Runtime authority model

Separate clocks and equipment authorities keep the virtual loop causal.

Beta.6 distinguishes TESTSET metrology, relay acquisition/protection, virtual contacts/wiring, WPF presentation, process-bus trust, transformer runtime, and AVR/MMS virtual-control authority.

Closed-loop feeder path

The external virtual-I/O chain owns TESTSET timing.

01TESTSET T0 + instantaneous source
02Virtual analog wiring
03Relay ADC/front-end model
04Causal measurement + protection
05Relay BO contact behavior
06Virtual binary wiring
07Independent TESTSET BI acceptance
Authority invariant: internal TripLatched can request BO1 but cannot become a TESTSET result directly. Accepted wired BI1 owns external trip timing and optional auto-stop.

Clock domains

Metrology, relay processing, and rendering are different clocks.

Domainbeta.6 reference behaviorAuthority
TESTSET metrologyMonotonic integer µs, 1 µs resolutionT0 and timing arithmetic
TESTSET BI10 kHz / 100 µs, 0.5 ms deglitch, 0 ms debounce holdoffAccepted external pickup/trip edges
Relay acquisition4 kHz / 250 µsTerminal sampling and protection progression
WPFSlower presentation cadenceDisplay only

Relay front end

Protection sees causal terminal samples, not source-side phasors.

Signed terminal samples

Virtual wiring produces instantaneous current/voltage samples at the relay boundary.

Input model

Signed clipping, 16-bit-equivalent quantization, 20 A/300 V RMS behavioral full scale, and 1.5 ms input/filter delay.

Rolling measurement

One nominal 50 Hz causal DFT begins from settled pre-fault history rather than an artificial empty window.

Timing semantics

Generic pickup is not the same as operated-element pickup.

RELAY ANY PU / BI2

BO2 is generic ANY PICKUP. Another enabled element can assert it before the element that later operates.

Operated-element P→T

Element timing is correlated to the element in the authoritative operation record. A representable 60 ms 50P definite delay is exactly 60.000 ms in the reference engine.

Relay trip / BI1

The live trip request drives BO1; BI1 acceptance is the external TESTSET result.

Reset and frozen evidence

Reset settles modeled feedback without erasing the completed run.

After BI1 auto-stop the source is OFF while the trip capture stays frozen. One reset transaction advances acquisition until stale pickup releases, clears the relay once, then waits for no pickup, BO1/BO2 LOW, and BI1/BI2 LOW before showing READY TO RE-ARM. The source is not restarted and completed timing evidence is retained.

Other product paths

Process bus, Transformer Differential, and AVR/MMS use separate shared runtimes.

Process bus

Live/Replay Ethernet frames flow through ARIEC61850 decode, identity, mapping/scaling, continuity/quality, complete windows, trust, and feeder/transformer runtime.

Transformer Differential

Internal synchronized HV/LV/NGR injection or paired external SV feeds the same transformer protection runtime for 87T, 87T-HS, and REF evidence.

AVR / OLTC + MMS

The simulated plant and controller expose browse/read, DataSets, reports, and modeled controls. Accepted commands terminate inside the virtual process.

Architecture boundary

Behavioral models do not imply hardware equivalence.

Exact commercial relay anti-alias filters, ADC topology, burdens, channel skew, electrical BI thresholds, statistical contact behavior, and calibrated timing uncertainty are not claimed by beta.6.