Topaz Low-Capacitance Isolation Transformers - for Affordably Clean Power
post-13659138
Post #16 of 532
Joined
Apr 28, 2007
Messages
5,493
Reaction score
1,059
Joined
Apr 28, 2007
Posts
5,493
Likes
1,059
For the most passionate of students, here's a very well-written article, which, among other things, provides very succinct definitions of the different types of noise.

The author, Marc Dekenah, all but carves 11th and 12th Commandments in stone.

Paraphrasing... Thou shall not use the term "Normal-Mode Noise," but rather "Differential Noise" and "Transverse Noise" is a superset of "Differential Noise;" they are not one-in-the-same.

Let the reader be schooled. LOL

http://www.marcspages.co.uk/pq/3270.htm

Here's an index to several other topics - all, equally well-written and in-depth:

http://www.marcspages.co.uk/pq/0000.htm
 
Last edited:
     Share This Post       
  • Like
Reactions: leeperry
post-13663849
Post #18 of 532
Joined
Apr 28, 2007
Messages
5,493
Reaction score
1,059
Joined
Apr 28, 2007
Posts
5,493
Likes
1,059
Erico_Differential_vs_Common-mode.jpg


Erico_Iso_Transformer_CMNR.jpg

https://www.erico.com/catalog/literature/TNCR016.pdf

Here's some very compelling discussion regarding that PDF's statement that Isolation Transformers cannot attenuate Normal-Mode Noise (Differential-Mode Noise):

https://electronics.stackexchange.c...attenuation-for-differential-transients/92015

This has me doing some serious head-scratching, as the first reply to the poster's question makes so much sense to me. :worried:

UPDATE on 27 Aug 2017: I've concluded, through open inspection of my 19095-32, that the Topaz Ultra-Isolators achieve -65 dB of "DM" noise reduction (Differential-Mode or Normal-Mode or Transverse-Mode noise reduction), not through the use of Y-Caps at the input, an X-Cap across the output, and/or surge suppressing MOVs, either before or after the coils, but rather through the Faraday shield design, which is most likely a double or triple shield, with grounding of the shields done as shown in some of the graphics that follow in my subsequent posts, below. I see no evidence of the Topaz 19095-32 using any caps or MOVs.
 
Last edited:
     Share This Post       
post-13664232
Post #19 of 532
Joined
Apr 28, 2007
Messages
5,493
Reaction score
1,059
Joined
Apr 28, 2007
Posts
5,493
Likes
1,059
More food for thought... Check out the last diagram, which talks about introducing capacitors to reduce differential-mode noise:

www.murata.com/~/media/webrenewal/products/emc/emifil/knowhow/26to30.ashx

This ties into the previously linked PDF from Topaz, where they mentioned the use of capacitors to deal with "extreme transverse noise":

http://www.surplussales.com/Transformers/Pdfs/tp-91092-31_N.pdf

Note, too, that in this Topaz PDF, they mention how the Topaz Ultra-Isolators use "line noise suppressors designed to eliminate virtually all transverse-mode noise that results from common-mode noise on the Primary."

I think what it's saying is that the common-mode noise coming into the Primary gets converted to normal-mode noise coming out of the Secondary (which I've read in several places happens with all Isolation Transformers), but they deal with it by adding "line noise suppressors" - whatever those are - which would explain how the Topaz/MGE/Daitron ITs magically manage to offer -60 dB (1000:1) of NMNR, in addition to the -146 dB (20,000,000:1) CMNR - which is the only kind of NR you're supposed to be able to get from a 1:1 transformer that's equipped with a Faraday shield (and the lower the inter-winding capacitance, the greater the NR at higher frequencies.)

But what exactly are these "line noise suppressors" that Topaz is claiming as capable of doing -60 dB of Normal-Mode NR?

UPDATE on 27 Aug 2017: Again, I've all but disassembled the windings of my 19095-32, but I don't see anything but wires coming out of the core - there are no caps, MOVs or other filter circuits, that I can see.
 
