Sunday, May 15, 2016

Choosing new UPS for the Living Room System

The Belkin AVR1500 UPS in the living room started beeping.  Fairly quickly I shut it off (along with the living room system, all except the Krell and Subs go through the Belkin).  When the "low battery" beeping starts, it means the batteries aren't holding any charge, and the charging circuit is running continuously.  This might (though it is claimed UPS's don't produce hydrogen because the charging voltage isn't high enough) produce a little hydrogen and I considered such UPS hydrogen generation a possible cause of my illness in 2014 after I returned from a week vacation.  When I got back home from vacation, the kitchen UPS fan was running continuously, and possibly had been for days.  I had the beeping turned off (since I didn't want to be bothered with brownout and outage beeps).  It took 6 days before I figured out why the fan was running: it was because that the battery charger and fan run continously when the battery won't charge anymore, and during that time I felt queasy enough to report it to a doctor.  I got an ultrasound which said my gall bladder had "sludge" and the specialist and my PP recommended removal.  But I replaced the UPS batteries and I haven't had any similar abdominal problem since, and no UPS problem either until those batteries from 2014 failed in 16 months, and I decertified the UPS itself and found that the backup UPS in storage was inoperative also.  So with the latest failure, all 3 Belkin UPS units I bought in 2010 have failed or become suspect and I obviously recommend everyone Stay Away from these long discontinued Belkin UPS units, as perhaps I should have done, though I'd gotten a total of 11 years usage from them now, for about the same cost as a single decent equivalent, and they seem to be fine as AC power conditioners as such, which I wrote articles on in 2010, if I could only disconnect their faulty UPS part, which I may investigate doing.)

This is hardly proof that the UPS caused my illness (tainted food had been my first guess and still is, travel stress my second guess) but I do strongly believe that it is best to avoid hydrogen, therefore avoid overcharging UPS batteries, just in case.  Since that incident I've always had the "beepers" turned on my UPS's so I can't ignore low battery warnings.  And so when the living room UPS started beeping last week, I fairly quickly shut it off--but now I know it must actually be unplugged also.

On the following Saturday I removed the battery box (I also finally marked the new right Acoustat position with masking tape because I had to move the speaker out of the way).   It had been hours since I actually unplugged the Belkin, and days since I shut it off, but the battery box was still warm, I estimated 110 degrees.  Both sets of batteries seemed to stick together a little, that seemed maybe only because of the adhesive holding the labels on, but the batteries themselves had a slightly warped look as if the plastic had melted a bit.

A little thinking about this led me to the conclusion, I won't be getting replacement batteries.  I won't be using this Belkin UPS anymore.  It could be unsafe, especially if you somehow missed the low battery warnings.  Though I might have prevented the warped batteries had I unplugged the unit at the first beeping, I don't want to take chances.  I think it could be dangerous, and I've known from day 1 that Belkin had quit making UPS's which was why I bought 3 at closeout prices for less than the price of one, and this suggested their UPS's might have serious issues which led to Belkin getting out of the UPS market.

But what to get now?  I do like the reliability and concept of the BrickWall UPS.  It uses no sacrificial and noisy MOV's, it uses a system of chokes and non-sacrificial capacitors to filter the power in such a way that surges can't happen.  It has a very long warranty and I believe the company says in over twenty years they've only replaced a few.  It's safe, works, and lasts a long time.

I figured this kind of system would also clean the noise on the AC power.  Just now I tested that, with the TV turned off in the bedroom, the AC power still sounded fairly noisy on my Audioquest Noise Sniffer.  It actually sounded quieter on the Cyberpower Metal Box Sinewave Interactive UPS in the kitchen.  I was very disappointed.

I though about my options.  I could use the PS Audio 1500W Power Plant Supreme that I already have.  I took that off the shelf and it's AC power was quietest of all.  A model of perfection.  But I don't entirely trust these units.  A few years ago it ended up blowing the fuse on two Parasound amplifiers after the PPS got into a tug of war with the APC UPS on the same circuit.  So under difficult circumstances the PPS can output dangerous AC that can at least blow fuses, if not worse.  This is sort of like the Belkin in that PS Audio made these "Premier" units in China, but for some reason after doing that decided to move their production back to the USA.  So there's reason to be suspicious of these units.

I think I'd trust a brand new PurePower UPS, but boy are they expensive!  The lowest power unit still in their model lineup is the 1500 and they run $3250.  If I'm going to plug everything in my system into it, as everyone recommends, I'd need the 3000 and they run $5850.

