HowY said:
I've been directly involved with several computer lab electrical installs
and DO NOT RUN A SEPERATE GROUND!!!!
All grounds need to terminate TOGETHER in the building...
From flats to mega multistory's ONE GROUND POINT period
Several rods perhaps but they all must be eletrically common
I can spin many a yarn about the "isolated ground" and
"independant ground" theories from just a few years back
and specifically through much heartache and cashola
all grounds MUST MEET at a common point (the rod)
You WILL introduce differential ground loop and so much
noise... and I can elaborate on some serious conditions
that arize from seperate grounds IN A DATA CENTER
for computers. From breaker failure to overvoltages
noise injection and on and on...
I can only imagine the effect upon audio
One should really do a little research on grounding and electrical
design in general before even suggesting a seperate rod.
FWIW
Yeah, I've done a few grounds in my time. A few years back we finished a 1.4 billion dollar plant with every type of electrical device you can imagine. From 13.5KV motors, two Cogeneration plants to data communications, DCS and PLC controls to the common light switch. Without a thorough understanding of grounds and the how the internals of the equipment is wired, you can end up with all kinds of strange voltages.
Most folks cannot fathom how an isolated ground works and the purpose. To most, ground is ground and that is how the isolated grond gets destroyed.
Shielding is meant to keep stray noise from entering the system and should always be on its own seperate grounding system. The same equipment that is using that shielding, should also be tied to that ground but only at a point that is closest to the final ground rod. The safety ground that is hooked to the wall plug should be on its on independent ground in the breaker box.
Now here is where the problem comes in, much of the older equipment would tie the ground for the power supply of the system to the chassis of the equipment. This in turn gets tied to the safety ground, instand loop. Newer equipment will float the ground in the power supply and not reference to the chassis. So you have to be aware of the internals of the equipment you are working with before you can design a grounding system.
Any where along the line, if you tie these seperate grounds together, you will induce current loops and if you tie the grounds together underground, you can also create a a path for noise.
During the commissioning of this new plant, we had noise on all the low level loops. After looking at the grounding scheem, I determined that the safety/equipment grounds were tied together under ground. After two months of arguing with the so called experts, they dug up the tie and cut it, all the noise in the system instantly went away.
We just finished a second plant, 230 million worth and again, the issue of correct grounding came up, the design had some bad tie points and again the tie points became a problem. Ties cut and problems vanished.
There is a right way and a wrong way to do grounding, I've been involved with noise reduction in electronics since before the computers came into our homes, way back when stereos were still tubes. I think I can qualify in saying that the correct way to ground a shield is to an isolated ground. But if you can't set one up, do the next best thing and run a seperate back to a good tie point, the best being the rod.
