When the Relay Misses the Point: A Field Story About Transformer Protection, Time Pressure, and What GE Multilin Taught Us

It Started with a Phone Call at 4 PM on a Thursday

I'm a field service coordinator for a midsize electrical equipment supplier specialized in transformer systems and substation gear. I've handled somewhere around 350 emergency orders over the last seven years—give or take, depending on how you count. Most of them are standard rush jobs: a blown distribution transformer, a damaged bushing, a failed cooling fan.

But this one was different.

A utility client called in a panic. One of their main step-down transformers at a critical substation had tripped offline unexpectedly. Not a fault—the transformer itself was fine. The issue was the protection relay. An aging electromechanical unit had misoperated. They needed a replacement fast. Normal lead time for a modern numerical relay with proper setup and testing was at least three weeks. They had a deadline: nine days. Maybe eight if the weather held.

That's when I got tapped for the job. My role is to triage these situations—figure out what's possible, what's not, and where the landmines are buried.

The Obvious Choice and the Uncomfortable Question

Our go-to for this kind of application is the GE Multilin 850 transformer differential relay. It's reliable, well-documented, and the integration with our existing monitoring setup is fairly straightforward. The 850 specifically handles two-winding transformer differential protection, which is exactly what this site needed. We'd installed maybe 40 of them over the past three years. I knew the specs by heart.

But here's where it got complicated.

The client's existing system used a different manufacturer's relay—I won't name which, but they're a major competitor. The wiring, the CT ratios, the control power... none of it was set up for the Multilin's input requirements. We were looking at a full re-termination of the CT circuits, a new control power transformer, and reprogramming the entire protection scheme from scratch.

I knew we should have done a full site survey before ordering anything. But time was the enemy. The client's operations team was breathing down our neck. The penalty clause in their contract with the local grid operator was substantial—I heard figures around $50,000 per day if the substation remained offline past the deadline.

So I made a call. I told the team to order the GE Multilin 850 and a matching control power transformer, based on the documentation the client emailed over. "What are the odds the drawings are wrong?" I thought. Well, that was my first mistake.

The Moment Everything Almost Fell Apart

The relay arrived in four days—GE's distribution network is pretty good for standard products. We put two technicians on the configuration and wiring prep. Everything was going smoothly. We were actually ahead of schedule.

Then came the site visit.

We arrived on a Tuesday morning. The client's engineer walked us to the panel. I opened the CT shorting terminal block and immediately felt my stomach drop. The CT wiring didn't match the drawings. Not even close. The ratio was different, the polarity markings were reversed on two phases, and the secondary wiring was a gauge thinner than spec'd.

We were standing there, tools in hand, and the clock was ticking. The client's engineer was visibly nervous. I had two technicians looking at me for a decision.

I had a choice. I could have pushed forward with a workaround—re-terminate the CTs on the fly, recalibrate the Multilin's settings based on field measurements, and hope the differential protection functioned correctly. That would have saved maybe two days. But I've learned the hard way that skipping the safety step on a transformer differential relay is a gamble with expensive consequences. I've seen a misconfigured relay fail to detect an internal fault. The result was a transformer fire and a six-month outage. Not on my watch.

So I called a timeout. We spent the next six hours re-wiring the CT circuits correctly, sourcing proper gauge wire from a local supplier, and re-certifying the ratios. It cost us an extra $1,200 in labor and materials. But we knew the GE Multilin 850 would see the current correctly.

The Payoff—and the Lesson

We finished the installation three days later. The relay was commissioned and tested. The differential protection operated within spec. The client's substation was re-energized with less than 24 hours to spare before the penalty clause kicked in.

There's something deeply satisfying about that moment—the final check, the green indicator lights, the quiet hum of a properly protected transformer. After all the stress, seeing it work correctly is the payoff.

But I have mixed feelings about the whole experience. On one hand, we delivered under extreme pressure. On the other hand, I made the initial decision to skip a proper site survey, and that directly created the near-miss. If I hadn't caught the CT wiring mismatch on site, we could have installed a relay that was effectively blind to a differential condition.

The GE Multilin 850 is an excellent relay—the differential element is robust, the communications are flexible, and the event logging is detailed enough to reconstruct events after a fault. But a good relay doesn't fix bad wiring. The quality of the installation directly determines the quality of the protection.

What I'd Do Differently (and What I'd Recommend)

Here's a short list of things I learned from this job, and from dozens of other emergency transformer protection upgrades:

• Never trust remote documentation. If possible, send someone to take photos of the existing panel. The drawings are often outdated.
• Verify CT polarity and ratio before the relay arrives. A quick field check can save you a day of rework.
• Rush orders on the relay are worth it. GE's lead times for the Multilin 850 are generally reliable, but confirm the exact model and firmware version before ordering. We once received a relay with different firmware than expected—cost us half a day to update.
• Build a 48-hour buffer. Our company now requires a minimum 48-hour contingency on all emergency substation projects. That policy came directly from this job.
• Test the differential function before energizing. The Multilin 850 has a built-in test mode that simulates through-fault conditions. Use it. Every time.

Looking back, the scariest part wasn't the technical difficulty. It was the moment of overconfidence. I knew I should have done a site survey before ordering, but I thought, "What are the odds the drawings are wrong?" Well, the odds caught up with me. The $1,200 we spent on re-wiring was cheap compared to the alternative—a misconfigured differential relay that could have failed to trip on a real fault.

As of early 2025, the GE Multilin 850 remains our preferred solution for transformer differential applications in this voltage class. The protection scheme flexibility and the quality of the internal diagnostics make it a solid choice for both new installations and retrofits. But no relay, no matter how good, can substitute for proper field verification and skilled installation.

The best part of finally systematizing our emergency process after this job: no more 3 AM worry sessions about whether the relay will operate correctly. The certainty of knowing the installation is right is worth more than any rush fee.