Wednesday, September 28, 2011

Lizard grows on me

OK, I confess when I first listened to Lizard (King Crimson) a few days ago, I hated it.

But I read a little about the album at Wikipedia.

Then I listened to it again last night.  I was going to just start with the title track-side called Lizard, but decided to flip through the earlier tunes as well.

Well, it's started to grow on me.  Of course the DVD-Audio fidelity is very nice, and it's well recorded, and interesting in various ways.

But I think music (and other things) can be like this.  If you can try it a second time, you may get hooked.

If it weren't for the great audio quality and interesting story-behind-the-story, I might not have bothered.

Tuesday, September 20, 2011

Nordost Baldur tested

I tried the Nordost Baldur 0.6m XLR interconnects on Monday night, connecting my Oppo BDP-95 and the Lavry AD10 digitizer that feeds my digital processors with 24/96.  I played the Turn of a Friendly Card DVD-Audio by Alan Parson's project.  I had listened to that disc a few months back, and it was still sitting in my player.  I immediately liked the sound.  It seemed less harsh than I remembered, but with greater spaciousness and resolution of detail.

I switched back the the 1 foot Belden 1800F pair made for me by Blue Jeans cable.  I immediately noticed that it sounded louder.  That was unexpected.  But it also sounded noisier, and I felt less able to resolve small details.

Other than the change in loudness, this was pretty much what I expected based on the use of solid core wire in the Baldur, and silver plating, and structured FEP insulation.  I think the wire stranding in the Belden is its worst feature...solid core wire is better for purist audio.  The Belden does use foamed polyethylene insulation, which is probably nearly as good as the spiral wound FEP.  (The Nordost wire is spiral wrapped with FEP threads so that 80% of the wire surface area is surrounded by air, a very clever design feature, probably better than foamed insulator for getting the closest approximation to air dielectric.)

I doubt I would be able to hear this difference in a blind test, and fear it could all be confirmation bias or something like that.  Probably the most solid conclusion I can make at this time is that the Nordost wire is NOT very obviously worse than the Belden.  Except that I do worry that the skinny Nordost wire may not be as robust over time.  But I bought brand new from an authorized dealer, so the factory warranty should apply.  I suspect Nordost would replace the cables if they ever went bad (Monster does, for example), but I actually don't find a warranty description on their website.

I'm leaning toward keeping the Baldur now, even though I ordered it by mistake.  The specs and design look excellent to me, though a friend thinks the wire gauge should have been 30gauge instead of 26g to nullify any possible skin effect (at 30g, it's all skin in frequencies significant to audio).  The Belden is 24 awg.  I can't understand his argument why 24 gauge is OK in uncoated wire, but you need 30 gauge in silver coated wire.  And it also doesn't seem by my reasoning that skin effect makes much difference in interconnects anyway, though other things being equal I'd go to 30gauge.

The more expensive Nordost interconnects differ mainly in having multiple conductors.  I don't think that's necessary in an interconnect, where inductance is not a big deal.  Also, contrary to what would be my friend's advice, the Valhalla $4000 interconnects use multiple 24g solid wires, which would be a step backwards by his reasoning, having more skin effect.  Actually skin effect probably isn't a huge deal in interconnects anyway, but silver coating as the Baldur has probably helps eliminate any small nonlinearities from copper oxides on the surface.  So I think of Baldur as a Goldilocks cable design, everything just as it has to be, good enough, without overdoing it.  If I could have everything I want, I'd have 30g solid silver plated wire and real foamed teflon insulation and braid+foil shielding, but that might make for a much less robust cable, and I haven't actually seen one like anywhere.  The closest I've seen is a Thermax wire with "fluorinated polyethylene" insulation (FEP?) and 30g silver plated solid copper wires, but lacking braided shield.

