Hi again,
An explanation in 4 parts:
PART 1:
To start: There is a difference between a "surge protector" and a "power conditioner".
At the simplest level, a "surge protector" is a fuse or a circuit breaker. If the filament inside the fuse had too much current going through it, it melts and the circuit is opened - therefore everything downstream stops getting current and the lights go out. In this case, the "surge" is in the current - typically caused by the fact that it's being provided with 110 or 220 V RMS, and you (1) put in a too-small resistor somewhere (like a fork in a toaster, for example) or (2) put a bunch of resistors in parallel (say, 10 toasters making toast while you're heating up something in the microwave and using a hair dryer at the same time). Since I = V/R, then if R is very small, I gets very big and fuses get very unhappy - hopefully quickly enough to prevent something bad from happening somewhere else... That way the little filament gets hot and melts before the wires in your wall get hot and burn the house down.
The basic job of a "power conditioner" (a.k.a. "line conditioner") is to remove noise from the incoming power - so you get a perfect sine wave at 50 or 60 Hz (depending on where you live) with an RMS voltage of some level (say, 110 or 220 V RMS, depending on where you live) within the tolerance of whatever the power company promises. So, it should behave more like a bandpass filter. For example: the power company (in theory) delivers a sine wave to your house, and then your refrigerator and your air conditioning unit turn on and cause the sine wave to distort. The reason this would happen is that some devices (like motors and old-fashioned light dimmers, for example) use only a portion of the full sine wave in time - say, the first 90º out of the full 360º to get 25% of the light coming out of your lamp. So the load on your incoming power is varying in time in an unexpected way - which can cause the sine wave to have a slightly different shape (= distortion). When a sine wave is distorted, you typically have extra high-frequency components that are generated. In theory a "power conditioner" will remove those high frequency components and get you back to a sine wave.
HOWEVER: this is only a portion of the story, so let's look at some of the other portions.
PART 2:
You wouldn't buy a surge protector or a power conditioner without some device to protect. For the purposes of this discussion (because it's the question that started the thread) we'll assume that that device is a B&O active loudspeaker, so let's look at what's inside there.
On a basic level, a B&O loudspeaker contains the following (in reverse order of signal flow):
- loudspeaker drivers
- "power amplifiers" (which would be better called "voltage amplifiers that can deliver as much current as the driver demands" - at least theoretically)
- digital signal processing (with an ADC at the input and DACs at the outputs. This also includes a little analogue circuitry before and after to make the ADC and DACs happier about "talking" to other devices)
To keep all this running, we either have
- a battery (that provides a number of DC voltage "rails" generaged by chemistry)
- a power supply (that provides a number of DC voltage "rails" by converting the mains AC input somehow)
Without knowing HOW they work, the amplifiers and the analogue inputs and outputs of the processing (in an older loudspeaker like a Beolab 8000, the processing itself is analogue) do the following:
- "look at" an incoming signal (a voltage that changes in time)
- create a new version of that signal (i.e. in the amplifiers, a bigger version) using a constant voltage source (the battery or power supply delivering the DC), modulating it, and sending that to the output.
The error that CAN occur is in the assumption in that last statement that the voltage supply really is DC. Poorly-designed audio circuitry (for example, an amplifier) is designed based on the assumption that the power supply is delivering DC (in this case, constant voltage - with as much current as is required as its needed). If you make this assumption, AND the power supply does not deliver "perfect" DC (for example: 15 V DC that modulates ± 0.15 V AC at 120 Hz - this is called "ripple" in the voltage rail, in one common example), then this modulation in the voltage rail MIGHT cause the audio output to also ripple by some amount, possibly at the same frequency.
So:
IF
you have a poorly-designed power supply
AND
the input of that power supply is anything other than a perfect 110 V RMS / 60 Hz (or 220 V RMS / 50 Hz) sine wave
THEN
the power supply's output MIGHT have a modulation in its "DC" output
AND
IF
you have an audio circuit (for example, an amplifier) that is intolerant of variations in the voltage supply (therefore it is poorly-designed)
AND
you have a power supply that does not deliver "perfect" DC (for some reason) (therefore it is poorly-designed)
THEN
You MIGHT be able to hear (or at least measure) some artefact at the audio output of the device.
PART 3
There are two basic types of AC-to-DC power supply
1. the good'ol' "linear" power supply, which used
- diode-based rectifiers to convert +/- sine-wave AC to positive-only (and/or negative-only) "rectified" (absolution value of a sine wave) AC at twice the frequency of the input
- big capacitors as tanks of electrons to smooth that wave to something that looked a little like DC with an AC ripple added to it.
- regulators to hold the output at some voltage that's lower than we had, burning off unnecessary power as heat
2. a switch-mode power supply which:
- generates a square wave that has an adjustable pulse width
- smoothes that pulse wave into a DC level with capacitors
- adjusts the width of that pulse wave so that its average voltage over time is the voltage you want
- optionally: some microprocessor-based monitoring of the output voltage and/or the incoming current, to make sure things are behaving, and the power supply doesn't ask for too much from the mains and blow a fuse in the basement.
In both cases, the power supply doesn't produce "perfect" DC - but the "error" (the AC component) is different between the two design philosophies - and will also be different from design to design, and with different loads on the supplies.
Note that the modulation on the DC output of a power supply can also occur because it can't deliver what's being required of it. So it's not ONLY the fault of the mains input "signal" quality.
PART 4
Having done all of the setup listed above, I'll make a bit of a blanket statement:
These days, if you have an audio device (say, a fully-active loudspeaker that has an analogue input, for example) that needs a power conditioner, then someone didn't do a very good job designing the device.
In other words:
- the audio circuit (i.e. the amplifier inside the fully-active loudspeaker) should be designed to accomodate some modulation of its power from the power supply
- the power supply should be designed to accomodate something other than a perfect sine wave from the power company (because it's a very safe assumption that it won't be a perfect sine wave)
So, if these are done well, then "conditioning" the incoming mains power with an external device is only removing something that would have been removed inside the device anyway...
This is especially true of a Bang & Olufsen loudspeaker, because we have designed the entire device, knowing the behaviours of the amplifiers, the analogue circuitry AND the power supply. So, we don't need to make any assumptions about how each of those things behaves (with respect to the voltage rails inside the product). Measuring those behaviours to their worst-case limits is on the basic list of the "things-the-engineers-do" to make sure that the audio output is not audibly infected by these issues. So, when I asked the electrical engineers here about whether you need a power conditioner for Beolab 17 or 18 (or 19 or 20) their paraphrased answer was "It probably won't make things worse, but why would you bother?"
(Don't get worried about the fact that I just said "audibly" in that last paragraph - I'm not trying to weasel out by saying "you can't hear it anyway, so it doesn't matter." I'm saying "any such artefact is measurable - but if it's at a level that is significantly below the signal level, or the noise floor of the system, then it is of no concern..." I'll never say "it's not there" - I'll just say "perhaps there are more important things to worry about than this particular issue.")
Cheers
-geoff
P.S. This doesn't answer the specific question in this thread about what's in the BL50 & 90 power supplies that could result in a "power conditioner" making things worse - but I think that I've written enough here - and besides, I wrote enough about that in the
