110202-1202 EST
tryinghard:
If correct equipment grounding exists between the circuit source and equipment as well as the source computer the only resistance will be the equipment grounding conductor itself, (should be extremely minimal if any – closest to zero potential). Unlike earth which will have high resistance for sure.
The assumptions of your statement do not fit reality. What is "correct equipment grounding"? EGCs are not essentially 0 resistance.
The discussion needs to be broken into two parts --- fault, and normal conditions.
First, an example of a fault condition in a very simple form. The power system as seen at the main panel will be assumed to be very stiff, a low impedance (1 V drop in supply voltage at 2000 A). This is a 120 V supply. Connect 100 ft of #12 Romex with ground. Today this means the EGC is also #12. The resistance of #12 is about 0.16 ohms for 100 ft. Total resistance of 200 ft of #12 is 0.32 ohms. 0.32 ohms across 120 V equals 375 A. A 375 load on this panel produces a voltage drop of about 0.2 V. Thus, ignore this drop.
At the end of this 100 ft run of Romex short the hot wire to the EGC. What is the voltage of this shorted point relative to the neutral point at the main panel? It is 60 V RMS or 85 V peak. This will destroy RS232 components.
Second, consider normal conditions. Do you really believe that in normal conditions that no current flows in EGC conductors? If you do, then you have never measured voltage drops along such conductors.
That the code says there should be no current flow on the EGC does not mean there will not be. At this moment, at home, if I measure the voltage drop on the EGC from my main panel to may work bench with a Fluke 27 the result is 0.021 V. If I activate my TED power monitor, then this jumps between 0.021 and about 0.060 V every second. TED uses a carrier current system for communication.
You can not just consider resistance. The inductive reactance is also important. At 115.2 kbaud the minimum bit period is 8.7 microseconds. Thus, the effect of inductive reactance is probably about 1000 times greater than at 60 Hz.
I had one customer with two HAAS machines side by side. Just enough room to walk between the machine enclosures. One to two years difference between date of manufacture. Otherwise essentially the same model machines. One machine worked fine in drip feed mode, the other was impossible.
HAAS said there was nothing wrong with the machine that had communication problems. The was based on using a laptop to drip feed to the problem machine, it worked. Thus, the HAAS dealer said the problem was the RS232 source. It should be noted the laptop was either run on battery or the power cord was a two conductor type.
Each machine had added to it a supplementary ground rod. The normal EGC was intact. Power to the machine was from an overhead bus. Power to the computer was from a wall outlet. RS232 cable length was about 100 ft.
With a Simpson 270 on AC about 1 V difference appeared between the two machines. This could mean the peak noise voltage might be 4 to 5 V. The EGC from the bus duct to the machine was about 30 ft. Thus, about 60 ft of EGC wire from one machine to the other. The two ground rods were about 6 to 8 ft apart.
It should be noted that with the problem machine servos off there was no RS232 communication problem. Also the problem machine did not interfere with the other machine.
I installed our RS232 isolator and the problem was solved and further operation was possible at 115.2 kbaud.
The real cause of the problem was noise generated by the brushless servos in the newer machine. The older machine had DC servos. So truly the HAAS machine had a problem, and likely a design problem.
There are all sorts of ground path noise sources and just by doing an installation per the NEC does not guarantee there won't be a problem.
If you measure a voltage with a normal voltmeter (not electrostatic), then there is a current related to that voltage.
.