Last edited:
     Share This Post       
post-13664597
Post #20 of 532

uncola

1000+ Head-Fier
Joined
Oct 5, 2007
Messages
1,263
Reaction score
143
Joined
Oct 5, 2007
Posts
1,263
Likes
143
I've seen a lot of IEC inlets with built in emi/rfi noise reduction using uh.. I guess caps and resistor networks? maybe it's that? I have one but it's 6 amp max input so not good for a big iso transformer
 
     Share This Post       
post-13665271
Post #22 of 532
Joined
Apr 28, 2007
Messages
5,493
Reaction score
1,059
Joined
Apr 28, 2007
Posts
5,493
Likes
1,059
I've seen a lot of IEC inlets with built in emi/rfi noise reduction using uh.. I guess caps and resistor networks? maybe it's that? I have one but it's 6 amp max input so not good for a big iso transformer
Bingo! I think you've guessed correctly! Yesterday, I posted a comment, to an existing thread at electronics.stackexchange.com, and got this response from a guy named Tony Stewart, whose answers I find difficult to follow, but...
CM = -146dB @ 100kHz , DM= -65dB @ 100kHz Considering attenuation for a Pi Filter at 100kHz, this means it has a good line filter included rated at 100kHz not 60 Hz. It is reasonable. Floating neutral ignores capacitive coupling of CM noise which does occur. . A CM choke and Pi filter Y caps are needed for CM Noise and X cap for DM noise. – Tony Stewart. EE since '75
I had referenced the MGE PDF (one of the first links I provided in this thread), where the 0.0005 pF models are claimed to offer -146 dB of Common-Mode noise reduction and -65 dB of Normal-Mode reduction. This Tony Stewart fellow, uses the acronym "DM" for Differential-Mode.

I had also asked what influence is had on noise reduction, if any, by floating the neutral of the secondary instead of having a grounded neutral secondary (the latter being what we have with the Topaz/MGE/Daitron transformers.)

My interpretation of his response is that first, it's "reasonable" for the Topaz transformers to claim both -146 dB of CM and -65 dB of DM noise reduction, but the DM NR can happen only by way of "a good line filter included rated at 100kHz not 60 Hz" - which is basically his interpretation of the MGE PDF's use of the phrase "line noise suppressors."

He goes on to add, somewhat cryptically for my thick-headedness, to say that a floating neutral [at the secondary] "ignores capacitive coupling of the CM noise."

Which begs the question: What does he mean by "ignores?" Is he saying that a floating-neutral Secondary, in and of itself, will remove all CM noise that manages to pass from the Primary to the Secondary by way of capacitive coupling? We know (I know) that the whole purpose of the Faraday shield is to combat that capacitive coupling of CM noise, some of which still gets through, of course, even if your inter-winding capacitance is as low as 0.0005 pF. But... What exactly does he mean when he says "a floating-neutral secondary ignores capacitive coupling of the CM noise?" I have to conclude that there would be no need for Faraday shields in a 1:1 transformer, if all you had to do is cut the ground from the Secondary's neutral - to remove CM noise that comes into the secondary, and since we don't see any such products out there, "ignores" does not mean "removes."

So what exactly does he mean by "a floating-neutral secondary ignores capacitive coupling of the CM noise?"

Unfortunately, they have an extraordinary policy at electronics.stackexchange.com, that prohibits the asking or questions beyond whatever question was posted by the OP of any thread! Give me a break! That's counterproductive! You can't have an expanding discussion. I was told I have to post any related questions, as brand new threads. As yet, I've not bothered to start a thread with the title: "What does Tony Stewart mean when he says, 'a floating-neutral secondary ignores capacitive coupling of the CM noise?' " LOL

He does go on, however, to say: "A CM choke and Pi filter Y caps are needed for CM Noise and X cap for DM noise," which reminded me of the Topaz Installation Instructions, previously mentioned, above:

Excerpt_Topaz_Installation_Instructions.jpg


THAT is what Tony Stewart calls an "X-cap" - it straddles the output (Line and Neutral) of the secondary - to deal with Normal-Mode/Differential-Mode/Transverse-Mode noise, but apparently, the Topaz/MGE/Daitron transformers already have some kind of X-cap in place to deal with transverse-mode noise. The PDF I've referenced above is suggesting that an additional X-caps can be added to deal with the "Special Noise Problem" of "Extreme Transverse-Mode Noise."

UPDATE on 27 Aug 2017: No - they don't already have any kind of internal Y-Caps or X-Caps! In fact, if you read the first sentence in the graphic immediately above, it says:

"Topaz Ultra-Isolator line noise suppressors are effective in removing both types of noise."

The context of this sentence suggests that the entire product (the whole transformer) is a "line noise suppressor" - not just some circuit within the product.

Similarly, further down, it reads:

"Topaz Ultra-Isolator line noise suppressors are designed to eliminate virtually all transverse-mode noise that results from common-mode noise on the primary."
Once again, the phrase "line noise suppressors" is obviously referring to the entire transformer, not to some unidentified components within the transformer.

Now read the part I had highlighted in yellow. They're explaining how an optional X-Cap can be installed across the output of the Secondary, to deal with extreme transverse-noise coming into the Primary, from the mains. This in itself, suggests that no caps are implemented in the design, as shipped.

Nope! The Topaz units must be achieving their specified -65 dB of Normal-Mode noise reduction by way of a proprietary Faraday shield design. (See my later posts, below.)
 
Last edited:
     Share This Post       
post-13665338
Post #23 of 532
Joined
Apr 28, 2007
Messages
5,493
Reaction score
1,059
Joined
Apr 28, 2007
Posts
5,493
Likes
1,059
I couldn't remember where I had seen the following diagram, but after a good deal of searching, I've found it again:

Diagram_of_Y-Cap_and_X-Cap.jpg

The entire article deserves examination...

http://www.murata.com/~/media/webrenewal/products/emc/emifil/knowhow/26to30.ashx

... but that last diagram clearly illustrates the difference between a Y-cap and an X-Cap, as mentioned by Tony Stewart.

Given that the 0.0005 pF Faraday shield of the Topaz transformer is already providing -146 dB of Common-Mode noise suppression, there's not much point in adding a Y-cap for additional Common-Mode noise attenuation, but... the -65 dB attenuation of Normal-Mode noise, claimed by the Topaz/MGE/Daitron tranfsormers, must be accomplished through use of an X-Cap (that is already factory-installed, across the Secondary output).

UPDATE 27 Aug 2017: Nope! I've looked. There are no caps inside the Topaz 19095-32, at either the Primary or after the Secondary. And, if there were, they wouldn't invite you to optionally install your own, as outlined in the instruction sheet, referenced in my previous post.
 
Last edited:
     Share This Post       
post-13665369
Post #24 of 532
Joined
Apr 28, 2007
Messages
5,493
Reaction score
1,059
Joined
Apr 28, 2007
Posts
5,493
Likes
1,059
     Share This Post       
post-13665393
Post #25 of 532
Joined
Apr 28, 2007
Messages
5,493
Reaction score
1,059
Joined
Apr 28, 2007
Posts
5,493
Likes
1,059
Wouldn't this be easier to find and cheaper (and looks better): http://emotiva.com/product/cmx-2/ http://emotiva.com/product/cmx-2/
That's an interesting product. It's certainly less expensive, smaller, lighter and easier to purchase than a used Topaz transformer, but I suspect it is much better at normal-mode noise suppression (surge protection) than common-mode noise suppression - which is what the Topaz 0.0005 pF transformers do, like nothing else can.

