We took the only IP68 router in our own catalogue out of the datasheet and put it where customers actually install it — on a pole, on a rooftop, in a coastal cabinet — then measured what the spec sheet leaves out. This is the full field review of our IP68 outdoor industrial 4G router, including the four datasheet lines that decide whether your installation survives its second winter.
Written by E-Lins Engineering Team
Here is a sentence we have heard from more customers than we would like to admit:
“The router is rated IP68. It was bolted to a pole. It still died.”
Then the RMA comes back to us, we power it up on the bench, and it runs for days without a problem.
The unit is not necessarily broken. Sometimes it was simply installed correctly according to the rating, but incorrectly according to the environment.
That distinction matters.
IP68 describes protection against dust and water under specified test conditions. It does not, by itself, describe solar heating, ultraviolet exposure, salt, thermal cycling, condensation, cable workmanship, voltage transients or lightning-induced surge.
That gap between enclosure rating and real outdoor conditions is the subject of this review.
The H820QO is a compact outdoor industrial 4G router with an IP68 enclosure, pole/wall/desktop/DIN-rail mounting, two Ethernet ports, PoE-PD, built-in high-gain cellular and Wi-Fi antennas with reserved connectors for external antennas, a 5–40 V DC input, and a stated operating range of −35 °C to +75 °C ambient air. It is currently the only router in the E-Lins range with an IP68 enclosure.
How We Review Our Own Hardware (and Which Parts You Should Discount)
Let us deal with the obvious objection first: a vendor reviewing its own product is not an independent test. We are not an accredited laboratory, and nothing in this article is a certification. What we do have is a bench, a rooftop, a coastal test frame, and a large number of customer sites that have been willing to tell us what broke. That is a different kind of evidence, and it has different limits, so here is exactly what is what:
| What we state | Where it comes from | How much weight to give it |
|---|---|---|
| Dimensions, ports, ratings, power consumption, temperature limits | The H820QO datasheet, quoted directly | Verifiable — check them against the current datasheet before you order |
| Thermal budget, power budget, dew point, feeder loss | Arithmetic, with assumptions printed alongside the result | Reusable — substitute your own numbers; the method is the deliverable |
| “In our experience”, “typically”, “what we see on customer sites” | Field observation across deployments, not controlled trials | Directional — treat as a hypothesis to test on your own site, not a guarantee |
| Anything about lightning, structural mounting, or mains wiring | General engineering practice and applicable standards | Out of scope for a router vendor — engage a qualified local specialist |
With that out of the way: the reason to read a vendor’s own field review is not the score. It is the failure taxonomy. We see the returns. We know which installations come back and which do not, and the pattern is consistent enough — and inconvenient enough for our own sales narrative — that it is worth publishing.
| The one-sentence version |
| Water almost never kills an outdoor router. Heat, condensation and workmanship do. Most “dead” outdoor routers we receive are physically fine, and most of the ones that are genuinely dead died of something the IP rating was never designed to measure. |
What Arrives, and What the Box Tells You Before You Open It
The H820QO ships with an Ethernet cable and a PoE power adapter; anything else — GPS antenna, additional antennas — depends on the options ordered. The unit itself is 185 × 172 × 82 mm and about 560 g without accessories, rising to roughly 750 g with them. For context, the PCBA inside is 125 × 105 × 20 mm, which means roughly two-thirds of the product’s volume is enclosure. That is the correct proportion for an outdoor device and the first thing to check on any product claiming a rugged outdoor router enclosure: if the case is a thin shell wrapped tightly around the board, there is no room for the sealing, the wall thickness, or the thermal mass that outdoor service actually needs.
