If you're about to sign a purchase order for transformer protection, stop and ask one question first: How will this relay be tested? That question determines whether you're buying a solution or a future emergency service call. I've spent six years managing electrical infrastructure purchases, and the most expensive line item wasn't the GE Multilin 845 or the GE dry type transformer. It was the relay that got installed without a test plan.
The cost controller's shortcut: buy protection that can be tested before it's energized, put the test requirements in the bid, and use a multimeter to check simple relays before you spend money on truck rolls and downtime. That's not a textbook rule. It's the result of my 2023 spending audit.
Why I'm the one writing this
I'm a procurement manager at a 200-person industrial services company. I've managed our electrical infrastructure parts and service budget, roughly $60,000 a year, for six years. I've negotiated with 30-plus vendors and documented every order in our cost tracking system. I'm not an engineer, and I don't pretend to be one. But I've read enough failed PO reviews and commissioning reports to know where the money disappears.
In Q2 2024, I nearly switched vendors because a relay quote was 14% cheaper. The spec looked the same. It wasn't. The lower price didn't include the configuration file, the communication module, or support for reading the settings after a trip. By the time I added the integration work and a site visit, the 'cheap' option was $1,100 more than the GE quote. That's the hidden cost I keep looking for.
Most people ask the wrong price question
The question everyone asks is, 'What's your best price?' The question they should ask is, 'What happens when this trips?' The GE Multilin 845 transformer protection functions are not a punch list for a sales brochure. They are the difference between replaying an event in a recording and standing in front of a burned transformer with no data.
Here's a misconception I see in nearly every project: cheap relays cause failures. Actually, untested relays cause failures. More precisely, the relay is often the victim, not the culprit. Of the 34 relay-related service calls I reviewed in the 2023 audit, 21 involved something else: a loose terminal, a blown fuse, a failed power supply, or a wiring error. The relay was the first thing everyone blamed, but it was rarely the root cause.
What the GE Multilin 845 actually does
If you're replacing an older relay, the 845 is not a drop-in equivalent. It is a digital transformer protection relay with a long menu of settings. The GE Multilin 845 transformer protection functions include transformer differential (87T), phase and ground overcurrent (50/51), directional overcurrent (67), voltage and frequency protection (27, 59, 81), and breaker failure supervision (50BF). It also records waveform captures and event logs, which is exactly what a procurement manager wants when a spare part demand becomes a warranty claim.
When I specified a GE dry type transformer for a facility wing that contains hospital bed control panels, the protection engineer didn't treat it as a small transformer problem. The low-voltage control panels share a power source with equipment that is sensitive to voltage dips. The 845's event recording helped us set the overcurrent elements so a control-panel fault wouldn't take out the whole transformer. Without that test data, we would have been guessing.
How to test a relay with a multimeter (and what it won't tell you)
Now the part that applies to almost any electrical problem, from a fuel pump relay fuse in a truck to a hospital bed control panel that stops moving: use a multimeter before you replace the relay. A relay that clicks is not necessarily good. A relay that doesn't click may still have a bad coil, but it might also be missing a ground or control voltage.
I run a simple test sequence before I authorize any replacement relay:
- Remove power. Check the fuse or supply voltage first. A blown fuse will make a good relay look bad.
- Find the coil terminals on the datasheet. Measure coil resistance. A typical 12V relay coil reads 50 to 120 ohms. An open reading means the coil is burned out.
- Measure the normally closed contact. It should read near zero ohms. Normally open contact should read open.
- Apply rated coil voltage and listen for the click. Then measure the normally open contact: it should now read near zero.
- If the contact resistance reads a half ohm or more under load, the contact is pitted or dirty. Replace the relay.
In my first year, I made the classic beginner mistake: I checked the coil resistance, saw 90 ohms, and called the relay good. It still failed under load because the contact mechanism was worn. Static testing with a multimeter can't guarantee dynamic operation. That lesson cost us a downtime event and a very loud phone call from our operations manager.
The fuel pump relay fuse example
Here's why I keep this simple. A fuel pump relay fuse issue is usually diagnosed by checking the fuse first. If the fuse is good, you check the relay coil and contacts. If the relay clicks but the fuel pump doesn't get power, the contact side is bad. If the relay doesn't click, the coil has no power or no ground. A $50 multimeter can narrow the fault to a $15 relay or a $150 diagnostic time. The same logic applies in an industrial control panel, except the price tags are bigger.
Old habits, new data
What was best practice in 2020 is not what I'd specify in 2025. The fundamentals haven't changed: protection should be tested before it's trusted. But the execution has transformed. A digital relay like the 845 gives you event data, self-diagnostics, and communications. An old electromechanical relay gives you a flag. That doesn't mean digital is automatically better—it means your commissioning process has to change. You don't just install the relay and close the panel; you upload a setting file, verify the analog inputs, and test each protection function with a relay test set.
That shift matters for the purchase order. I now ask vendors to include commissioning support in the quote. I don't want a 'great price' on hardware and a surprise invoice for the software session.
Boundary conditions
Now the honest limits. A multimeter cannot test a high-voltage transformer protection system. For a GE transformer, you need the factory test report, a site-specific energization plan, and qualified commissioning personnel. A relay test set is required to verify the 845's differential elements. And if the application is medical equipment, like a hospital bed control panel, electrical safety standards override my cost-saving shortcut. A hospital facility isn't the place to test a relay with a generic multimeter while the patient is in the bed. That requires proper isolation and qualified biomedical technicians.
Before you pay for a GE dry type transformer, ask for the test certificate and the standard it was designed to meet. My reference for dry-type transformers is IEEE C57.12.01; for relay performance, it's IEEE C37.90. Those standards should be in your project file, even if they never come up in the sales meeting.
I also stay away from vendors who use the word 'maintenance-free.' Per FTC guidelines (ftc.gov), claims need to be truthful and substantiated. No transformer is maintenance-free. A GE dry type transformer is lower maintenance than a liquid-filled unit, but 'lower' isn't 'zero.' If a vendor tells you otherwise, they're not giving you a cost advantage. They're giving you a future repair order.
At the end of the day, my job is not to pick the cheapest part. It's to avoid the most expensive failure. Testing is the cheapest insurance I can buy. A $40 multimeter, a $20 relay, and a 10-minute test sequence are usually enough to avoid a $4,000 service call. The switch to digital protection doesn't remove that need; it raises the stakes because the data is right there. You just have to read it.
Leave a Reply