I will jump in here and give you my thoughts because I specialize in performance grounding for sensitive electronic equipment.
CNC manufactures are misguided. They do not understand IGR principles or how electrical noise and interference are generated. The root cause of RFI/EMI is forcing or allowing common-mode currents to flow on Ground Circuits. I mean any current, DC, AC Power frequencies, or RF..
So the question becomes how to prevent current from flowing on equipment grounds? It is pretty simple; current cannot flow in an open circuit. For current to flow requires a complete circuit, a node for current to enter, and another point to exit. You have to create a true Single Point Ground void of ground loops.
There are 3-ways to induce common-mode currents onto equipment grounds, and they are:
1. Starting with the easy one is 60-Hz line currents. Please think of how most 120 VAC AC branch circuits are wired in a home or office; they are daisy-chained unless dedicated. Now think of what we plug into the circuits like TVs, AV systems, USB chargers, or anything electronic with Line Filters installed between Line and Ground. For example, a .common practice is placing a 01 ufd capacitor installed between Line & Ground. That capacitor will leak or intentionally inject 60-Hz line current into the EGC. It will inject 4.6 milli-amps into the EGC. Although the current is small, it flows in common-mode without an equal and opposite circuit conductor closely coupled to cancel out the mutual inductance resulting in high impedance creating voltage drop along the length of the conductor. Those slight voltage differences are picked by sensitive equipment and amplified, causing data or signal errors. The simple solution is a dedicated circuit so no upstream or downstream devices can corrupt the EGC..
2. This is the primary concern of CNC and any sensitive equipment fear, the father of Ground Loops, multi-point ground. A common myth believed by many engineers, technicians, and sparkies believe, ground rods and other ground electrodes are at a 0-volt potential. Nothing could be further from the truth. Get that garbage out of your mind; it leads you to the wrong conclusion.
If you drive two ground rods 20 feet apart, they will be at different voltage potentials. You can measure it with a DMM. There will be 60-Hz voltage, DC if you have any cathodic protection systems in the area, cosmic noise, and RF if you have a radio, TV, or cellular provider nearby. Bond those two rods together with a jumper or equipment, and you have forced outside common-mode currents to flow in your ground conductor. Ham radio operators are famous for doing that. They love to spend the rest of their lives trying to fix all the RFI and EMI issues they created by placing a ground rod outside the shack, placing themselves in a nasty ground loop.
With that said, think of a commercial or industrial application like a large fabrication shop spread out over a 40,000/ft2 concrete and steel building. The whole building is a ground electrode, thousands of them bonded together with common-mode current flowing. You run an equipment supply circuit in RMT supported by building steel and concrete anchors. That construction method has multiple unplanned
incidental-grounds contacts with earth-ground electrodes. Once you arrive at the terminal equipment, the EGC is bonded to the frame/enclosure secured to the concrete/steel floor with floor anchors. You have placed the equipment in a ground loop. To complicate matters, that piece of equipment has to communicate with other equipment on the opposite side of the building. If it uses copper wire signal circuits, you can have a problem using any Ground Referenced Signal Topology.
Isolated Ground supply circuits can resolve the problem if
used correctly. They must originate at the MBJ junction, insulated to prevent incidental contact (electrically isolated) with ground, continuous without splices or junctions, and terminate to the IG bus inside the terminal equipment. Electrically the Isolated ground wire is nothing more than a long insulated wire with 10-feet of insulation removed, stuck in the mud, and runs inside going nowhere. It is an open circuit with no current. No current means no voltage and no NOISE.
It is essential to understand the IG wire is not an EGC or a safety ground. It does not carry fault currents. It is a SIGNAL GROUND or DC Equipment Ground for electronic processing hiding inside the magic black box. The fastest way to corrupt your IG is by following manufactures directions driving a ground rod and bonding with a jumper to the IG bus. You just put yourself right back in a ground loop after you spent all the money, time, and material trying to prevent it. DOH! NEC 250.54 is a meaningless article not worth the ink and paper wasted on it. It is the only article with no requirements put in NEC 250 to shut CNC manufactures and ham radio operators up. It is the only article I know of that can make an installation unsafe and inoperable.
3. The last mechanism to inject common-mode currents is magnetic induction from strong EMI fields. Easily controlled with proper installation practices, which include:
a. Run your CNC or sensitive circuits in ferrous metal raceways like RMT to provide EMI shielding.
b. Do not mix circuits with sensitive equipment circuits. Run CNC supply circuits in dedicated raceways. It is OK to run more than one CNC supply circuit in a raceway, do not mix other loads like lighting, HVAC, motors, etc.
c. Keep CNC raceways as far from other raceways as possible, and cross at right angles to the furthest extent possible.
I hope that helps.
EDIT NOTE: FWIW if you have a stick-built building like a house, all the circuits are IGR. If you want a squeaky clean AC outlet, just run a dedicated circuit. The construction method makes it IGR. Romex running through wood framing terminate in a plastic box nailed to wood framing. Well unless you have your HVAC unit sitting outside on a concrete pad.
