Engineering Insights

The Day I Realized Voltage Drop Could Kill an IoT Deployment (and What I Did About It)

The Setup: When a 'Simple' Order Goes Sideways

In early 2023, my boss from operations handed me a requisition that looked simple enough. We needed a batch of Sierra Wireless Airlink MG90 gateways for a new fleet management pilot. I'd ordered networking gear before—routers, switches, the usual IT stuff. This was just another PO to process, right?

I pulled up the spec sheet for the MG90. It's a solid piece of hardware, built for vehicles. I confirmed the part number, checked the price against our budget, and sent the order to our distributor. Standard procedure. About three weeks later—or rather, closer to four when you count the shipping delays—the boxes arrived. I was ready to close out the purchase order and move on. But my phone rang the next day. It was our lead installer, and he was not happy.

The Crisis: What's a 'Voltage Drop Calculator'?

"These gateways are rebooting," he said. "Every time the truck's engine starts. We've lost data three times today."

I didn't have a technical answer. I'm an admin buyer, not an electrical engineer. I manage vendors and process orders. My first instinct was to blame the hardware. "Maybe we got a bad batch?" I asked.

"No," he said. "The voltage drop calculator we used says the cable run is too long for the current the MG90 draws during startup. We need a different power supply setup or shorter cables."

He'd used a term I'd never even seen on a spec sheet: voltage drop calculator. I felt a knot in my stomach. I'd approved a purchase order for nearly $15,000 worth of equipment, and I'd never once asked the crucial question: "What's the power draw under load?" I assumed the spec sheet was enough.

Honestly, looking back, I should have asked the install team to run a power budget before I ever hit 'buy.' But given what I knew then—that a router is a router—my choice was reasonable, if incomplete. The MG90 spec sheet listed an input voltage range and typical power consumption. I didn't realize that 'typical' and 'peak' were two different animals.

The Search: My Hunt for a Better Answer

So, I did what any admin would do: I started googling. I dug into the Sierra Wireless support forums and the product documentation for the MG90. I found a community post where a system integrator was complaining about a similar issue with the older MP70. Another guy said, "You need to factor in the inrush current when the cellular module boots. Look at the startup specs for the EM9191 module."

"In my five years of buying industrial IoT gear, I had never stopped to think about inrush current. It was a massive blind spot."

I don't have hard data on industry-wide failure rates due to power issues, but based on my subsequent research, my sense is that at least 15-20% of first-time industrial router deployments have a power-related hiccup. It's not a hardware defect. It's a specification mismatch. The voltage drop calculator became my new best friend. I found a simple online tool and started plugging in numbers: 12V source, 18 AWG wire, 25 feet of cable. The results were eye-opening. The voltage at the end of the line was under 10.5V—below the MG90's guaranteed operating threshold.

I spent the next two days on the phone with our distributor and a Sierra Wireless application engineer. They confirmed the issue and walked me through the solution: use a shorter, heavier-gauge power cable, or add a local DC-DC converter near the gateway.

The Fix: Standardizing the 'What Ifs'

We ended up re-pulling the cables in the test vehicles—a cost of about $400 in labor and materials. But the gateways ran flawlessly after that. The pilot data started flowing, and my boss was happy. But I was still chewing on the lesson.

I created a new column in my RFQ template. Right next to "Unit Price" and "Delivery Date," I added a column for "Peak Power Draw (Startup)" and a link to a voltage drop calculator. I also started adding a 25% margin to any power requirement estimate if the manufacturer didn't provide clear peak figures. It's not scientific, but it's cautious.

Before I send out an RFP for any cellular router now—whether it's an Airlink MG90, an LX40, or an FX30—I always ask:

  • What is the peak inrush current?
  • What is the recommended wire gauge for a [X] foot run?
  • Can you provide a wiring diagram that accounts for vehicle power rail fluctuations?

If a vendor can't answer those questions quickly, it's a red flag. I tell our team: "I recommend this gateway for 80% of truck rollouts. But if you're running power longer than 15 feet, you might need a different plan." This solution works for most cases. Here's how to know if you're in the other 20%: if your installation involves pre-existing, long, thin-gauge wiring. If so, what is the cost of a local power conditioner vs. the cost of a field failure? Usually, the conditioner wins.

The Lesson: Honesty is the Shortest Path

I used to think buying hardware was just about the best price. Now I know it's about asking the right dumb questions. The voltage drop calculator isn't just a tool for electricians; it's a tool for procurement. It's the difference between a deployment that "looks good on paper" and one that works in the real world.

If you are an admin buyer, a system integrator, or anyone specifying Sierra Wireless gear for a mobile application, take my advice: Don't just look at the spec sheet for the EM9191 module. Calculate the power at the endpoint. It can save you a night of angry phone calls.

Per FTC guidelines (ftc.gov), I should note that this is my personal experience managing procurement for a mid-size fleet operator. Your mileage may vary depending on vehicle electrical architecture. Our success rate improved immensely after standardizing these checks, but no single fix is universal.

Leave a Comment

Your email address will not be published. Required fields are marked