Separately derived systems, and the one reading that decides every item built on them

The test that settles it

Article 100 defines a separately derived system as an electrical source other than a service that has no direct connection to the circuit conductors of any other source. A service is excluded by name, so a service is never a separately derived system however it's wired.

Then comes the carve-out that answers most items: connections established by grounding and bonding are excluded from that test. The enclosure is bonded to the same metal as everything else, an equipment grounding conductor runs back to the panel, and the earth sits under both. Every one of those is a grounding or bonding connection, and a circuit conductor is a different thing.

So the question to ask is always the same. Does a circuit conductor run through, uninterrupted, from the other source? An uninterrupted circuit conductor makes the two one system. Interrupt all of them and you have a new source.

How the exam asks it

Our bank of 500 original Texas practice items puts 66 in the grounding and bonding area, and the phrase separately derived appears in items spread across two of the eight areas we tag, which tells you the topic travels. On the PSI content outline the home area is heavy by Texas standards: services, service equipment and separately derived systems carries 6 of the 56 scored items on the journeyman knowledge portion and 4 of the 24 scored items on the calculations portion, where it ties with branch circuit calculations for the largest block.

Two forms account for nearly all of them.

• Classify the source. A standby generator feeds a building through a transfer switch and you say whether it's separately derived. The right answer turns on whether the switch opens the grounded conductor along with the ungrounded ones.
• Follow the consequence. A fault happens on a 120 volt branch circuit downstream of a 480 to 208Y/120 volt step-down unit, and you say where the current returns to. It returns to the secondary winding that supplied it, across the system bonding jumper.

The distractors on the classify items are all plausible facts about the installation that have no part in the definition: the kilowatt rating of the generator, whether it's permanently installed, and whether an electrode conductor was run from it. A candidate who reaches for size or permanence is answering from the nameplate while Article 100 asks about the conductors.

Pole count is where it breaks

The shortcut in circulation is that a four-pole transfer switch makes a generator separately derived and a three-pole switch leaves it tied to the service. On an ordinary three-phase, four-wire system that shortcut lands on the right answer. It inverts on the system most Texas candidates actually wire.

What the code keys on is the switching action. The informational note traveling with 250.30 says an on-site generator sits outside the definition where its grounded conductor stays solidly interconnected with the service-supplied grounded conductor, and the example it gives is transfer equipment with no switching action in the grounded conductor. Read the voltage first, work out how many conductors that system runs on, and the pole count then tells you something.

A portable machine with its neutral bonded to the frame, fed into a house through an inlet and a switch that passes the neutral through, is the same error wearing different clothes. The generator stays part of the service-supplied system, so the bond inside it becomes a second neutral to ground connection on the premises wiring, which is the condition worked through on the neutral to ground bond.

One bond, at one point

The system bonding jumper ties the grounded conductor of the derived system to the equipment grounding conductors at any single point on the system, from the source to the first system disconnecting means or overcurrent protective device (NEC 250.30(A)(1)). At the source, at the first disconnect, or anywhere between the two.

Install one at the source and a second at the first disconnecting means, calling it redundancy, and the two close a loop. Derived neutral current then comes back partly on the metal between the two enclosures, which is the same parallel path the single-point rule exists to prevent. A permission to choose a location stays a long way from a permission to use both of them.

One exception belongs in place. Where a building is supplied by a feeder from an outdoor separately derived system, a jumper at both the source and the first disconnecting means is permitted, so long as no parallel path for the grounded conductor results (NEC 250.30(A)(1), Exception No. 2). That is a narrow allowance written for one arrangement.

Past the bonding point the grounded conductor goes back to being an ordinary circuit conductor, with the same separation that applies downstream of a service. So a panel fed from the derived system is separated, and a panel fed from that panel is separated too.

Where a fault goes after that

A derived system has its own source, so a fault on it returns to that source. The path runs out on the equipment grounding conductors to the first disconnect or back to the winding, crosses the system bonding jumper, and returns on the derived grounded conductor. The grounded service conductor upstream has no part in it, and the grounding conductor of the primary circuit protects the primary circuit.

Leave that bond out and the fault path terminates nowhere useful, so the overcurrent device sees a fraction of the current it needs while the metal stays energized. The conductor carrying that fault out to the bonding point is covered on the equipment grounding conductor, and it's sized from the rating of the overcurrent device ahead of the circuit under 250.122.

Where the source and the first disconnecting means sit in separate enclosures, a supply-side bonding jumper runs with the circuit conductors between them (NEC 250.30(A)(2)). It does the job on the supply side of the first overcurrent device that an equipment grounding conductor does on the load side, and it's one of the four jumpers sorted out on bonding jumpers.

The system also gets its own connection to earth. The grounding electrode conductor runs from the same point where the system bonding jumper is made, out to the grounding electrode (NEC 250.30(A)(5)).

The metal in the area served

This is the requirement candidates skip, and it makes a clean item because the intuition says the building was bonded at the service years ago.

NEC 250.30(A)(8) bonds the metal water piping system and the exposed structural metal in the area served by a separately derived system, and the connection is made at the same point on the system where the system bonding jumper is installed. A unit on the third floor feeding panels on that floor pulls the piping and steel in that area onto its own fault path.

