The 4:47 PM Danfoss Oil Pressure Switch Call That Almost Went Sideways

4:47 PM, Friday. March 2024.

The phone rang just as I was closing my laptop for the weekend. Facility manager at a cold storage plant, about 90 minutes from our warehouse. His main refrigeration compressor had tripped on an oil pressure fault. He had roughly $180,000 worth of frozen product sitting at risk, and the weekend forecast was warm.

"Can you help us or not?" he asked. Not rude — just scared.

I'm the technical support lead at an industrial controls distributor. In my role coordinating emergency orders for HVAC and refrigeration clients, I've handled over 200 rush jobs in the past six years. Most of those are straightforward: customer needs a replacement Danfoss pressure switch, we ship it, done. But this one felt different from the first thirty seconds.

The first diagnosis (from my desk)

I asked him to send a photo of the switch. It was a Danfoss oil differential pressure switch — RT series, older model, probably installed when the compressor was commissioned. Classic workhorse. The setting can drift over time, especially on a unit that's been running for a decade.

"Does the unit have a sight glass on the oil line?" I asked.

"Yeah. Oil level looks fine."

"What's the pressure reading on the gauge?"

Pause. "I don't have a gauge on it right now."

That was my first clue. You can't diagnose a pressure switch problem without knowing the actual pressure. I walked him through checking the setpoint adjustment on the switch — on the RT series, it's a screw mechanism under the front cover. If it had drifted too low, the switch could trip even when oil pressure was fine.

"Turn the setpoint screw clockwise about a quarter turn," I said. "See if the compressor restarts."

Click. Whir. It started.

"It's running!" he said.

Where it got complicated

I should have stopped there. Honestly. But something bothered me. A quarter turn on the setpoint shouldn't fix a switch that's been working fine for years — not permanently, anyway. That suggests either the switch itself is drifting, or something else is going on in the system.

"Can you check the PLC?" I asked. "What fault code is showing?"

He read it off: "Oil pressure fault, input 12."

"And the VFD? Is the drive showing any faults?"

"No, VFD says normal."

That's when I knew I needed to go out there. If the pressure switch was tripping, but the VFD was fine — and the actual oil pressure was fine — the problem might not be the switch at all. It could be the signal path between the switch and the PLC.

Looking back, I should have asked about the full control architecture before suggesting any adjustments. I assumed it was a simple switch-only system. It wasn't. The compressor had a VFD for speed control, a PLC for sequencing and protection, and the Danfoss pressure switch was just one input in a larger safety chain.

The drive out (and the doubt)

I loaded a spare RT series switch, a calibrated pressure gauge, and my multimeter. Even after deciding to go, I kept second-guessing. What if the switch wasn't the problem? What if I drove 90 minutes each way and couldn't fix it? The two hours until I got there were stressful.

I also thought about the alternative: ship a replacement switch overnight and hope. But if the issue was in the PLC input card — which I suspected — a new switch wouldn't solve anything. And this guy had $180,000 in product on the line.

So I drove.

At the site

Got there a little before 7 PM. First thing I did was put a mechanical gauge on the oil line. Reading was steady, right in spec. So the actual oil pressure was fine — the switch was either mis-set or the signal wasn't getting through correctly.

I pulled out my multimeter and checked the switch contacts. They were opening and closing correctly when I tested with the gauge. So the switch itself was fine (which, honestly, I was a little surprised by — usually these things fail closed after enough cycles).

Then I traced the wiring. The VFD's fault relay was wired to a PLC input module — the same module that was receiving the pressure switch signal. The PLC input card was configured for 24V DC, but the VFD's dry contact output was pulling through a shared 120V circuit.

There it was. The PLC input card was seeing voltage on an input it wasn't expecting. That's why the PLC threw an oil pressure fault even when the switch was fine — the input module was misreading a signal from the VFD, not the switch.

This is where my training kicked in. I've been through the troubleshooting sequences at PLCs at Work Institute — not as a student, but as a service provider who's had to reverse-engineer their methods from the field. The key lesson: don't trust a single fault code. Trace the signal from source to controller.

