This is very true. I've been working on this issue for some time, seems class II power connects are easier to work with than class I in regards to switching noise. In class I, lots of noise can couple into the ground such exceeding regulatory requirements. Trying to filter those is a pain, especially when space is limited. However, the better they are decoupled from the mains, the better they work in audio devices.
When I first got into contact with SMPS in audio this was at times when Compact Cassette was at its top and people doing home recording trashed their expensive Revox and Tandberg engines to follow the hype of audio recording on video tapes (VHS).
Below cute „professional“ mobile recording machine (Sony TCD-5M) used SMPS to up-step the low voltage of two 1.5V cells and did an impressive job in terms of clean and dynamic sound for the given low voltage rails in use.
On top of that, there came along a sonic quality that was new to me and which could possibly best be described as „spot on in the lower department“.
This I specifically relate to the operation of switched power supplies.
Generally I didn't find HF pollution drawbacks of SMPS – when implemented correct – an issue, also SMPS is used in quite some pro gear with no hassle.
--------------
Some measurements regarding the draw backs of linear power supply for power amps - not many are aware of - I've shown along the thread discussing crossover distortion with Earl:
Here we see the low frequency hash of a well respected NAIM NAP 140
And in comparison here is the considerably cleaner „cheap“ KENWOOD car amp with SMPS
Interesting, no?
---------------
IMO not correlated to the above measurements there is a shift in audio perception related to the behaviour of source resistance of the PSU at the bottom frequencies.
This is most obvious when experimenting with the „Bryston curcuit", basically a capacitor in front of the transformer of a linear PSU to get rid of any DC in the mains supply which cause odd behaviour especially when using toroidal transformers.
Quite any resistance in front of PSUs immediately result in softening bass attack of power amp's.
This is most obvious with SS amps.
With tube amps some like to insert those “new” and fancy current limiters device that limit start up peak current and drop in resistance (due to heating up) in normal operation.
Same sonic behaviour here – in addition – a subtle pattern of “dynamic enhancement” due to slight resistance shifts with load demands (if not operating class A).
Michael
Below cute „professional“ mobile recording machine (Sony TCD-5M) used SMPS to up-step the low voltage of two 1.5V cells and did an impressive job in terms of clean and dynamic sound for the given low voltage rails in use.
On top of that, there came along a sonic quality that was new to me and which could possibly best be described as „spot on in the lower department“.
This I specifically relate to the operation of switched power supplies.
Generally I didn't find HF pollution drawbacks of SMPS – when implemented correct – an issue, also SMPS is used in quite some pro gear with no hassle.
--------------
Some measurements regarding the draw backs of linear power supply for power amps - not many are aware of - I've shown along the thread discussing crossover distortion with Earl:
Here we see the low frequency hash of a well respected NAIM NAP 140
An externally hosted image should be here but it was not working when we last tested it.
And in comparison here is the considerably cleaner „cheap“ KENWOOD car amp with SMPS
An externally hosted image should be here but it was not working when we last tested it.
Interesting, no?
---------------
IMO not correlated to the above measurements there is a shift in audio perception related to the behaviour of source resistance of the PSU at the bottom frequencies.
This is most obvious when experimenting with the „Bryston curcuit", basically a capacitor in front of the transformer of a linear PSU to get rid of any DC in the mains supply which cause odd behaviour especially when using toroidal transformers.
Quite any resistance in front of PSUs immediately result in softening bass attack of power amp's.
This is most obvious with SS amps.
With tube amps some like to insert those “new” and fancy current limiters device that limit start up peak current and drop in resistance (due to heating up) in normal operation.
Same sonic behaviour here – in addition – a subtle pattern of “dynamic enhancement” due to slight resistance shifts with load demands (if not operating class A).
Michael
Last edited:
I don't normally endorse audiophile gizmos, but the Van den Hul Polarity Checker looks like a neat way to identify leakage currents from the AC line getting into the chassis or system grounds. I don't know how useful it is to use your own body as the ground reference, but at least you could chase out any shocks when you touch the knobs!
What didn't get mentioned in the SWPS comments above is the leakage path from the AC line to the chassis or system grounds. This is not a trivial problem, and is the direct cause of "ground loops" and HF buzz in multi-component systems. Increase the isolation, and system noise goes down in proportion.
The most comprehensive solution is medical-grade high-isolation power transformers with the low-capacitance side of the primary connected to the HOT side of the AC line, combined with transformer-coupled balanced interconnects between each piece of gear. Each chassis will have the minimum possible leakage current, and they do not electrically connect to each other. No ground loops, with the noise floor set by the electrical circuits, not unwanted interactions between components. This is the quietest possible way to build the system, unless you want to go to fully-floating battery power. (Remember, even with battery power, the charger must be completely disconnected from AC when listening.)
I should add that leakage-noise problems do NOT always appear on the test bench, where the interconnections with test gear are completely different than a functioning hifi system. As often as not, test gear is "floated" from the AC line, or the DUT (device under test) is run off isolation transformers or Variacs to minimize shock hazard.
What didn't get mentioned in the SWPS comments above is the leakage path from the AC line to the chassis or system grounds. This is not a trivial problem, and is the direct cause of "ground loops" and HF buzz in multi-component systems. Increase the isolation, and system noise goes down in proportion.
The most comprehensive solution is medical-grade high-isolation power transformers with the low-capacitance side of the primary connected to the HOT side of the AC line, combined with transformer-coupled balanced interconnects between each piece of gear. Each chassis will have the minimum possible leakage current, and they do not electrically connect to each other. No ground loops, with the noise floor set by the electrical circuits, not unwanted interactions between components. This is the quietest possible way to build the system, unless you want to go to fully-floating battery power. (Remember, even with battery power, the charger must be completely disconnected from AC when listening.)
I should add that leakage-noise problems do NOT always appear on the test bench, where the interconnections with test gear are completely different than a functioning hifi system. As often as not, test gear is "floated" from the AC line, or the DUT (device under test) is run off isolation transformers or Variacs to minimize shock hazard.
Last edited:
To summarize briefly, thanks to leakage currents (between the AC line and chassis), a complete audio system will always have more hum, noise, and buzz than the measurements of the individual components on a test bench. More importantly, sub-threshold levels of buzz can mimic tonal colorations, affect the perception of space (since buzz is pinned to the speakers or a single center location), and the you-are-there impression of delicacy, air, and realism. (If a piano fades into a point-source location of buzz instead of an enveloping acoustic space, the perception of the piano is greatly altered, even if the buzz isn't directly audible as such.)
If your primary choice of music are recordings of powerfully amplified live rock concerts, this won't matter much, since the recordings already have plenty of hum and buzz in them - an unavoidable consequence of complex PA systems with high gain factors. Same for recordings of amplified jazz or blues groups, or anything where a PA system was involved.
Reduction of hum and buzz is most noticeable on recordings where you don't expect to hear any electronic artifacts in the first place - acoustic instruments in a physical performing space. When we listen to PA-amplified music, we accept the inevitable hum and buzz as part of the experience, so a minor addition from the system isn't as important. Of course, if the rock concert was recorded in 50 Hz Europe, and you're listening in 60 Hz North America, there should be an interesting 20 Hz growl (the difference between 100 and 120 Hz rectified AC) that wasn't there in the original recording, not to mention a large number of sum-and-difference sidebands at higher frequencies. I think you can see where the low-level spectrum can get cluttered pretty fast - all from those tiny little leakage currents that go unmeasured on the test bench.
Just a little stray capacitance inside various power transformers, combined with shared grounds between components, causes all the trouble mentioned above. That's the difference between the idealized model on the schematic and the real thing.
If your primary choice of music are recordings of powerfully amplified live rock concerts, this won't matter much, since the recordings already have plenty of hum and buzz in them - an unavoidable consequence of complex PA systems with high gain factors. Same for recordings of amplified jazz or blues groups, or anything where a PA system was involved.
