My experience with absolute polarity is that 90%+ of people that I demonstrate this to on their own system can point out exactly what they hear to be different. I could not hear the difference until I treated drivers such that the CSD reach a certain level of fast decay. Of course all this also depends on how the recording and mixing process was conducted. There are some recordings during which I can hear a difference, but cannot clearly tell which might be the preferred polarity; there are also some where some instruments sound right and some instruments sound inverted; in the majority of CDs that I listen to, absolute polarity preference is very clear.
One thing to look for is bass punch. If the bass punch is sort of muffled, then the polarity is not correct, and generally swapping the polariy will provide good bass performance; however, if the enclosure is too small or overstuffed, then the polarity change may not improve things.
One thing to look for is bass punch. If the bass punch is sort of muffled, then the polarity is not correct, and generally swapping the polariy will provide good bass performance; however, if the enclosure is too small or overstuffed, then the polarity change may not improve things.
In order to be an apples-to-apples comparison of stranded versus non-stranded, I would think that you would need to use comparable wire gauge sizes in your comparison. In other words if you're comparing a four strand wire to a single strand wire, then the single strand wire should be equal in diameter and current passing ability to the four strand wire. Ergo if you're using a 64 strand wire, then the single strand wire that you're comparing it to should be of comparable size and current passing ability. For arguments sake let's say that the 64 strand wire is 22 ga, then the single strand wire should also be 22 ga. in order for it to be a valid comparison.
I would suspect that the improved coherency and focus that was experienced with the stranded wire has more to do with current passing ability than with the strand count.
I've experimented with solid core interconnects, speaker wire, and headphone cabling, and to my ears coherency and focus are noticeably improved with the solid core, and I no longer use any kind of stranded wire in my system.
Solid
I choose solid core as well. Belden 1701a for signal level and digital. 12ga copper foil or Nordost Flatline (individually insulated rectangular core) for speaker wires. Litz is also very good but is a hassle to terminate. Either is better than multi strands that touch each other.
I would think it appropriate to at least match the linear components, resistance, capacitance and inductance. And I would also suggest similar termination, i.e. gold plated connectors, etc. Also check the base metal of the connectors, brass? Steel? More of the wires/interconnects I have tested and for which I heard a difference all had some measurable difference in one or more of those parameters. That is not to say that different cables didn't sound different, but there is often a very straight forward reason.
I just wanted to add, in order to make everything as equal as possible in a stranded versus non-stranded comparison, both wires should have 50 or so hours of run-in time on them. There are those of us that very clearly hear differences between a burned-in versus a non-burned-in wire, and if the stranded wire in the comparison was burned in and the non-stranded wires were not, then that could explain the lack of coherency/focus that was heard in the non-stranded wire.
Well, I kind of knew that any discussion of wire would really drag things off-track - but in this thread, I'm hardly the one to complain about it.
The comparison Pimm and I did was at line-level, with transformer coupling on both ends, XLR's on both ends, and power-amp input impedances in the 50~100K ohm range. That works out to no more than 20 microamps of audio-frequency current flow - and we were listening at levels well below clipping, so more likely 10 microamps at most.
With all the zillions of turns in the transformers, frankly, I didn't expect the wires to sound different at all, and the XLR floating/balanced connection (no galvanic connection between the various amplifiers, thus no ground loops) was absolutely silent, with no side issues of RFI pickup to confuse things.
With RCA connections to solid-state or RC-coupled tub gear, results could be completely different, since ground loops and RF pickup then become real issues, and not trivial problems to solve. Noise and RFI rectification at the -80 to -110 dB level can easily masquerade as tone coloration and differences in transparency. I'm serious about this; merely reducing noise and RFI buzz to nearly inaudible levels, as is common in many high-end systems, still leaves lots of room for audible coloration and lack of transparency. That's the main reason that Gary Pimm and I like transformer coupling - noise and RFI reductions in the 60 to 90 dB range are easily attained, taking the problems far below the noise thresholds of the electronics.
Since these are dry circuits, I feel that surface-layer corrosion can be more serious than generally realized, particularly in the -80 dB range. These are currents in the 2 nanoamp range for the 50~100K impedances mentioned above - extraordinarily low currents by any standard. For a moving-magnet phono cartridge looking into 47K load, of course, it's even lower - down in the picoamp range.
