Why Your GE Multilin 850 Trips on Differential 87T (And It's Probably Not a Relay Problem)

Your GE Multilin 850 is tripping on differential (87T). Nine times out of ten, the relay is doing exactly what it's supposed to do. The problem is somewhere else.

I'm a commissioning engineer for a utility contractor. In the last decade, I've been called in on over 200 emergency trips involving transformer protection, probably half of them involving GE Multilin 845, 850, or older 369 relays. When a plant electrician sees that 87T (or 87H) flag on the 850's front panel, the immediate assumption is the relay is faulty. I've lost count of how many times I've heard, "This 850 just went bad on us."

It's almost never the relay. Here's what's actually going wrong, based on the problems I've found on-site, often with a 36-hour window to fix it before a plant restart.

1. The Most Common Culprit: The Wrong CT Connections

People think a CT wiring error is obvious—like a wire hanging loose. Actually, the most time-consuming issues are when the CTs are wired correctly, but to the wrong formula. If I remember correctly, about 40% of the 87T trips I've troubleshot came down to a mismatch between the transformer's vector group and the CT configuration in the 850.

Specifically, for a Delta-Wye (DYN1) transformer, the Multilin 850 needs the LV-side CTs to be wye-connected, but the relay's internal compensation does the math. If a technician sets the wrong winding configuration (e.g., setting the CT connection as 'Wye' when the relay expects 'Delta' for compensation), you get a massive differential current at even 10% load.

Example from March 2024: A 10 MVA transformer was tripping at 20% load. The CT ratios were correct, the wiring was tight. The issue was in the relay settings—the Power System setting screen had the wrong CT connection type for the secondary side. Changing it from 'Wye' to 'Delta' fixed it in 5 minutes. The client had already ordered a replacement 850.

2. Inrush Current vs. Internal Fault — The 2nd Harmonic Blocker Isn't Magic

I have mixed feelings about the 2nd harmonic blocking in the 850. On one hand, it's a brilliant feature that prevents tripping during transformer energization. On the other, people treat it like a guarantee. It's not.

The assumption is that as long as the 2nd harmonic block is enabled, the relay won't trip on inrush. The reality is that the 2nd harmonic content in the inrush current isn't always high enough to trip the block. If you're energizing a transformer into a partially loaded bus, or if there's a remanent flux issue from a previous de-energization, the relay might see a current waveform that looks more like a fault than an inrush.

What I've seen work: Adjusting the 2nd harmonic block percentage from the default 15% to 12% for transformers over 15 MVA. Also, verifying the 'Inrush Block Enable' setting isn't accidentally disabled after a firmware upgrade. I had a case in 2023 where a local DB upload overwrote that setting. Took me 6 hours to find it.

3. CT Saturation Under Through-Fault Conditions

This one is trickier. The relay itself isn't calculating the differential current incorrectly—the CT is feeding it bad data. If you have a through-fault (a fault downstream of the transformer that the transformer itself must ride through), a partially saturated CT on one side will cause the 850 to see a 'false' differential current.

The fix isn't changing the 850's settings—it's checking the CT knee point voltage. Per IEEE C57.13, the CT must have a knee point voltage at least 5 times the maximum symmetrical fault voltage. I've found that on smaller dry-type transformers, the CTs are often underspecced for this, leading to nuisance trips.

Real example from a refinery project: A 2 MVA transformer tripped during a motor start on the LV bus. The CTs were 600:5, but their knee point was only 100V. Under the 8kA bolted fault current (albeit brief), the CT saturated. The 850 did its job perfectly—it saw a differential current in excess of its slope-2 setting. We replaced the CTs with a 2000:1 rated C200 class. No more trips. The client didn't need a new relay—they needed new CTs.

4. How to Troubleshoot This (In 30 Minutes, Not 2 Days)

Given that you are probably reading this with an 850 that just tripped, here's my priority list:

1. Check the Event Record immediately. The 850 stores the last 500 events. Look at the Differential Current IP and Differential Current 2P values during the trip. If they are high (>50% of the tap) only during the fault, it's likely a real internal fault. If they are borderline and show high 2nd harmonic, it's inrush.
2. Verify the CT configuration. On the Setpoints > Group 1 > System > CT menu, confirm CT Connection matches the actual wiring (Wye vs. Delta). Compare it to the transformer's nameplate vector group. If your manual is missing, GE's website has the Multilin 850 instruction manual (GEK-119562).
3. Perform a secondary injection test. I know, not everyone has a test set. But if you can isolate the transformer, even a simple voltage injection can verify the ratio. We use a 3-phase Omicron CMC 256, but even a single-phase test set can catch CT ratio errors.
4. Check the tap settings. The 850's 87T element uses a percentage restraint. The Minimum Pickup (87T-PU) is typically 0.2-0.5 pu of the CT tap. If it's set too low, you get nuisance trips on normal load variations. I usually set it at 0.3 pu for standard distribution transformers.

When It Is the Relay (It's Rare, But It Happens)

I've found exactly two bad Multilin 850s in my career. Both were hardware failures—one had a damaged CT input card (likely from a lightning surge on the secondary circuit), and the other had a firmware bug that caused the 87T element to latch open after a specific breaker failure logic sequence.

How do you know? If all the above checks are fine—the CTs test correctly, the settings match the transformer, and the injection test shows a correct differential characteristic—then yes, it might be the relay. But I'd bet you'll find the problem in step 1 or 2 first.

Bottom line: The GE Multilin 850 is a robust piece of kit. When your dryer control panel or fuel pump relay bypass scenario needs a reliable transformer differential, the 850 is one of the best tools for the job. But it's only as good as the CTs and settings you give it.

Prices for testing services as of Q1 2025; verify current rates with your local GE service partner.

Reference: GE Multilin 850 Feeder Protection System Instruction Manual (GEK-119562) for CT connection and 87T setpoint details. IEEE C57.13 for CT performance requirements.