Bonding done back at the service ties that metal to the utility source and does nothing for a fault on the secondary, because the two are different sources with different return paths. An item on this offers you real field practices as wrong answers: counting the service bonding as sufficient, or landing the bond on the equipment grounding bar of the nearest panel.

The third one on offer is an electrode driven on that floor, and it repays a moment, because an electrode clears no fault on any system. That is the split drawn at 250.4(A)(5) and worked through on grounding and bonding as two jobs.

Where it sits in the book

250.30 holds the requirements and four subsections carry most of the items: (A)(1) for the system bonding jumper, (A)(2) for the supply-side bonding jumper, (A)(5) for the grounding electrode conductor, and (A)(8) for the piping and structural metal.

Two sections outside 250.30 are worth a tab of their own. NEC 250.64(C) asks the grounding electrode conductor to run in one continuous length with no splices, then names four places where that gives way, and three of those four are provisions of the separately derived systems section. A building can hold several derived systems on several floors, so the code permits a tapped arrangement in place of an unspliced run from each of them.

NEC 250.21(A) opens a short list of ac systems between 50 and 1000 volts that are permitted to be grounded while carrying no obligation to be, and two of the systems on that list are named as separately derived. Both are industrial. The sentence carries a permission and a release from an obligation at the same time, so an answer choice will hand you one of the two.

Texas adopts the 2026 edition on 1 September 2026 under 16 TAC 73.100, and the examinations reference it from the same day. Article 250 kept its number while other parts of the book moved, so read 250.30 in the 2026 copy. Our page on which code edition applies covers that date from the licensing side.

What to do before test day

The identification question comes first, so put your preparation there.

1. Read the Article 100 definition, including the words excluding connections established by grounding and bonding. Two minutes, and it's the whole topic.
2. Read 250.30 in place once and mark two things: where the system bonding jumper may go, and where the grounding electrode conductor connects.
3. Draw the same generator twice, once with the grounded conductor switched and once with it solid, and write the system voltage beside each drawing. Then work out how many poles that switch actually needs.
4. Answer thirty mixed items under a clock at our free timed practice set, which prints the section behind every one.

A single-winding unit with a tap, the kind used in a buck-boost arrangement, shares conductors between its input and its output, so a circuit conductor is common to both sides and the test fails at the first question.

Where you want the sizing worked line by line, with the figures set out and the tables traced, that lives in the question set. This page is the Texas candidate's view: which reading decides the item, and what the classification costs you on the truck when you get it backwards.

Questions people ask

What makes a system separately derived?

An electrical source other than a service, with no direct connection to the circuit conductors of any other source. That is the Article 100 definition and the whole test. The sentence then excludes connections established by grounding and bonding, so an equipment grounding conductor running back to the panel, a bonded enclosure and the earth underneath both leave the classification untouched. Ask one question of any arrangement: does a circuit conductor run through, uninterrupted, from the other source? An uninterrupted one makes the two a single system, and interrupting all of them gives you a new source.

Is a standby generator a separately derived system?

It depends on what the transfer switch does to the grounded conductor. A switch that opens the grounded conductor along with the ungrounded ones leaves the generator with no circuit conductor in common with the utility system while it carries the load, so the generator becomes the source of a separately derived system and gets its own bond and its own electrode connection. A switch that passes the neutral straight through keeps the generator tied to the service, and the bond stays back at the service. The kilowatt rating and permanence have no part in the definition.

Where does the system bonding jumper belong?

At any single point on the system from the source to the first system disconnecting means or overcurrent protective device (NEC 250.30(A)(1)). At the source, at the first disconnect, or anywhere between the two, and at one of those places only. Installing one at each end closes a loop and puts derived neutral current on the metal between the two enclosures. Exception No. 2 permits a jumper at both ends where a building is supplied by a feeder from an outdoor separately derived system and no parallel path for the grounded conductor results, which is one narrow arrangement.

Does a single-winding tapped unit create a separately derived system?

A unit built with one winding and a tap shares conductors between its input and its output, so a circuit conductor is common to both sides and the Article 100 test fails at once. The system on the output side stays part of the system feeding it. An ordinary two-winding unit is the opposite case, because power crosses by magnetic field and nothing on the secondary is electrically continuous with the primary, which makes the secondary a new system. The tell in an exam item is usually a description of separate windings, or the pointed absence of one.

What happens if a system bonding jumper is installed at both ends?

Neutral current appears on the metal between the two bonding points. The derived system now has two connections between its grounded conductor and the equipment grounding conductors, so the return current from the load splits between the grounded conductor and the raceway, the enclosures and everything else bonded along the way. Nothing trips and nothing fails a visual check, which is why the arrangement survives. NEC 250.30(A)(1) permits a choice of location and permits one jumper, and our items put that wrong answer in the words an installer would use to defend it.

How heavily is this area weighted on the Texas journeyman exam?

Services, service equipment and separately derived systems carries 6 of the 56 scored items on the journeyman knowledge portion and 4 of the 24 scored items on the calculations portion, where it ties with branch circuit calculations for the largest block on that outline. Those counts come from the PSI content outline, which prints an item count beside every subject area, and they sum to the scored total. In our own bank of 500 Texas practice items, 66 sit in the grounding and bonding area that this topic anchors.