One approach comes up repeatedly: a phone hotspot cannot carry several devices, so the phone gets cabled into a portable travel router, which broadcasts Wi-Fi and can hold a VPN tunnel on its own side. A more practical answer usually arrives in the same conversation: fine for temporary use, but for something left running all the time, a unit that takes a data SIM is the better buy. Both are fair. The difference sits in how the words "all the time" actually get honored.
Staying online is not the same as being switched on. Four things have to keep happening: a dropped link has to be brought back, the devices behind it have to keep receiving addresses and a way out, the box itself has to survive being powered continuously, and someone far away has to be able to see that it is still there. A phone plus a portable router splits the first two across two devices and leaves the last two to nobody. A dedicated cellular wifi router keeps all four on one box.
In the two-device setup, the cellular leg still belongs to the phone. Recovery after a drop is manual: restart the hotspot, or send a text message to move onto the next data bundle. With nobody on site, neither happens, and afterwards there is no record of how long the link stayed down. On a box built to stay online, that leg handles itself: automatic redial and multi-layer link detection pull the connection back up, and wired, cellular and Wi-Fi paths back each other up, with an automatic switch when the primary path stops working. One item causes the most rework at SIM activation: IoT and private-network SIMs need an APN, and when it is left blank the symptom is endless failure to register, which is often misread as no local coverage. Acceptance: unplug the primary link once, time the recovery, and write that number into the handover document, along with whether site data was being buffered for retransmission or simply stopped during the wait. Antennas should not be left inside a metal cabinet; logging signal strength at one position before moving the unit is quicker than repeated trial and error.
Handing Wi-Fi over to a portable router does raise the number of terminals that can associate, but there is still only one way out, and client capacity, coverage and address management on consumer gear are all sized for a few people going online for a while. A site with many devices needs something else: Ethernet and wireless distributing at the same time, scanners, handheld terminals, cameras and PLCs each holding a fixed address, and outbound rules written in one place on the box that owns the exit, naming which unit may talk to whom and denying everything else by default. Once addresses are bound to MAC addresses, moving to a new customer means changing the destination address and port only, with the rule itself left alone. The gap at this layer is usually temporary equipment brought in for commissioning: not on the list, still sending data outward. Glancing at the list of currently connected terminals from time to time is easier than reading a bill afterwards. Coverage has a reference value as well: figures for this class of device are usually given for two conditions, roughly 200 meters outdoors in the open and roughly 40 meters indoors with obstacles, discounted further in a metal-dense workshop, with an on-site measurement deciding whether that is enough.
A phone is not built to sit in a cabinet year round. Charge cycles, room for heat to escape, and reboots after a system update are none of them meant for continuous operation. Portable routers mostly run on USB power inside plastic housings, with temperature ranges set for being carried around. The unit inside the cabinet has to take a different set of conditions: a wide-voltage terminal block taking DC straight from the cabinet, where 24 V dips at the moment motors start and stop, with reverse polarity protection and two supply inputs on separate circuits belonging to the same layer; a metal housing carrying heat away, mounted clear of the space directly above inverters and heat-generating modules; hardware and software watchdogs resetting the box if it hangs, with continuous operation written into the parameter sheet rather than hoped for. Acceptance: pick the hottest stretch of the year, measure intake air temperature and housing surface temperature, and keep more than twenty degrees Celsius of margin below the stated upper limit.
This is the weakest point of the two-device setup. Nobody hears about it at the moment the link drops, and days often pass before anyone notices. A box built for continuous duty turns three conditions into pushed messages: device offline, signal weakening, data usage past a threshold. They go out by email and text message and can repeat, so there is no need to wait for a bill to arrive. Link switching triggers on availability rather than on signal bars, and a weak signal counts as its own alarm class instead of waiting for a complete drop. Routine changes do not need a site visit either; a remote management platform can open the built-in web page of the unit from a distance to change parameters, reboot and upgrade firmware. Before the SIM goes in, one month of usage comes from reporting interval multiplied by message size multiplied by point count, with the threshold set at seventy or eighty percent of the plan, so the alert arrives while allowance remains.
The USR-G806w matches sites short of cabinet space with few devices: three Fast Ethernet ports with one WAN port configurable as LAN, 802.11b/g/n wireless with dual 5 dBi antennas and two FEM paths, 104 x 102 x 28 mm for DIN rail or wall mounting, terminal block and DC jack as two redundant supply inputs, -20 to 70 degrees Celsius, hardware and software watchdogs, and three alarm classes plus remote management all included. The USR-G816 takes remote sites with no broadband and mobile assets: 5G SA and NSA on a Qualcomm X62 module, dual SIM with the second slot optionally an embedded eSIM, RS232/RS485 terminal block for hanging legacy instruments directly, a metal housing at -35 to 75 degrees Celsius, and automatic failover between primary and backup links. The USR-G809s covers sites with a lot of devices: two ports configurable as WAN or LAN plus six LAN ports and two SFP fiber ports, dual-band Wi-Fi 6, DC 9-60 V wide input with reverse polarity protection, DIN rail mounting, and no extra switch needed when subnets have to be split.
A phone plus a portable router solves having network access where people are present. A dedicated cellular wifi router solves having network access where nobody is.
1. How is a 5G cellular router different from a regular Wi-Fi router?
The dividing line is whether there is a SIM slot. A regular Wi-Fi router needs an incoming broadband line; a cellular wifi router carries a SIM of its own and reaches the network by itself. Another frequent mix-up is 5G versus 5 GHz: the first is the fifth generation of cellular network, the second is a Wi-Fi band. The two have nothing to do with each other.
2. Should the choice be 4G or 5G?
Decide on upstream demand, not on downstream speed. Status values, energy readings and minute-level telemetry run fine on 4G. 5G comes in when video has to be sent back, firmware has to be distributed in batches, or one site carries several sustained streams. Where 5G coverage is missing, the module falls back to 4G, so actual local coverage deserves a question at selection time.
3. What matters when getting a data SIM?
Three things to settle in advance. The data SIM has to be permitted on a router rather than restricted to phones. Some carriers require the device IMEI to be registered. The APN decides how the connection is made, and an IoT SIM and a VPDN private-network SIM do not share one, so it belongs in hand at SIM activation; left blank, the symptom is endless failure to register.
4. Without a public IP address, can the device still be reached remotely?
Cellular lines usually receive a private address behind carrier-grade NAT, so inbound connections from the public internet do not arrive and port forwarding is basically closed. The workable approach is to have the device initiate the connection outward: a VPN tunnel, or opening the built-in web page of the unit through a remote management platform. Neither exposes a port to the public internet.
5. Is cellular access secure?
Security comes from the tunnel and the rules, not from the medium. The tunnel answers whether a destination can be reached; the rules decide what can be seen after arrival, with destinations limited to the IP address and port of specific devices and everything else denied by default. Compared with shared Wi-Fi, cellular removes one layer: knowing the passphrase is no longer enough to get in.
6. What is dual SIM for?
Two uses. Primary and backup, with traffic switching to the second card when the usual carrier drops. And changing carriers without changing hardware. Dual SIM models generally treat the external card as primary, with the second optionally an embedded eSIM, which saves a batch of on-site card swaps on cross-border deliveries.
7. Is data lost while the link is down?
It depends on who buffers during the switch. Some designs cache on the device side and retransmit after recovery, others handle it on the platform side, and which applies should be settled before ordering. At acceptance, unplug the primary link once, time the recovery, and confirm whether data caught up afterwards, rather than discovering during a real incident that nobody was covering that window.
8. How should client capacity and data usage be estimated?
Capacity involves two numbers: the count of wireless clients, and the number of simultaneous connections. The two are not the same order of magnitude. Data usage comes from reporting interval multiplied by message size multiplied by point count, with continuous sources such as cameras counted separately. Set the threshold at seventy or eighty percent of the plan and the alert arrives before the allo
