Generator Synchronization Requirements: Evidence to Review Before Breaker Closure
Generator synchronization requirements cover electrical matching, reliable sensing, approved close permissions and breaker timing. For a connection to an energized bus, the review must connect those requirements to the instant the breaker main contacts close.
A controller screenshot can show acceptable conditions before that instant. It does not, by itself, establish which conductors were measured or what happened during the closing delay. Buyers therefore need a traceable evidence package alongside the proposed control equipment.
Illustrative scenario: A proposal includes a green “sync ready” indication but no timing record. The reviewer can identify the controller, yet cannot identify the event behind the indication. The next useful request is for the source labels, approved criteria and command-to-contact evidence. This is an example of an open document question, not a reported customer failure.
If you are reviewing a package, you need a practical way to separate evidence already supplied from checks still outstanding. This guide follows that chain through the one-line drawing, electrical measurements, closing sequence and acceptance records. It focuses on document review; qualified project personnel define and perform the electrical tests.
To clarify an offer, discuss the project scope with ZC Power. Provide the intended operating mode and your open document questions, and ask which records belong to the proposed supply.
Key Takeaways
- Review voltage, frequency difference, phase sequence and phase relationship together, using criteria approved for the selected equipment and operating mode.
- Trace each sensing input to the physical source and breaker. A screen label does not establish that correspondence.
- Distinguish close permission, close command, main-contact closure and status feedback. They describe different events in the evidence chain.
- Keep live-bus synchronization and first closure onto a dead bus as separate review cases, with defined interlocks and responsibility.
- Accept a test result only within its recorded configuration and scope. Simulated factory inputs leave installed interfaces to be addressed separately.
Define generator synchronization requirements for the intended bus
Identify the source, bus and synchronizing breaker
Start with the one-line drawing. Identify the incoming generator, the bus it will join and the breaker that makes the connection. Record drawing number, revision and equipment identifiers together. “Generator breaker” is too vague when a package contains several incomers or bus sections.
Specify the operating mode for that connection. A generator joining an islanded generator bus and a generator connecting to a utility involve different project boundaries. Utility-parallel operation needs the applicable utility requirements and approval responsibility identified; a controller feature cannot establish that approval.
Ask what is energized on each side before closure, which source supplies each voltage input and which control owns the close request. Tie these answers to the drawing rather than relying on words such as “incoming” and “running.” Those words describe roles, which may change with the selected breaker or mode.
For the broader system boundary, use the generator paralleling specification. That review covers operating modes, interfaces and acceptance responsibilities. This article follows the narrower evidence chain before connection; it does not specify post-close load sharing or a complete plant commissioning sequence.
An integrated standby power system also has interfaces beyond synchronization. Keep those responsibilities visible, but avoid treating every system feature as evidence that this particular breaker can close correctly. A package description and a breaker-specific record answer different questions.
Separate synchronizer, sync-check and load sharing
An active synchronizer can bring a generator toward the bus conditions. It may send speed commands to the engine governor and voltage commands to the automatic voltage regulator, or AVR. The AVR controls the alternator field, also called excitation. Check the selected design to see which adjustments the synchronizer makes and whether it also issues the close command.
A sync-check function evaluates whether its measured conditions satisfy its configured permissive logic. Permission and active matching are different functions. A device may combine them, so two function names in a brochure do not prove independent hardware, separate sensing or an independent close-inhibit path.
Ask the supplier to mark those functions on the control drawing. Identify which device changes speed or voltage, which provides permission and where the final command passes. Where independence is a project requirement, request evidence of the actual architecture rather than inferring it from the product name.
The functional distinction is explained in Cummins T-016, Paralleling Application Manual, particularly its synchronizing discussion. Use the manual to understand roles, then obtain the selected equipment documentation and approved project logic. It is not a universal setting schedule.
After closure, real-power and reactive-power sharing become separate control tasks. Successful synchronization does not prove stable kW or kVAr sharing, equal loading or adequate redundancy. Keep those acceptance records separate so that a successful close is not reported as proof of the entire parallel system.