The comparison Pimm and I did was at line-level, with transformer coupling on both ends, XLR's on both ends, and power-amp input impedances in the 50~100K ohm range. That works out to no more than 20 microamps of audio-frequency current flow - and we were listening at levels well below clipping, so more likely 10 microamps at most.
With all the zillions of turns in the transformers, frankly, I didn't expect the wires to sound different at all, and the XLR floating/balanced connection (no galvanic connection between the various amplifiers, thus no ground loops) was absolutely silent, with no side issues of RFI pickup to confuse things.
With RCA connections to solid-state or RC-coupled tub gear, results could be completely different, since ground loops and RF pickup then become real issues, and not trivial problems to solve. Noise and RFI rectification at the -80 to -110 dB level can easily masquerade as tone coloration and differences in transparency. I'm serious about this; merely reducing noise and RFI buzz to nearly inaudible levels, as is common in many high-end systems, still leaves lots of room for audible coloration and lack of transparency. That's the main reason that Gary Pimm and I like transformer coupling - noise and RFI reductions in the 60 to 90 dB range are easily attained, taking the problems far below the noise thresholds of the electronics.
Since these are dry circuits, I feel that surface-layer corrosion can be more serious than generally realized, particularly in the -80 dB range. These are currents in the 2 nanoamp range for the 50~100K impedances mentioned above - extraordinarily low currents by any standard. For a moving-magnet phono cartridge looking into 47K load, of course, it's even lower - down in the picoamp range.
Last edited:
Hmm - yes and no. At those current levels, bimetallic thermal generator effects, along with triboelectric mechanical noise induction, start to be significant. Most gold-plated connectors are not directly plated on copper, since gold is soluble in copper over the long term, so an intermediate "strike" layer is needed to physically isolate the gold from the copper. The "strike" layer is typically nickel - which, of course, is magnetic. The other problem with gold is that it can be easily wiped off in the small, high-contact-pressure regions where the connectors slide against each other, so heavy plating is desirable.
It is my understanding that copper oxide is a rectifier (a PN junction), while silver oxide is resistive (but not a rectifier). With current densities as low as they are in line-level connections, and signal integrity important down to -100 dB levels, I suspect corrosion at the single atomic layer of the surface, and between crystal grains deep within the bulk copper, is more important than generally acknowledged.
One little tidbit that John Atwood passed on to me was that mobile phone transmitters require extremely low levels of distortion - in the -120 to -140 dB range in order to maximize data transmission rates. (Distortion acts to "fill in" empty portions of the spectrum, and can thus degrade bit rates, which are an expensive commodity in the mobile-phone world.) I was informed that mobile phone transmitters use quite exotic technology to minimize nonlinear distortion in the physical cable between the transmitter and the antenna. This is something I'd like to look into further.
It bothers me that any of this is audible at all; the measurable effects of these artifacts well past -100 dB, and should be inaudible. But what I heard at Gary Pimm's place was easily audible on a repeated basis - but again, this was a system with no galvanic connection between devices, and fully balanced connections with transformer coupling providing more than 60 dB of noise rejection (compared to direct-connected RCA unbalanced connections). Thus, a very quiet system, with the only hum and noise coming from internal power supplies.
All bets would be off with conventional direct-coupled unbalanced connections, with all chassis electrically hard-connected to each other. Stray capacitance in the power transformers leaves the various chassis at different AC potentials (this can be measured with equipment turned on but electrically disconnected from each other), and small voltage drops in the ground-return side of the unbalanced connectors induces noise into the audio signal. One hint is that buzz, as opposed to pure sinewave hum, is almost always stray capacitance where it doesn't belong, or occasionally rectified RFI pickup.
The reason that reversing AC connections on the power plug makes a difference (and it does) is that power transformers do not have symmetrically wound primaries. One side of the primary has a much lower core-to-winding capacitance than the other (a 1:5 difference is common). The low-capacitance side is the one that belongs on the "hot" side of the AC plug for best isolation from AC line noise (and is part of the transformer specification). Unfortunately, the low-capacitance side is not usually marked, so when an audio component is manufactured, it's purely random whether or not the power transformer has the optimum connection or not. If you have a capacitance meter, the capacitance of each side of the incoming AC line can be measured against the component case, and the low-capacitance side of the primary determined. Too bad any modification of the component (like reversing and resoldering the incoming AC connections on the power-transformer primary) voids the warranty.
