Show Notes
Why power quality problems get misdiagnosed
This episode opens with a familiar operations story: servers keep rebooting during the busiest hour, every clue seems to point to software, and the real breakthrough comes only when someone notices the hallway lights dipping whenever a rooftop unit cycles. That framing sets up the core lesson of the conversation: many failures that look like IT issues are actually power quality issues.
Alex Morgan and Michael Harrington define “dirty power” in practical terms. They are not talking about dramatic blackouts. They are talking about the quieter electrical problems that rarely get highlighted on an incident board but still wear down equipment, create intermittent faults, and trigger outages that feel random to tenants and operators. The examples they call out include:
- Voltage sags
- Spikes and transients
- Harmonics
- Grounding and bonding oddities
The operational challenge is that these issues often show up as symptoms far away from the electrical source. A half-second sag may be enough to upset a server power supply, reset an access controller, or trigger an alarm condition without leaving behind a simple, obvious cause. Tenants may notice application issues first. Facilities teams may notice lighting flicker first. The answer usually requires both viewpoints.
The red flags facilities and IT teams should watch for
One of the most useful parts of the episode is the translation from electrical theory into field symptoms. Michael lays out the signs that should make a facilities leader pause and bring in IT or an electrical contractor:
- Intermittent equipment reboots
- Error logs even though nothing recently changed in software or firmware
- Nuisance breaker trips
- Transient alarm activations
- Lights dimming or flickering briefly
- Problems that correlate with time of day or mechanical cycles
That last point matters. Patterns are often the clue that separates “mystery outage” from a solvable power-quality problem. If an event repeatedly lines up with a motor start, a lab device cycle, or evening utility behavior, the troubleshooting path changes immediately.
Common root causes in real buildings
Rather than presenting one silver-bullet explanation, the episode stresses that these failures usually come from a mix of conditions. The common suspects discussed include shared tenant loads, large motors or welders cycling on and off, aging UPS systems that behave strangely under partial loads, poor grounding and bonding, and even utility or transformer interactions involving resonance and voltage regulation.
That is an important operational takeaway for commercial properties and campus environments. There is often no single villain. Multiple building systems may contribute to the same symptom set, which is why replacing one piece of equipment without data can waste time and budget.
How to prioritize fixes instead of overreacting
A major theme in the conversation is sequencing. Michael recommends starting with impact, not hardware. Before buying new equipment, teams should ask which loads are truly critical, which failures are acceptable for a short period, and which disruptions create immediate business or safety consequences.
The episode lays out a practical order of operations:
- Monitoring first, so you have timestamps and evidence
- Conditioning second, where targeted fixes can solve specific trouble spots
- UPS strategy next, with real-world testing instead of nameplate assumptions
- Larger capital solutions like generators or microgrids only when the business case supports them
This framework keeps the response proportional. It also prevents teams from jumping straight to expensive infrastructure when a smaller, more focused fix would address the real problem.
Why a bigger UPS is not automatically the answer
One of the most useful clarifications in the episode is the warning that bigger is not always better when it comes to UPS planning. Oversizing can hide battery health issues, and low-load operation can create transfer behavior that still causes momentary blips when the system switches sources.
The recommendation is clear: do not judge a UPS strategy by capacity alone. Test battery runtime and transfer performance under real load. Watch for the micro-interruptions that do not show up well in paperwork but absolutely show up in production. In other words, validate field behavior, not just specifications.
Three short case examples that bring it home
The episode includes three concise examples that show how misleading symptoms can be:
- A clinic experienced intermittent electronic health record freezes. The root cause was a lab centrifuge on the same panel creating sags. The fix was a dedicated conditioned circuit for the clinic plus basic power logging.
- A campus faced repeated false fire alarms at odd hours. The root cause was harmonic distortion from VFDs in the mechanical room. The fix was harmonic filters and load rebalancing.
- An office tower saw nightly connectivity losses in network closets. The root cause was utility voltage regulation drift during peak conditions. The fix was power quality meters, staged utility coordination, and local conditioning for the most sensitive floors.
