On a Tuesday morning in September 2023, I drove three hours to a pump station because the customer said their Sierra Wireless RV50X kept dropping offline. They had already replaced the first gateway. The second one did exactly what the ticket described: it connected to LTE, worked for maybe forty seconds, then rebooted. The cycle repeated every few minutes.
I checked the configuration. Nothing wrong. I moved the antenna. Same result. I opened the router's event log and saw clean restarts, which made me suspect firmware, which made me want to blame the box.
Then I put a multimeter on the power line.
That should have been step one. The little AC adapter was rated at 12V, and it showed 12.1V while the router sat idle. The moment the RV50X started transmitting, the voltage at the input terminal dropped to about 7.4V. The gateway wasn't dropping off the network. It was losing power under load and resetting itself. Not the network. Not the router. The power supply.
I've been handling Sierra Wireless gateway deployments for seven years, mostly for remote monitoring sites: water, energy, agriculture. I've personally made, and documented, sixteen significant installation mistakes. Together they cost roughly $18,000 in wasted trips, RMAs, and downtime. After the third one, I started keeping a checklist so the next person doesn't repeat my errors. This article is basically that checklist. The frustrating part is that most mistakes were preventable in under five minutes.
A note on limits: my experience is with outdoor field deployments. I can't speak to gateways sitting in clean network closets with conditioned power. If that's your situation, some of this won't apply. For the rest of us, it applies a lot.
The first clue: a gateway that 'drops' but logs a clean restart
A genuine cellular network dropout doesn't normally make a router reboot. It makes the router try to reconnect, but it stays powered on. So when a site reports random drops and the logs show the unit restarting, the question shouldn't be 'why is the network bad?' It should be 'why is the router losing power?'
In my experience, the answer is usually one of these: an undersized AC adapter, a thin DC cable, a corroded terminal, or a shared power feed that dips when something bigger switches on. You can't see any of that from the web dashboard. The router only reports its internal voltages. To know what's actually arriving, you have to measure it at the input.
I'm not going to write a full how to use multimeter guide here—just the field test that matters. Set your multimeter to DC voltage, red probe on the positive input terminal, black probe on negative. Watch the reading while the router boots, connects to LTE, and sends data. Not just at idle. A small dip is normal. A big dip, or a reboot during transmission, means the power path is the problem.
To be fair, this is easy to miss. Intermittent power problems are the worst kind because the router looks perfectly healthy when you arrive. That's exactly why I now test under load before touching any other setting.
The layer below that: not every 'drop' is a connection drop
The second lesson took me longer. Some of our RV50X gateways showed as offline in the monitoring dashboard while the router itself had a perfectly good cellular connection. We could see it in the management interface: connected, with an active session, responding to pings. But the dashboard said the site was down.
Why? Because the dashboard marked a device offline if it didn't receive a heartbeat in 45 seconds. The gateway had been configured to report once per minute. The router was fine. The monitoring platform was impatient. We created an outage by not matching those two numbers.
I call those phantom disconnects. They usually trigger at 2:00 AM, resolve themselves by 2:12, and by the time a technician arrives, everything looks normal. If you've chased one, you know how frustrating that is.
The fix is boring: make sure the device's heartbeat interval is shorter than the platform's offline threshold. If you manage more than a few field gateways, write it down. We created a one-page reference we call the HeartGuide. It lists each gateway type we deploy, its heartbeat interval, and the matching timeout on the monitoring side. Before commissioning, a technician checks the device against that table. Simple.
Since we started doing that, we've caught more than twenty potential false alarms before they became actual tickets.
What these mistakes actually cost
Let me put a number on it. One unnecessary truck roll to a remote site runs $600 to $1,200 where I work. Add the diagnosis time, the expedited shipping for a replacement that wasn't needed, and the credibility damage when the customer watches you shrug and say 'seems fine now.' It adds up fast.
This is also where the price-focused purchasing conversation gets dangerous. I've used Sierra Wireless USB modems and adapters, and they're fine for what they are: temporary connectivity, laptop-based setups, quick deployments. But if someone specs a USB modem because it's $300 cheaper than an industrial gateway, and it lives behind a PC that reboots for updates at 3:00 AM, you're not saving $300. You're buying a future truck roll.
That's not a criticism of the USB product. It's a criticism of matching the wrong tool to the environment. Same logic applies when you move to a newer model. If you're evaluating the G310 5G for the next round, the specs matter, but the installation matters just as much. Put a good 5G gateway on a weak power feed and you'll still get a 5G outage.
The short version of what we check now
The pre-deployment checklist I maintain has three items that catch most of the problems I used to find in the field:
- Measure power at the gateway's input while the LTE radio is transmitting. If the voltage sags badly or the unit resets, fix the power path before you install the router.
- Check the device's heartbeat setting against the monitoring platform's timeout. Use the reference doc, not memory.
- Physically inspect the antenna connector, ground, and enclosure before leaving the site. Software can't tell you a connector is loose.
If you ask me, the cheapest upgrade we made as a team was putting a multimeter in every technician's bag. It cost less than one wasted trip. The second cheapest was writing down the heartbeat settings so we stop guessing.
Your Sierra Wireless RV50X probably isn't as unreliable as the tickets make it look. It might just be starving for power, or waiting too long to say 'I'm still here.' Check those two things before you swap hardware. Most of the time, that's the whole answer.