The second thing to check is colour, and it is not an aesthetic point. The H820QO case is light grey. A dark grey or black enclosure of the same material in full sun will run measurably hotter inside — for the same polymer, surface absorptivity in the visible and near-infrared is the single largest driver of solar gain, and it is a choice the manufacturer makes for you. If you are specifying an outdoor enclosure and nobody can tell you why it is a particular colour, that is worth a question.
| 185×172×82 |
| Millimetres with case; PCBA is 125×105×20 |
| 560 g |
| Without accessories; ~750 g with |
| 2 |
| Ethernet LAN/WAN ports — and that is all |
| IP68 |
| The only IP68 enclosure in the E-Lins range |
Third, and this one is genuinely useful: the port count. The H820QO has two Ethernet LAN/WAN ports and an on-board console port. Two ports is the right number for a pole-mounted device doing one job — uplink in, one device out — and the wrong number if you imagined it as a site aggregation point. Any review that does not tell you the port count up front is not a review.
Reading the Enclosure the Way an Installer Does
An IP68 rating on an outdoor router is only as good as the four interfaces that carry it, and the datasheet line that matters most here is short enough to miss:
| From the H820QO datasheet |
| “IP68 grade anti-dust and anti-water enclosure for outdoor installation; PC+anti-UV+anti-oxidation OSP; Light-gray color; Wall-mount, Pole Mount.” |
Three things are hiding in that sentence. The first is the polymer: polycarbonate, not the commodity ABS that a lot of “outdoor” plastic enclosures are made from. The second is the anti-UV package — polycarbonate without UV stabilisers chalks, yellows and loses impact strength, and it does so fastest exactly where you would deploy it. The third is anti-oxidation OSP on the metal parts, which is the least glamorous and most consequential item on the list, because corrosion does not start at the connector you can see.
Now the part that a rating never tells you: the enclosure has four independent sealing interfaces, and the IP68 rating belongs to all of them working together. The lid gasket is factory-fitted and, in our experience, essentially never the problem. The other three — the waterproof cable gland for outdoor router installation, the antenna connectors, and the power terminal — are finished on site, by whoever is standing on the ladder.
| Sealing interface | Who finishes it | What actually goes wrong | Control |
|---|---|---|---|
| Lid gasket | Factory | Compression set after years of thermal cycling; a lid re-seated with a cable trapped across the gasket | Re-torque lid screws in a cross pattern; never route a cable across the gasket land |
| Cable glands | Installer | Gland sized for the cable diameter but not the jacket; unused holes left open; glands on the top or side face instead of the bottom | IP68-rated glands matched to measured cable OD; blanking plugs in every unused hole; all entries on the bottom face |
| Antenna connectors | Installer | N-type left finger-tight; no self-amalgamating tape over the joint; water wicking down the feeder into the connector | Torque to the connector manufacturer’s value; wrap the joint; form a drip loop below it |
| Power terminal | Installer | Two-pin terminal over-tightened and cracked, or under-tightened and fretting; unsealed entry | Torque to spec; strain-relieve the cable; seal the entry with the supplied gland, not with silicone alone |
Note what is not in the table: the IP68 rating itself. Once the lid is closed and the glands are correctly populated, the enclosure does its job extremely well. Every ingress failure we have traced at a customer site was traceable to one of the three installer-finished interfaces.
Test 1 — Water: the Hose, the Tank, and the Thing IP68 Does Not Test
Let us be precise about what the second digit of an IP rating means, because it is widely misquoted in both directions. IP68 is not “waterproof at any depth forever”. Under IEC 60529, the first digit (6) means dust-tight. The second digit (8) means the enclosure is suitable for continuous immersion in water under conditions specified by the manufacturer, which are normally more severe than IP67’s 30 minutes at 1 m. The manufacturer must state the duration and depth. That is the whole test.
So: a hose, a pressure washer at a sensible distance, a fortnight under a monsoon, a flooded pit — the enclosure handles these. What it does not address is the mechanism that actually puts water on the inside of outdoor electronics.
The 5 a.m. dew point
A sealed outdoor box contains a fixed charge of air, sealed at whatever humidity prevailed on the day it was closed. During the day, the sun heats the enclosure wall well above the air temperature. Overnight the wall cools, and it keeps cooling until shortly after sunrise, which is when the daily minimum temperature usually occurs — somewhere between 05:00 and 06:30 in most temperate locations. The wall is now the coldest surface in the system.
If the wall temperature falls below the dew point of the air sealed inside the box, water condenses on the inside of the lid and runs down onto the board. No rain is required. No seal has failed. In fact, the better the seal, the worse the problem, because a well-sealed box cannot exchange its moist internal air for drier outside air.
Run a worked example. Take a temperate coastal site: afternoon air at 20 °C and 65% relative humidity. The dew point of that air is roughly 13.5 °C. The box was closed in those conditions. Overnight the wall falls to 11 °C. Condensation forms. Now take a Gulf site: 32 °C at 80% RH gives a dew point near 28 °C, and while the wall rarely falls that far, it does not have to in an air-conditioned-adjacent or shaded north-face installation — and it certainly does in a box that spent the night radiating to a clear desert sky.
Here is the part of the datasheet that we think deserves more attention than it gets:
| Read the humidity line carefully |
| The H820QO is rated 5% to 95% operating (non-condensing), and 5% to 95% storage. That parenthetical is not boilerplate — it is the manufacturer telling you the rating does not extend to the condition your enclosure will produce at dawn. Every serious industrial router we know of carries the same caveat. It is not a weakness of this product; it is a property of the environment that the rating system does not cover. |
The mitigations for IP68 vs condensation in outdoor enclosures are unglamorous, and they work: a drain/breather plug at the lowest point of the enclosure so any liquid that forms has somewhere to go; a hydrophobic vent membrane so the box can equalise pressure and humidity without admitting liquid water; a desiccant sachet sized to the enclosure volume and replaced on a maintenance interval; and mounting the router high in the enclosure, never at the bottom where condensate pools. What does not work is “seal it harder”. Sealing harder is what caused the problem.
Test 2 — Heat: 75 °C Is the Air Around the Box, Not Inside It
The industrial router operating temperature range is the most quoted specification in this category and the most misread. The line on the datasheet reads “−35 °C to 75 °C ambient air operating”. Look at the two words that carry all the weight: ambient air. That is the temperature of the air immediately surrounding the device. If the device is bolted to a pole in free air, that is the weather. If the device is inside a cabinet, that is the air inside the cabinet, and it is your job to know what that number is.
Here is the arithmetic we use for outdoor enclosure thermal derating, with the assumptions printed so you can argue with them.
| Input | Value used | Why this value |
|---|---|---|
| Peak solar irradiance, clear sky, sea level | ≈ 1000 W/m² | Standard peak-sun figure; use your site’s design value |
| Effective projected area of a small enclosure | ≈ 0.15 m² | One large face plus the lid of a ~400 × 300 × 200 mm box |
| Solar absorptivity, light grey polycarbonate | ≈ 0.6 | Dark grey would be ~0.9; white ~0.25 |
| Fraction of absorbed solar energy that drives internal air | ≈ 20–30% | Thin-walled box, still air, radiation trapping under the lid |
| Enclosure thermal conductance, sealed, no fan | UA ≈ 1.0–1.5 W/K | Typical for a small sealed non-metallic box with internal film resistance |
| Internal dissipation — H820QO | 1.2 W idle / 3.6 W typical Tx | Datasheet: 100 mA and 300 mA typical at 12 V DC |
Worked through: absorbed solar power ≈ 1000 × 0.15 × 0.6 ≈ 90 W, of which roughly 20 W reaches the internal air. Add the router’s own 1.2–3.6 W and a companion device or two — a small PoE switch at 8 W, a power supply at 5 W — and you are at roughly 30 W of internal load. Against UA = 1.2 W/K, that is a steady-state rise of about 25 K above ambient.
At a 40 °C Gulf summer afternoon, that puts the average internal air near 65 °C — and the air at the top of the enclosure, which is where a wall-mounted router tends to sit, runs another 5–8 K hotter. You are now at the datasheet ceiling with nothing defective anywhere in the system.
