The harmonic was the operator’s.

§1Facts of the dispute

Northwestern Paper’s finishing line — a 1991 Beloit-built corrugating operation extensively retrofitted in 2017 — runs eight 250 HP variable-frequency-drive-driven motors on the finishing rolls, two 100 HP VFDs on the slitter-rewinder, and a 600 HP VFD on the main drive. All eleven drives are ABB ACS880 series installed during the 2017 retrofit. The drives ship from ABB with integral DC-link chokes (a base-level harmonic mitigation feature) but without active front-end converters or line-side LCL filters — these are options the customer can specify but did not in 2017.

From late 2024 through early 2026 the finishing line experienced eleven separate drive-trip events on assorted drives, with downtime costs averaging $28,000 per event. The plant’s electrical maintenance group attributed the trips to “dirty utility supply” — a phrase familiar to anyone who has spent time in an industrial power-quality investigation — and the plant’s general counsel approached the utility with a request for compensation. The utility’s power-quality engineer (Karen Voltawicz of FirstEnergy’s Erie regional office, named in the sealed report) responded with a one-week SCADA log review that showed no events on the supplying 23 kV feeder coincident with the customer’s drive-trip log. The customer engaged Presque Isle PFE to perform an independent measurement.

The PCC was agreed before measurement began: at the customer’s primary metering point on the 23 kV side of the customer-owned 23 kV / 4.16 kV step-down transformer. The customer’s one-line drawing — dated 2017 — was confirmed against the as-installed configuration on 14 March 2026; no significant changes.

§2Measurement campaign

Three instruments were deployed: a Hioki PQ3198 at the agreed PCC (IEC 61000-4-30 Class A — direct connection to the customer’s 23 kV PT and CT secondaries through dedicated test taps); a Dranetz PowerVisa PX5 at the 4.16 kV bus directly downstream of the customer’s service transformer; and a Fluke 1773 at the 480 V finishing-line MCC bus. The measurement window ran for fourteen consecutive days from 17 March through 30 March 2026, capturing a full operational cycle including two finishing-line shutdowns and one drive-trip event.

Data was aggregated to 10-minute and 3-second windows per IEEE 519-2022 §5.6. The weekly 95th-percentile and 99th-percentile values were calculated for voltage THD, individual voltage harmonics h2 through h50, current TDD, and individual current harmonics h2 through h50.

The 23 kV PCC measurement showed voltage THD at the 95th-percentile of 4.6% (within the 5.0% IEEE 519-2022 §6 limit for systems >1 kV ≤ 69 kV). Individual voltage harmonics — h5 at 2.6%, h7 at 1.9%, h11 at 0.9% — were all within the §6 individual-harmonic limits of 3.0%. The utility’s supply at the PCC was compliant.

The current TDD at the same PCC told a different story.

§3Finding the source

The combination of voltage THD inside §6 limits but current TDD substantially over §10 limits is the classical fingerprint of customer-side harmonic injection on a supply with adequate stiffness to absorb most of it as current rather than letting it manifest as voltage. The customer was the source of the harmonic distortion they were complaining about.

The fingerprint was confirmed by the downstream 4.16 kV measurement, which showed the same individual harmonic spectrum amplified — h5 voltage at 4.1%, h7 at 3.1%, h11 at 1.7%. The amplification corresponds to the impedance ratio between the PCC and the 4.16 kV bus (essentially the customer’s step-down transformer impedance), and the consistency of the harmonic spectrum across both measurement points fixed the source upstream of the 4.16 kV measurement but downstream of the 23 kV measurement — that is, on the 4.16 kV bus and below, which is to say the customer’s drives.

The drive trips were then traced to the ACS880 drives’ integral input filter overheating thermistors. ABB ACS880 series drives ship with overtemperature protection on the input DC-link choke; when the harmonic spectrum on the supply side of the drive includes substantial 5th, 7th, 11th, and 13th harmonics that the choke must absorb, the choke runs hot. The drives’ trip log codes — recovered from the drive’s onboard memory via ABB’s Drive Composer Pro — read “F2310 — Input filter overtemperature” for nine of the eleven events, not the “F3210 — Supply voltage” or “F3220 — Supply phase imbalance” codes that a utility-side problem would generate.

§4Finding

The voltage and current distortion at Northwestern Paper’s 23 kV point of common coupling over the 14-day measurement window of March 2026 is consistent with the customer’s own harmonic injection from the eleven ABB ACS880 VFDs on the finishing line, without active front-end or line-side LCL filtering. Voltage distortion at the PCC is within IEEE 519-2022 §6 limits; the customer’s current TDD at the PCC is approximately 18%, against an applicable §10 limit of 12% for the measured I_sc/I_L ratio of 50. The eleven drive-trip events recorded between November 2024 and February 2026 are consistent in fault-code pattern with input-filter overtemperature trips caused by the harmonic spectrum the customer’s own drives inject into their own supply bus. Sealed PE report · 8 May 2026 · Margaret W. Trzaska, PhD, PE

The finding ended Northwestern Paper’s claim against FirstEnergy / Penelec. The customer’s engineering staff opened a separate retrofit project to install passive line-side LCL filters at the 4.16 kV bus; the firm declined to participate in the design (conflict with the matter), and the design was let to a separate consulting engineer in Cleveland. Northwestern Paper’s subsequent IEEE 519 compliance, after the filter installation in October 2026, will be filed under CF-2026-094.

§5Aftermath

This Case File became the subject of an evening lecture by Dr. Trzaska at the Penn State Behrend School of Engineering on 15 April 2026, attended by approximately forty undergraduate and graduate students plus seven practitioners from regional industrial firms. The lecture title — “The Harmonic Was the Operator’s” — borrowed from a 1995 IEEE Industry Applications Society paper of the same construction; the connection was acknowledged.

The Northwestern Paper finding has, in the firm’s experience, a generalizable lesson: when a customer’s harmonic complaint arrives and the voltage THD looks suspiciously close to but below limit while the same customer is running heavy unfiltered VFD load, the harmonic is almost certainly the customer’s. The utility, in those circumstances, is doing its job — its supply impedance is providing the buffer that prevents the customer’s current distortion from manifesting as bad voltage at the PCC. The customer’s drives are tripping on what the drives are themselves generating.

§6References