Discussion of AC mains isolation transformers

[skatbelt]

I've received the Jensen transformer. It is the manufacturer's advice to place the transformer as close to the subwoofer input as possible. Below their full explanation. This made me think: does this also apply to the output of the subject of this thread, the AC isolation transformer? I mean, is it best to place the IT as close to the equipment as possible?

 

----

There is no “free lunch” in transformer design. In ISO-MAX models with an “I” in the model number, the Faraday shields responsible for extremely high noise rejection unavoidably makes their output sensitive to capacitive loading. Since the cable that connects isolator output to equipment input is the main source of capacitance, it must be kept reasonably short. Capacitance up to 200 pF, about 3 feet (1 meter) of standard cable, will preserve the rated high-frequency bandwidth. There is no restriction on length for the cable that connects equipment output to isolator input

----

[Speedskater]

On 3/1/2018 at 2:08 PM, skatbelt said:

I've received the Jensen transformer. It is the manufacturer's advice to place the transformer as close to the subwoofer input as possible. Below their full explanation. This made me think: does this also apply to the output of the subject of this thread, the AC isolation transformer? I mean, is it best to place the IT as close to the equipment as possible?

For a large I.T. wired as a Separately Derived System it is best to install it reasonably close to the audio equipment. But too close and it may add audible buzz or electrical hum.

[wwc]

After, skimming this long thread, I thought I might try the Topaz, Eaton or Elgar with the low capacitance (.ooo5 pf).  

 

My first question: Is there any sound quality reasoning to choose one over another?

(I know the Topaz can produce some audible hum.)

 

And could I get advice on my proposed set up?  That is:

 

Cisco Switch (with Uptone Lps 1.2) > Roon Nucleus Server (with SBooster lps)  >  DAC  >  Topaz (and power strip)  > Richard Gray 1200c >  Dedicated 20 amp wall outlet #1 .   That Amp bypasses everything and goes to dedicated wall line #2.

 

The Richard Gray1200c uses two large parallel chokes to supply a generous supply of on demand power.

 

My question:  Is the RG 1200c compatible with the Topaz iso setup, or is it in any way undermining

it's leakage loop reduction function?

 

Thanks much!

 

 

 

[wwc]

Follow up to my question above:  I see now that the RG 1200 (which has balanced surge protection) will probably mess with the Topaz/power strip leakage loop application.  That is, if it has MOV's for surge protection...

 

Any ideas how I could integrate the RG into the circuit upstream of the Topaz?

 

Thanks

 

[sockpit]

Here's a theoretical puzzle that's been keeping me up at night.  I'm hoping @JohnSwenson or one of his better informed understudies might offer their views and suggestions.

 

I have my entire office system powered by a small Topaz (91095-12) rated at a mere 500VA (.005 capacitance).  It's all pretty simple:  dedicated outlet => Topaz => both a JS-2 and my power amp (100 watts).  JS-2 => my DAC and to an LPS-1.  LPS-1 => microRendu 1.4 => DAC.

 

In this setup, I am able to plug both the power amp and the JS-2 directly into the duplex receptacle on the Topaz and everything is covered (thanks to two rails on the JS-2 feeding DAC and mR).  And, according to John's theory, that's the best way to do it for reasons I only vaguely understand.  That's ok, I come to CA to just do what I'm told.

 

Now here's the twist and quandary:  My DAC is an Exogal Comet and the amp is an Exogal Ion.  The latter is powered, believe it or not, by a large 24V. 9.2A Meanwell SMPS (#GST220A24-R7B).  The worry in such a set up, I've learned from reading all these threads, is that the big ol' Meanwell is presumably pumping noise back into my Topaz, which could, in turn, show up again in my DAC and mR via the JS-2.  Is this correct?

 

There is an alternative to the above setup: there is a separate receptacle (not a dedicated one) very nearby.  If I plugged the Meanwell into that receptacle, presumably it would add its garbage to a totally separate circuit and not pollute my DAC.  (It would also be sure to get all the power it needs, if somehow the Topaz's 500VA wasn't up to the demands.)  But would there be a downside to separating these two components from the same receptacle?

 

To make things even more interesting, the Comet and Ion have a unique relationship to each other.  They are like two interacting computers more than like a preamp sending a signal to an amp for analogue amplification.  In fact, their "interconnect" is a single HDMI cable that permits them to communicate as one unit in processing the signal digitally--including a fully digital amplification process that is apparently quite unique.  Might this make them less liable to ground loop problems than typical equipment discussed in this thread? Is that irrelevant because of the LPS-1 anyway?

 

Here is the best description of this new and supposedly totally unique technology I've found for anyone interested in the technical details behind my question.  http://highfidelity.pl/@main-732&lang=en

 

Thanks in advance for anyone willing to give this and educated guess.