Three conclusions, and the third is the actionable one:
- Vent the enclosure, do not just seal it. A shaded, vented, light-coloured cabinet with a vent at the bottom and top will run dramatically cooler than a sealed dark one. Venting with a hydrophobic membrane preserves the ingress rating for liquid water.
- Specify the internal limit, not the weather. The number to write into a tender is “internal enclosure air shall not exceed 70 °C at design ambient”, not “router shall be rated −35 to 75 °C”. The first is a requirement you can verify; the second is a marketing claim you cannot.
- Budget for the whole box, not the router. The router is the smallest heat source in most outdoor cabinets. The PoE switch and the power supply are the ones that push you over.
One more, because it is the most expensive mistake on this list: outdoor enclosure thermal derating is not a linear risk. Electrolytic capacitors, the usual first component to go, follow an Arrhenius-ish rule of thumb — every 10 K rise roughly halves useful life. A cabinet running 20 K hotter than it needs to is not “a bit worse”. It is a quarter of the service life.
Test 3 — Cold: the −35 °C End Matters More Than the 75 °C One
Everyone specifies the hot end. In our experience the cold end is where commercial-grade hardware quietly fails, for three reasons that have nothing to do with the silicon.
First, there is no self-heating to help you. In the thermal example above, the router’s own dissipation contributed to the problem at the hot end. At −35 °C, that same 1.2 W spread across a 1.2 W/K enclosure buys you about one kelvin. The device cold-soaks to true ambient. For any industrial router cold start test, that is the pass criterion: a commercial router rated 0–40 °C does not “run slowly” at −20 °C, it does not start.
Second, the cold start is also a condensation event. A unit that has been cold-soaked and then powered up warms its own internals while the case is still cold — a transient where the board can be below the dew point of the air immediately around it. The same applies to maintenance: opening a cold enclosure on a humid day is the single most reliable way to put water on a board. If you must open a cold box in humid conditions, let it equalise first, or bring a dry air source.
Third, voltage drop gets worse, not better. If you are feeding the router over a long DC run or a long PoE run — and pole and rooftop installations are exactly where runs get long — conductor resistance is one thing at 25 °C and another altogether when the cable has been at −30 °C overnight, and the current surge at power-up is when the drop is largest. The H820QO’s 5–40 V DC input range (5–60 V as an option) gives you real margin here, but margin is only useful if you calculate the drop rather than assume it.
| What to write into the tender |
| Not “operating temperature −35 to 75 °C”. Instead: “Equipment shall cold-start unaided at the site design minimum ambient, ≤ −30 °C, after a minimum 8-hour cold soak, with no heater and no operator intervention.” That sentence is testable, and it is the one that distinguishes industrial hardware from a ruggedised consumer product with a wide-temperature sticker. |
Test 4 — Sun and Salt: the Ten-Year Enclosure
Ingress is instantaneous. Ultraviolet and salt are cumulative, and they are the two mechanisms that decide whether an enclosure survives its second half of life rather than its first week.
Ultraviolet
Unstabilised polycarbonate yellows, chalks and loses impact strength under UV. The failure is not cosmetic: a chalked surface is a micro-crazed surface, and a micro-crazed surface is where a gasket stops sealing. This is why the H820QO spec calls out PC + anti-UV + anti-oxidation rather than just “plastic”, and why the UV resistant outdoor router enclosure question should be asked of every outdoor product you buy, including the cable glands, the antenna radomes and the cable jackets — all of which are usually the least UV-resistant parts of an installation.
Colour matters twice over here: it sets the solar absorptivity (Section 05) and it sets the rate of visible ageing. If you are tempted to specify a dark enclosure for aesthetics, understand that you are trading internal temperature and enclosure life for it.