Update June 2017.  It appears that original Nordost Baldur can exaggerate very high frequency energy.  Playing at fairly low level, my Krell FPB 300 (with new Audio GD Master 7 Singularity driving the Baldur XLR 0.6m to the Krell) shut down in two subsequent attempts to play Taquita Militar on the Crystal Cables label (which has very intense HF energy) at very reasonable -15dB level.  Error message is insufficient AC power: apparently the Krell is kicking up to the maximum 400W (factory update) bias level and my system 20A circuit isn't up to the task, but that shouldn't happen at -15dB. Replacing with a standard Beldon 1800F (AES 110 ohm) eliminated the problem up to 12dB higher in level, about the highest I dared test.  Now I can guess why the original Baldur was discontinued by Nordost.  A bit too hot.  I'm thinking that merely reducing inductance to nil isn't necessarily the thing to do.  Inductance and capacitance and resistance have to stay in balance in each section of interconnect.  This suggests to me that the standard tried-and-true forumations, such as 75 ohms and 50 ohms and 110 ohms are pretty much the way to go.  They maintain pulse coherence up to UHF and beyond.  Try to optimize just one parameter and you may get a dangerous Ghz oscillator or something that may wreak havok on some systems.

Friday, September 16, 2011

Cable effects

I've engaged in lengthy discussion with a friend of mine about cables this week.  To a first approximation the result is I now think various cable features such as dielectric, skin effect, resonance, stored energy, etc., can be important, and these are not always easily known as say, capacitance, which is usually specified (and which I can easily measure if it isn't specified because I have a pf capacitance meter).

High end cables such as the original Blue Heaven flatline cable may be somewhat flawed because of lack of shielding and underdamped high frequency LCR resonance in the 100Mhz region.  While you can't hear 100Mhz, it may make your equipment complain audibly.  That cable was very low capacitance, but rather high inductance, leaving me to believe it is not an interleaved cable as I had thought, but a flat cable with polarities on opposite sides.  The lack of interleaving further increases susceptibility to RF pickup, and makes for rather high inductance.  It might make a nice FM or HDTV antenna.  If that's the price for super low capacitance, I think I'll pass.  Nordost has completely changed the design in Blue Heaven II, making it a more conventional shielded design now, like all their other current interconnects, and you can guess the reasons.  But the original Blue Heaven got huge praise in the audio press.  Well, it was different.  I'm thinking of trying the Baldur XLR 0.6m interconnects I ordered by mistake; it looks reasonable.

Silver is a better conductor than copper at DC and low frequencies, but worse at higher frequencies (above about 60Hz) because it is more magnetic.  The effect of this is that less silver may be better than lots of it.  Very very thin silver plating may be better than pure copper because copper oxides are non-conductive much less than silver oxides.  My friend claims that for coax you want center conductor, if silver plated, no thicker than 30 gauge, but pure copper can be larger.  I haven't worked the details on that out in my mind yet.

Nordost cables are now using FEP for dielectric, my friend insists actual Teflon (PTFE) is still the best, with FEP and various PE formulations about comparable.  Teflon is rare in audio cables (it's not easy to make a cable with Teflon because it's not very strong) except very expensive ones, but can actually be found in various relatively inexpensive RF cables.

Belden tends to make cables with foamed PE, and their coaxes and balanced cables used for audio by Blue Jeans cable have respectable capacitance and inductance and presumably little tendency toward extreme resonance in RF because they were originally designed for video and digital audio.  They tend to use unplated copper, in stranded form for the balanced cable, and I wonder if that's actually the best.

For what little it may be worth, here's a white paper by a company that makes pure silver cables with some but not all teflon in the dielectric.

Friend suggests LL120 From Harbour Industries, but it has 21 gauge silver plated conductor, not the 30 gauge he recommends.

Thursday, September 8, 2011

Fancy Cables and Dielectrics

Perhaps it's no surprise that each high end cable company has a different story.  Most typically, it's a story about they have discovered something unique that no other cable company pays attention to.  Thus there are an infinite variety of different labels and buzzwords used by each company, no two the same.  They want to make you a true believer, not a careful comparison shopper.

Meanwhile it's often very hard to find out the basic facts about each kind of cable.  For starters, the facts of interest to real audio engineers would be:

cable impedance: resistance, capacitance, and inductance
shielding: none, single, double; braided vs foil; coverage percentage

Some additional facts are generally recognized by audiophiles and some engineers:

dielectric material: vinyl, polyethylene, teflon (from worst to best...this affects a parameter known as dielectric absorption, which is not ordinarily specified, and may even defy a single number specification, but if you know the dielectric you can make some predictions about dielectric absortion, teflon being the best and better than vinyl by 10-100x, and polyethylene being closer to teflon than vinyl)

general geometry: twisted pair, coaxial, shielded twisted pair, multi-conductor ribbon, hyperlitz, etc. (but the virtues of the geometry would also be revealed by a full accounting of the impedance and shielding effectiveness anyway).