 

[look&listen]

8 minutes ago, sockpit said:

If I plugged the Meanwell into that receptacle, presumably it would add its garbage to a totally separate circuit and not pollute my DAC.  (It would also be sure to get all the power it needs, if somehow the Topaz's 500VA wasn't up to the demands.)  But would there be a downside to separating these two components from the same receptacle?

Try and listen. Easy.

[wwc]

Follow up  to follow up:

 

I've been told that while the Richard Gray has MOV's they are not connected to the signal path, rather they simply send a signal to trip the circuit breaker in the case of an overload.  So that's not an issue.

 

Also, the RG is not filtering, rather it   stores and releases power on demand through induction rather than capacitance so it will correct line impedence not create  impedance issues.

 

So it was suggested to me to go ahead and connect all components into the RG 1200c in lieu of the power strip.

[wwc]

Can the Topaz power cable be switched plug and play

(I'd like to use my own PC) or does it need to be re wired?

[sockpit]

1 hour ago, look&listen said:

Try and listen. Easy.

But that’s no fun. 

[wwc]

Correction, I now know the RG 1200 is compatible.  It's MOV's are not part of the signal chain.

[One and a half]

13 hours ago, wwc said:

After, skimming this long thread, I thought I might try the Topaz, Eaton or Elgar with the low capacitance (.ooo5 pf).  

 

My first question: Is there any sound quality reasoning to choose one over another?

(I know the Topaz can produce some audible hum.)

 

It totally depends what's available as a priority. If you choose either, you won't be disadvantaged.

 

13 hours ago, wwc said:

 

And could I get advice on my proposed set up?  That is:

 

Cisco Switch (with Uptone Lps 1.2) > Roon Nucleus Server (with SBooster lps)  >  DAC  >  Topaz (and power strip)  > Richard Gray 1200c >  Dedicated 20 amp wall outlet #1 .   That Amp bypasses everything and goes to dedicated wall line #2.

 

The Richard Gray1200c uses two large parallel chokes to supply a generous supply of on demand power.

 

My question:  Is the RG 1200c compatible with the Topaz iso setup, or is it in any way undermining

it's leakage loop reduction function?

 

Thanks much!

 

 

 

I would keep all the IT gear on the one source and from the Roon Server down to the amp on another source.

The Topaz removes common mode noise very well, but the IT gear produces a lot of it. So use the Topaz from the server down, and the RG for the IT gear.

[One and a half]

10 hours ago, sockpit said:

Here's a theoretical puzzle that's been keeping me up at night.  I'm hoping @JohnSwenson or one of his better informed understudies might offer their views and suggestions.

 

I have my entire office system powered by a small Topaz (91095-12) rated at a mere 500VA (.005 capacitance).  It's all pretty simple:  dedicated outlet => Topaz => both a JS-2 and my power amp (100 watts).  JS-2 => my DAC and to an LPS-1.  LPS-1 => microRendu 1.4 => DAC.

 

In this setup, I am able to plug both the power amp and the JS-2 directly into the duplex receptacle on the Topaz and everything is covered (thanks to two rails on the JS-2 feeding DAC and mR).  And, according to John's theory, that's the best way to do it for reasons I only vaguely understand.  That's ok, I come to CA to just do what I'm told.

 

Looks OK so far.

 

10 hours ago, sockpit said:

 

Now here's the twist and quandary:  My DAC is an Exogal Comet and the amp is an Exogal Ion.  The latter is powered, believe it or not, by a large 24V. 9.2A Meanwell SMPS (#GST220A24-R7B).  The worry in such a set up, I've learned from reading all these threads, is that the big ol' Meanwell is presumably pumping noise back into my Topaz, which could, in turn, show up again in my DAC and mR via the JS-2.  Is this correct?

 

If you ground/earth the 0V of the Meanwell's output stage. The plug they use is that 4 way PS2 looking thing, not so simple to get to the 0V from that.  The high impedance leakage currents from the Meanwell will be shunted out. The garbage on the Meanwell's input AC side, however will travel wherever it likes. 

A suggested alternative is to remove the Meanwell and replace it with one of these, or a Teddy Pardo, what suits your fancy. Well not suggested, but mandatory. The new power supply needs to have the same regulation properties as the Meanwell which according to the data sheet is +/- 2%, which is easy enough to achieve.

 

Cut the cable at the Meanwell end, fish out the wires and connect it to the linear supply. Might have four conductors, so it will take some logic to work out which is which. You will need to invest in a small digital multimeter of modest cost. Perhaps the cables are available as flying leads, not sure, Mouser is a good start to search, that way the original Meanwell PSU is intact.

 

10 hours ago, sockpit said:

 

There is an alternative to the above setup: there is a separate receptacle (not a dedicated one) very nearby.  If I plugged the Meanwell into that receptacle, presumably it would add its garbage to a totally separate circuit and not pollute my DAC.  (It would also be sure to get all the power it needs, if somehow the Topaz's 500VA wasn't up to the demands.)  But would there be a downside to separating these two components from the same receptacle?

 

To make things even more interesting, the Comet and Ion have a unique relationship to each other.  They are like two interacting computers more than like a preamp sending a signal to an amp for analogue amplification.  In fact, their "interconnect" is a single HDMI cable that permits them to communicate as one unit in processing the signal digitally--including a fully digital amplification process that is apparently quite unique.  Might this make them less liable to ground loop problems than typical equipment discussed in this thread? Is that irrelevant because of the LPS-1 anyway?

 

Sorry the link to Poland went belly up. This situation is quite complex and there's a compromise. With the isolation provided by the JS2/LPS that feeds clean to the Exogal Comet, the HDMI cable that wanders downstream will have cause an unwanted current to from from the ION to the COMET via the HDMI cable since the power supplies are sourced from different potentials.

The best we could do in this case, is to join the 0V from the JS2 to the 0V of the new linear for the Comet. Ideally the 0V from the output of the LPS should be joined to the 0V on the linear, but this may introduce a back door to the ION which won't be desirable.

 

If it's possible, connect the cases of the ION and the COMET together with a 12 guage Wire, use a case screw with a star washer to bite through the paint.

 

10 hours ago, sockpit said:

 

Here is the best description of this new and supposedly totally unique technology I've found for anyone interested in the technical details behind my question.  http://highfidelity.pl/@main-732&lang=en

 

Thanks in advance for anyone willing to give this and educated guess.

 

 

[sockpit]

Thanks @One and a half for taking time to comment.

 

Given my lack of electronics knowledge and the fact that Exogal is working on a better power supply for the Ion, I won’t be trying some of these daring moves   I’m just a chicken ‘cause they’d immediately void the warranties.  Whether that new supply will also run the Comet remains to be seen.  (Exogal already sells a Comet ps upgrade, which I don’t think is linear; they do a lot of things differently.) They have a streamer in development, too, so I imagine one very good supply to power them all would be the optimal solution in terms of common grounding,etc.

 

in the meantime, two questions:  1) what thinks you about plugging the big meanwell into a separate receptacle?  Did I understand those trade offs? (Sorry, I am an idiot about such things)

 

And 2) can you remind me what advantage the LPS1 actually  provides in my set up?  I ask only because I’ve been considering simplifying my chain and taking it out, on the assumption that the JS2 is plenty good to power the mR (and powering an LPS1 with a JS2 is madness.

 

Now if you ask me to trust my ears, I’d say the differences between plugging the Ion into a separate outlet or not, or taking out the LPS1 or not, are so subtle to my middle aged ears that I’m likely to be the victim of confirmation bias.  I think the LPS1 makes a small difference in lowering noise floor, but am sure I’d flunk a double blind test.

 

Here is the link again.  The concept is purported to be quite unique.  This is as forthcoming about it as I’ve seen Haagenstad be.  The link immediately above works on my iPad as well.

 

 http://highfidelity.pl/@main-732&lang=en

 

Thanks again for your time.

[wwc]

I've been looking in the Eaton Power Suppress 100 (the evolution of the Topaz we're talking about here).  The specs on interwinding capacitance aren't listed, so I thought I would make sure they were in the .005 - .0005 pf range spoken about here.

 

The Eaton engineers said it rates at .90 pf.  I'm going to follow up tomorrow

to confirm that.

 

As a non engineer, that sounds like a big difference to me.  Is it?

[Qhwoeprktiyns]

Questions to @JohnSwenson (but anyone else with some understanding of the issue please feel free to answer as well).

 

- I use hypex Ncore amplifiers, that contain switch mode power supplies. Hypex recommends tightly twisted cables internally to avoid picking up noise from the supplies. Would this noise not travel back to the rest of the system through the ground? Should the hypex amps be powered before or after the isolation transformer? 

 

- if I am using the recommended NetGear switch with grounding to filter out upstream noise conducted by lan cables, should this switch be connected before the isolation transformer (I have a daitron model). 

 

Thanks! 

 

[One and a half]

6 hours ago, wwc said:

I've been looking in the Eaton Power Suppress 100 (the evolution of the Topaz we're talking about here).  The specs on interwinding capacitance aren't listed, so I thought I would make sure they were in the .005 - .0005 pf range spoken about here.

 

The Eaton engineers said it rates at .90 pf.  I'm going to follow up tomorrow

to confirm that.

 

As a non engineer, that sounds like a big difference to me.  Is it?

This can be worked out by a couple of formulas, Ohms law and capacitive reactance.