Setting the Danfoss switch properly

While I was in there, I reset the Danfoss pressure switch anyway — no point leaving a drifted switch in the system. Here's the basic process for an RT series oil differential pressure switch:

• Verify actual pressure with a calibrated gauge at the test port. Don't trust the system gauge — it may be off by 10-15%, especially on older units.
• Locate the setpoint adjustment. On the RT series, it's under the front cover — a screw mechanism that moves the setpoint within the switch's range.
• Set the switching point. Turn the screw until the switch trips at the correct differential pressure. For most refrigeration compressor oil systems, that's typically around 10-15 PSI differential, but check your compressor's manual — this varies by manufacturer and model.
• Adjust the differential if the switch has that feature. The differential controls how much pressure change is needed to reset the switch. Too narrow, and you get nuisance trips. Too wide, and you lose protection.
• Verify the reset. Run the compressor through a full cycle and confirm the switch resets at the correct pressure.

I also checked the VFD diagram — the wiring schematic on the inside of the drive enclosure — and confirmed the fault relay output configuration. That's where I caught the voltage mismatch. Per Danfoss product documentation (danfoss.com), the RT series switches are designed for specific voltage and current ranges, but the issue wasn't the switch — it was the shared circuit between the VFD and the PLC input module.

The generator question

While I was working, the facility manager asked me something that seemed random at first: "You work on generators too? How does a whole house generator work?"

Turns out their backup generator had been acting up, and he was wondering if the pressure switch issue was related. It wasn't — different systems, different components — but I gave him the short version.

A whole house generator works by monitoring the incoming utility power. When the voltage drops or cuts out, a transfer switch disconnects the house from the grid and connects it to the generator. The generator itself has its own safety switches — oil pressure, coolant temperature, over-speed — that protect the engine. If oil pressure drops too low, the generator shuts down to prevent damage, just like your refrigeration compressor.

"So the same kind of switch?" he asked.

"Same principle," I said. "Different setpoint, different size, but yeah — pressure switches are everywhere."

The difference, I explained, is that a whole house generator is usually air-cooled or liquid-cooled with its own dedicated control board. The pressure switch is built into the engine's protection circuit, not a separate component you'd replace. But the concept — monitoring pressure and tripping a switch when it drops — is the same across HVAC, refrigeration, and generator systems. According to the U.S. Department of Energy (energy.gov), modern standby generators typically run on natural gas or propane and can power a whole home within seconds of an outage, but they rely on those internal safety switches to prevent engine damage.

Fixed by 10 PM

I reconfigured the PLC input module, replaced the VFD relay wiring with the correct voltage setup, and reset the Danfoss switch to spec. Compressor restarted. Oil pressure held steady. PLC showed "normal" across the board.

The facility manager looked like he'd aged five years in five hours. "So what was it?"

"Not the switch," I said. "Well — the switch was drifting, and I reset it. But the actual fault was a wiring issue between your VFD and PLC. The switch was doing its job — the PLC just wasn't reading the signal correctly."

There's something satisfying about that moment. After all the stress and the drive and the uncertainty, seeing the compressor run and the product safe — that's the payoff. But there was also a humbling part: I almost replaced a perfectly good switch because I assumed the fault code was accurate. It wasn't.

Three things I took away from that night

First, hindsight is brutal: I should have asked about the full control system before suggesting any adjustments over the phone. A quarter-turn on a setpoint screw is a temporary fix at best, and at worst it masks a bigger problem. If I'd known there was a VFD and PLC in the system, I would have approached the diagnosis completely differently.

Second, there's real value in knowing your boundaries. I'm good with pressure switches. I can set a Danfoss RT series to within a fraction of a PSI and tell you why it's tripping. But the PLC input card issue? That's a controls engineer's territory. I knew enough to find the problem, but a dedicated controls specialist would have spotted the voltage mismatch in half the time. That's not a weakness — that's just how it works. A specialist who knows their limits is more trustworthy than a generalist who says they can fix anything.

Third — and this is the one I keep coming back to — the fault code doesn't tell you what's wrong. It tells you where the system noticed a problem. A $180,000 lesson in not trusting the first answer.

The compressor ran fine through the weekend. The facility manager called me Monday to say thanks. I sent him a spare RT series switch anyway — because that drifting setpoint was a preview of coming attractions, and I'd rather he have a backup on the shelf than call me at 4:47 PM on another Friday.