[Speedskater]

While this paper is from back in 2008, I guess that I never noticed it before. The first five pages are excellent (the 6th page is about their product).

Power Line Noise: How Series Filters work (and why they don't always)

"Most power-line treatment products depend upon series filters to correct AC line noise problems in the audio/video system. However, these series filters, unlike MIT’s parallel filter system, have inherent flaws that make them ineffective at many vital frequencies in A/V use and cause them to add more noise and distortions than they remove."

http://www.mitcables.com/pdf/powerline_noise.pdf

[Folsom]

Wow, I hope no one cared when that paper was released at MIT.

There is no "filter" design in it.

A monkey could conclude that parallel is best. But just because you know that's better, doesn't mean you have a design that works that well. Pi audio's PFC network might be close, maybe not, I haven't tried one.

Also not everyone uses their exact configuration they described. Plus real world application varies a bit more. Again sometimes you have to listen to know. I can tell you for sure that while I use a "traditional" filter in a sense, it doesn't have notable problems they describe; and even if it did no one would care at all because it's not affecting music in a bad way.

If they think it's traditional then their idea of inductors is probably back to the days of the Grey monstrosities. Furthermore their ground examples are ludicrous. Safety wise there are serious limits to have much inductance and resistance you can place there before a breaker won't trip. 

[Speedskater]

While I'm very surprised that MIT Cables would publish a good engineering paper, I'm not surprised that you would be unhappy with a good engineering paper from any source.

[jtwrace]

While I'm very surprised that MIT Cables would publish a good engineering paper, I'm not surprised that you would be unhappy with a good engineering paper from any source.

[Folsom]

While I'm very surprised that MIT Cables would publish a good engineering paper, I'm not surprised that you would be unhappy with a good engineering paper from any source.

Good? That paper is a worthless joke. Maybe it makes you feel good because it confirms some sort of moral-ego investment you have, but it's about as interesting as a bogus anti-gmo blog post by a yogi.

[Folsom]

I re-read some of it, and there is a lot of wrong concepts represented in it. It's criticizing the real word use, and yet it fails to make an applicable example.

"The series filter is best designed
for a constant load (current) and
a particular, known load impedance."

This is only true to a point. First common CMC/DMC's these days are too low of resistance, it's totally negligible for the point that they can slow current when sized right. But more importantly they are forgetting about how this stuff works. They are saying a 50 ohm source and 50 ohm load isn't good to model, that's true. But in the real world source impedance (noise) rises all the time, and it's often considerable. What happens when it rises? The Q of the filter goes down. The type of filter they don't like improves under real world testing, not comes to a fault because it's not under a somewhat applicable model that is often used.

They've overlooked the biggest benefit to balanced connections, too.

I might be able to go on for awhile. However, what's the point? Try filters and pick the one you like.

[sts9fan]

Are you an engineer by training or is your opinion based on Internet learnin?

[Speedskater]

'Salis' go to a technical library and borrow a copy of:

"Electromagnetic Compatibility Engineering"
by Henry W. Ott,
publisher: John Wiley & Sons, hardcover 872 pages, 566 figures, 65 tables.
Publication date: August 2009, ISBN#: 978-0-470-18930-6.

http://www.hottconsultants.com/EMCE_book_files/emce_book.html

Parts of the book are visible on-line. Or better yet, spend $100 and buy a copy, it will be the best reference book you own. (on this topic)

It's a real big book so don't try to read it cover to cover.
You will find the same information as the above paper, but with a lot more words.

[Folsom]

Are you an engineer by training or is your opinion based on Internet learnin?

I haven't gone to engineering school. But I have gone to school for the basics (not the net).

What I'm talking about now I learned from a much more formidable mind than my own on the subject.

MIT's paper here is heads in the cloud, and not well thought out theory to disprove "series" devices. There are slower things in a system than the ability of modern CMC/DMC usage; which can be oversized anyway. Take a look at Audience Adept conditioners for example. No reported loss in dynamics, only increased. They do exactly what the MIT paper says not to do. Plus people upgrade from MIT's product all the time. I'm sure it does something very beneficial, but I'd just buy a Pi unit that uses the same idea for way less (probably better too). But then again I'd rather just go bigger and better anyway.

It seems like they could just tell you that their product is good, show you how it works, instead of banking on fact that idealism concepts should over-rule real world use, by writing a paper that is rather wrong by todays standards. I mean they say that "series" filters look good on paper and with 50/50ohm tests but actually they tend to look worse under those circumstances than real world situations. What would they say about power transformers?

Speedskater, there is no local copy here; you're welcome to buy me a copy.