 

The leakage current passed by the transformer is I = V/Z, I is in amps, V volts, Z = capacitive reactance

 

V = 230, 120 US, Canada

Z = (2 x pi x f x C)

pi =3.14159

f = 50Hz or 60Hz Canada US

C= capcitance in Farads. so 0.9pf = 0.9 x 10^-12 F

 

Even without going through the formulas, the leakage current from 0.9 to 0.0005 is a factor of 1800:1. Significant enough to not want to use the new Eaton transformers.

 

The above assumes the source supply is a grounded TN system (Neutral bonded to ground).

[flkin]

What’s the purpose of having low capacitance in an isolation transformer? How does this relate to losses or transient abilities?

 

Don’t understand the formular above... since the C is very small so anything divided by very small is very large. Surely the leakage current isn’t that large?

[JohnSwenson]

34 minutes ago, flkin said:

What’s the purpose of having low capacitance in an isolation transformer? How does this relate to losses or transient abilities?

 

Don’t understand the formular above... since the C is very small so anything divided by very small is very large. Surely the leakage current isn’t that large?

Transformers have two methods for transferring AC signals from one side to the other: magnetic and capacitive. 

 

Power transformers are designed so the primary line frequencies and a few harmonics are transferred magnetically. Higher frequencies are transferred capacitively. The lower the capacitance the less well it transfers high frequencies. With the Topaz the magnetic part is good up to about 400Hz, with the extremely low capacitance anything above that gets dramatically attenuated. The higher capacitance doesn't block high frequency noise as much as the lower capacitance.

 

One very nice effect of that is transient suppression. A voltage spike typically has a rise time much faster than 400Hz, the low capacitance prevents it from rising extremely quickly, which spreads the energy over much longer time, which has a much lower peak amplitude, thus "transient suppression".

 

Note: contrary to popular belief a Topaz or similar does NOT block leakage current, leakage current is made up of line frequency and harmonics. Since the Topaz passes up to 400Hz, the primary components of leakage current sail right on through. It DOES block the high frequency components from an SMPS, but those are relatively low in level compared with the line frequency and lower harmonics.

 

John S.

[flkin]

John, does the Topaz impede current flow in any way in comparison to direct from wall socket? Does the magnetic way of transferring AC signals reduce the ability to send transient current demands instantaneously? 

[wwc]

So, assuming I was given accurate info from Eaton, (that it's capacitance is .90 pf), we should avoid the Eaton (which, higher up in this long thread was recommended as having the same specs as the Topaz)  in favor of the Topaz, no?

 

Am I right that the three iso transformers with low capacitance specs of .005 to .0005 are the Topaz 30 series, the Elgar and the Xentek?

 

From what I've seen online they look like the identical units with different names?

[Solstice380]

58 minutes ago, wwc said:

So, assuming I was given accurate info from Eaton, (that it's capacitance is .90 pf), we should avoid the Eaton (which, higher up in this long thread was recommended as having the same specs as the Topaz)  in favor of the Topaz, no?

 

Am I right that the three iso transformers with low capacitance specs of .005 to .0005 are the Topaz 30 series, the Elgar and the Xentek?

 

From what I've seen online they look like the identical units with different names?

 

Are you sure they gave you the value for inter-winding cap?  That seems really high. 

 

The 30 series Topaz are .0005 and the 20 series is .005.  Not sure on the other brands.   

[wwc]

Does anyone here use the Eaton Power Suppress 100?  If so, did you confirm it's interwinding capacitance pf #?  

[wwc]

Another question on these iso transformers?

 

I'm not sure if I want to run my subs (combined 3000 Watts RMS)

through the transformer or into the wall outlets.  I only need a 1800 VA

transformer for everything else, but would need at least 5000VA if 

using the Subs.

 

First question, if I do not use the Subs, is there any downside to having

an oversized transformer (other than probably more expensive)?

 

For the very low frequencies Subs are working at, is the iso transformer of 

significant benefit?

 

Thanks

[Cornan]

3 minutes ago, wwc said:

Another question on these iso transformers?

 

I'm not sure if I want to run my subs (combined 3000 Watts RMS)

through the transformer or into the wall outlets.  I only need a 1800 VA

transformer for everything else, but would need at least 5000VA if 

using the Subs.

 

First question, if I do not use the Subs, is there any downside to having

an oversized transformer (other than probably more expensive)?

 

For the very low frequencies Subs are working at, is the iso transformer of 

significant benefit?

 

Thanks

A word of advice. Get a 500-750mA balanced IT with floating secondary. Special order but cheaper and a lot more power overhead than any other IT versions out there IME. Otherwise dont mind and upgrade much later! ?

[wwc]

 Cornan,

 

Thanks for the advice.  Could you be more specific as to what you're talking about?  

Torus, Plintus...?

 

A make, model example would help get me on the trail!

 

And are you advising against the low capacitance IT's ?  

 

Any elaboration you want to go into, please feel free!

 

Thanks