Reduction of hum and buzz is most noticeable on recordings where you don't expect to hear any electronic artifacts in the first place - acoustic instruments in a physical performing space. When we listen to PA-amplified music, we accept the inevitable hum and buzz as part of the experience, so a minor addition from the system isn't as important. Of course, if the rock concert was recorded in 50 Hz Europe, and you're listening in 60 Hz North America, there should be an interesting 20 Hz growl (the difference between 100 and 120 Hz rectified AC) that wasn't there in the original recording, not to mention a large number of sum-and-difference sidebands at higher frequencies. I think you can see where the low-level spectrum can get cluttered pretty fast - all from those tiny little leakage currents that go unmeasured on the test bench.
Just a little stray capacitance inside various power transformers, combined with shared grounds between components, causes all the trouble mentioned above. That's the difference between the idealized model on the schematic and the real thing.
Last edited:
I'm not sure how one can conclude that transformer stray capacitance needs to be lower. Haven't seen any measurements that support that theory. I've used different transformers on the power supply of an NCD with significant vairation in sound quality. As mentioned before, using an R core vs C core transformers of same rating supplied by the same factory just sounded totally different. One thing that was noted is that the turn on surge current were significantly different using both transformers. One would always burn a fuse while the other would not. Perhaps someone with real understanding of the issues can explain. Maybe Bud?
Some of my tests seem to suggest that stray capacitance in SMPS can improve performance.
Some of my tests seem to suggest that stray capacitance in SMPS can improve performance.
Last edited:
Lynn"s statement:
Just a little stray capacitance inside various power transformers, combined with shared grounds between components, causes all the trouble mentioned above. That's the difference between the idealized model on the schematic and the real thing.
Is perfectly true! Just how big the problem may be under different circumstances is open to discussion.
Each of the following papers, devotes pages to the problem.
The Bill Whitlock of Jensen Transformers Seminar paper
http://www.jensen-transformers.com/an/generic seminar.pdf
The Jim Brown of Audio Systems Group white paper
"Power and Grounding for Audio and Audio/Video Systems"
http://www.audiosystemsgroup.com/SurgeXPowerGround.pdf
"The TRUTH" from ExactPower of Middle Atlantic Products
ExactPower(tm) - Home of the Residential Power Integrity System
or a different version of the same paper
"Power White Paper" from Middle Atlantic.com
Middle Atlantic Products - White Paper - Integrating Electronic Equipment and Power into Rack Enclosures
Just a little stray capacitance inside various power transformers, combined with shared grounds between components, causes all the trouble mentioned above. That's the difference between the idealized model on the schematic and the real thing.
Is perfectly true! Just how big the problem may be under different circumstances is open to discussion.
Each of the following papers, devotes pages to the problem.
The Bill Whitlock of Jensen Transformers Seminar paper
http://www.jensen-transformers.com/an/generic seminar.pdf
The Jim Brown of Audio Systems Group white paper
"Power and Grounding for Audio and Audio/Video Systems"
http://www.audiosystemsgroup.com/SurgeXPowerGround.pdf
"The TRUTH" from ExactPower of Middle Atlantic Products
ExactPower(tm) - Home of the Residential Power Integrity System
or a different version of the same paper
"Power White Paper" from Middle Atlantic.com
Middle Atlantic Products - White Paper - Integrating Electronic Equipment and Power into Rack Enclosures
Last edited:
Stray capacitance is of course not directly related to turn on inrush current, but they can conspire to blow fuses..
Many many years ago Ordean Kiltie and I had the same brain storm, just about the same time. He developed an E/I core stacking method that absolutely set the core to B/H zero after every saturation peak. This meant that no matter how far into saturation the core was driven, it never "hung" at b max, until H crossed zero and collapsed the polar magnetic field, thus drawing huge currents. When a transformer of any kind can stop it's activities with the core pinned to saturation, usually because high dielectric constant film plastics have been used in the high voltage section, when the transformer turns on next, the input voltage half wave can easily be of opposite polarity and a 6 to 10 times ordinary inrush current can occur.
Ordean was only interested in power and I was more interested in audio, but we did discuss the mechanism a bit and one of the results was what is called long finger E only core stack, where the flux is forced to hop from lamination to lamination to complete the path as there is no crossing I. Significantly lower distortion is the result, within telephone voice pass frequencies. Ordean actually worked on this aspect, which I found amusing since he didn't care about audio at all.
Application of a derivation of this stacking method is not really a good idea in audio reproduction power transformers due to the much greater stray flux field and greater physical vibration caused. However, for Guitar Amps it is just perfect, allowing a properly managed power, choke and low storage factor output transformer to provide the tubes with enough freedom to allow three times their rated power for 100 to 150 ms. to be provided from the push pull tubes.
Different sounds in amps due to power transformers is usually due to poor power supply rejection, with SE being far more susceptible than PP. When the power transformer is as exposed as it normally is in SE amps it's capacitive characteristics in the coil are every bit as important as those in the output transformer, and dielectric materials and their usage in coils is the mother of tone, in all amplifiers, guitar or reproduction.
Transformer stray field coupling is an interesting problem. You have the capacitive coupled E Field charge, that devolves into current flow and you have the inductive stray field, as a contribution from the coil, before the wave form peak and a contribution from the core after the wave form peak. All three will couple into chassis, inappropriately aligned circuit traces, metal film and foil capacitors, any nickle core anything, non heat treated steel chassis bends and poorly constructed grounds. All of these inputs and targets have different characteristics and relationships and not all of them can be corrected for, without a deep understanding of their sources and merit. And if you think transformer design is a black art, the reduction of noise from the things is high magic indeed.
I suggest Ralph Morrison's "Grounding and Shielding Techniques" as your first introduction. Now there is a scarily deeply knowledgeable, high magician
Bud
Many many years ago Ordean Kiltie and I had the same brain storm, just about the same time. He developed an E/I core stacking method that absolutely set the core to B/H zero after every saturation peak. This meant that no matter how far into saturation the core was driven, it never "hung" at b max, until H crossed zero and collapsed the polar magnetic field, thus drawing huge currents. When a transformer of any kind can stop it's activities with the core pinned to saturation, usually because high dielectric constant film plastics have been used in the high voltage section, when the transformer turns on next, the input voltage half wave can easily be of opposite polarity and a 6 to 10 times ordinary inrush current can occur.
Ordean was only interested in power and I was more interested in audio, but we did discuss the mechanism a bit and one of the results was what is called long finger E only core stack, where the flux is forced to hop from lamination to lamination to complete the path as there is no crossing I. Significantly lower distortion is the result, within telephone voice pass frequencies. Ordean actually worked on this aspect, which I found amusing since he didn't care about audio at all.
Application of a derivation of this stacking method is not really a good idea in audio reproduction power transformers due to the much greater stray flux field and greater physical vibration caused. However, for Guitar Amps it is just perfect, allowing a properly managed power, choke and low storage factor output transformer to provide the tubes with enough freedom to allow three times their rated power for 100 to 150 ms. to be provided from the push pull tubes.
Different sounds in amps due to power transformers is usually due to poor power supply rejection, with SE being far more susceptible than PP. When the power transformer is as exposed as it normally is in SE amps it's capacitive characteristics in the coil are every bit as important as those in the output transformer, and dielectric materials and their usage in coils is the mother of tone, in all amplifiers, guitar or reproduction.
Transformer stray field coupling is an interesting problem. You have the capacitive coupled E Field charge, that devolves into current flow and you have the inductive stray field, as a contribution from the coil, before the wave form peak and a contribution from the core after the wave form peak. All three will couple into chassis, inappropriately aligned circuit traces, metal film and foil capacitors, any nickle core anything, non heat treated steel chassis bends and poorly constructed grounds. All of these inputs and targets have different characteristics and relationships and not all of them can be corrected for, without a deep understanding of their sources and merit. And if you think transformer design is a black art, the reduction of noise from the things is high magic indeed.
I suggest Ralph Morrison's "Grounding and Shielding Techniques" as your first introduction. Now there is a scarily deeply knowledgeable, high magician
Bud
Last edited:
I am totally convinced that inproper ground layout, and any component layout as a matter of fact, will effect equipment performance; however, I do not see any correlation with stray capacitance. This is based on my experience and measurements. I am curious of any specific case of stray capacitance and measureable data correlated with listening experience. As a matter of fact, I have experienced improvement in audio reproduction due to adding what could be called stray capacitance. Personally, I think stray capacitance issue is blindly blamed because it can improve performance if it occurred at the right location.