Salt
Coastal and de-iced-road sites are the hard case, because salt mist corrosion on outdoor electronics does not work the way people expect — the mechanism is not the dramatic one. Salt does not usually eat through a polycarbonate lid. It attacks the things attached to it: connector threads, screw heads, bracket bolts, antenna ferrules, and the cut end of a cable armour. We have seen N-type connectors that could not be unscrewed after two winters within a kilometre of the surf, and pole brackets where the bolt, not the bracket, was the failure.
Practical controls, in rough order of value per euro: stainless or hot-dip galvanised fasteners rather than zinc-plated; dielectric grease on every outdoor RF connector before mating; self-amalgamating tape over every outdoor RF joint; siting the enclosure on the sheltered side of a pole or structure where the site allows; and a fresh-water rinse in the maintenance schedule for coastal installations, which sounds absurd and is genuinely one of the most effective things you can do.
Test 5 — Power: 5–40 V, ISO 7637-2, and the 48 V PoE Ordering Trap
This is the section with the most expensive single sentence in the datasheet, so let us put it in front of you verbatim:
| The ordering trap |
| “POE (Power over Ethernet, support 5-40VDC default, if the POE voltage is 48V, please order 5-60VDC version).” |
This is the trap hidden in a PoE PD 48V industrial router order: standard 802.3af/at PoE is nominally 48 V, and the H820QO ships with a 5–40 V input by default. If you install a default unit and feed it from a standard PoE switch or injector, you are outside the specified input range. This is not a defect — it is a configuration choice the datasheet states clearly — but it is the single most common ordering mistake we see on this product, and it is entirely avoidable: if your site uses PoE, order the 5–60 VDC version.
The rest of the power story
| Parameter | H820QO specification | What it means on site |
|---|---|---|
| Input range | 5–40 V DC (5–60 V DC option); dual power inputs with failover | Works from a 12 V battery, a 24 V vehicle system, or PoE — if you ordered the right version |
| Protection | Reverse polarity and transient voltage protection per ISO 7637-2; over-current and over-voltage | Survives a reversed connection and automotive transients — see the caveat below |
| Fusing | Inline fuse required for vehicle installations; 3 A for 9–24 V DC, 2.5 A above 24 V DC | Not optional. Size the fuse to the supply voltage, not to the router |
| Consumption | Idle ~100 mA @ 12 V; Tx/Rx ~300 mA; max 500 mA idle / 800 mA Tx; 12 V 2 A adapter recommended | ≈ 1.2 W idle, ≈ 3.6 W typical transmit, ≈ 6 W peak |
| Ignition sensing | Automatic ON and time-delay OFF | For vehicles: powers up with the ignition, and — critically — stays up long enough to finish a session and shut down cleanly |
| ISO 7637-2 is not lightning protection |
| This distinction gets lost constantly. ISO 7637-2 — the standard most often quoted for transient voltage protection on industrial routers — covers transients in 12 V and 24 V vehicle electrical systems — load dump, inductive switching, alternator behaviour. It does not address lightning or lightning-induced surge. A pole-mounted router is a lightning target, and no router input stage changes that. Lightning mitigation is site-level work: bonding, surge protective devices on the DC and Ethernet lines, antenna feeder grounding at the point of entry, and an equipotential reference. That is a job for a qualified local specialist, not a line item on a router datasheet. |
Solar and battery sites
Because the consumption figures are published, an off-grid site is a calculation rather than a guess. At typical transmit draw of 300 mA at 12 V, the router is about 3.6 W; over 24 hours that is roughly 86 Wh per day. With four peak-sun-hours — a reasonable mid-latitude annual average, and a number you should replace with your own site data — the panel must deliver about 22 W during the sun hours, which after margins for temperature, dirt, ageing and charging inefficiency means a 60–80 W panel for a router-only site, plus battery autonomy for your worst-case cloudy stretch. The solar powered outdoor router power budget is dominated by autonomy days, not by average consumption, and that is where off-grid designs usually go wrong.
Test 6 — Antenna: the Only Part That Is Genuinely Outside
The H820QO ships with built-in high-gain internal cellular and Wi-Fi antennas, with reserved connectors for external high-gain antennas (N-type on the option list). This is the right default for most pole installations and the wrong default for one specific case, and knowing which you are is worth more than any other RF decision on this page.
| The rule |
| If the router is mounted in free air, the built-in antennas are fine and you save a feeder, a connector pair and a leak path. If the router is mounted inside anything conductive — a metal cabinet, a kiosk, a control cubicle, a steel shelter — an internal antenna is not “a bit worse”, it is behind a Faraday cage. In that case you must use the external antenna option and route the feeder out. |
When you do go external with an N-type high gain antenna for outdoor routers, feeder loss dominates the budget and it is routinely underestimated. Order-of-magnitude figures for common coax: RG-58 loses roughly 0.6 dB/m at 900 MHz and about 1 dB/m at 2.1 GHz; LMR-400 is closer to 0.2 and 0.35 dB/m respectively. On a 6 m run at 2.1 GHz, that is the difference between losing about 2 dB and losing about 6 dB — which is the difference between a marginal and a solid link, and it costs less to fix at the design stage than after the first truck roll.
Three installation details that decide whether the RF stays good for years: keep the feeder short rather than routing it neatly; form a drip loop below every outdoor connector so water runs off the low point rather than into the joint; and torque and tape every N-type joint — finger-tight is not sealed, and an untaped joint is a corrosion site waiting for its first winter.
Test 7 — Network Behaviour: Three Lines In, and What Failover Actually Costs You
The H820QO supports cellular / RJ45 WAN / Wi-Fi client three-line redundancy with failover, failback and load balancing, with dual SIM failover for outdoor routers available as an option (single SIM is standard). VLAN 802.1Q, QoS with DSCP and priority queuing, DDNS, and management through NMS, TR-069, SNMP, SSH/Telnet, web and SMS.
The review observation here is not about features, it is about time. A link-down event is detected quickly. A cellular re-registration is not: depending on the network, the band plan and whether the SIM has to re-attach, bringing a modem back can take tens of seconds, and that is before any VPN tunnel re-establishes and any routing protocol re-converges. If your application has a 10-second supervision timeout, you will see a “router failure” that is really just a network behaving normally.
| Event | Typical order of magnitude | Design response |
|---|---|---|
| Ethernet link loss detection | < 1 s | Trust it; fail fast |
| Cellular re-registration / re-attach | Tens of seconds | Set application timeouts above this, not below it |
| VPN tunnel re-establishment | Seconds to tens of seconds | Use dead-peer detection tuned to the tunnel, not to the link |
| Routing convergence | Seconds to minutes | Prefer simple, predictable failover over elegant dynamic routing at remote sites |
One feature that earns its place specifically at outdoor remote sites: SMS control and status. When a device is two hours’ drive away and its IP path is exactly what you suspect is broken, being able to query status, reboot, or force a link change by text message is not a convenience. It is the difference between a five-minute fix and a truck roll. Combined with per-device data-usage alerting in the NMS, it also catches the failure mode that no redundancy scheme handles: the SIM that ran out of data.
| Field note · Coastal telemetry estate |
| “The pattern we see is not the dramatic one. It is the site that goes quiet on the same few mornings a month, always around dawn, always clearing by mid-morning. Nobody raises a ticket for a fault that fixes itself. Then one day it does not clear.” |
| E-Lins field engineering — condensation-driven intermittency is the most under-reported outdoor failure mode we encounter. |
What the H820QO Is Not: the Honest Part of the Review
A review that only lists strengths is marketing. Here is where this router is the wrong choice, and in each case what we would recommend instead.