Each example reinforces the same lesson: symptoms that first appear as application, network, or alarm issues can begin at the power layer.
A quarter-ready checklist for operators
For listeners who need immediate action, the episode closes with a practical checklist:
- Enable timestamped power logging on panels serving critical loads
- Correlate those timestamps with IT and alarm logs before replacing equipment
- Inspect grounding and bonding during maintenance windows
- Identify large single-source loads on shared panels and evaluate dedicated circuits
- Review UPS age, runtime, and transfer behavior, and test under load
- Add targeted conditioning at chronic trouble spots before replacing major assets
- Only after that, evaluate generators or microgrids if long-duration autonomy is truly required
The episode also adds a practical vendor-management note: when bringing in an electrical contractor, provide correlated event logs and describe the operational impact. That changes the conversation from a vague suspicion to a defined pattern that can be tested.
The big takeaway
This conversation is a strong reminder that reliable buildings are not just about uptime during major outages. They are also about managing the smaller electrical quality problems that shorten equipment life, trigger false alarms, and create frustrating intermittent failures. The path forward is disciplined and practical: monitor first, correlate across teams, apply targeted fixes, and spend capital only where the risk truly justifies it.
Dirty power rarely looks like a power problem at first
In commercial buildings, some of the most disruptive failures do not arrive as obvious outages. They show up as application hangs, intermittent reboots, nuisance alarms, and strange network instability that seems impossible to pin down. That is what makes power quality such an important operational topic. It often hides behind symptoms that get assigned to software, networking, or aging devices long before anyone looks upstream at the electrical environment.
In this episode of Built, Wired & Secured, Alex Morgan speaks with Michael Harrington about how “dirty power” quietly drives failures in real facilities. The opening example captures the issue perfectly: servers keep rebooting during the busiest hour, the troubleshooting path focuses on the application, and only later does someone notice the hallway lights dipping whenever a rooftop unit cycles. The real clue was not in a software log. It was in the building behavior.
That is the central value of this discussion for facilities leaders, IT teams, and owners. It turns electrical nuance into operational decision-making without requiring deep design expertise.
What “dirty power” actually means in operations
The episode keeps the definition practical. Dirty power is not limited to a full blackout or a catastrophic utility event. It includes the quieter disturbances that happen every day in occupied buildings and mixed-use environments:
- Voltage sags
- Spikes and transients
- Harmonics
- Grounding and bonding irregularities
These issues matter because many devices are more sensitive than operators assume. A brief sag when a large motor starts may be enough to upset a server power supply, reset an access control panel, or trigger false alarms. Those events can be short, intermittent, and difficult to reproduce, which is why they often evade simple troubleshooting.
The result is operational confusion. Tenants may report that “the app is down.” IT may see unexplained errors. Facilities may notice a brief flicker in lighting. Each team is seeing part of the same event, but without coordination, nobody has the full picture.
The signs that should trigger a deeper look
One of the most useful parts of the conversation is its focus on symptoms. Michael does not bury the audience in theory. Instead, he outlines the field signs that should make a facilities leader or IT manager stop and ask harder questions.
Those red flags include intermittent equipment reboots, error logs when no recent software changes have been made, nuisance breaker trips, transient alarm activations, and lighting that dims for a blink. Just as important, he emphasizes patterns. If problems correlate with a known mechanical cycle, a time of day, or a recurring load event, that correlation may be more valuable than any single alarm or ticket.
That point is critical in multi-tenant and campus settings. The issue may not sit directly beside the equipment that fails. A device on a shared panel or a system elsewhere in the building may be creating enough disturbance to cause downstream operational pain.
Why root cause is often a combination, not a single failure
Another strength of the episode is its refusal to oversimplify. There is no single villain behind most power-quality incidents. Instead, Michael describes a handful of recurring contributors:
- Shared tenant loads
- Large motors or welders cycling
- Aging or poorly maintained UPS systems
- Bad grounding and bonding
- Utility or transformer interactions tied to resonance or voltage regulation
That matters because it changes how leaders should authorize repairs. If the symptom is caused by a combination of shared loads, transfer behavior, and poor grounding, replacing one device may not resolve the problem. This is exactly where teams lose budget: they swap components, assume the issue is fixed, and then see the same intermittent fault return weeks later.