Bud, thanks for filling us in on all this knowledge. I quite understand the importance of not having stray fields, but the issue between the R-core and O-core (I think I got it wrongly written as C core in a previous post) still baffles me. As a matter of fact, the O-core required a bigger fuse, and sounded much better than the R-core with the same setup. Since currently I have no continuing project using these, the issue had been put aside for a while. But I think a C-core was developed to improve distortion issues, was it not?
Last edited:
I've never understood the logic in this.
Having a 6 foot long "special" power cord plugged into a receptacle with ~ 60 feet of 14-2 NMD7 - solid copper, parallel conductor, cabling back to the main distribution panel in a home.
Sure it can reduce radiated RF/EMI in the last 6 feet before the amp, but what about the ~ 60 feet of ordinary cable in the wall?
I'm willing to learn and listen to reason.
Can someone please explain in technical terms why the last 6 feet is so much more important than the distribution cabling?
The high inrush problem (= blowing fuses) is not exactly correlated to stray capacitance.
Also - as said before - different "stiffness" of the mains and any resistance variation upstreams the charging capacitors will result in audible differences for any but the most constant load amplifiers.
Hence HF-filters, DC blockers, symmetry transformers etc in the mains are mostly a no-no for some amp developers.
Some "addicted" also swear different mains fuses also to result in audible difference - guess why...
So you possibly have to be more specific, I suggest.
Go and try to add low value resistors in the mains and report back your findings...
But thats of course completely a different point to what Lynn was pointing at.
-----------
What Bud outlined regarding increased inrush current at turn on due to some "magnetic hang on" when switched off last time can be addressed by specialized circuits used in power gear, if one likes to experiment with:
Function of TSRL and TSRLF.: EMEKO, Michael Konstanzer
Michael
Last edited:
Fair enough. Conventional bridge-cap power supplies, nearly universal in transistor gear, and pretty common in tube electronics, only conduct current during the brief interval when the rectified AC waveform is 0.7V higher than the slowly falling voltage of the main power supply capacitors. The "ON" interval is typically only a millisecond or less, rather than the 8.33 mS you might expect from one period of a 120 Hz fullwave rectified AC waveform. In addition, the pulse width is modulated by the current draw on the PS capacitors, which is another way of saying program content in a Class AB amplifier changes the spectra of the PS switch noise.
Typical EI-core power transformers are flat to 5 kHz or higher - this is not a common power transformer spec, so transformers are all over the place in terms of bandwidth. Toroids are typically flat to much higher frequencies, 30 kHz or more. So there is nothing to impede the transmission of very narrow duration current pulses all the way out to the step-down transformer of the medium-tension distribution system.
Why does the six feet of the power cord matter more than all the house wiring? In the general sense, it doesn't. But that six feet is closer to the rest of the analog equipment than the rest of the (unshielded) house wiring. I suspect much of the effect of expensive audiophile power cords and power conditioners comes down to how they act as antennas for the radiated switch-noise. Shield the first six feet of the antenna, and it behaves a bit differently than a completely unshielded antenna, particularly in the nearfield.
Power conditioners that are small AC regenerators are in a special class, since they rectify incoming AC to DC, filter the DC, create low-distortion sinewave AC (typically less than 1% distortion), and amplify the clean AC. Aside from the limitations of power, this is a good way to lower the noise floor of a system - at a rather high cost.
Fancy power cords and "power conditioners" are another matter. Yes, they filter the incoming AC to a degree (although far less than a regenerator), but the filter circuits also modify the pulse shape of those sub-millisecond current pulses, which in turn changes the noise spectra emitted in free air and into adjacent equipment. Since the current pulses of transistor amplifiers with very large cap banks are much larger and narrower than the current pulses of a tube amplifier with tube rectification (which has a smoother on-off transition than a solid-state diode), the "power conditioner" may sound quite different on different types of amplifiers.