Verify electrical matching and the measurements behind it
Review phase sequence, voltage, slip and angle together
Phase sequence describes the order of the three phases. Phase angle describes the relationship between corresponding waveforms at a particular time. Matching sequence is necessary, but it does not establish the angle at which the breaker will connect the sources.
Frequency difference is often called slip. If two sources have equal frequency, their relative angle need not be zero; it can remain offset. If their frequencies differ, the relative angle moves. This is why equal nominal frequency on two displays cannot replace a synchronization assessment.
Voltage comparisons also need a defined basis. Record whether values are line-to-line or line-to-neutral, where they are measured and how they are scaled. A voltage transformer, commonly called a VT or PT, supplies a scaled measurement. Incorrect ratios or source selection can make a plausible displayed value misleading.
Use a review table with project-specific criteria rather than copying acceptance numbers from an unrelated manual:
| Condition | Evidence to request | Question the record must answer |
|---|---|---|
| Phase sequence | Approved drawing and commissioning correspondence record | Do the identified phases on both sides correspond to the intended circuit? |
| Voltage magnitude | Measurement basis, VT ratio and approved limit | Are equivalent quantities being compared at the correct source locations? |
| Frequency difference | Selected control criteria and time-based record | What slip existed during the relevant approach to closure? |
| Phase relationship | Identified measurement channels and closing-event record | What relationship applies to the corresponding source voltages? |
| Stability and supervision | Required observation period and implemented logic | Was the permissive based on suitable sustained conditions rather than one displayed instant? |
The STAMFORD/AvK AGN022 guidance on parallel operation discusses synchronization windows and the significance of unstable speed. Its equipment guidance should not be averaged with another manufacturer’s limits. Have the responsible designer reconcile applicable alternator, engine, control and project requirements.
Trace sensing from the drawing to the record
Illustrative scenario: A panel drawing labels one input “incoming” and another “running,” while a test sheet lists only channel A and channel B. The reviewer cannot yet connect either channel to a named source. Request the channel mapping and drawing revision before accepting the readings as evidence for the intended breaker.
Trace that mapping through source selection, sensing protection, VT ratios, phase identification and controller configuration. The record should identify the actual installed arrangement and distinguish it from a simulated input. This is a request for qualified commissioning evidence, not an instruction to make live measurements.
Then check whether the recording method can resolve the event being claimed. A slowly refreshed display may help explain a sequence without proving the instantaneous relation at closure. Cummins T-037, PowerCommand Control Application Manual specifically cautions against using its InPower service tool to confirm sync-check accuracy because of sampling latency.
That warning applies to the named tool. For any proposed method, ask what each channel measures and how the records share a time reference. Request the sampling or event resolution and an explanation of why it can resolve the claimed event. Keep the instrument identity and channel labels with the record so another reviewer can follow the evidence.
Review breaker closing time against the synchronization window
Distinguish command, pulse, contacts and feedback
A close command begins a chain; it is not the physical connection itself. Depending on the architecture, the chain can include controller outputs, interposing devices, the closing coil and the breaker mechanism. Define the timing start and end points before comparing any delay values.
Close pulse duration is how long the command output stays on. Breaker closing time is an interval with a defined start and end.
Does it start at the controller output, the coil supply or another point? Does it end at the main contacts or an auxiliary contact? Ask the integrator to mark those boundaries on the close-chain drawing.
A status input reports its own defined signal; its timestamp needs that same scrutiny.
For example, a controller log timestamp for “close output active” cannot automatically be paired with a later “breaker closed” message as an accurate main-contact measurement. Ask how each message is generated and timestamped. Auxiliary status and main-contact evidence need a documented relationship if one is used to support the other.
The approved timing record should identify the relevant breaker, control supply conditions, configuration and test method. It should show whether the configured delay represents the intended part of the close chain. A default parameter or an input field’s configurable range does not establish the physical delay of the supplied equipment.
