24. System Grounding

Differences in ground potentials on interconnected HiFi units are arguably the most significant cause of signal infidelity in sound reproduction and constitute one reason for the manufacturer’s THD levels not to be met. Ground potentials may vary due to small amounts of voltage leakage from the transformer within a unit, or similar issues, but they are even more likely to differ between units, due to internal design choices and—most importantly—due to differences in the grounding that is provided by the AC power source. Furthermore, external antennas and installations from cable network providers constitute additional noise sources.

While it is true that the power transformers in our HiFi units regulate the AC current supplied from the wall socket and isolate the chassis, transformers will unavoidably leak small amounts of current, resulting in chassis voltage. Measured on our Harman Kardon HK730 receiver, for instance, the current leakage was confirmed as varying between 0.3 and 0.6 volts. Transformers in audiophile units will most likely be of the conventional linear type, with an iron core and wound primary and secondary coils. Since this winding has a direction by definition, the input resistance is likely not the same at both primary ends. Since in AC power supplies only the phase oscillates, e.g. between +230V and -230V in Germany, there will be a different transformer input resistance depending on which end is driven and different levels of leakage voltage, even if the resulting secondary voltage is the same.

On an all-in-one HiFi unit, slight deviations from ground level are normal and will not affect the sound, but when two or more units are interconnected, differences in chassis potential will be harmonised via the RCA/cinch connectors and form a ground loop with the house installation. It is a loop, because the HiFi units are connected with each other on two levels: the power supply and the RCA/cinch connectors. And even in cases where the noise itself is not audible via the speakers, the integrity of the music signal is compromised, because it carries the additional burden of energy leakage. Human ears tend to be unforgiving to technical noise, perhaps, because it is not a natural occurrence to which our species has had millions of years to adapt. To address the issue of varying chassis potentials, the professional audio industry has long since introduced symmetrical and optical connections that do not connect chassis grounds. There are, however, acoustic downsides that result from the addition of couplers which have made most audiophile listeners shy away from this technology.

In audiophile circles, therefore, it is common practice to minimise differences in ground potential between the units by turning the power plug of each unit and comparing the resulting audio output though listening tests. When there is minimal difference in potential between chassis, the music will sound fuller and lusher with better imaging and tonal balance. When there is greater difference between potentials and more leakage current is carried via the RCA/cinch connectors, both lower bass and transients will be compromised with the tonal balance leaning towards brighter and yet somehow duller highs. To identify the lowest amount of difference, it is important to start from a minimum number of devices and then check the polarity of each new unit. Companies such as Oehlbach offer a polarity tester, however, I have found my ears to be the better judge of polarity and have at times disagreed with the unit’s findings. In recent years, I have achieved the most satisfactory results by checking the polarity by ear.

Since polarity makes such an audible difference on grounded units that are connected with RCA/cinch, finding out the preferable polarity has become a common first step in setting up HiFi systems, especially among vintage audiophiles, who do not follow the path of correcting the signal retrospectively, by running it through a digital sound processor (DSP). In fact, when signal integrity is maintained throughout the unit, the sound will be superior to retrospective DSP corrections, simply because a DSP cannot correct what is not there and will in the end be kept busy filtering and flattening the effects of a hum or of HF interference, without being able to differentiate between the signals. Maintaining signal integrity is similar to fly fishing, whereas a DSP is more the practice of throwing dynamite into the stream. Sure, skilled people will catch fish either way, but I doubt that the pleasure of the achievement will be the same.

My initial reason for embarking on the subject of system grounding was of a different nature, however. For many months, I had a humming issue on our Philips turntable that I could not explain. I had already removed the antenna cable, optimised system polarity, and still the humming continued to be present at a level that was highly annoying. The turntable signal was compromised to the point where it was lacking both dynamics and soundstage. Since I could not find the solution, I decided to write about my sorrows in a vintage audio forum, describing our system setup from power sockets to our Martin Logan electrostatic speakers. The response was imminent: Where did you plug in the power cords of the speakers‘ high voltage supply?

The mistake I had made by plugging in the speakers into wall outlets (right where they stood) was instantly clear to me, but only with the little nudge from a forum member. By using separate sockets, I had created a difference in ground potential between our high voltage powered Martin Logan SL3 speakers and our B&K ST140 amplifier that was now trying to harmonise potential through the speaker cables in the form of an audible hum. Since this only occurred when phono was selected as input, I had simply failed to understand the situation. With hindsight, the reason for this is simple: since phono is more sensitive to interferences, this was the most obvious point of the error to become visible. This does not mean, however, that other sources were not affected. Connecting the speakers via the same PowerStar rail as the rest of our system brought phono noise down by an audible 70%, at least.