Anyway, these are generally not the kinds of facts you get in the descriptions of most high end cables, even if you go to the "specifications" page.

I have even found two significantly different descriptions of the same cable, the Audioquest Copperhead.  The description at claims the dielectric is polyethylene, and the description at AudioAdvisor says the dielectric is a specially chosen PVC (aka vinyl).

Sorry, but I won't pay a high cable price without the cable company being very straightforward with me about what I am buying.  I might be impressed by Audioquest's "PSC" (TM, perfectly smooth copper) as an interesting treat, but if it's merely used on plain old vinyl wire, I'll pass.

I'm now looking to upgrade some if not many of my old vinyl and not well shielded Radio Shack cables (which have now become quite expensive) with better shielded cables using polyethylene and teflon.

Teflon cables, it turns out, are often very expensive.  Leading manufacturers of audio cables with teflon dielectric are Cardas and Nordost.  These are companies known for making multi-kilobuck cables, but they also make relatively affordable ones also, if you consider a $150 cable affordable.

Yesterday I finally ordered my first Cardas cable, and 0.5 meter Crosslink interconnect, priced at $128.  It uses teflon dielectric (actually the entire insulator is teflon) and Cardas' trademark golden ratio litz (litz prevents noise causing strand interactions and decreases self inductance and skin effect), in a multiconductor inductance cancelling geometry, with good shielding.  I plan to use this for one of my few actual analog sources, my Pioneer F-26 tuner.  If I like it, I'll get more.  Unfortunately, it's on back order.

I also another audio cable from Blue Jeans cable.  Blue Jeans generally uses standard designs from Belden and Canare (the world's two leading industrial cable manufacturers) with very nice RCA's from Taversoe or Canare with gripping metal fingers inside.  Now they also offer some exclusive cable designs, typically made by Beldon under special order.  Most of their cables use polyethylene dielectric, which is common among industrial cables (for example, broadcast use).  Many of their cables use solid conductors (which improves impedance consistency under bending--the official reason, and also reduces noise from strand interactions--a more tweako concern.)  I like their cables a lot, but they are often fairly (or even extremely) stiff, stiff enough to lift light equipment like my home built notch-blend filter, which is one reason why I'm checking out the Cardas.  Plus, the Cardas has Teflon.  But I also ordered some 2 foot long LC-1, their latest custom audio cable design built by Belden to be more flexible.

Kimber makes a less expensive interconnect cable with teflon insulators called the Hero, but it is unshielded.

Mogami is another industrial cable manufacturer like Canare and Belden.  They make cheap interconnects (now actually cheaper than the cheapest from Radio Shack, and clearly for home use...) and various midlines.  The midline versions use polyethylene. I'm ordering some WR-01, WR-03, and WR-06.

Sunday, September 4, 2011

More thoughts on Power Plant Premier

On the plus side, I tested the Tensor lamp plugged directly into wall while starting the vacuum cleaner on same outlet.  The tensor lamp did indeed briefly dim.  On the Premier, it was rock solid.  That's a big score for the Premier, and I expect that within it's power ratings (which most audio systems don't come close to exceeding) it actually makes the AC voltage more stable, just as Paul McGowen has said, for better transient response.  I think of it as much like converting your power amp into one with regulated HV rails, like Krell FPB amplifiers.  As I said before, the whole purpose of this design is that you can (most likely) plug your entire system into it, with the power amp quite likely getting the biggest advantage, unlike low power regenerators like the PowerPlant 300.

The downside is first that I'm not even sure this can really justifiably called a "regenerator."  Yes, it can, but actually it's more like an instantaneously tracking corrector.  From what I've read, the entire power supply follows the incoming AC voltage, so all the power plant amplifier has to do is make up the difference to ideal.  The first consequence of that is you can't play games with the output frequency, as on earlier (and later?) PowerPlants.  The second it that in some sense you are not as completely removed from the incoming AC as you would be with a full regenerator, it doesn't seem to me it would give as much isolation.  And the third might be that under some circumstances, the tracking doesn't work perfectly or interacts with the unit in some way, making the line noise in certain situations worse.