Use a timing illustration without prescribing settings
At constant slip, relative phase movement during a delay follows:
Phase movement in degrees = 360 × slip in hertz × delay in seconds.
For a hypothetical slip of 0.1 Hz and a hypothetical delay of 0.10 seconds, the phase movement is 3.6 degrees. Hertz means cycles per second, so the units cancel to degrees. These values illustrate timing sensitivity; they are not recommended settings, measured package performance or an acceptable closing angle.
The timing principle appears in Michael J. Thompson’s SEL paper on synchronizing systems, including its advance-angle equation. Predictive synchronizing can account for closing delay, while check functions and their supervision depend on the implemented device logic. Do not assume every sync-check relay predicts the future contact instant.
The example also assumes constant slip. If speed is oscillating, the angle movement over the interval may differ from a simple calculation using one displayed frequency value. Request evidence that the selected algorithm, supervision and test method address the actual operating conditions. The designer determines the accepted limits and required stability.
Keep the allowed window, any required time within that window and closing-time compensation distinct in the documentation. One describes acceptable conditions, another describes duration, and another addresses the delay to physical closure. A report should explain which of these the test observed and which were only settings readbacks.
Return to the opening Illustrative scenario: the missing record now has clear fields. Ask for the command event, contact event, time reference, measured interval and corresponding source conditions. Leave the result blank until evidence arrives; adding a timing field to a worksheet does not establish a successful test.
For an offered package, ask ZC Power about the timing-record scope. Identify the proposed breaker and controller, and ask which records come from the genset supplier and which require the switchgear integrator. Keep any exclusions in the project evidence list.
Check live-bus and dead-bus close permissions separately
Identify first-close selection and sensing-failure handling
Live-bus synchronization compares an incoming source with an energized bus. First closure onto a dead bus is a different decision. A conventional sequential scheme selects a source to establish the bus, then synchronizes additional sources to it. Special engineered schemes need their own documented logic and acceptance basis.
Ask how the selected design determines that the bus is dead and which device grants the first close. A low or absent sensed voltage is a measurement; its cause and validity still matter. The review should address how the scheme distinguishes its intended dead-bus condition from relevant sensing failures.
Request the first-close selection logic, breaker selection and interlocks for the defined modes. Identify what prevents competing close decisions under the approved design. Cummins T-016 discusses first-close control, but its description does not prove that the same arrangement is installed in a proposed package.
Use the cause-and-effect schedule to find the response to invalid measurements or missing permission. For a lost-sensing case, ask which input or diagnostic identifies the condition and what the approved logic then allows. Record the blocked or allowed action, indication and reviewer. This asks for evidence of the selected design; it does not assume that the package has detected every possible sensing fault.
Define denied-close and command/status responses
The sequence should say what happens when synchronization is not achieved, permission is denied or an issued command does not produce the expected breaker status. Ask who defines timeout, retry and recovery behavior, and which records demonstrate it for the actual configuration.
Keep manual and test modes in the same review. Request evidence of the permissions and interlocks that apply in each mode rather than assuming that “manual” either preserves or bypasses every automatic check. The accepted arrangement must come from the project design and selected equipment documentation.
A useful open item might read: “Evidence of the approved response to lost bus sensing is outstanding.” That wording identifies the missing proof without claiming the installation has a particular fault. Assign an owner and closure record instead of recording an unsupported pass or proposing wider synchronization limits.
This matters to the wider availability review too. N+1 capacity and shared failure paths should be considered separately from the number of available generators. A successful close alone does not establish system redundancy. Detailed communication-loss recovery, protection settings and neutral arrangements need their own review scope.
Request a traceable synchronization evidence package
Link drawings, settings and evidence owners
Build an evidence ledger that lets another reviewer follow each requirement to a named record. A practical entry contains the requirement ID, operating mode, source and breaker IDs, drawing revision, approved criterion and its owner. Add the method, actual or simulated inputs, result, record location and open finding.
Identify who supplies the record and who accepts closure. Those may be different parties. The controller supplier can explain the implemented function; the switchgear integrator can identify the close chain; the project designer owns the approved criteria within the assigned design scope. Confirm those assignments in the contract rather than assuming them.
| Evidence item | Required identity or boundary | Open action if evidence is absent |
|---|---|---|
| Sensing correspondence | Sources, phases, VT ratios and drawing revision | Obtain the qualified commissioning record |
| Control configuration | Device model, firmware and settings export revision | Reconcile the supplied and tested versions |
| Closing timing | Start event, contact event and test method | Obtain the applicable timing record |
| Mode permissions | Live/dead bus, selected breaker and interlocks | Confirm logic and accepted test scope |
| Acceptance closure | Finding, corrective action, retest and approving owner | Keep the item open until authorized closure |
This is a request structure, not a completed test report. Record “awaiting evidence” where appropriate. If the configuration changes, have the responsible test owner identify which evidence remains applicable and which checks require repetition.
Separate factory evidence and installed checks
Illustrative scenario: A factory acceptance test (FAT) uses simulated bus-voltage inputs. Its record can describe the control response to those inputs, but it does not establish the installed VT wiring or the behavior of a breaker absent from the test. Keep those interfaces listed for later verification rather than extending the factory result to them.
A factory acceptance dossier should identify the tested equipment and exclusions. Ask which devices were present, what was simulated and whether actual contact timing was recorded. Have the qualified test owner state which startup conditions were covered, including cold-start behavior where relevant to the approved scope.
If a breaker, sensing circuit or settings revision changes later, ask that owner to identify the affected records and required retests. Keep the decision with the changed revision.
Retain the report, configuration export and supporting traces together. Avoid a summary that says only “synchronization passed.” A useful handoff states what was demonstrated, the criteria applied, the open interfaces and the person authorized to close each action. That supports a later acceptance decision without inventing one now.
Frequently asked questions and project handoff
What conditions must be checked before generators are synchronized?
For connection to an energized bus, review phase sequence, voltage magnitude, frequency difference and phase relationship against the approved criteria. Also establish that the sensing represents the intended sources, the required permissions apply and breaker timing is addressed. The selected equipment and project design determine the accepted limits and evidence.
Does matching frequency mean the sources are in phase?
No. Equal frequency can leave a fixed phase offset between corresponding voltages. Different frequencies cause that relative angle to move. A matching frequency reading therefore does not establish phase alignment at main-contact closure. Review the angle evidence and closing timing as well as frequency, using the approved measurement method.
What is the difference between a synchronizer and sync-check?
A synchronizer actively brings source conditions toward the required relationship and may initiate closing, depending on its design. A sync-check function supplies permission according to its measurement and logic criteria. Both can exist in one device, so their names alone do not establish independent sensing, hardware or close-inhibit paths.
Why does breaker closing time matter?
The sources can move in relative phase between a command and main-contact closure when slip exists. Closing-time compensation and supervision must therefore be understood for the selected architecture. Ask how the delay was defined and verified; a command pulse duration or a status message alone does not establish contact timing.
Does a FAT report replace site synchronization checks?
A factory acceptance test establishes evidence only for its recorded configuration, inputs and scope. Simulated sensing or absent site switchgear leaves interfaces unproven by that test. Use its exclusions to identify the remaining installed checks, their owners and acceptance criteria. A factory pass does not automatically establish site acceptance.
Prepare the evidence request
Bring the review back to one breaker, one mode and identifiable records. Keep open actions visible, with named owners and a defined acceptance authority. The factory-simulation example remains incomplete until the outstanding installed checks have their own evidence; the worksheet does not turn those gaps into results.
Use the parent generator paralleling specification for the wider package boundary. Keep post-close sharing, load management and protection-setting work in their appropriate scopes. This review supplies a focused question: what supports the intended connection at the actual contact event?
To discuss a proposed supply, send ZC Power your project questions with the one-line, source and controller identities, intended operating mode and open evidence list. Ask for the included documentation and exclusions so the project team can assign the remaining review work.
