That sounds fine, I think you were getting confused by us techs babbling.
You can connect an rca jack to jack with ground thats fine, nearly every stereo does.
In some circumstances it might be better to use different grounding methods, if your system works as its connected up then use it that way.
It seems that your amp blew up because there was a cable fault, ie broken.
My opinion on home audio violating the grounding at one end rule is that not everyone connecting up a stereo is not an electronics technician, if I were to connect up mine with seperate grounds and loads of jazz then I'm sure something would end up in smoke.
You are ever so correct in wanting separate grounds. When we did our 100 Watt amplifier for the AirCraft Carrier announcing system (2500 Watts total system), 100W amp chips were available, but not in mil spec, so we had to do it the old fashioned way. Our audio design engineer wanted each circuit on the board to have it's ground come back to a single point connection with ALL of the other circuit grounds on the board. Of course, with limited space available, this was not practical, so we had to work out a compromise that worked. Basically, the most sensitive stuff goes on the end of the ground, with the heavier current circuits located closer to the main ground point.
Having done circuit board design for over 20 years now, a ground plane is best, but when you're only doing a 2 layer board, you may not be able to dedicate one layer to a ground, so a ground grid is your next bet. You want as many ground to ground connections as possible. This does not violate your shielding rules, as this is not a shield.
You are correct when you say shields should be connected at one point only, and yes, it is usually at the source but this is not always the case or fesible. In 1987 I built a digital tachometer (Digitach®). I used a divide by n counter (makes incoming clock a two rpm period) and a LUT (look up table), without using a micro as they were fairly expensive at the time. My buddy put it in his car, and I asked him how it was working, and he said, "Lously". So I got in his car and looked at it. It was all over the place (it had a digital display and a bar graph, it read 4 bytes from the clock count, so it had 3 digits = 12 bits and 20 bits for bar graph info). I asked him, "did you ground the shield?" "Yes" "which end" "both ends"... "oh," I said, and cut the shield from ground on the board, and it fell right into place, working perfectly...
There is a book I read in the 80's, "Grounding and Shielding Techniques for Instrumentation", author unknown as the book has disappeared from my library (just googled - Ralph Morrison). I only read through it once and in 2002 was able to troubleshoot a ground loop problem in a video cockpit display system over the phone (I built the power supply/video amp board for Flight Systems Inc out of Florida) and they swore my pcb was 'noisy', but I had determined it was their system wiring and was correct.
Last edited: