Thanks for all your responses and discussing it amongst yourselves. I hope we continue along those lines. There are some good questions here and rather than quote each one I will take the subject on as a whole. If I miss a point you particularly want to discuss please post it.
In the beginning, nothing was grounded. Not radios, not amplifiers, nothin. I cannot say when 3 prong outlets came into vogue. I didn't have them in the house I grew up in which was built in VA in 1956. None of the classic Marantz, Fisher, Harmon Kardon or Dynaco... etc, etc had them. Adding one to these may make a quiet system into a hummer.
I recommend grounding only the power amp for the simple reason that it has the largest power transformer in the system. If your preamp had the largest transformer then I would ground it and not the power amp. The reason is quite simple: Every power transformer has capacitance between primary and secondary. The larger the transformer, the larger the capacitance. This capacitance is returned to the chassis making it have enough AC voltage that you can feel it. A high impedance digital AC voltmeter can often read several hundred volts between an ungrounded chassis and ground. But, if all is well, there is no current behind that voltage (Thevenin can be applied here). Most DVM (digital volt meters) have 10 megohm impedance. When I do these tests I put a 1 meg resistor right across the input terminals. If your meter has ¾ inch spaced bananas you can put a 1 meg resistor on the screws of a double banana plug and insert it in parallel with your probes. If you have any of the recent Fluke meters you will have to cut off that pesky plastic that hides the metal part of the banana. I’ve cut them off all of mine. With the 1 meg resistor in place the reading should go down by a factor of 10 and now every volt is a microamp. I typically see under 100 microamps, which presents no danger, but you might get a tingle.
Now we can answer the question of plug polarity. The Stereophile article, which I am not sure I have ever seen and would love to read it should a copy get posted here, likely refers to the following practice which many audiophiles became aware of in the late 70’s. The trick was to disconnect all amps, preamps, tuners, tape decks, etc from each other and from the wall plug. So you have everything out in front of you, nothing connected to anything else. Then you found a good ground, a water pipe if no 3rd pin was available, and put the AC meter between the ground and the chassis of each component starting with the power amp as it should be the highest. Then you plugged the amp in the wall, noted the voltage (current), rotated the plug 180 degrees and noted the voltage in that polarity. You would choose the lower figure and mark the plug accordingly. Then on to the preamp, then on to the signal sources finding the lower current polarity for each and noting that the readings were getting smaller as you progressed toward the sources. A tape deck might have 1/10 the current of a power amp. It is also possible to get voltages from chassis to ground greater than the 120 volt line when the meter is 10 megs.
Now that the stage is set we can answer the question properly. If all this technical and measurement talk leaves you cold I encourage to take in what you can. As your knowledge builds these things will become more obvious. Often two engineers (or people skilled in the field) can say all of what I am explaining here in a sentence or two. One day when Bruce DePalma (I’ve mentioned him in previous posts) said to me: “I called the cable company to complain that their signal was rather weak causing noise (snow) in my picture. I asked the engineer “Hey, why don’t you just turn up the gain a bit and give me a little more signal? The cable engineer replied “cross modulation”. Bruce chuckled, I chuckled as that’s all we had to hear. That problem had never occurred to us. The unsaid part, that the cable engineer felt would me immediately understood from his one word answer, was that higher gain in the amplifiers would allow cable channels to interfere with each other causing greater problems than a little noise in the picture. End of story. I hope many of you may gather the knowledge necessary to have such conversations. I also hope you will find that talking amongst each other about these matters makes conversations like “This capacitor made my system really wonderful, you just have to try it”. There’s really nothing to discuss in such an exchange is there? It’s just a preference like white or whole wheat bread.
Since almost everything we buy these days has a 3 pin plug (which I usually cut off) we have a new situation to contend with. We can’t rotate that plug. The person who advised Steve to cut off the pin, rotate the plug and, I assume, leave it flipped over, makes no sense at all. That implies that EVERY grounded plug we have is backwards. If his trick does reduce the hum, it is because he eliminated the problem when he cut off the ground pin. One could confirm that by flipping the plug back to the way it was. Nowadays most gear has IEC power cords so you can take a cheap one and cut the ground pin for your experiments or use a ground lifter (cheater they are also called) to make your experiments. I have a box of IEC cords from computers with all their ground pins cut off. I reach for them more often than a grounded one. I have over a dozen instruments on my bench all missing but one missing their ground pins (or never had them). The one unit I left the ground pin on has that ground brought out to a front panel binding post so I can use it or not without having to fool with plugs. Many home theater amps have a “ground lift” switch. Lifting the ground via the switch is the same as floating it at the plug.
Something that is often overlooked is a tuner or TV that is connected to cable or a roof antenna. At Beveridge we had a rooftop antenna only 30 feet below some high voltage power lines. The 60 Hz coming down that cable was very strong. TV cable running along with AC distribution can pick up very strong 60 cycle currents. Even if you take my advice and ground just the power amp you can have very loud hum even with the TV off and listening to CD.. The best way to cure this is a simple cable isolator. It’s basically a 1:1 isolation transformer for cable frequencies. In a pinch I’ve made one by taking two 300 ohm to 75 ohm transformers connecting the 300 ohm leads back to back leaving a 75 ohm input and output. Then you get double isolation but a little more signal loss. This hum often intrudes on all inputs as it finds its way to ground via whatever path is available. This offending cable current is there whether the TV is on or not. Whether it is plugged into the wall or not. But this hum is easy to find, just pull the TV audio cables out of the back of you preamp and the hum goes away. For a quick fix the other choice it to float all the grounds, but I advise against that because I have measured (and felt) some sizable voltages between TV cable and ground. Anyone ever get shocked hooking the coax to the tuner RF jack while holding the tuner? Don’t be standing barefoot on concrete when you fool with the TV cable either.
Here is why I contend the power amp and only the power amp should be grounded. Remember that we measured or through logic decided that the power amp has the largest AC (60 Hz) current to ground. The preamp less and the sources even less. If we ground the preamp and not the power amp that current has to flow through the interconnects to the preamp to find ground via the preamp power cord. That’s where the hum enters the system. The current causes a small voltage drop across the shield wire of the interconnect which, to the power amp, looks just like the audio signal and gets amplified and gets sent to your speaker. Many years ago I had loaned a power amp to a speaker maker and he was having more than a little problem at the CES with hum. I walked in the room and heard it, looked at the plugs, saw the CD player (of all things) was grounded. From a hum point of view a CD player should not be grounded. Let it and the preamp find ground through the interconnects. If you still have hum due to these currents a heavy wire tied along the interconnect from chassis to chassis will help. The cable should be shielded. Some cables have this third wire and it should be at least 18 gauge; thicker for longer runs.
So far none of this is about loops. Often the problem is not a ground loop but chassis current’s finding their way to ground. When something electrical doesn’t work, I find most people say “there must be a short”. In actually, 9 times out of 10 when something won’t work it’s an open. In the same way, the audio community has made the term “ ground loop” as the catch-all for hum problems. It is often not the case. In a simple system with a CD player, preamp and power amp (all with ground pins) you might find that arbitrary grounding only the preamp reduces the hum considerably, you might as well find that the system is even more hum free if just the power amp is grounded. That is the logical choice.
After we kick this one around for a bit, we can get into balanced lines.
