AC Connection Strategy?
default -- Wed, 01/28/2009 - 12:59
i was wondering if anyone could offer any sound (!) advice on optimizing the way that devices are plugged into the AC lines. for instance, should care be taken to isolate components with traditional power transformers from those with switching power supplies, if one has access to different filter banks? does it matter if switching power supplies are isolated from each other, or does their means of operation already make them relatively immune to the noise generated by other switching power supplies? my guess would be that the switching power supplies would be more immune to line noise that traditional transformers, and that isolating the two types of supplies would be an important rule of thumb for powering equipment (and obviously as much further isolation from component to component as is feasible would be desirable too).
i would be very interested to hear thoughts on the matter.
thank you,
agim
All content, design, and layout are Copyright © 1999 - 2011 NextScreen. All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
Switching power supplies will throw noise back onto the AC line. This noise will pollute the AC powering your other components. If you have an AC conditioner with banks of outlets that are isolated from each other, you should segregate the components with switching supplies to the same isolated bank. Even better would be smaller, completely isolated AC conditioners for each component with a switching supply.
I don't have direct experience with systems using multiple compoments with switching supplies, but I suspect that switching supplies are more susceptible to noise than linear supplies. The Linn CD12 CD player (one of the three greatest CD players ever made) used a switching supply, and it was very fussy about the quality of the AC feeding it.
in other words, the recommended connection strategy could be summarized as:
priority 1: isolate non-switching supplies from switching supplies
priority 2: isolate switching supplies from each other
priority 3: isolate non-switching supplies from each other
is this correct?
which leads me to a follow-up question: since switching supplies are notorious for throwing noise back onto the AC line, why not build them with an integrated noise-suppression capacitor across the line to mitigate their impact on AC line noise? wouldn't the extra $2 worth of noise suppression be money well spent? does anyone know whether there are any plug-thru noise-suppression filters on the market that would effectively do this? i know audioprism's quietlines exist, but those eat up an outlet -- and since they are parallet devices anyway, i wish they came in plug-thru form, so they wouldn't waste an outlet.
so it would seem that the general principle would be to start with an approach like:
1) isolate switching supplies from non-switching supplies
2) isolate switching supplies from each other
3) isolate non-switching supplies from each other
i decided to do a $40 experiment to see if i could hear any difference. i purchase a 12 outlet tripp-lite strip, which has no filters, chokes, surge-protectors, etc in it -- only 12 snug-fitting receptacles in an aluminum housing. i plugged it into one outlet of my power conditioner, and plugged all of the low-wattage components that i believe have linear power supplies into it. i next plugged my power amp (also linear) into a second outlet on the conditioner. i then plugged the 3 items that i know (or suspect) to have switching supplies into a separate set of outlets, which are filtered from the others. 2 of those 3 items are not in the audio chain for the main room, but power my outdoor deck speakers -- the third is a squeezebox, which is in the audio chain, but only via its digital output.
this arrangement replaced the prior one -- in which the devices were all connected to the power conditioner in a more or less random way. while switching of power connections, no other interconnects were changed or even removed at all -- so any effect is solely due to the power connection arrangement.
the effect was positive -- no change in tonal balance, but the tails of the high frequency content seem a little longer, and the system seems to possess slightly better dynamics (as if a bit of compression were removed). well worth the investment in time and money.
while i was re-connecting all the power, i also decide to use one of those "Kill-A-Watt" devices to measure (a) watts consumed (b) amps and (c) power factor for each component. the wattage and amperage yielded no suprises -- but the power factors were interesting, though i do not know how to interpret them.
for instance, my classe power amp had a power factor of 0.65, and it was firmly planted at that level. but a mini-amp that powers my outdoor speakers had a power factor that fluctuated wildly from second to second -- between 0.20 and 1.00 -- which seems like it can't be a good thing. at the opposite end of the range, my upsampler had a power factor of 0.93.
is power factor something that could be further considered when choosing an optimal AC connection strategy?
so it sounds like you might say that the priorities for AC connections would be:
1) separate linear from switching supplies
2) separate switching supplies from each other
3) separate linear supplies from each other
my starting point was everything connected to my power conditioner, in more or less a random fashion. to experiment, i spent about $40 on a 12 outlet tripp-lite power strip that offers no filtering, chokes or surge protection, and i used it to extend one filter bank to serve up to 12 components. i hooked up every low-wattage component with a linear supply to it, hooked up the power amp to a second bank on the conditioner, and hooked up the 2 components that i believe have switching supplies to a third bank. no audio cables were changed or even disconnected/reconnected during the process, so that any effect could be isolated to the AC connections alone.
there was an immediate improvement in the sound -- not a change in tonal balance, but seemingly longer tails on high frequency content, and a subtle sense of improved dynamics (as if there was a slight compression before the change). it was well worth the small amount of time and money spent.
while i was doing all this, i decided to use one of those "Kill-A-Watt" devices to record some specs for each component -- watt usage, amp usage and power factor. the watts and amps were generally unsurprising, but the power factor numbers were interesting -- though i don't know how to interpret them.
for instance, my power amp had a power factor of 0.65, and it was very consistent on that number. but my mini-amp (class-T) which powers my outdoor deck speakers and has a switching power supply gave inconsistent power factor readings, which fluctuated between 0.20 and 0.90 -- which perhaps could be interpreted as wreaking havoc with the AC line? but i'm not sure.
at the opposite end of the spectrum, my upsampler held steady with a power factor of 0.93 -- which i think represents a very well-behaved load that is almost purely resistive in nature.
my follow-up question then, on the topic of optimizing AC connections, is whether power factor is a useful or relevant number in helping to arrive at the optimal conection to a power conditioner. the first step of separating the linear from the switching power supplies was very worthwhile -- can consideration of the power factor of each component be used to take it to the next level? thoughts?
so it sounds like you might say that the priorities for AC connections would be:
1) separate linear from switching supplies
2) separate switching supplies from each other
3) separate linear supplies from each other
my starting point was everything connected to my power conditioner, in more or less a random fashion. to experiment, i spent about $40 on a 12 outlet tripp-lite power strip that offers no filtering, chokes or surge protection, and i used it to extend one filter bank to serve up to 12 components. i hooked up every low-wattage component with a linear supply to it, hooked up the power amp to a second bank on the conditioner, and hooked up the 2 components that i believe have switching supplies to a third bank. no audio cables were changed or even disconnected/reconnected during the process, so that any effect could be isolated to the AC connections alone.
there was an immediate improvement in the sound -- not a change in tonal balance, but seemingly longer tails on high frequency content, and a subtle sense of improved dynamics (as if there was a slight compression before the change). it was well worth the small amount of time and money spent.
sorry if this is a repost -- but 2 or 3 attempts to post seem to have failed...
so it sounds like you might say that the priorities for AC connections would be:
1) separate linear from switching supplies
2) separate switching supplies from each other
3) separate linear supplies from each other
my starting point was everything connected to my power conditioner, in more or less a random fashion. to experiment, i spent about $40 on a 12 outlet tripp-lite power strip that offers no filtering, chokes or surge protection, and i used it to extend one filter bank to serve up to 12 components. i hooked up every low-wattage component with a linear supply to it, hooked up the power amp to a second bank on the conditioner, and hooked up the 2 components that i believe have switching supplies to a third bank. no audio cables were changed or even disconnected/reconnected during the process, so that any effect could be isolated to the AC connections alone.
there was an immediate improvement in the sound -- not a change in tonal balance, but seemingly longer tails on high frequency content, and a subtle sense of improved dynamics (as if there was a slight compression before the change). it was well worth the small amount of time and money spent.