A custom-engineered burn-in test bench that drives ten PumpLogic controllers in parallel — exhaustively verifying every wire, every line of firmware, and every piece of hardware before a single unit ships to a customer.
The Orchestrator is a custom-engineered burn-in test bench, designed and built in-house at MicroControl Systems for one purpose: making sure every PumpLogic that leaves our shop is exactly the unit the customer paid for.
It's a Docker-containerized async test harness that connects directly to ten PumpLogic controllers over the bench network and drives every API endpoint, every relay, every sensor pathway — all ten units, all at the same time. No shortcuts. No spot-checks. No skipped units.
Because it was designed for PumpLogic specifically — not adapted from a generic test framework — it knows exactly what PumpLogic should do in every situation, and exactly how to make it prove it.
These figures are for one single PumpLogic — your PumpLogic. The Orchestrator runs all of this on up to ten devices in parallel, but every number below reflects what gets verified on the unit you'll receive.
Every PumpLogic that ships is hand-assembled. The Orchestrator hunts the three things that can go wrong on the bench, before they ever get a chance to surprise you in the sugarbush.
A miswired sensor or a swapped relay lead won't hide from the threshold-flip pattern. The Orchestrator drives every signal path until it has to respond.
Every code path that decides when your pump runs is exercised — comparators, hysteresis, schedules, blackouts, safety lockouts. Nothing ships untested.
The 24-hour soak is where weak hardware reveals itself — drifting sensors, flaky power regulation, marginal WiFi radios. We find it. You don't.
All ten devices run all twelve scenarios in parallel. Each scenario produces structured pass/fail results tied to the device's MAC address.
Export → import → re-export. Byte-equal verification of full configuration persistence.
Time-warp through 48 hours of schedule windows and blackout overlaps in seconds.
Direct relay engage / disengage commands verified through state and live status.
Threshold-flip drives lockout/recovery transitions; every comparator path exercised.
Max-runtime cutoffs and auto-disable counter validated with simulated PSI events.
Chronological ordering, no truncation, no corruption — across rapid event bursts.
24-hour battery, PSI, and temperature history endpoints validated end-to-end.
Soft reboot triggered, then settings, schedule, and live state verified intact.
180-second safety net verified — bad credentials roll back automatically.
Full reset, verify defaults, restore custom config, verify again.
Malformed JSON, out-of-range values, missing fields — all must be rejected cleanly.
Realistic settings, real ambient conditions, your unit polled every 60 seconds for a full day. Pass requires zero unexpected reboots, stable sensors, no event-log corruption, no heap leaks, and RTC drift inside tolerance. The final, unforgiving check before your unit ships.
Ten freshly-built PumpLogics power up and join the bench network. The Orchestrator scans the subnet, identifies each unit by MAC address, and registers it for the run. No manual setup. No mismatched IDs.
Operator selects which devices and which scenarios to run from the test bench dashboard. Schedules can be queued — the 24-hour soak typically kicks off Friday at 5 PM and runs straight through the weekend.
The async test runner drives every device's HTTP API at once. Time-warp replays days of schedule logic in seconds. Threshold-flip exercises every sensor decision path. Each scenario, each assertion, logged as it happens.
Live dashboard shows green checks and red X's as they land. A failure shows exactly which assertion broke, what was expected, and what was returned. When the batch is done, every passing unit gets its own digital test certificate — and only then does it leave the shop.
The Orchestrator exists for one reason: so the controller you wire up in your sugarhouse is the one we already shook down on the bench. Here's what that means in practice.
No batch-sampling. No "tested at the factory" generality. Every single PumpLogic that ships has its own pass record, tied to its own MAC address.
A bad relay or a flaky regulator is a bad day in our shop and a bad season in your sugarbush. We'd rather fail it on the bench, every time.
A full 24-hour soak weeds out the units that "work for an hour" — the most common failure mode in any embedded device. If it survives the soak, it survives the season.
Every test run is logged with timestamps, assertions, and outcomes. If you ever need the full test record for your unit, we have it on file.
Generic test frameworks check generic things. The Orchestrator was designed from scratch around PumpLogic's exact API and behavior. It catches what matters.
When you receive a PumpLogic from MicroControl Systems, it has already answered 340+ questions correctly. You can wire it up knowing it works.
When your PumpLogic leaves our shop, it's because it earned a record like this: