|
|
Mains filter test results
|
Ok, as promised, results of tests with various filters, power cables etc...
First test on my own system, Arcam DV-79, AVR-250, Mordaunt-Short Avant
908/905C/903S speakers, B&W ASW-1000 sub, Audio Innovations Silver
Bi-Wire speaker cable.
As promised, I've been loaned an Isotek power cable (£60) in order to
carry out these tests. I was also loaned an Isotek power distribution
block (not sure of price, I think they're about £150).
For comparison, I also used a standard IEC mains cable, and an IEC cable
with a filtered plug.
I normally have my system plugged in via a Masterplug surge protected
and RFI-filtered 4-way distribution block, £8 trade or £20 retail.
|
O dear!, in the advent of a lightning strike make sure your insurance
covers your gear!....
Course going to church on Sunday might appease the almighty so your
protected!.....
|
|
Yep.. no brains at all.....
|
For a test track I decided to use "So Far Away", from Dire Straits'
"Brothers In Arms" CD. The version in question is the SBM remaster.
Two other people were also present. One person was unsure, the other was
a complete sceptic.
First things first, I plugged the DVD player and amplifier into an
unfiltered block. I then played the test track.
I then switched power cables on the DVD player to use the Isotek, and
played the track again. This made no difference whatsoever - which is to
be expected, as the DVD player is merely acting as a transport. I
switched cables back again and played the test track to be sure, and
indeed there was no difference.
|
You presumably were aware thoughout of what arrangement was in use. You
also seem to have only done this a few times, not many times to form
abody of data on which any meaningful statistical analysis could be
carried out.
|
We didn't worry about setting up a blind test at this stage, as my
concern was simply to find out whether there was a difference,
perceived, psychological, whatever.
|
Anybody with a clue about experimental design or psychology could predict
the outcome of your test without bothering to do it.
|
The implication of what I wrote before was that if there appears to be a
difference (whether real or psychological) then it's worth the effort
and hassle of setting up a blind test. If there's no difference apparent
using a straight test then there's no point wasting time with blind testing.
However, as a straight test concluded that there was indeed a difference
(whether that difference was real or psychological is a matter for
debate, but the fact that three of us all heard the same effects would
suggest that the difference was indeed there) then it's perhaps worth
setting up a blind test.
As I appear to have won an Isotek power cable, at least for the time
being, I may well set up a blind test next time I've got a few hours to
waste. If I do so then I will of course post the results here.
|
|
From my point of view I'm satisfied with the results. Using an RFI
|
That would be fine for you - if you had not then posted your opinions here
as if your 'test' had any reliability or value for anyone else. :-)
Alas, since your 'test' was not conducted in an appropriate manner, your
report will be unlikely to have eithe 'satisfied' or been of any use to
anyone else.
[snip]
|
filtered 4-way strip in combination with a generic filtered IEC lead
will get rid of the RF hash on the mains supply, and thereby lower the
noise floor (and improve dynamics). These bits can be had at trade price
for about £8 for the 4-way block, and less than a fiver for a bare-ended
IEC cable and a filtered 13A plug. Use the two together and it's just as
effective as an Isotek cable (£60).
When you have the amount of PC hardware on your ring main as I do (WinXP
and NetBSD workstations in the dining room/office, 10 servers in the
attic, RISC PC in one bedroom, A3020 in another bedroom, plus a couple
of laptops, then all the Ethernet switches, wireless access point etc
all spewing RF onto the mains) then decent RFI filtering on the mains
does make a difference. And even if you're living out in the middle of
nowhere, with no PCs or SMPSUs for miles, it's worth having a surge
protected strip anyway.
Quick question - in the event of a lightning strike, what would you
rather have fried? £8 worth of surge protected mains distribution block,
or £1,000 worth of amplifier?
Bit of a no-brainer really, that one.
|
Indeed. :-) Particulary since a lightning strike may well crisp the
'surge protection' *and* the rest of the equipment locally connected. I'm
afraid that a small VDR isn't likely to be able to absorb the energy of a
nearby lightning strike. :-)
Nor is a VDR likely to be much use against RF spikes as it laregly converts
a voltage spike into a current pulse. So may still allow 'clicks' though if
the units you wish to 'protect' don't have decen PSU, grounding, etc.
If you wish to avoid clicks and pops from audio gear, then go for decent
kit, and perhaps suppliment it with *filters*, not a 'surge protection'
unit.
|
|
|
I then switched the power cable on the amplifier to use the Isotek
cable. Immediately there was an improvement in dynamics, percussion in
|
Which would make perfect sense, as by taking out RF interference what
you've done is lowered the noise floor. You may not be able to hear RF,
but if your amplifier is trying to reproduce it then it's increased the
noise floor, which will result in the dynamic range of the system being
reduced.
|
Glenn,
do you live somewhere like Brookmans Park or Holme Moss?. Never known a
bloke to be troubled by so much RF!......
|
|
You give no explanation of how the amplifier would be "trying to reproduce"
the RF and this then leads to it "inceasing the noise floor".
Thus you make a speculation, but give no mechanism.