It is my understanding that copper oxide is a rectifier (a PN junction), while silver oxide is resistive (but not a rectifier). With current densities as low as they are in line-level connections, and signal integrity important down to -100 dB levels, I suspect corrosion at the single atomic layer of the surface, and between crystal grains deep within the bulk copper, is more important than generally acknowledged.
One little tidbit that John Atwood passed on to me was that mobile phone transmitters require extremely low levels of distortion - in the -120 to -140 dB range in order to maximize data transmission rates. (Distortion acts to "fill in" empty portions of the spectrum, and can thus degrade bit rates, which are an expensive commodity in the mobile-phone world.) I was informed that mobile phone transmitters use quite exotic technology to minimize nonlinear distortion in the physical cable between the transmitter and the antenna. This is something I'd like to look into further.
It bothers me that any of this is audible at all; the measurable effects of these artifacts well past -100 dB, and should be inaudible. But what I heard at Gary Pimm's place was easily audible on a repeated basis - but again, this was a system with no galvanic connection between devices, and fully balanced connections with transformer coupling providing more than 60 dB of noise rejection (compared to direct-connected RCA unbalanced connections). Thus, a very quiet system, with the only hum and noise coming from internal power supplies.
All bets would be off with conventional direct-coupled unbalanced connections, with all chassis electrically hard-connected to each other. Stray capacitance in the power transformers leaves the various chassis at different AC potentials (this can be measured with equipment turned on but electrically disconnected from each other), and small voltage drops in the ground-return side of the unbalanced connectors induces noise into the audio signal. One hint is that buzz, as opposed to pure sinewave hum, is almost always stray capacitance where it doesn't belong, or occasionally rectified RFI pickup.
The reason that reversing AC connections on the power plug makes a difference (and it does) is that power transformers do not have symmetrically wound primaries. One side of the primary has a much lower core-to-winding capacitance than the other (a 1:5 difference is common). The low-capacitance side is the one that belongs on the "hot" side of the AC plug for best isolation from AC line noise (and is part of the transformer specification). Unfortunately, the low-capacitance side is not usually marked, so when an audio component is manufactured, it's purely random whether or not the power transformer has the optimum connection or not. If you have a capacitance meter, the capacitance of each side of the incoming AC line can be measured against the component case, and the low-capacitance side of the primary determined. Too bad any modification of the component (like reversing and resoldering the incoming AC connections on the power-transformer primary) voids the warranty.
Last edited:
Trannies?
I am a transformer newbie but am starting to appreciate them more and more for coupling the analog directly out of my dac chips. I did some listening today between my Jensen JT-11-FMCF and EMCF as configured for balanced to single ended direct out from the CS43122 and AK4395 dac chips. The extra headroom of the larger EM is readily apparent in the crack and slam of the cymbals during the intro to No God from the Extreme, Waiting for the Punchline disc. Darn it. The $66 trannie sounds better than the $50. I was hoping that the maximum levels of the chips wouldn't challenge the capacities of the FMCF or something even smaller and cheaper. Now I am intrigued to try something even better. What are your favorite output transformers? Input transformers?
I am a transformer newbie but am starting to appreciate them more and more for coupling the analog directly out of my dac chips. I did some listening today between my Jensen JT-11-FMCF and EMCF as configured for balanced to single ended direct out from the CS43122 and AK4395 dac chips. The extra headroom of the larger EM is readily apparent in the crack and slam of the cymbals during the intro to No God from the Extreme, Waiting for the Punchline disc. Darn it. The $66 trannie sounds better than the $50. I was hoping that the maximum levels of the chips wouldn't challenge the capacities of the FMCF or something even smaller and cheaper. Now I am intrigued to try something even better. What are your favorite output transformers? Input transformers?
I should add that for all equipment that has a hard ground connection between different components, all power transformers should have the low-capacitance side of primary connected to the "hot" side of the incoming AC line.