There are power supplies which have a steady current draw on the AC line, but they are the fairly expensive "choke-fed" type, and almost never seen in the transistor world.
I agree with the general thrust of the original question - selection of power cords and power conditioners are not a first-order problem. In practice, playing around with these acts like a mild tone control (since it changes the noise spectra, and that is perceived as a spectral change) with subtle changes in dynamics (since the pulse characteristics of the power supply are being affected).
The more serious issue is reduction of ugly-sounding hash, which is mainly the result of stray capacitance between power transformer primaries and the chassis-grounded core of the transformer. As long as equipment is interconnected by wires with hard grounds between different chassis, though, it is desirable to reduce the stray capacitance of all power transformers in a system, since noise from one chassis is readily conducted to another, as well as DC drops in the interconnect transforming current flows into small voltage differences which are then amplified by the input stage of the following electronics.
I am also not a fan of bridge-cap supplies, since the narrow current pulses are only made worse by increasing the PS capacitance, or decreasing the DCR of the power transformer. The "choke-fed" type of supply, although requiring a plenty expensive (and heavy) choke, has a much smoother current delivery, since the rectifier conducts for the full duration of the positive or negative waveform. Reducing EMI at the source is always much more effective than filtering after the fact.
Last edited:
In the beginning, that is what I thought as well. In the process of developing filters for a power supply, that is also what I concluded as well. However, it did not make sense to me; but after I got the interconnects where I want them, adding the filters actually improved the sound cleaness and detailedness which provided a better sound location focus. So it made sense to see proper filtering improved the sound in a way that defections in other areas emerg.
Note that the full solution for noise reduction is quite expensive: choke-fed supplies, medical-grade high-isolation power transformers, and transformer-isolated balanced interconnects between all components. Jensen will gladly sell you studio-grade isolation transformers, but redesigning power supplies in commercial audiophile equipment is no small task.
Playing around with interconnects, power cords, or power conditioners does not address the fundamental design flaws of most audio equipment. In practice, it is "good enough", but that is a long way from designing for the lowest possible noise level in a complete system.
Bill Whitlock's paper is an excellent starting point for addressing system-wide noise. For any of you using a cable box that is connected to a home-theater setup, you need an isolator for the cable coming into the cable box - cable grounds are notoriously dirty.
Playing around with interconnects, power cords, or power conditioners does not address the fundamental design flaws of most audio equipment. In practice, it is "good enough", but that is a long way from designing for the lowest possible noise level in a complete system.
Bill Whitlock's paper is an excellent starting point for addressing system-wide noise. For any of you using a cable box that is connected to a home-theater setup, you need an isolator for the cable coming into the cable box - cable grounds are notoriously dirty.
Last edited:
So what is exactly studio-grade vs medical grade transformers? What are the specifications that separate these? Let's not forget that most companies will use certain names to try and differentiate themselves, but it's the actual performance that count's, which should be explainable in numbers. Most of the requirements of medical equipment are there for safety reasons. The price differentiation is mostly due to the efforts to pass those requirements to get certified, and the specialized limited quantities sold, but by no means will be directly identifiable with audible improvements.
Last edited:
There are two "isolation" types of transformers. One is for isolation of medical equipment that is patient connected and has specific activities performed on it, like being potted in a plastic container, to limit the surface currents and voltages to below 5 mv. These are not necessarily shielded from capacitive coupled noise in the power line. This is Hospital grade shielding.
There is a variant of this that has a copper shield wound between primary and secondary for a 70db reduction in capacitive coupled noise from the power grid. These are also potted and have other activities performed for the 5 mv limits on surface voltage. These are Hospital grade data safe units.
Then there are noise isolation transformers that can allow as much as 250 mv of surface voltage to propagate. They will have a copper shield, for 70 db of noise rejection above 250 Hz, or they will be wound in adjacent wells for 80 db of noise rejection. Another level of this is adjacent windings, with the secondary winding section painted with copper conductive paint, over a stranded wire that has been glued up the bobbin wall. An insulated turn copper shield with drain wire is placed before and after the winding and the painted copper drain wire is connected to the inner shield. These also have a copper strap around both core and coil, oriented in the same direction as the coil windings and placed over a high permeability ferrous metal shield wrapped tightly around the core perimeter and extending beyond the edges of the core stack by 0.050 inch on either side. This is a studio grade isolation transformer and if it is also potted in a plastic container, it can be used in situations where coupling to any chassis material is prohibited, since the surface voltages will be below 5 mv. The flux density of these transformers must be kept below 8 k gauss for all of the above to work properly. The result is better than 100 db of isolation for air borne, circuit borne or surface borne noise.
These are all E/I transformers. I use a 1.7 kva Hospital grade Data safe isolation toroid, with a common mode rejection toroid in front of it to keep DC from the isolation toroid, to isolate my system from the power grid. Mostly because it was free, provided by dumpster diving at Fluke, and would conveniently fit on a shelf, in a box with required outlets, and was built by someone else. An equivalent E/I device, with all bells and whistles, would be immense, hot and physically noisy, unless potted in a material specifically designed for noise abatement, typically with a Reynolds number lower than 30 and ideally with a gradient to it's hardness.
I still have noise in my drivers, but it is 80 db down, with 95 db efficient speakers and so I can pretend it isn't there and that I don't care about it etc. Part of this noise comes from the power transformer in the Nikko SS preamp I am using and I will eventually build another power transformer to get rid of it. Power line junk is no longer a problem and a galvanic isolation transformer between the pre and power amp does isolate the pre amp power transformer noise from the speakers, leaving only the power amp power transformer needing to be changed out. I don't particularly care for the sound of the isolation transformer, or any others I have either borrowed or built, so I just don't stress about the noise I have.
Bud
There is a variant of this that has a copper shield wound between primary and secondary for a 70db reduction in capacitive coupled noise from the power grid. These are also potted and have other activities performed for the 5 mv limits on surface voltage. These are Hospital grade data safe units.
Then there are noise isolation transformers that can allow as much as 250 mv of surface voltage to propagate. They will have a copper shield, for 70 db of noise rejection above 250 Hz, or they will be wound in adjacent wells for 80 db of noise rejection. Another level of this is adjacent windings, with the secondary winding section painted with copper conductive paint, over a stranded wire that has been glued up the bobbin wall. An insulated turn copper shield with drain wire is placed before and after the winding and the painted copper drain wire is connected to the inner shield. These also have a copper strap around both core and coil, oriented in the same direction as the coil windings and placed over a high permeability ferrous metal shield wrapped tightly around the core perimeter and extending beyond the edges of the core stack by 0.050 inch on either side. This is a studio grade isolation transformer and if it is also potted in a plastic container, it can be used in situations where coupling to any chassis material is prohibited, since the surface voltages will be below 5 mv. The flux density of these transformers must be kept below 8 k gauss for all of the above to work properly. The result is better than 100 db of isolation for air borne, circuit borne or surface borne noise.
These are all E/I transformers. I use a 1.7 kva Hospital grade Data safe isolation toroid, with a common mode rejection toroid in front of it to keep DC from the isolation toroid, to isolate my system from the power grid. Mostly because it was free, provided by dumpster diving at Fluke, and would conveniently fit on a shelf, in a box with required outlets, and was built by someone else. An equivalent E/I device, with all bells and whistles, would be immense, hot and physically noisy, unless potted in a material specifically designed for noise abatement, typically with a Reynolds number lower than 30 and ideally with a gradient to it's hardness.
I still have noise in my drivers, but it is 80 db down, with 95 db efficient speakers and so I can pretend it isn't there and that I don't care about it etc. Part of this noise comes from the power transformer in the Nikko SS preamp I am using and I will eventually build another power transformer to get rid of it. Power line junk is no longer a problem and a galvanic isolation transformer between the pre and power amp does isolate the pre amp power transformer noise from the speakers, leaving only the power amp power transformer needing to be changed out. I don't particularly care for the sound of the isolation transformer, or any others I have either borrowed or built, so I just don't stress about the noise I have.
Bud
Last edited: