Advanced Manufacturing
The Hidden Cost of Power Quality in Manufacturing
Outages and power quality problems cost U.S. businesses an estimated $119 billion to $188 billion a year in 2001 (EPRI), and manufacturers carry much of it as downtime, scrap, and equipment wear. In a semiconductor fab, a voltage sag lasting a few cycles can stop a process tool and scrap the wafers inside it, at a cost of $0.5 million to $2 million per event by Texas Instruments’ 1997 estimate, and up to $3.6 million per sag in Chinese semiconductor plants, from a 2016 survey that a 2021 review summarizes. A disturbance that does not stop the tool can still create defects that surface weeks later, after the processing cost is spent.
Industry 4.0 has changed the electrical environment. Variable frequency drives, servo drives, LED lighting, and power supplies on every line distort the supply that sensitive tools share, and on-site solar and battery storage add their own disturbances. Monitoring specified for linear loads and gross violations does not see what these loads do to each other.

Every Cycle, Tied to the Process
Most power monitoring records interval averages, or captures a waveform only when a trigger fires, so the subtle disturbances that accumulate into failures go unrecorded. Energy Quotient’s Continuous Point-on-Wave (CPOW) monitoring records every cycle, so each power event can be traced to its equipment and process consequence. With that record, facilities get:
- Root-cause analysis that connects power events to tool faults and quality excursions previously labeled “unexplained”
- Early warning from the trend before a failure, the basis of predictive maintenance, which the DOE’s O&M Best Practices Guide credits with cutting maintenance costs 25–30% and downtime 35–45%
- Tool health read from each tool’s own electrical signature and tracked against its history, since the current a tool draws changes as its motors, pumps, heaters, and power supplies wear
- The stress that power disturbances put on each tool, from process tools worth a few million dollars to High-NA EUV lithography scanners at about $370 million each (Reuters, 2024)
- Process insight from power quality correlated with production metrics
- Voltage sag immunity evaluated against SEMI F47, the semiconductor industry’s ride-through standard, which requires tools to ride through sags to 50% of nominal for 200 ms, 70% for 0.5 s, and 80% for 1 s
- Energy and power quality losses found with waveform evidence, at the tool and the line
The platform connects to Manufacturing Execution Systems (MES) and existing infrastructure through its REST API, Modbus TCP, and MQTT, with OPC UA available on request.
In the Field
Energy Quotient has run in semiconductor manufacturing proofs-of-concept across multiple fabs. At JIFX 26-3, a Department of Defense field-experimentation event, we measured a U.S. Army-owned re:3D GigabotX 2 XLT 900 industrial 3D printer from a single sensor on its AC supply. Over a continuous 24.8-hour run, the platform resolved the printer’s operating modes, verified its heater switching cycle by cycle, characterized its supply, and attributed energy to each part, and the first report was delivered within 72 hours. The same workflow scales from one machine on a 120 V outlet to a 480 V three-phase production line. Read the assessment.
The same record shows when a circuit stops drawing what it drew last month, an early sign of substituted or unauthorized equipment on the plant floor (Power as a Security Sensor).
When a problem needs a dedicated investigation, EQ Forensics™ traces it to its root cause with waveform evidence.
Questions
What causes unexplained tool faults and trips in a fab or factory?
Often a sub-cycle voltage sag, a transient, or harmonic and interharmonic distortion from drives and power supplies on the same feeder. These events are too short for interval meters and often too small to trigger a power quality analyzer, so they go unrecorded. A continuous waveform record shows them, along with the equipment that caused them.
How do you check equipment against SEMI F47?
The platform measures every voltage sag continuously and evaluates its depth and duration against the SEMI F47 ride-through curve, so you can see which events a tool should have ridden through and which fell outside the curve.
Can it connect to our MES and plant systems?
Yes. Each gateway serves its data through REST and WebSocket APIs, Modbus TCP, and MQTT, with OPC UA available on request, and our open-source Python toolkit reads the full-resolution record directly.
See It Live in The Control Room
Book a session and work with live waveform data from our lab, with an engineer to answer your questions.