If your equipment has an HDMI connection between the AVR, cable/sat DVR, and TV, yup, that means all of the gear needs to be optimized for the lowest-capacitance side of the incoming AC line. Same for USB or coaxial S/PDIF connections between audio gear and music-server computers. The more gear that is physically interconnected with hard grounds, the more transformers have to optimized for the low-capacitance connection. One single "wrong" transformer will contaminate all the rest of the system, even if the offending component is turned off.
This is why transformers come in so handy for breaking ground connections. Each component can find it's own ground, and ground currents do not flow between components. Otherwise, when you compare cables, you might be comparing little more than the DC resistance of the ground side of the cable, which is where the translation between ground currents and induction of audio noise occurs.
If your equipment has an HDMI connection between the AVR, cable/sat DVR, and TV, yup, that means all of the gear needs to be optimized for the lowest-capacitance side of the incoming AC line. Same for USB or coaxial S/PDIF connections between audio gear and music-server computers. The more gear that is physically interconnected with hard grounds, the more transformers have to optimized for the low-capacitance connection. One single "wrong" transformer will contaminate all the rest of the system, even if the offending component is turned off.
This is why transformers come in so handy for breaking ground connections. Each component can find it's own ground, and ground currents do not flow between components. Otherwise, when you compare cables, you might be comparing little more than the DC resistance of the ground side of the cable, which is where the translation between ground currents and induction of audio noise occurs.
Some measurements of input impedance with various off-the-shelf cables connected.
Some measurements of input impedance with one test cable connected under different input impedance condition
Both graphs are not of same scale, but are intended to show difference in trend. Hope someone else is willing to show their data.
An externally hosted image should be here but it was not working when we last tested it.
Some measurements of input impedance with one test cable connected under different input impedance condition
An externally hosted image should be here but it was not working when we last tested it.
Both graphs are not of same scale, but are intended to show difference in trend. Hope someone else is willing to show their data.
On a Burmester based system, in previous RCA vs Balanced comparison using remote controlled switching, I could not hear a difference, the other person says he heard a difference, but was so small that he felt it was of no concern.
Can you give an example what kind of system you did comparisons on? While I also always though balanced is better, all experience I come across seems to tell me that proper implementation and application in the simplest manner is far more important.
First a note on Lynn's exactly correct portrayal of power transformers and chassis circulating currents. According to a 70 year old EIA agreement, all transformers, power and signal, are to have their winding starts, both primary and secondary, marked by a round black spot adjacent to the lead connection point, on a schematic. This is the lowest capacitance point of an unbalanced winding. This also means that the primary and secondary lowest capacitance points are out of phase, by 180 degrees.
sendler
This is sort of the wrong question to ask of Lynn. You are standing at the tip point of "aquired musical taste" in transformer usage and Lynn stood there, probably, 40 years ago and has learned many lessons that you may not have yet encountered.
So, let me help by providing a thumbnail outline.
When I say standing at the tipping point, I mean that you are using transformers from a company celebrated for it's low distortion, exact specification limits, minimum size and complexity to meet a specific rms signal goal. This means that they are designing for a completely linear signal, one that has no impulses that are not contained within the rms specification (typically 0 db or 1 volt AC rms). This does not mean that all of their devices are limited to linear signals, but you have to know enough about transformer specification to read between the lines of the test data sheet. Jensen makes great transformers, but you can grow in your learning curve quite a bit beyond their self imposed, very technically correct limitations. You have found the difference between one product that does not meet your taste and one that does, this might not be true of someone with other tastes or more technically limited needs. The cost increase you noted is really rather small, considering what it could grow to.
Other companies relax the distortion clamps a bit, and allow for more information transformation. Others add quite extended non linear capability to their offerings, since much of music is comprised of non linear aspects, and design to allow those impulses to cause a "soft" saturation, one that doesn't splatter signal artifacts around the frequency bandwidth when it occurs. Neither of these two paths will have as low a measured distortion as does a Jensen design, at the carefully described test points of the Jensen transformers in question. However, they may offer a more musical non linearity and in some cases quite a bit more information. Perhaps in a slightly larger size and probably at greater cost.
Then their are the audio specialists, designers who take a particular characteristic and spend quite a bit of manufacturing dollars and materials dollars to extend their favorite category, one appreciated by purchaser's who's taste in reproduced music, has matured to the point that they want to pursue one of those specialist's paths. At this point, performance is strictly a matter of taste. All of the choices, including those of Jensen, will fall within the boundary of 99.993% perfect signal reproduction. The differences will be in what is deliberately given up in certain categories, to attain maximum performance in other categories.
I really am not trying to be smug and audio lofty here. These really are the choice vectors when you begin to look beyond simple measurement, in the putative "black art" of transformer design.
What Lynn likes, and I do know some of it, might not suit your taste at all. Best to ask a lot of people what transformer manufacturers they prefer and why. What characteristics are they expecting to receive and have they received them in the past from more than one manufacturer. Then begin to develop a nose for specifics, in the comments of the people you ask. A major category is going to be price. Another will be measured specifications. Another will be grace under the demands of very non linear music and a fourth might be information content and musical values, as espoused by musicians, not electrical engineers or transformer designers. Also a good idea to get some idea of the musical genre's that someone you have asked, considers acceptable.
Bud
sendler
This is sort of the wrong question to ask of Lynn. You are standing at the tip point of "aquired musical taste" in transformer usage and Lynn stood there, probably, 40 years ago and has learned many lessons that you may not have yet encountered.
So, let me help by providing a thumbnail outline.
When I say standing at the tipping point, I mean that you are using transformers from a company celebrated for it's low distortion, exact specification limits, minimum size and complexity to meet a specific rms signal goal. This means that they are designing for a completely linear signal, one that has no impulses that are not contained within the rms specification (typically 0 db or 1 volt AC rms). This does not mean that all of their devices are limited to linear signals, but you have to know enough about transformer specification to read between the lines of the test data sheet. Jensen makes great transformers, but you can grow in your learning curve quite a bit beyond their self imposed, very technically correct limitations. You have found the difference between one product that does not meet your taste and one that does, this might not be true of someone with other tastes or more technically limited needs. The cost increase you noted is really rather small, considering what it could grow to.
Other companies relax the distortion clamps a bit, and allow for more information transformation. Others add quite extended non linear capability to their offerings, since much of music is comprised of non linear aspects, and design to allow those impulses to cause a "soft" saturation, one that doesn't splatter signal artifacts around the frequency bandwidth when it occurs. Neither of these two paths will have as low a measured distortion as does a Jensen design, at the carefully described test points of the Jensen transformers in question. However, they may offer a more musical non linearity and in some cases quite a bit more information. Perhaps in a slightly larger size and probably at greater cost.
Then their are the audio specialists, designers who take a particular characteristic and spend quite a bit of manufacturing dollars and materials dollars to extend their favorite category, one appreciated by purchaser's who's taste in reproduced music, has matured to the point that they want to pursue one of those specialist's paths. At this point, performance is strictly a matter of taste. All of the choices, including those of Jensen, will fall within the boundary of 99.993% perfect signal reproduction. The differences will be in what is deliberately given up in certain categories, to attain maximum performance in other categories.
I really am not trying to be smug and audio lofty here. These really are the choice vectors when you begin to look beyond simple measurement, in the putative "black art" of transformer design.
What Lynn likes, and I do know some of it, might not suit your taste at all. Best to ask a lot of people what transformer manufacturers they prefer and why. What characteristics are they expecting to receive and have they received them in the past from more than one manufacturer. Then begin to develop a nose for specifics, in the comments of the people you ask. A major category is going to be price. Another will be measured specifications. Another will be grace under the demands of very non linear music and a fourth might be information content and musical values, as espoused by musicians, not electrical engineers or transformer designers. Also a good idea to get some idea of the musical genre's that someone you have asked, considers acceptable.
Bud
Last edited:
OK, interesting info. But what makes the "hot" side of the power (in North America) special? It's just one side of the distribution transformer up on the pole. The neutral line is the center tap of that transformer. It gets tied to ground.
So if the grounded center tap of the house current is tied to the high capacitance side of the device power transformer, what does that do?
.
So if the grounded center tap of the house current is tied to the high capacitance side of the device power transformer, what does that do?
.
The absolute value is not accurate, that is why I did not show the scaling.