The better approach is to treat power quality as an evidence problem first. Gather data, correlate events, and narrow the conditions under which failures happen. Only then should corrective work begin.
Start with impact, then sequence the response
One of the most practical frameworks in the episode is the idea that every decision should begin with a simple question: what breaks if this goes down? That question forces teams to rank loads by operational importance instead of treating every disturbance the same way.
From there, the discussion lays out a clear order for action. Monitoring comes first because it gives teams timestamps they can compare against IT logs and alarm history. Conditioning comes next because selective fixes are often faster and more cost-effective than major infrastructure changes. UPS strategy follows, but with a warning: do not assume a bigger system is automatically a better system. Finally, larger capital projects such as generators or microgrids should be reserved for cases where the business case truly demands long-duration autonomy.
This sequencing is one of the most business-useful ideas in the episode. It keeps spending aligned with impact. It also helps property teams avoid overcorrecting based on fear or incomplete diagnostics.
The UPS lesson: capacity is not the same as resilience
Many operators assume that if power quality is causing instability, a larger UPS is the obvious fix. Michael challenges that directly. Oversizing can mask battery health issues, and a UPS operating at low load may still create momentary blips during transfer events. In other words, a system that looks impressive on paper may still leave sensitive equipment exposed to exactly the kind of micro-interruption that causes field failures.
The recommendation is to test under real load, verify runtime, time the transfer, and actively watch for micro-interruptions. This is where operations teams need real behavior, not just procurement specs or installer assumptions. The episode makes the point clearly: the half-second events are the ones that hide from paperwork but break things in practice.
Short examples with long-term lessons
The episode also stands out because it uses concise, relatable examples instead of abstract theory. In a clinic, intermittent electronic health record freezes were traced back to a lab centrifuge on the same panel causing voltage sags. The fix was a dedicated conditioned circuit and basic logging. In a campus setting, repeated false fire alarms turned out to be harmonic distortion from VFDs in the mechanical room, solved through harmonic filters and load rebalancing. In an office tower, nightly drops in network closet connectivity were linked to utility voltage regulation drift, addressed with power-quality meters, staged utility work, and local conditioning on the most sensitive floors.
These examples matter because they show the cost of misclassification. A clinical application issue, a life-safety alarm issue, and a network issue each looked like a problem in its own domain. In every case, the real path to resolution started with the electrical environment.
A practical checklist teams can run now
The closing checklist is especially useful because it translates the discussion into near-term action. The priorities are straightforward:
- Enable timestamped power logging on panels serving critical loads
- Correlate power events with IT and alarm logs before replacing devices
- Inspect grounding and bonding during maintenance windows
- Identify large single-source loads on shared panels and evaluate dedicated circuits
- Review UPS age, runtime, and transfer behavior, and test under load
- Add targeted conditioning in chronic trouble areas
- Only then evaluate long-autonomy solutions such as generators or microgrids
That checklist is valuable because it supports better capital planning as well as better incident review. It helps teams separate what needs monitoring, what needs targeted remediation, and what justifies a larger resiliency investment.
Why clear handoffs matter when vendors get involved
The episode closes with one more operationally important point: when involving an electrical contractor, bring correlated event logs and clearly describe the operational impact. That changes the engagement from a vague concern to a testable pattern. It also makes it far more likely that the right scenarios will be checked during the visit.
For property teams and owners, that is the broader lesson of the conversation. Power quality does not just require technical fixes. It requires better coordination between facilities, IT, and outside vendors so the investigation starts with evidence instead of assumptions.
Listen with an operations mindset
If your building experiences intermittent reboots, unexplained alarms, or strange failures that seem to move between systems, this episode provides a strong framework for thinking about the problem. It is not a call to overspend on infrastructure. It is a call to monitor, correlate, and fix what the data actually shows. That is how teams reduce downtime, protect equipment life, and avoid expensive misdiagnosis. For operators responsible for critical systems in occupied buildings, it is a practical conversation worth hearing in full.