| Limitation | Why it matters | Choose instead |
|---|---|---|
| Only 2 Ethernet ports LAN/WAN × 2 | Fine for a single camera or a single controller on a pole. Not a site aggregation point. | H820 — five Gigabit Ethernet ports, IP30, for in-cabinet aggregation |
| No serial port | No RS232 or RS485. If you need to read a meter, a PLC or a legacy controller over serial, this is not the device. | H820 (DB9 RS232/RS485) or H900f (terminal socket serial) inside a cabinet |
| Single SIM standard; dual SIM is optional | Single-carrier sites have single-carrier outages. If the site matters, specify dual SIM. | Order the dual SIM online option, or use H900f with dual modem |
| Modest compute Qualcomm chipset, 1 Gb DDR2, 128 Mb flash | Right-sized for routing, VPN and telemetry. Not an edge-compute platform for containers and local analytics. | Consider a separate edge compute node alongside the router |
| Humidity rating is non-condensing | As Section 04 explains — this is an industry-wide caveat, not a unique weakness, but it is a real design constraint. | Same caveat applies across the range; solve it at the enclosure level |
Worth stating plainly: every other router in the E-Lins range is IP30, including the H820, H900, H900f, H900pf, H900t6 and H720. IP30 means protected against solid objects larger than 2.5 mm and no protection against water at all. Those devices are excellent inside a cabinet and wrong bolted to a pole. If your design puts a router outdoors with no enclosure, the H820QO is currently the only device we make that is intended for it.
Specifying the Work, Not Just the Box
The most useful thing we can give you is not a product recommendation. It is the acceptance checklist we wish every outdoor router pole mount installation used, because in our experience it separates installations that come back from installations that do not.
- Every cable enters the bottom face. Not the side, not the top. Gravity is on your side only at the bottom.
- IP68-rated cable glands, sized to the measured jacket diameter. Not the nominal cable size — measure it. And a blanking plug in every unused hole.
- A drip loop on every external cable, formed below the entry point so water leaves at the low point.
- Torque to specification, on the lid, the glands and the RF connectors. “Tight” is not a value. Under-torqued glands leak; over-torqued terminals crack.
- A drain or breather plug at the lowest point of the enclosure, plus a hydrophobic vent membrane so the box can breathe without admitting liquid water.
- Mount the router high in the enclosure, never at the bottom, and away from other heat sources.
- Size the fuse to the supply voltage: 3 A for 9–24 V DC, 2.5 A above 24 V DC, and always inline for vehicle installations.
- Dielectric grease and self-amalgamating tape on every outdoor RF joint.
- Stainless or hot-dip galvanised fasteners at coastal and de-iced-road sites. Zinc-plated is not a ten-year fastener outdoors.
- Photograph every sealing interface before closing the box. It costs ninety seconds and it is the only record you will have when something leaks in year three.
Point ten deserves emphasis because it is the cheapest and most-skipped item on the list. When an outdoor installation fails after the warranty conversation has become awkward, the question is always “was it sealed correctly?” and by then nobody can answer it. Photographs answer it.
Verdict: Who Should Buy It, and Who Should Not
| Assessment area | Our read | Comment |
|---|---|---|
| Ingress protection (IP68) | Strong | The only IP68 enclosure in the range, with UV-stabilised PC and anti-oxidation treatment. Real outdoor credentials. |
| Operating temperature | Strong | −35 to +75 °C ambient air, −40 to +85 °C storage. Remember “ambient” means at the device, not at the weather station. |
| Condensation behaviour | Caveat | Rated non-condensing, as is normal. Manage it at the enclosure level — vent, drain, desiccant, mount high. |
| Power flexibility | Strong | 5–40 V (5–60 V option), dual input with failover, ISO 7637-2, ignition sensing. Order the 5–60 V version if you use PoE. |
| RF flexibility | Good | Built-in antennas suit free-air mounting; N-type external option for cabinet or weak-coverage sites. MIMO supported. |
| Port count and I/O | Limited | Two Ethernet ports, no serial, no DI/DO. Deliberately single-purpose. |
| Cellular generation | Caveat | 4G LTE only. Correct for telemetry and CCTV backhaul; wrong for a ten-year high-throughput plan. |
| Manageability | Strong | NMS, TR-069, SNMP, SSH, SMS control and per-site data alerts — the features that actually prevent truck rolls. |
Buy the H820QO if you need a cellular gateway mounted where there is no cabinet and you cannot build one: a pole, a wall, a rooftop parapet, a roadside kiosk exterior, a temporary or rapidly deployed site, a single camera or a single controller, or a coastal or industrial site where the enclosure itself is the risk.
