear the blindest bit of a difference.  I spoke to one of those vendors, and they were happy to tell me in great detail exactly how and where they perform their Cryo treatments (the facility they use is one that I am actually familiar with), so I am quite convinced that they actually do what they claim to be doing.

The bottom line for me is that the effect a power cord will have on your system will depend to a large extent on how “clean” your power source is to begin with.  For example, I spent some time with a MIT power conditioner.  First in a downtown hotel room at a Hi-Fi show, where it undoubtedly cleaned up the sound quite appreciably.  Second, in my listening room at home, where my main power is apparently a lot cleaner to begin with, and the MIT made hardly any audible difference.  I have also used an industrial UPS as a power conditioner with mixed results.  It gave a smoother sound, but also a notably less dynamic one.  I used it for nearly a year, but when it expired I chose not to replace it.  If you live in an urban environment, a high quality power conditioner such as those sold by PS Audio, MIT, or Transparent Audio should be considered seriously before you spend a similar amount on power cords.

I think money spent on buying good quality wall sockets is money well spent.  Mine cost me about $10 each from my local “Reno Depot” (a big-box hardware store).  What you are looking for is a “K-Line Contact”.  Peer into a wall socket, and you will see that when the blade of the plug is inserted it slides parallel to one flat surface, and is pressed into contact by an angled blade.  In a cheap socket there is one angled blade, which makes point contact, not line contact.  What you want to see is two angled blades in a “K” shape, each making line contact.  When you insert a plug into this, it is going to take quite a bit of force to pull it back out again.  You can spend a lot more, and buy sockets with Rhodium- or Gold-plated connectors, but I don’t see the point unless you will be plugging Rhodium- or Gold-plated plugs into them.  Whatever you do, don’t spend hundreds of dollars on power cords without installing decent wall sockets first.

When you finally do buy power cords, be sure to work with a reputable dealer who will let you try before you buy.  Don’t, don’t, repeat DON’T buy one on-line because it was favorably reviewed or recommended.  Power cords are all about synergy, and finding something that works well in your system is paramount.  Transparent Audio in particular, while expensive, are very good about letting you try them out for quite a long time without obligation.

If you go to a HiFi show, Nordost can be relied upon to perform a through, detailed, and quite convincing demonstration of their wares, and their products are of reputable quality.

Finally, another aspect of power cord performance - indeed of almost all after-market cable performance - is the break-in period.  This is another place where the skeptics love to roll about the floor laughing.  Yes, cables do require a break-in period.  Power cords are quite easy - you can leave them plugged into a running CD player or something for an extended period - loudspeaker cables less so.  You need to plan for a minimum of 24 hours of continuous running to break in a cable, and some will need up to 300 hours.  My experience with power cords is that most of them will be in the 24-100 hour range.  My new PS Audio C7 Jewel cable, which was sounding a bit “breathy” straight out of the box is now sounding a lot more composed 48 hours layer.

Good luck :)