Herzig Engineering

10/31/2026

"We scan annually" is a schedule, not a program.

NFPA 70B ties maintenance intervals to equipment condition and criticality rather than to the calendar. A switchboard feeding a process that cannot be interrupted, in a dusty environment, at 80% load, does not belong on the same interval as a lightly loaded panel in a clean office.

The useful version of the question is which assets would hurt most if they failed unannounced, and what the last few scans told you about them. A breaker that has shown a rising delta across three consecutive scans is telling you something an annual cadence will not act on for another eleven months.

Frequency should follow the equipment. Not the fiscal year.

Learn more: https://vist.ly/5qtjd

10/30/2026

The 2027 NFPA 70E is already in effect, and some of the changes will reach your permit forms, training records, and audit calendar.

Our new blog goes through it section by section: the additional person rule, task-specific qualification, documented training, hazards below 50 volts, the split lockout/tagout audit, and the new Chapter 3 rules for batteries, solar, and dc systems. It also covers what didn’t change, and a few “changes” you may have read about that never made it into the final standard.

Read it here: https://vist.ly/5qqgr

The 2027 edition of NFPA 70E, Standard for Electrical Safety in the Workplace, became effective on May 6, 2026, replacing the 2024 edition. If you want the short version, our NFPA 70E 2027 changes page covers the handful of updates most facilities will feel first. This post goes further. It walks th...

10/30/2026

The electrical room is not storage. It keeps becoming storage anyway.

NEC 110.26 requires clear working space in front of electrical equipment: a defined depth depending on voltage and what is across from it, a minimum width, and headroom. Not "mostly clear." Not "clear once you move the pallet." Dedicated, maintained, and not used for anything else.

What actually accumulates there is rarely dramatic. Spare parts that needed a home. A ladder leaning where nobody walks. Boxes staged for a project that ended in spring. Each item arrived reasonably and the space degraded gradually, which is why nobody can point to the day it stopped complying.

The reason the clearance exists is movement under pressure. Someone needs to open a panel wearing arc-rated gear that restricts their field of view. Someone needs to get to a disconnect quickly. Someone needs to step back fast. All three assume floor space that is there every time, not space that was there last month.

It is also among the easiest findings to close. An afternoon and a floor marking.

Electrical safety training: https://vist.ly/5qnj3

10/29/2026

A contractor inherits your electrical system and none of your institutional memory.

They do not know the panel in the north bay is fed from two directions. They have not heard the story about the breaker that did not open. They are working from your labels, your drawings, and whatever the site contact said in the first ten minutes.

That makes the study and the labels matter more for contractor work, not less. Your own crew carries context that partly covers for a missing label. A contractor has only what is written on the gear.

One thing that never transfers: whoever puts hands in tests for absence of voltage themselves, at the point of work, no matter who told them it was dead.

Arc flash studies: https://vist.ly/5qje5

10/29/2026

Most lockout procedures are written for one machine at a time. Electrical systems are not built that way.

A single motor has an obvious disconnect. A line, a substation section, or a piece of gear fed from two directions does not. Tie breakers, alternate feeds, standby generation, and transfer switches all mean a circuit can be dead from one source and live from another, and the procedure that names only the local disconnect will happily walk someone into that.

Getting this right starts upstream of the lock. It needs a single-line diagram that matches the building, every isolation point identified, and the boundaries of what a given procedure actually covers written down rather than assumed.

The question a procedure has to survive is not "is the disconnect open." It is "what else can feed this."

Learn more: https://vist.ly/5qhhb

10/28/2026

A thermal scan taken at 20% load is not a thermal scan. It is a photograph of a system that is not working hard enough to show you anything.

Heating from a resistive fault rises with the square of the current. Halve the load and the temperature rise over ambient drops to roughly a quarter of what it would be. A termination that would run 40°C over reference at full production might sit at 10°C over at light load, which reads as unremarkable and gets no comment in the report.

That is why the scan window matters as much as the scan. Industry practice generally looks for a minimum of around 40% of rated load, and more is better. Scheduling the inspection for a quiet week, when the plant is easiest to walk, systematically removes the evidence you are paying to find.

The awkward version of this: a scan done during a shutdown can come back clean on equipment that is months from failing.

Infrared inspections: https://vist.ly/5qea6

10/28/2026

Most facilities have electrical safety practices. Fewer have an electrical safety program, and the difference shows up the moment someone asks to see it.

A program is a written document that states, in one place, how this organisation controls electrical risk. Who is authorised for what, and on which equipment. What the de-energizing procedure is, and what has to be true before energized work is permitted. How the arc flash study is maintained and what triggers a review. What training each role gets and how often. How contractors are controlled. How the program itself is audited.

Practices live in people's heads and in habits. They work until the person who holds them retires, or until two crews with different habits work the same job, or until an investigator asks what the rule was and gets four answers.

Writing it down is tedious and it is also the cheapest part of the whole thing. The practices mostly already exist. The work is deciding, once, which version is the rule.

Electrical safety training: https://vist.ly/5qc69

10/27/2026

"We had a study done" is a statement about the past. It is not a statement about today.

NFPA 70E sets the review interval at a maximum of five years. In most facilities that interval is doing far more work than it was meant to, because it was never intended as the only trigger. The other trigger is change, and change is continuous.

A breaker setting adjusted during a nuisance-trip investigation. A fuse replaced with the closest thing on the shelf. A new transformer with different impedance. A utility upgrade that raises available fault current at your service. Any one of those can move incident energy at a piece of gear, and none of them generate paperwork that reaches the person who holds the study.

The practical fix is not a shorter calendar. It is a rule that says electrical changes get flagged for study review the same way they get flagged for maintenance records, so the five-year interval becomes a backstop rather than the whole program.

Arc flash studies: https://vist.ly/5q8kk

10/27/2026

The single-line diagram on your electrical room wall is a historical document.

Every plant has one. Most were drawn at commissioning and have been accurate for shrinking portions of the system ever since. A transformer gets swapped for a larger one during an expansion. A tie breaker is added. A feeder is abandoned in place but never erased. A panel is fed from somewhere new after a renovation, and the change lives in one electrician's memory.

This is not a records problem. It is a safety-calculation problem. An arc flash study models the system the drawing describes. If the drawing is five years behind the steel, the incident energy numbers on your labels describe a plant that no longer exists, and the direction of the error is not predictable: it can be optimistic just as easily as conservative.

Field verification is the unglamorous half of a study and the half that determines whether the numbers mean anything. Before the next one, it is worth asking when the one-line was last walked rather than last printed.

Arc flash studies: https://vist.ly/5q6jz

10/26/2026

Arc-rated PPE that is worn wrong is not arc-rated PPE. The rating assumes the garment is closed.

Sleeves pushed up above the elbow. A coverall unbuttoned at the chest because the room is hot. A hood left open for better visibility during a fiddly step. Gloves removed for the last connection because the fingers are clumsy. Every one of those is a reasonable response to real discomfort, and every one of them opens the exact path an arc flash takes.

This is almost never a knowledge problem. People know the suit should be closed. It comes apart under the small frictions: gear that does not fit, a task that takes longer than the comfort window, a job where nobody wants to be the one to stop and re-dress.

Which means the fix is not another slide about PPE categories. It is properly fitted gear, hands-on practice in it before it matters, and a crew culture where saying "close your hood" is routine rather than a reprimand.

Electrical safety training: https://vist.ly/5q38h