Cyberpower makes a online UPS like the PurePower but it has a very noisy continuous fan.  Likewise for the online UPS's made by APC and Tripp Lite.

Given all the expensive and unobtainium equipment in the Living Room, good protection is needed.  I think the UPS protection is somewhat better than the Brickwall protection because of the continuity of power (doesn't go on and off repeatedly) and brownout protection.  And it's annoying when power outages happen and all the Sonos units don't come back up correctly, for example.  The Behringer units don't always come back up correctly.  And often power outages have multiple jerks on restart.  I think repeated starting and stopping can lead to failures even on purely electronic units.

So I had decided to get another Metal Box Sinewave Cyberpower UPS (PR1500LCD).  It seems to have quieter power (Cyberpower does advertise EMI and RFI filters, FWIW, but I would have never expected it to be quieter than the BrickWall, though my current tests may not be definitive on that) and I think better protection overall.  It may not last as long though (and there will be replacement batteries to get every few years, hopefully more like 5 than the 3-2 I was getting with the Belkin).  In the kichen system, I've never noticed the fan to come on, and it's powering a lot more stuff.

As I was writing some of this article, and thinking how good the system sounded, powered by the new Cyberpower.

After a few more days of looking around online, I decided to get a UPS actually designed for audio equipment, the Panamax MB1500.  This looks to be identical to the Furman F1500 (Furman and Panamax are the same company now) and typically sells for the same price as the Furman ($1299) but I managed to find one, brand new, for $811, from a well regarded online store Newegg.  Every other online vender wanted the list price, $1299, or even higher (!!!).  Now I see that Newegg is NOT an authorized dealer for Panamax (Panamax has very few Authorized online dealers, hardly any of them are big names), and Panamax won't honor either the product or attached-equipment warranties because of that.  Oh, well, if I had known that I might or might not have made the same decision.  Except that I simply would not buy this for $1299, which seems to me like a rip-off when I can get the Cyberpower PR1500LCD for only $389 from an authorized Cyberpower dealer, and it's very similar as far as being a line-interactive UPS with sine wave generated power and some EMI/RFI filtering.  I am just hoping that being a UPS designed for audio purposes like the Panamax or Furman the EMI/RFI filtering will be a little better.  But without hard facts, or knowing if it would even make a difference, I wouldn't spend $890 more, but I was willing to spend $422 more.  For $890 more, I'd like the kind of filtering that used to come with the Monster HTPS 7000.

I had been a little inspired by this review of the Furman F1500.  Unlike my old Belkin, this looks like a very serious UPS (I suspect Panamax/Furman gets an OEM from the likes of Cyberpower as the features seem very much like Cyberpower features).  And as a "conditioner" they show the effect of the EMI/RFI filtering, at least at audio frequencies.  The noise (or is it distortion?) is highly and visibly reduced on the spectrum graph above 3kHz, though strangely it's slightly increased below 3Khz by a very small amount.  (This is typical, from what I've read elsewhere, and some even say the tiny increase at lower frequencies makes the improvement at higher frequencies moot…I'm going to guess and hope this is not so and that that the mixed improvement is still an improvement overall.)

Why not just do the "try it and see" thing?  I don't trust subjective tests other than true DBT with statistical analysis, that's much more hard work than audiophiles are used to, and I can guess in advance from my previous experience testing things that objectophiles say don't make a difference that I would be almost certain to get a negative result in such tests.  Any single sighted test, or even a year of sighted usage, actually proves nothing and can only create superstitions (which many audiophiles are know are cursed by).  But still even if something can't be proved to be better isn't proof that it can't be.  So I'm doing this on a "it might be better" basis based on objective considerations (though I realize the particular objective considerations I'm using, as describe here, are themselves subjective) and basically "it might be better" is worth something but not a lot.

Here is a review of the Panamax MB 1500.  Note it looks absolutely identical to the Furman F1500.  But in this review they also show pictures of the inside.  Yes this does look like a much more serious UPS than the Belkin.  And it does have fairly serious looking AC filter parts as you can see behind the outlets in the back.*  These look just a little less impressive than the very impressive loooking AC filter parts in the Belkin I have now (which would have been great if it also had a good UPS instead of the flaky and potentially dangerous UPS it does have).  But the filters look a lot more than just hype, and likely more impressive than the filters in a typical computer UPS like the Cyberpower PR1500LCD.  Now I don't know what difference it would make, but the typical prepackaged "AC line filter" parts you can buy off the shelf from Mouser or Newark such as the Corcom RFI filters don't have much effect below 100kHz, and interestingly the Cyberpower specs list 150kHz as the lowest frequency.  I'd guess the filter parts in the Cyberpower are similar to the Corcom parts, single part filters, though probably not as expensive.

