Okay, let's break down your scenario and questions regarding the grounding and bonding of this separately derived system.1. Is my interpretation and application of the code correct here?

Yes, your interpretation is generally correct, but there are nuances and things to consider. Let's dissect each scenario: Without the Disconnect (Transformer Directly to Switchboard)

You are correct that since the System Bonding Jumper (SBJ) is at the transformer (as it should be in a separately derived system), and you have multiple raceways from the transformer to the switchboard, the bonding jumper in each raceway is a

Supply-Side Bonding Jumper (SSBJ). 250.102(C)(2) is the relevant section, and it allows the SSBJ to be sized based on the ungrounded conductors in each raceway. With one 500 kcmil Cu conductor per phase in each raceway, Table 250.102(C)(1) dictates a minimum 2/0 CU SSBJ. Note: This is the minimum. Important Note

While 2/0 Cu is the minimum, practical considerations and engineering best practices often lead to the use of larger SSBJs. Think about impedance and fault current return paths. A larger conductor reduces impedance, which can improve the effectiveness of the ground fault path.

With the Disconnect (Transformer to Disconnect to Switchboard)

This is where the change in grounding nomenclature and sizing requirements occur. Once you pass the load side of the overcurrent protective device (the fuses in the disconnect), the bonding conductor is now an

Equipment Grounding Conductor (EGC), not a SSBJ. You are correct that 250.122 now applies. 250.122(B) states that where ungrounded conductors are installed in parallel, the equipment grounding conductor shall be installed in parallel. Each parallel equipment grounding conductor shall be sized on the basis of the overcurrent protective device protecting the circuit conductors in the raceway or cable in accordance with Table 250.
122. The 4000A overcurrent protective device protecting the parallel conductors would therefore require 500 kcmil Cu ground per raceway.2. If #1 is "yes", how in the world does this make sense?

This is the crux of your question, and it highlights a common point of confusion and frustration with the NEC. It does seem counterintuitive that adding overcurrent protection increases the minimum required EGC size.Here's the (sometimes unsatisfying) reasoning behind it: Fault Current Magnitude and Duration

The fundamental principle is this: The EGC must be able to safely carry the fault current that it is expected to conduct long enough for the overcurrent device to operate.

While the available fault current may be similar on both sides of the disconnect, the presence of the fuses limits the let-through current and clearing time during a fault. The magnitude of the let-through current is what drives the EGC size. If the overcurrent protection were a large, slow-acting circuit breaker, the EGC might need to be sized even larger to handle a higher potential current flow for a longer duration than with the fuses.

Protection Coordination and Selective Coordination

A larger EGC helps ensure that the fault impedance is low enough to allow the overcurrent device to operate as intended. This is particularly important for systems that are designed for selective coordination, where the upstream overcurrent device should only trip if a fault occurs downstream of it. Proper EGC sizing plays a role in maintaining the impedance that is needed to make sure the system properly coordinates.

Fault Clearing Time

The faster an overcurrent device clears a fault, the less energy (I²t) is let through. Fuses are generally faster than circuit breakers at interrupting high fault currents. Even though the clearing time is faster with fuses, the EGC still needs to be appropriately sized to withstand the energy let through, and Table 250.122 reflects this.

NEC is a Minimum Standard

It's crucial to remember that the NEC is a minimum safety standard. It doesn't necessarily dictate the optimal grounding and bonding system. Engineers often exceed the NEC's minimum requirements to improve reliability, reduce voltage drop, and minimize the potential for equipment damage during faults.

Why it Still Seems Illogical

You are right; it can feel odd because it implies that the presence of protection increases the grounding conductor size needed. It is, in a way, an artifact of how the NEC is structured and how it aims to ensure adequate fault current carrying capacity based on the overcurrent protection provided. The NEC doesn't necessarily try to optimize cost or ease of installation; its primary goal is safety.

In Summary

Your understanding of the code sections and their application is generally correct. The logic behind the EGC sizing increase with the disconnect and fuses is related to fault current magnitude, protection coordination, and the NEC's focus on minimum safety requirements.

Recommendations


Consider Engineering Analysis

For a system of this size, a thorough engineering analysis (including short-circuit studies and coordination studies) is highly recommended. This will help determine the optimal grounding and bonding system, not just the NEC's minimum requirements.2.

Don't Skimp on Grounding

Err on the side of caution when it comes to grounding and bonding. A larger EGC (beyond the minimum required) can significantly improve system reliability and reduce the potential for equipment damage.3.

Consult with a Qualified Electrical Engineer

This type of installation requires the expertise of a qualified electrical engineer who is familiar with the NEC and best practices for grounding and bonding.4.

Consider Impedance:* Especially with long parallel runs, consider the impedance of the grounding path. A larger conductor will lower the impedance and allow for higher fault currents to flow, potentially resulting in faster clearing times.Hopefully, this detailed explanation clarifies the code requirements and the underlying (albeit sometimes confusing) logic. Remember that electrical codes are complex, and professional guidance is always recommended for critical installations. Flag for review