Nor did you report any measurement or period of listening that showed that
you could detect a change in the audible noise floor. If this was occuring,
it would show most clearly when you were not playing any music, and the
effect would then be obvious if it was occurring.
Thus you did some 'tests' but obtained no evidence that the effect your
speculation assumes actually occurred.
|
|
Which is why you shouldn't design an amplifier with a ridiculously
extended frequency range.
|
|
Roderick doesn't like that, you know. It get's him mad and all.
|
|
particular had much more presence and depth.
|
...or at least that was the impression of those involved.
[snip]
|
I switched back to the standard IEC cable and listened again to ensure
that there was in fact a difference and that I wasn't imagining things.
The sound seemed flat after switching back.
Next I tried the standard cable with filtered plug. This again made
quite a dramatic difference to dynamics. Perhaps not quite so much as
the Isotek cable, but that may be the psychological bit coming into
play. But maybe not, as later tests showed.
I then brought filtered mains distribution blocks into play, and tried
|
Apparently with these Masterplug surge protected strips, Masterplug
themselves will guarantee to replace up to £4,000 worth of connected
equipment in the event of a power spike causing damage.
And as I've had a VCR get fried by a mains spike before (an old Ferguson
3V48... fortunately not a 3V43!) I tend to err on the side of caution
when it comes to power surges.
|
various combinations of esoteric and generic branded filters, unfiltered
cables etc.
What I discovered (and the two other people in the room also heard quite
clearly) was that a filtered distribution block makes a difference, but
there did not appear to be any audible difference between the Isotek
block at £150 or the Masterplug block. Both have RFI filtering and surge
protection, the Isotek has a metal case, the Masterplug is plastic. But
none of us could hear a difference between them.
|
The problem is that no-one else at this point can tell if your opinions
have any weight or not since your test method may not be reliable, and
you give no data which others could uses to assess the validity or
statistical reliability of what you report.
If you wish to obtain valid results which can be used as a reliable guide
then you would at least have to:
A) Do some tests where the results can be shown to depend *only* on the
sounds being produced. i.e. ensure that those involved don't know which of
the items being compared are in use at any point when having to make
decisions.
B) Do the tests enough times to allow the results to be assessed for level
of statistical significance.
It might also help if you'd tried actually measuring or examining the
output from the system to see if you could find any changes that might
account for the opinions you formed.
It would also help to take measures to check for, or control the risk
of interpreting an effect due to some other cause as being due to
what you may incorrectly assume. e.g. ensure that you have enough
test runs to deal with alterations in listening position (acoustics)
and reverse ordering to deal with psycho- or physio- adaptions due
to immdiately previous experiences.
In this particular case, it would have been useful to repeat the tests
in your friend's house, with someone else in another room either having
the PC (alleged to be the source of assumed 'interference') on or off,
and have you say which was the case simply by listening. If you could
do this repeatedly and get the answer correct then that would be a
useful result. But if you know the PC is on or off when listening, then
it is not really very useful.
Alas, in the absence of the above, the report you make tells us about the
opinions of those involved, but does not tell us if the cables, etc,
actually did anything to the output of the system. Nor if this had
anything to do with 'filtering'.
IIUC it has been explained on more than one occasion in the past that such
'tests' do need to be carried out in an appropriate manner if we wish the
results to have any reliability as evidence. It therefore seems a shame
that you seem not to have done this. The result is that your opinions
may or may not be well-founded, and no-one can tell which since either
possibility will seem equally probable to anyone other than yourself.
|
|
What you discovered was that a particular amplifier sounded better to three
people under certain circumstances, no more, no less.
|
Not quite. It was 'discovered' that they had this opinion... However the
'test' as reported shows no signs of being reliable beyond that.
|
|
Also, the Isotek cable sounded exactly the same when plugged into an
unfiltered or filtered block. The generic filtered cable sounded
slightly better when connected via an RFI filtered block.
The combination of generic filtered cable and Masterplug filtered block
sounded as good as using the £60 Isotek cable. Both sounded
significantly better than using an unfiltered cable.
So on this basis I would recommend the Masterplug filtered 4-way block,
£8 from CPC or anything up to £20 retail, along with a filtered plug on
the IEC cable (costs about £3.50 from CPC, not sure about retail).
However, the next part of the test was interesting. We then went around
to a friend's house, who has a similar setup to mine - Arcam CD73, A90,
B&W DM603 speakers, QED speaker cables (not sure which one, but nice and
thick).
We then performed the same series of tests, and obtained a similar set
of results. However, the differences on his system weren't anywhere near
as dramatic as on mine.
So we then shut down and unplugged his PC, in the next room but on the
same ring main, and repeated the experiments.
This time the differences were negligibly small. We heard a tiny
improvement when using either the Isotek or generic filtered cables, and
a similar improvement when using the filtered mains blocks. This time we
didn't hear any difference between a filtered and unfiltered IEC cable
when used with a filtered block.
We plugged the PC back in and fired it up again, and repeated several
times for consistency. Once again the results with filtered mains were
much more noticeable when the PC was turned on.
|
So this friend who believes his ears demand fancy speaker cables had never
noticed that his computer was degrading the sound of his system?
With two friends, you had the opportunity to do the test blind. Why didn't you?
Does anyone know what frequencies are removed by mains filters? How should I
choose between them? Is there a standard? Do they all offer the same
attenuation?
If not then, as with fancy speaker cables, what explains the variety when they
all have the same purpose?
cheers, Ian
|
So, in summary:
If you don't have any computer equipment (or anything with switch-mode
power supplies) on the same ring main as your hi-fi, buy a Masterplug
RFI filtered 4-way strip, if for no other reason than it incorporates
surge protection, so will protect your kit against voltage spikes.
(Believe me, I've seen quite a few PSUs fried by this type of thing.) If
it makes it sound better then that's a bonus. If it doesn't, well, at
least you know your equipment is protected.
If you have one or two PCs in the house that aren't left on all the
time, get the RFI filtered block. Try using a filtered IEC cable and
listen, if it makes a difference then keep using it.
If you've got a miniature version of Docklands Telehouse in your attic
or spare room, stick an RFI filtered 4-way block in, then use filtered
power cables from that. It's also probably worth putting filtered cables
on your PCs and other items with SMPSUs, as then you'll tackle the
problem at source (and additionally protect *that* expensive equipment
from spikes).
In the meantime, I appear to have "won" an Isotek power cable worth £60.
They want the distribution block back, but they've said I can hang on to
the power cable for now. Might try opening up the plug and see exactly
what type of filter they've put in...
|
Yes, see what's adding in extra resistance and or inductance.
Don't fancy doing it all again with a storage scope across the mains to
see what's happening to the waveform and voltage?, to see if the
filters/power cords etc are seriously affecting the mains power
increasing the source impedance etc, into the units under test?........
|
I was about to say exactly the same thing as Tony. I don't doubt that Glenn
and his chums thought that they heard some difference - and perhaps they
did. I'm certainly not going to join the " it's all just a load of bollocks
" brigade on this BUT, if adding in filtering ahead of the equipment really
DID produce a difference, then in order to evaluate the true validity of the
perceived improvement, we would need to know exactly what the filters were
filtering against, and how effectively. A high bandwidth storage scope would
be a satisfactory way of doing this. If the filters are indeed removing
mains-injected hash from a single computer's power supply, then firstly, you
would have to question if that computer's power supply and input filtering,
were working correctly. If there is sufficient level to have a siginificant
enough effect to hear, being introduced by this computer, then it would
almost certainly be operating outside of it's CE approval ( assuming that
it's got one ). Strictly speaking, if it hasn't, and is introducing hash to
the mains supply, you render yourself liable to prosecution ...
Secondly, if the computer is within spec regarding its emmissions, then you
would have to question why an expensive amplifier is being affected to a
degree that you can hear, by such a low level of mains borne noise ? If this
|
Could be a high-end amp with capacitors bypassing the psu diodes to reduce
'diode switching noise'
and let through RF - thus creating a market for special filtered mains cables.
|
If you use a set of bypass caps on the diodes it has the effect of blocking
any differential mode (Live-Neutral) noise via that route. If properly
referenced to RF ground it will also deal with common mode above ground. In
effect you can use it as part of ensuring all the power wiring is RF
shorted together and to local ground.
It also defines the RF impedance at the relevant points rather than
allowing it to change during each mains cycle and with current demand.
Hence in contrast with the implication of what you write, using such caps
to shunt the diodes can be a very effective way to block/reduce some of the
RF noise, clicks, etc. Provided it is done appropriately, of course.
|
|
is indeed the case, and the amp is CE approved, then it should not respond
to such low levels, introduced through its mains port. Indeed, if you
contact the manufacturer and ask nicely, they might be able to produce a
copy of the CE test report for you, or at least tell you what figures were
achieved for noise immunity.
Either way, if you want to progress this discussion further, in a sensible
way, we need to see some objective test results as to what the filters were
doing, otherwise, this thread will just degenerate into another long string
of insults and taunts and challenges, that are meaningless, and time
wasting, and do nothing for the credibility of the group.
|
|
|
Well, this would all be a little more convincing if the changes you
"observed" were ones concerned with symptoms that might possibly be
addressed by cleaner power. These would be essentially (as Dave
Plowman pointed out), an absence of clicks from devices like
thermostats.
|
And an absence of mains-bourne RF hash from the switching PSU in a PC...
(BTW, did you have to do an AOL and quote my entire post with only a
dozen lines of text? It's only one step away from top-posted HTML...)
|
Far better than clipping the part that actually needed a response.
And it is mains-borne, not mains-bourne - this has nothing to do with
small streams.
And once again, RF is NOT audio.
d
Pearce Consulting
http://www.pearce.uk.com
|
|
What you present here amounts to "I used Cillit Bang to clean my sink,
and yes, I was three inches taller as a result".
So no, yet again you have presented us with some nonsensical
pseudo-science to support something imaginary that you and your
friends convinced yourselves had happened.
d
Pearce Consulting
http://www.pearce.uk.com
|
|
|