*I could imagine a standalone power conditioner with parts like the Panamax/Furman in a cheap box for $200, or a nice box and outlets like the ones it has for $500, or at an audiophile price of $2000 in an even nicer box.

So it is very much looking like when doing AC line filtering, you can do a lot for cheap at the highest frequencies, it's the lower frequencies where it starts costing real money.  For comparison, look at the incredibly complex filtering stuff you used to be able to get with the likes of the Monster Power HTPS 7000.  (If you can't see the pictures in that blog, look here.)

The entire Monster 7000 HTPS chassis is filled with cool looking parts, big yellow box film capacitors that look like Wimas (probably aren't actual Wimas) and heavy wound chokes, and two large isolation transformers (which Richard Marsh himself says were very expensive and high performance isolation transformers for the low current outlets).  This is not a UPS at all, just a "power conditioner" which really means that it filters noise and distortion from the AC 60Hz, and yet it fills a very large box with parts.  Anyway, I don't know the actual performance of this unit, just that a fairly serious designer (Marsh) did the best that he could with the funds available, but it probably shows the minimum amount of "stuff" you would actually need to filter out the noise and distortion lower than the 3kHz cut off of the Furman and Panamax UPS's.   You can see that just having a few parts behind the outlets just isn't going to do it, you need to fill a whole box with filter parts if you want to filter the lower frequencies well.

I believe Monster called the HTPS 7000 "stage 5" or maybe "stage 7" filtering, and you can see that each set of outlets does have about 5 or 7 major "parts" that the AC power must flow through.  By this standard of rough measurement, the Panamax and Furman UPS's are something like "Stage 3", just like my old Monster 3000 power strips (which all died after less than 5 years btw, but it probably was not the filter parts which died but instead the stupid MOV's and the voltage/current display).

Now I'd spend the extra $$$$$ for a good filter if I really knew it was important, but I don't.  I tend toward believing the objectophile view that well designed equipment, like pretty much everything I already have, was already designed to handle AC power as it actually exists and doesn't need any more "conditioning."  Further, my entire living room system runs off a 20A dedicated isolated ground circuit.  It is often said to be a better thing to get a dedicated circuit than to buy a fancy AC conditioner, and that's my starting point also.  I also have the power company transformer for the houses adjacent to me in my own back yard with about 50 foot buried lines capable of over 600 amps.  I know of no specific problems with my AC power.  It consistently measures 120V within a volt or two, and about 2% unfiltered THD on the PS Audio Power Plant Premier.  That's about as good as utility power gets.  Long long ago I had ground loops in my bedroom system, which was my first "audiophile" system in this house.  But those were not related to the power as such but to how the old cable TV system was grounded.  I worked around those by removing all ground loops from my system and isolating everything related to TV from things related to audio.  That was a great learning experience which I think all audiophiles should have.  Don't fix hum with band aid but by actually removing ground loops and equipment with faulty power supplies (which were the second problem I discovered, many 20 YO units have faulty power supplies which produce audible hum even when said equipment is disconnected from everything else).

Much later my electrician discovered that my house grounding was broken, and I had him fix that right then and there with a new heavy duty copper grounding rod near the AC panel.  The broken house ground probably had something to do with the cable TV ground loops, actually, the cable was grounded to the plumbing (and later the panel) which should have been grounded to the panel which should have had a working ground rod, instead I could measure 120V between the ungrounded panel and the water-pipe grounded CATV wire.

IF I really knew this conditioning thing was hugely important, I'd spring for the hugely expensive PurePower 2000 or something like that.  Or at least I might.  Or not, I mean that's actually more than I spent on my minty used Krell power amplifier including the first factory repair.  There's also a particular problem with large regenerators.  You apparently MUST plug EVERYTHING in your system into it.  That's because while the regenerator gives you pristine clean power at its output, it compounds the amount of grunge at the wall outlet itself.  So with one of those, for the best results, you must plug in the power amplifier to it also.  I'm disinclined to do that with power amplifiers generally, and it would also mean that the PurePower 2000 wouldn't be sufficient, and I'd need to get the 3000 or still larger!

Speaking of regenerators based on On-Line UPS's, you can get those far cheaper than the ones made by PurePower.  You can get a nice 1500VA online UPS from Cyberpower for about $560, or similarly from Tripp Lite or APC.  The problem with online UPS's made for computer server rooms is that they consistently have noisy fans.  I've read numerous blogs where people change out the typically small fans inside such units for 5" or larger fans that more quietly run at low speeds.  I have some experience with that sort of modification, I did something like that on my Amiga 2000 computer about 23 years ago.  But it's tricky and very time consuming (I spent months modifying the Amiga 2000) and I just don't have the time for that style of equipment redesign anymore, and especially for something I'm suspicious is not very important anyway.

If the line AC noise and distortion at lower frequencies is really all that important, I might opt for the basically passive approach of filters (like that Monster 7000, it's been discontinued but you can get used ones cheap on eBay) or use an isolation transformer.   The isolation transformer thing doesn't look bad, if you shut your system off when not in use and don't really need the UPS part.  Typical isolation transformers tie the output neutral to ground, I've read that should be cut for use in an audio system, and the neutral left "floating".

WRT the Monster, however, I'm suspicious of their long term reliability.  I bought three very fancy Monster 3000 power strips (about 10 pounds of solid metal) that looked indestructable around 2004.  By 2010 they had all failed after 4-5 years usage.  Perhaps the 7000 being a more expensive and serious product would do be better.  But the fact that Monster has discontinued the 7000 and many of the higher end models doesn't look good.  There are many online stories about failed Monster Power products too, and it seems like many of the upper tier models have been discontinued and not replaced with anything similar.  In the story I linked above, one 7000 owner asked Monster if his failed unit could be repaired and Monster said no, it had sacrificed itself to save his equipment.  Well it would be nice if it could save equipment and NOT sacrifice itself in the process, as it appears the BrickWall filters can do, for example, and Panamax/Furman seems to use a similar series suppression technique in their UPS's.  Of all the non-Monster surge strips and supressors I've bought over the years, only about 1 other unit has lost its surge capability.  All the others, regardless of age, still show the "surge working" indicator light.  With all it's nice looking filter parts, it might have been nice if Monster could have made the 7000 repairable, for example by having a pluggable MOV or avalanche diode section.

Anyway I like the idea of UPS and active voltage monitoring to protect my expensive equipment from all the possible kinds of power events, and I like not having to manually reset stuff after a power failure, and the living room hosts my two Tuner +Sonos nodes that I use elsewhere in the house at any time of day, so having an actual UPS is an actual and not an imaginary benefit, though perhaps not a huge one.  I've used a UPS/conditioner in the living room for 6 years now and I'm not sure I want to go back to a non-UPS solution (though, unlike the Kitchen system, which also has DVR's and security cameras, it's not an absolute necessity).  It's likely nothing else could be proved to sound better, but convenience is convenience.

(Although it IS inconvenient to have to replace the batteries every 3 years or so, and that has given me some pause already.)

I would like to see an online UPS which absorbs, rather than creates, further noise and distortion back at the wall outlet.  With such a UPS I could plug my line level equipment into it, and the power amplifier as always straight into the wall, and have it be all plusses.

In fact, the effect of any given piece of equipment on the wall power should be an important design consideration.  Devices should not only be themselves immune to powerline noise and distortion, they should tend to absorb rather than create them on the power line.

This is similar to my room equalization concept, where modes present anywhere must be absorbed, if not completely at least partly.  It's not all about me the one optimal listening position.

I believe generally speaking an incandescent light bulb is said to have such a property: cleaning up the powerline it is plugged into.  That's not what we should use, they simply waste power, but power supplies should be designed to be like that.

Powerline noise and distortion is a fun thing to think about, and I think for some a ready excuse for perceptual variability, but not really the hugest factor for most people I think, and especially me (because of my high quality utility power, dedicated circuit, and well designed equipment).

It seems, actually, so long as I wanted the UPS features (and...I could possibly give them up for this system) there aren't many choices.  Either mega expensive PurePower, or Furman/Panamax, or some server grade unit with quiet research last time made me pick the Cyberpower PR1500LCD for the kitchen and I'm very happy with it.  Not only is it perfectly quiet, it seems to have low output noise.  It's highly programmable just from the detachable LCD screen (with same functions available in computer app but I haven't tried that).

Now I haven't researched APC much.  My old server grade APC was noisy, provided noisy power, and pulled the bedroom circuit to it's knees just to provide "stable power" when there was hardly any actual problem.  Since then APC has made a series of UPS's where they at least suggest Audio/Video usage: the J, H, and S series but I can't find any but J anymore (I recall the S as the sinewave series and quite expensive like $1500).  The J is stepped not sinewave and amazingly cheap.  Strangely it looks physically like the Panamax/Furman.  Perhaps the same factory makes both models, the Panamax/Furman being the sinewave upper end model which might have been the APC S if APC sold the S anymore which they don't.

Now about why a line filter, smoother, etc might be useful...

It's always said "well designed equipment doesn't need one" which is of course begging the question.

But why would, in principle, it be advantageous.  In non-audiophiledoublespeak.  I'm not going to say it's because the power is everything, everthing else is just modulated power.  That's audiophiledoublespeak.

Well power like crap could cause some sort of peculiar problems, like power supply motorboating.  But I'll leave those aside for the moment, and just focus on the obvious.  Any piece of equipment can only reject the line noise to a particular degree, what you might call the AC Noise Rejection.  This is sort of reflected in the Hum and Noise specification, but some of the Noise, and even some of the hum, may be generated subsequent to the power supply.  AC Noise Rejection is the specification we actually want.  And we can imagine for much equipment you might have an A weighted figure not much different from Hum and Noise, say -90dB.  This is caused by electromagnetic chassis coupling and by failure of the power supply circuitry to eliminate all the noise.  Even sophisticated regulated power supplies do not have infinite line noise rejection.  But for "well designed" equipment it's in the same low range as the audio material itself.

But this hides the fact that the actual noise generated can be very spotty in it's spectrum.  And furthermore not all noise is equally bothersome.  And the AC Noise Rejection itself can have irregular frequency response.  So if a burst of particularly annoying noise occurs at a particular high spot in the frequency response curve of the AC Noise Rejection, it can, in principle, poke through to perceptible levels, even if -90dB of say pink noise would not be a problem.

In a real life complex system, the vulnerabilities to line noise are increased.

So it's reasonable to believe, at reasonable expense, pre-filtering ac line noise is a good idea.  Etc.

I have yet to be convinced of the need for crystals, quantum field generators, and the like.

The objectophile objection is written succinctly by Ethan Winer.

Ethan says it's all just electricity, which is easily measured.  "All competent audio gear rejects the normal amount of noise on the AC line."  You can measure the noise at the output of your power amplifier, with and without conditioner in place.  If you have a good amplifier, there will be no difference, which means the power conditioner may be cleaning the AC power to the power amplifier, but it makes no audible difference.

Well that's exactly the sort of thing I created this blog for.  I did at one time make such a measurement, and I decided that the conditioned power reduced noise measurably.  By now I completely discount that measurement, I believe now it must have been error.

It's actually not easy to do such measurements.  I would agree with Ethan that a single number noise measurement is likely not to show any difference.  But the best way would be to produce a full spectrum.  I think there is likely to be a visible difference in the noise spectrum somewhere below -100dB from full signal level.  Does that make a difference?

I suspect it would not make the kind of consistent difference that would be required for DBT.

But it might make audible differences in two kinds of ways:

1) Just enough difference at some moment of music to change the momentary perception, though not necessarily in a repeatable way.

2) Difference at some moment when the power is unusually noisy and the music unusually quiet at the same time.

*****  Update

I've tested the new Panamax MB1500 on a noisy circuit, using AudioQuest NoiseSniffer.  With nothing plugged into it, it's as quiet as the PS Audio Power Plant Premier, quieter than the Cyberpower UPS and the Brickwall AV...though those both had stuff running on them so perhaps not fair comparison.  But I don't ever remember having as clean output with anything other than a regenerator before.  So I think this may be better than Cyberpower for the Living Room System, as hoped.

While more of the weight may come from chassis metal than doesn't appear so on the outside.  The case is not ostentatiously heavy like something from Monster or even the old Belkins.  I can see on the inside that there is shielding around both the battery and the inverter parts.  That's a fine use of metal.  So this is 72 pounds of mostly real stuff, not ostentation.

If I need UPS, and I'm not sure, this was probably the thing to get (or identical Furman), for less cost than the PurePower (and they're out of my league completely, except possible I could have used a used 500 which I've seen for around $1k).  Other than the amps, and the TV, the power is probably under 200W.  With the TV, 360W.  The TV has to be on the same circuit, which is only possible if it's somehow plugged into the conditioner/UPS.

PS Audio still makes its regenerators, priced in the same range as PurePower.  They are now made in USA and I think they also have significant design improvements from my Power Plant Premier.  If I were spending this kind of money, I think I'd get the PurePower if I were planning to power the entire system (including Krell amplifier), or PS Audio if I were just powering front end items.

No comments:

Post a Comment