So I finally had enough time to finish editing this final tome about power cables and such.
To wit…
Why DO power cables and such make ANY difference to our headphone gear?
Yet another experiment focusing on power delivery and some not generally understood factors to consider.
Part A Why Would I Want To Do This In The First Place?
https://www.head-fi.org/threads/the-diyrs-cookbook.781268/page-85#post-14294952
Part 1 Theory and Expectations Based Upon My Initial Research
https://www.head-fi.org/threads/the-diyrs-cookbook.781268/page-89#post-14593223
Part 2 Measured Results and Observations.
https://www.head-fi.org/threads/the-diyrs-cookbook.781268/page-91#post-14672408
Part 3 Analysis and Conclusions.
So these measurements have led me to see a picture that points clearly at how sensitive our gear can be while taking in power to use and deliver our music to us.
It shows us that the vast majority of the actual electrical power that our gear gets, is delivered in ≈ 1+ms pulses separated by ≈ 7+ms of ‘Off’ time.
This goes a long way in explaining, for me, why cables, fuses, metallurgy, connectors, wire geometry, etc can make audible differences.
In my experience dealing with peak and/or the extremes of the ranges of the voltages, currents, frequencies etc, that we deal with in our audio pursuits, is where I usually find there are gains to be realized in terms of SQ, especially when these demands for peak signals (voltage and current) are synchronized and occur in very short time windows.
In situations where there are demands for peak performance, we can make improvements to accommodate them, such as using bigger gauge wire for speakers that have big amps driving them, or again using bigger gauge wire in our ac power delivery as another example.
In addition where we find up stream limitations limiting the ability to dump large amounts of current in a short time windows of current delivery (1ms), and then stop, repeatedly, and with ≈ 7ms of off time (which is a 12.5% duty cycle), tells me that there are gains to be realized in all of this.
This type of pulsed activity puts ‘stress’ on the whole electrical distribution system (and especially the wires and receptacles in our particular branch circuit) that runs back to the panel, and the rest of the ac power load in the house and even further back up the line, where all of the existing load can ‘contribute’ their own electrical signatures.
The bottom line is the ac power system is essentially spiking current into our gear, all of our audio gear (with a few exceptions), right along with the rest of the electrical load the main breaker panel feeds power to.
And as my previous
ASCC (Available Short Circuit Current) experiments showed, having greater current dumping ability feeding our gear DOES result in improved SQ.
And when you think of it, how could it not?
Because of these sorts of ‘needs’ for such short duration pulses of current, any and all upstream limitations can (and do) combine, and so, can limit the delivery of these peak current demands.
Including such aspects as the ‘power factor’ from the pole outside,
the existing demand on the rest of the in home power distribution system, including the main circuit breaker panel itself, and
ALL the functioning circuit breakers,
the length of the branch circuit being used, and
condition of the wire in the walls, and
health of all of the receptacles, and
the health of all the inter-connections on that branch circuit, etc. especially those that feed our systems directly.
And remember those receptacles you plug into are contractor grade (meaning as cheap as possible yet will still pass the safety inspection, when 1st installed!), that is unless they have already been replaced AND the entire branch circuit feeding the gear has been ‘tightened up’, in which case there is a very good reason to believe you have already heard the impact that the removal of these
CP’s (Choke Points) can provide.
And those ‘contractor grade’ receptacles they used?
They used to cost 49¢.
I mention this because one definition of a
CP is where current or voltage is un-wantedly hindered or altered as it passes thru a passive device or component.
So a ‘standard’ LPS (Linear Power Supply) wants the power delivered in very short pulses of relatively larger amounts of current by a factor of ≈ 470% greater than the rated power would suggest. And even if you use the stock fuse as a means of knowing how much current the amp ‘needs’, the actual ≈ 3+ms current pulses are many times greater.
This is also where those I2T fuse rating come into play, as was mentioned previously.
And when the
ASCC test results are combined with these findings, a deeper look at what is happening emerges. Namely when the
ASCC results show ‘better’ results in terms of being able to dump greater amounts of current quickly, (I measured a peak current of between 1 to 2 KA’s on a 20A breaker) are combined with the results in this 2nd series of tests, it becomes very apparent that being able to ‘quickly dump current’ is significant.
The
ASCC Tests start here and continue for several posts over several pages.
https://www.head-fi.org/threads/the-diyrs-cookbook.781268/page-43#post-12738982
What stands out to me is that these results point back to the assumptions we often make on how things operate, that also have missing elements or aspects buried or obscured within them.
Such as, the way in which current actually flows into our gear.
Which includes a key factor, namely time.
As in how much time is actually spent flowing current.
So for me this single factor plays a key role in what is missing from the ‘debate’ over power cables and such.
And these 2 key factors, the ability to dump current quickly, and the short slice of time (relatively speaking) available for peak current flow, and which has a (relatively speaking) long pause between ‘ON’ pulses, means any hindrance in this ability to deliver fast rise time short duration peak current loads will likely affect the power supplies ability to meet it’s needs, not to mention the added demand the music adds to the picture.
This dynamic portion of this situation is beyond the scope of this inquiry.
These 2 key factors both point at the same conclusion, even on very low powered gear, such as our headphone gear.
The conclusion is that being able to deliver current pulses, QUICKLY yields better results, and ALL upstream ac power components can influence this ability to, in a very short amount of time (≈ 1ms), be able to dump maximal current, then stop.
Conversely ANY component that in any way restricts these short peak current flows will affect the downstream power supply in some way.
Just how this affects the rest of the power supply is also outside of this inquiry.
But I can suggest at least one such possible interaction.
That being, voltage starvation (sag) due to the inability to supply all of the power needed by the downstream load, at any moment in time.
These cumulative findings further imply that the gauge of the wire must play a role in this since these pulses are several times larger than what a 18w load would suggest.
IOW these pulsed flows of current are brief and are less than the duration of the ‘off’ window slices of time, AND are many times larger than would be commonly assumed.
Further, if I do some head scratching and figur’n, it looks like,
of that ≈ 3ms ‘on time’ window,
≈ 1+ms is spent rising up to the peak,
≈ 1+ms is spent at or near the peak current flow, and
≈ 1-ms is spent falling back to zero.
Which upon even further scribbling could be thought of as ≈1.6ms± of near peak current flow.
And this time slice of ≈ 1.6ms± must provide for ALL of the power needs for the entire ≈ 8ms window, which comprises the previous ≈ 5ms of the ‘off cycle’ and the ≈ 3ms of the ‘on cycle’.
As such any degree of limitation to the delivery of current flow will affect the amount of power delivered, to some degree.
How much would be very difficult to answer especially if we take into account the complexity of a musical wave form added to the load. Sine wave test tones are usually too regular to reflect the dynamics of music so we’d not likely see the peak demand for current, nor would I be able to measure it repeatedly, let alone accurately.
AND it also suggests that as the demand for more current rises (dynamic bass peak for instance) that the characteristics of that ≈ 3ms window will change, and that the rise and fall times will try to adjust as well, in order to have more time to transfer more current to meet the greater demand which also implies that the demand for peak current will rise, as well.
And it also suggests that during such times of greater demand for power, that the voltage to turn on the diodes lowers and allows current to flow sooner thus increasing the duration for additional current to flow during that ≈ 3ms window, and to a lessor degree the turn off voltage will adjust as well.
This is where the asymmetric wave form enters the picture for the current pulse.
(Note what follows is a thought experiment, so consider this one set of possible interactions with how the PS can influence the rest of the device’s functionality.)
ASSUMING
(Case #1) That the amount of current delivered is ‘sufficient’ to meet the demand of the downstream load or ‘reserve’, before the incoming AC voltage itself drops below the now replenished ‘reserve’ voltage.
(Case #2) That the current flow was insufficient to fully top off the ‘reservoir’, so in this case the PS is relatively ‘starved’, until the demand for current lowers enough so that the reserve can get fully ‘topped off’.
In case #1 as long as the downstream demand for power is met, any existing restrictions to quickly delivering current shouldn’t have any effect on the PSU and the rest of the device.
But in case #2 where there the PSU isn’t able to receive all of the power it ‘needs’, this ’shortage’ will be carried over to the next ≈ 1ms current dump window, where in turn these ’shortages’ can accumulate if the demand exceeds the ability to actually deliver the ‘needed’ current continues thru time.
Therefore, by removing
CP’s, which enable greater amounts of current to be delivered, means the ‘threshold’ where the PSU shifts from case #1 to case #2 (or visa versa) is altered such that case #2 happens less frequently and presumably for shorter periods of time as well.
This is where ‘
better’ ‘audio grade’ components can make a difference.
By enabling greater amounts of current to flow quickly, and as quickly ( if not more so), stop flowing.
So if a fuse, which is designed to be a ‘controlled’ restrictive safety device, actually impedes the flow of current then it will have an effect upon the ability to quickly dump current.
As can wire gauge, and even the ability of the plugs and receptacles to be able to deliver peak current pulses, not to mention the additive effect of all of them AND the wire in the wall, and the branch circuit breaker, and the rest of the load the panel is dealing with, all the way to the step down transformer outside the house.
All of the cumulative effects of the entire power distribution system can, and in multiple ways, restrict the ability to quickly deliver current.
And as I have continued my investigations and experiments, I have proved to myself, repeatedly, that removing these (
CP’s) does result in ‘
Better’ SQ.
So in total it isn’t JUST the power cables nor JUST the fuses, nor any one factor, but the interaction of all the components of the entire power distribution system that feeds our gear.
And while power cables and fuses are 2 of the easiest ways to make changes, there are others, such as ‘audio grade’ duplex receptacles, bigger gauge wiring, short (or as short as possible) dedicated runs from the breaker panel, tight connections everywhere, among other factors.
A case in point, I soldered the wires of the romex extension I had added to my existing dedicated branch circuit, AFTER I had already installed and used it for several months, and I noticed an immediate change for the ‘
better’.
In addition, the rest of the audio system’s existing
CP’s also play a role in how much of a noticeable effect being able to deliver these peak power demands will have.
Such that if there is a MAJOR
CP built into the system (say a dirty or worn out connector, or a damaged cable connection) this will also limit the system from delivering peak performance as well. And in my investigations until any/all MAJOR
CP’s ARE ameliorated, the system is operating with less than optimal results.
I call this busting thru a major log jam, where eradicating those MAJOR
CP’s, of which a poorly performing ac power feed is definitely one that can be an obvious detriment to the overall SQ, can lead to significant increases in SQ.
These findings also help to explain why I prefer hard wiring, where possible, as it further reduces both the number of possible
CP’s as well as the effects that all remaining
CP’s can make.
I hope that these findings are helpful in yielding a better understanding of the actual dynamics of the power being fed to our gear.
And why many do notice improvements after upgrading, improving, reducing/eliminating
CP’s from their power distribution that feeds their systems.
JJ
Fin.