If I had a Topaz that was making lots of acoustic noise, as some people have reported with theirs, I would definitely consider front-ending the humming Topaz transformer with this Emotiva product - specifically to elimate the DC offset on the mains power, that is attributed to transformer hum.

Transformers can also remove any DC offset before it reaches the load, but if the transformer is humming loudly (and running hot) due to DC offset, I'd rather block the DC offset with something like this Emotiva product, before the transformer sees it (instead of putting the transformer inside a foam-filled box, the way John Swenson does with his very noisy Topaz, so that he can enjoy listening to his music!)

Here's a paper on the subject of transformers creating acoustic noise. DC offset and its sources are discussed:

http://www.idc-online.com/technical...MEASURING_ACOUSTIC_NOISE_EMITTED_BY_POWER.pdf
 
Last edited:
     Share This Post       
post-13665893
Post #26 of 532

foreverzer0

100+ Head-Fier
Joined
Feb 11, 2008
Messages
418
Reaction score
53
Location
Bay Area, CA
Joined
Feb 11, 2008
Location
Bay Area, CA
Posts
418
Likes
53
That's an interesting product. It's certainly less expensive, smaller, lighter and easier to purchase than a used Topaz transformer, but I suspect it is much better at normal-mode noise suppression (surge protection) than common-mode noise suppression - which is what the Topaz 0.0005 pF transformers do, like nothing else can.

If I had a Topaz that was making lots of acoustic noise, as some people have reported with theirs, I would definitely consider front-ending the humming Topaz transformer with this Emotiva product - specifically to elimate the DC offset on the mains power, that is attributed to transformer hum.

Transformers can also remove any DC offset before it reaches the load, but if the transformer is humming loudly (and running hot) due to DC offset, I'd rather block the DC offset with something like this Emotiva product, before the transformer sees it (instead of putting the transformer inside a foam-filled box, the way John Swenson does with his very noisy Topaz, so that he can enjoy listening to his music!)
Actually, if you take a closer look it's supposed to do both. They also have a 6 outlet that's only common mode line filter and specifically states the output is free of "high-frequency line noise" whereas the 2 outlet product I linked to before does this plus dc offset correction.
 
     Share This Post       
  • Like
Reactions: zilch0md
post-13669010
post-13671424
Post #28 of 532
Joined
Apr 28, 2007
Messages
5,493
Reaction score
1,059
Joined
Apr 28, 2007
Posts
5,493
Likes
1,059
Last edited:
     Share This Post       
post-13671793
Post #29 of 532
Joined
Apr 28, 2007
Messages
5,493
Reaction score
1,059
Joined
Apr 28, 2007
Posts
5,493
Likes
1,059
LinearCircuitDesignHandbook_ITs copy.jpg

UPDATE 27 Aug 2017: I strongly suspect, but really can't tell by examining my partially disassembled Topaz 19095-32, that they were designed with a double Faraday shield - not shown in any of the examples in this graphic.
  • I can only conjecture that the first of two Faraday shields is connected to the Primary's Neutral - providing some Normal-Mode noise reduction.
  • I must also conjecture that second of two Faraday shields is connected to ground (as with most 1:1 isolation transformers).
  • I know from testing continuity between the outlet prongs that the Secondary's Neutral is grounded - which greatly reduces the probability, though still possible, that they have a third shield connected to Secondary's Neutral. (What's the point, if you've already got a shield connected to ground, as is the Secondary's Neutral - unless, of course, doing so, somehow diverts just a little bit more Normal-Mode noise to Ground before it hits the Secondary's outputs?)

Topaz_Shield_Design_Conjecture.jpg
 
Last edited:
     Share This Post       
post-13683420
Post #30 of 532
Joined
Apr 28, 2007
Messages
5,493
Reaction score
1,059
Joined
Apr 28, 2007
Posts
5,493
Likes
1,059
Last edited:
     Share This Post       

Users Who Are Viewing This Thread (Users: 0, Guests: 1)

Top