I might have found such a situation.  It seems like when I use the dimmer on my kitchen light, the noise as revealed by the Noise Sniffer is actually greater through the PowerPlant than straight on the wall.  Perhaps it isn't really greater, but the Noise Sniffer is simply more sensitive to the noise frequencies coming from the PowerPlant in this situation which have been shifted somehow.  I haven't yet done a full sweep test using the "volume" control on the sniffer yet.

Now even if this is true, I have fanatically avoided using a dimmer in the bedroom area.  I don't trust dimmers anywhere near audio, but capitulated to convenience in the kitchen.  Dimmers generate horrible noise and distortion on the powerline.  So this might not matter much for me, it is merely a small disappointment.

On retesting today, using a full sweep of the noise sniffer in both cases, it's clear the the PowerPlant is reducing the dimmer noise after all, plus some other noise that is very annoying now.  (In fact, over several days of testing as of this edit, the reduction is quite remarkable...the PowerPlant is actually performing like a very good regenerator.)  The difference around 3pm on the volume control is enormous, with a very nasty buzz straight from the wall, and sounding like ordinary clean power through the PowerPlant.  But at max volume, they are both about the same, with the wall socket possibly having a deeper quality that is more easily ignored from the other room.  Also, the X10 dimmer control signals are very audible at the wall socket, but completely hidden through the power plant, making the background sound at the wall socket seem softer by comparison.  Still, the PowerPlant obviously does not eliminate the dimmer noise completely, you can hear a very phasey sound on the noise sniffer as you change the dimming.  That same phasey sound is also present at the wall socket, but less obvious amidst the high level of noise generally.

One thing I've noticed that is rather despicable about my Smarthome remote dimmer.  Even at full output, it makes a lot of line noise.  I would have expected line noise to pretty much disappear at full output.

Finally, in spite of my thinking this is not truly a regenerator, I think the general design is a good one, whatever it is called, perhaps the best one possible.  The earlier approach to PowerPlants wastes huge amounts of energy, and even then is hard to scale to meet the needs of an entire audio system, with possibly the power amp needing the most stabilization of all.  And it should be noted that the earlier approach didn't, nor could any approach, perfectly limit the influence of incoming noise.  If Class AB amplifiers eliminated the effect of incoming noise, it wouldn't be necessary to use the PowerPlant at all.  The only reason why power conditioning is needed is precisely because amplifiers are not perfect in rejecting incoming line noise.  Having a full regenerator merely adds one more degree of Class AB amplifier separation from the line noise.  And possibly adding an instantaneously tracking corrector based on a smaller Class AB amplifier adds about 1/2 a degree of Class AB amplifier separation, in principle.  But as always, the results can vary by implementation.

I suspect, however, that the fully regenerating power plants like P300 limit the dimmer noise better, and possibly don't make the line they are plugged into noisier.  But they don't provide much power, run incredibly hot, and have a constantly running fan.  It's the fan noise issue that has kept me from using my P300 much.  I tried using a quieter fan and it didn't help much.  My plan has been to use an even quieter fan and change to a lower speed, but the project has been stalled for awhile.  But I never even imagined using the P300 to power an entire system, mainly it was just for powering standalone equipment like my CD duplicator and electronic synthesizer.

Initial tests of PS Audio Power Plant Premier

The distortion indicator is very interesting, and if it is to be believed the Premier is reducing line distortion somewhere from 5 fold to 10 fold.  I measure line distortion in the range of 1.4% to 2.2% on a continuous basis and the Premier reduces that to 0.2% to 0.4%.  I wonder how important this is.  Given that the Multiwave feature actually increases the THD by adding back in select harmonics (though not much actually) much in the way that minerals are added to purified water for flavor, one might well like the sound of slightly greater line distortion more than less, and I wouldn't be surprised if many GESR's (Golden Eared Subjective Reviers) are responding to something like that when condemning power conditioners of various or all kinds.

Tests with the noise sniffer suggest that the Premier can reduce line noise, just as the PureAV conditioner can.  As to which works better, I can't yet say.  In both cases they aren't perfect, and they seem different in what kinds of noise they may reject, but I have not and will probably not bother soon to test both on the same outlet under identical conditions.

The Premier does a great job at stabilizing voltage over the small fluctuations I've tested.  I plugged in an old incandescent Tensor Lamp and turn a vacuum cleaner on the same wall outlet on and off.  Zero change is noticed in the lamp.  There is a slight fluctuation in the THD meters, both incoming and outgoing.  Obviously the PureAV does not do this, it can correct voltage somewhat but not small changes.

I have one observation not noted in the reviews I have read, in fact it contradicts one of them.  The Premier does not make the line it is plugged into quieter, in fact it makes the line it is plugged into noisier.  On the other hand, the PureAV does seem to make the line it is plugged into, or at least the outlet it is plugged into, quieter.  Filter based conditioners have a large parallel effect, as if they were largely similar to a big capacitor plugged into the wall.  Now actually they have series as well as parallel elements (the PureAV uses pi filters) so you would expect the outlets on the conditioner itself to have the most filtering, and the wall less, but not none.

I conclude from all this that it would be counterproductive to use the Premier on low power equipment and then plug a power amplifier into the same line.  The best use of the Premier is clearly using it for ALL equipment plugged into the same line.  And I think this works fine with amplification used on speakers of reasonable sensitivity, such as the Revel M20's in my bedroom powered by a Parasound HCA-1000A (or HCA-1500A again in future).   It might even be OK with the 1500A powering Acoustats in the living room, but I'm not going to try, nor would I even think about plugging the Krell FPB 300 into it.

As the Premier does actually stabilize output voltage under all but the most extreme conditions (in which it automatically bypasses itself) I think it is probably fine to use on many audiophile systems, though you'd have a hard time convincing most audiophiles of that.  They seem to have been convinced by the need for infinite peak current from the wall outlet.  Actually, the main source of fast peak current in a power amplifier is the amplifiers own power supply, the wall outlet isn't providing much of that anyway.

The Premier is clearly intended for this kind of "full system" use, and full system use has a huge advantage.  Power amplfiers usually have only lightly filtered (and not regulated) DC voltage on the main power rails.  AC line noise and distortion can easily contaminate this, and thereby contaminate the output of the power amplifier (though the amplifier feedback will do considerable correction).  On the other hand, most preamplification equipment and audio sources use heavily filtered and regulated power supplies, and it is hard to impossible to see how they would be much affected by AC line noise and distortion.

Thoughts about the Noise Sniffer

I've been testing the PS Audio Power Plant Premier in various technical ways, mainly with the AudioPrism noise sniffer.  Or should I say I've been testing the noise sniffer?

First let me suggest the Noise Sniffer does have some limitations:

1) You cannot simply leave the "volume control" at one position.  The "volume control" somehow tunes the sniffer to different spectra.  So the best use involves turning the volume control from minimum to maximum to back and noting the various levels, at least the maxima.  At maximum "volume" the volume doesn't necessarily increase much if any by itself on clean AC.

2) The Noise Sniffer does seem to be affected somewhat by RFI not in the power line.  Under some circumstances I could pick up lots of noise, including an AM radio station, or make it go away, by simply moving the Noise Sniffer around while it was plugged into wall.  Interestingly, the worst noise (and the AM radio station) was picked up most while sitting on the Simmons Hide-a-Bed convertible sofa, perhaps as a result of all the old metal inside the sofa, including springs.

3) The Noise Sniffer, as its name even suggests, does not seem to be particularly affected by harmonic distortion in the AC line.  The PowerPlant Premier goes to great lengths to reduce this distortion, and I believe the nice indicator when it says I have reduced THD from 1.8% to 0.2%.  But the Noise Sniffer doesn't notice this difference at all.  And of course it doesn't detect differences in AC voltage level, though it might detect abrupt changes.

4) I don't necessarily assume the Noise Sniffer detects the all the noise having audible effect on audio equipment (if there is any!) nor that it weighs the noise that it does detect on an audibility scale.  The noise sniffer is not an "A Weighted" instrument.  Actually it might be interesting to construct such an instrument; one could for example extract the noise and then apply "A weighting" to it.  But then there is additional assumption that equipment is affected in similar fashion.  It might well be that equipment is most effected by subsonic and supersonic noise which somehow affects audible performance of the equipment through intermodulation or similar processes.

In other words, it's crude, but it does give a quick sense of garbage on the line, that is difficult to the level of impossibility by other quick means, barring the availability of even more sophisticated equipment.  It's the best thing I have for a quick test, I just don't trust it completely.