Do not buy it if you need more than two Ethernet ports, or any serial connectivity, or 5G, or a device that will aggregate a whole site. In those cases the right answer is an enclosure plus a cabinet router: H820 for aggregation with serial and five Gigabit ports, H900pf where the cabinet must supply PoE to cameras or terminals, H900f for dual-modem 5G throughput, H685frc — a 5G RedCap module at just 100 × 60 × 21 mm and 220 g — where the router has to disappear inside someone else’s equipment, and M400 at 83 × 53 × 25 mm where you need RS232/RS485/USB and GNSS in a very small space.
The bottom line
The H820QO is a well-executed answer to a narrow, real question: how do you put a cellular router outdoors with no cabinet? Its enclosure, temperature range, power input flexibility and management features are all genuinely fit for that job, and it is the IP68 outdoor industrial 4G router we would shortlist for it. It is not a site router, it has no serial, and it is not a 5G platform.
The larger point of this review is not about one product. Ratings are necessary and insufficient. The enclosure is not the environment; the installation is. If you take one thing away, take the tender line, not the spec sheet: internal enclosure air shall not exceed 70 °C at design ambient, and equipment shall cold-start unaided at the site minimum after an eight-hour soak. Write those two sentences and most of this article becomes unnecessary.
Frequently Asked Questions
Q1：Is IP68 enough to mount the router in the open with no cabinet at all?
For the enclosure itself, yes — that is exactly what the H820QO is built for, and it is the only device in the E-Lins range with an IP68 enclosure. What IP68 does not remove is the need for correct cable glands, correct torque, UV-appropriate cable jackets, and surge protection on the DC and antenna lines. The rating covers the box; the installation is still yours.
Q2：Can I power it from a standard 48 V PoE switch?
Only on the 5–60 VDC version. The default input range is 5–40 V DC, and 802.3af/at PoE is nominally 48 V. The datasheet states this explicitly — “if the POE voltage is 48V, please order 5-60VDC version” — and it is the single most common ordering error on this product. If your site uses PoE, specify the 5–60 V version when you order.
Q3：Does it have RS485 for my energy meter or PLC?
No. The H820QO has two Ethernet LAN/WAN ports, an on-board console port, and no serial interface. If you need RS232 or RS485, use an H820 (DB9 RS232/RS485, five Gigabit ports) or an H900f (serial on a terminal socket) inside an enclosure, or an M400 where you need a compact serial-to-cellular device with GNSS.
Q4：Do I need an external antenna?
Not if the router is mounted in free air — the built-in high-gain cellular and Wi-Fi antennas are designed for that. You do need the external option if the router sits inside anything conductive, or if the site has weak coverage and you want a high-gain directional antenna. If you go external, budget for feeder loss: LMR-400 rather than RG-58 on anything but a very short run.
Q5：What fuse should I use on a 12 V or 24 V vehicle installation?
Per the datasheet: an inline fuse is required for vehicle installations, 3 A for 9–24 V DC and 2.5 A above 24 V DC. The device also has ignition sensing with automatic ON and time-delay OFF, so it starts with the vehicle and shuts down on a delay rather than abruptly.
Q6：How cold can it really start?
The rating is −35 °C ambient air for operation and −40 °C for storage. The practical caveats are that the device cold-soaks to true ambient (its own dissipation buys roughly a kelvin at that temperature), that DC voltage drop along a long cable run is worst exactly at cold-start, and that opening a cold enclosure on a humid day is a reliable way to put condensation on the board.
Specifying an outdoor or harsh-environment site?
Send us your location, design ambient range, mounting situation and what the router has to connect to. We will come back with an enclosure recommendation, a thermal and power budget, and an installation checklist you can put straight into a tender.
