Industrial Radio Telemetry. The invisible backbone of every remote operation.
Cellular has signal. Fiber has reach. Satellite has resilience. But across millions of remote tank batteries, wellheads, lift stations, and substations, the data is moving over licensed industrial radio — point-to-multipoint links that have been quietly carrying telemetry for forty years. This article is the practical guide: what the bands are, what choosing between them costs, how degradation actually happens, and why predicting it matters more than chasing redundancy.
The 30-second version
Industrial radio telemetry is the use of RF links — licensed or unlicensed, point-to-point or point-to-multipoint — to move data between remote field equipment and a central SCADA host. It's the dominant comms layer for upstream oil and gas, water utilities, pipelines, distribution-grid SCADA, and any operation where field sites are too dispersed for fiber and too remote for reliable cellular.
The defining tradeoffs: licensed bands give you interference protection and long range at the cost of FCC paperwork and per-site fees. Unlicensed bands(900 MHz ISM, 2.4 GHz, 5 GHz) are zero-paperwork but you share the spectrum with consumer gear. Cellular is faster and easier but you don't own the link.Satellite reaches anywhere but costs more per byte and adds latency.
The operator's question is rarely "which is best?" It's almost always "which combination gives me primary + failover with acceptable cost-per-site and acceptable degradation visibility?" Mixed-mode is the operational reality.
The bands you'll actually see
US-centric. International allocations differ — Canada and Mexico track US bands closely; Europe and Asia have their own ITU regional allocations.
The four cost dimensions that actually matter
Comparing radio comms options on the spec sheet ("VHF range 30 miles vs UHF range 20 miles") misses the operational economics. Four cost dimensions matter:
| Dimension | What it covers | Where it bites |
|---|---|---|
| CapEx per site | Radio hardware, antenna, mount, install labor | Licensed bands need higher-power radios + better antennas: $1.5k-$5k/site vs $400-$1,500 for cellular gateway |
| OpEx per site per month | Spectrum fees, cellular plans, satellite minutes | Licensed: $10-$50/yr per frequency. Cellular: $15-$40/mo. Satellite: $50-$300/mo. Compounds at fleet scale. |
| Tower / infrastructure | Repeater sites, master stations, tower leases, backhaul | Privately-licensed SCADA: you own/lease towers. Cellular: carrier owns. Satellite: no infrastructure burden. |
| Outage cost | Lost production, regulatory exposure, dispatch costs | The hidden killer. A 6-hour radio outage on a custody-transfer site means lost flow measurement and a PHMSA reportable event. The radio with the lowest CapEx is rarely the cheapest after outage costs. |
How radio links actually degrade
Radio failures rarely happen all at once. They happen in slow motion — over weeks or months — and the difference between "we caught it" and "we got a 2 a.m. call from dispatch" comes down to whether anyone is watching the right metrics.
RSSI drift
Received Signal Strength Indicator. The single most-watched radio metric. Falls when the antenna deteriorates (corrosion at the connector, water in the line), the antenna shifts (wind, vehicle impact, lightning), nearby foliage grows in the Fresnel zone, or multipath conditions change seasonally. A 6 dB drop is "noticeable." A 15 dB drop is "you're losing the link soon."
SNR collapse
Signal-to-Noise Ratio. RSSI can be holding steady while SNR collapses because a new interference source appeared nearby — a neighboring operator on the same frequency, a cellular booster install down the road, a defective LED highway sign. SNR is the metric that catches interference; RSSI alone won't.
Retry / error-rate climb
Most industrial radios surface retry counts or packet-error rates. A climb from "0 retries per day" to "150 retries per day" doesn't lose the link — but it doubles the effective channel occupancy and signals degradation. Operators who only look at comms-status booleans miss this completely.
Latency / jitter creep
On cellular links especially, average round-trip latency drifting from 80ms to 200ms over a month signals carrier-side issues that will become outages. Hardly anyone instruments this proactively.
The pattern operators see again and again: a remote site that was "fine yesterday" goes hard-offline today, and the post-mortem shows RSSI was drifting -2 dB per week for the previous two months. Nobody was looking at the trend. The radio was telling them. The SCADA was rendering only "comms OK / comms FAIL."
Redundancy patterns that actually work
"Redundant comms" is one of the more deceptive phrases in field SCADA. Four patterns in increasing order of effectiveness:
1. Cold standby (most common, weakest)
Two radios at the site, one active. If the primary fails, an operator dispatches and re-cables. Counts as "redundant" on a spec sheet but provides ~hours of recovery time and zero protection against the failure that downed the primary (lightning, antenna loss).
2. Hot standby with watchdog
Two radios, watchdog circuit failovers within seconds. Works for hardware failures. Still vulnerable to band-wide interference and tower outages.
3. Dual-path (different bands)
Primary licensed UHF + cellular failover. Path diversity. Costs more in OpEx but covers both hardware and band-wide interference. The most common production-quality pattern.
4. Dual-path with predictive degradation
Same as above, plus continuous monitoring of RSSI/SNR/retry trends on the primary, with automatic preemptive failover when degradation crosses a threshold rather than waiting for hard failure. The threshold and the analytics aren't shipped with the radio — they're an analytics layer added on top.
That fourth pattern is what Aevus enables. We ingest the radio's RSSI, SNR, retry, and link-state telemetry, model degradation across the fleet, and flag sites entering a failure window two to four weeks before they go hard-offline. The operator gets time to dispatch a planned maintenance visit instead of a 2 a.m. emergency.
Spectrum hygiene — the operations no one talks about
Three operational practices that separate a healthy radio fleet from a chronically unstable one:
- Annual spectrum survey. Walk a portable spectrum analyzer (or SDR-based replacement) across your operating bands at each major site at least annually. Catches new interferers before they cause outages.
- Cable / connector recertification. Antenna systems lose performance from corrosion at the connector, water ingress through cable cuts, and lightning damage. Pull return-loss measurements every 2-3 years. Replace before failure, not after.
- Licensed-frequency renewal calendar. Part 90 licenses are 10-year. FCC paperwork failure does not result in a "renewal reminder" — it results in non-renewal and forfeiture of frequency. The number of operators running on expired licenses is non-trivial.
Where Aevus reads radio telemetry
Aevus's radio-health monitoring is one of the most differentiated capabilities in our platform — and the focus group explicitly named it as a top operator pain point. Four inputs Aevus correlates:
- Radio diagnostics (RSSI, SNR, retry, link-up). Pulled from the radio's management interface, the gateway it's connected to, or the SCADA front-end that polls it.
- Field-device polling latency. If the same RTU's poll responses are taking longer over time, that's a radio path issue manifesting at the SCADA layer before the radio itself reports.
- Environmental data. Weather (heavy moisture in line of sight), local cellular tower congestion (carriers publish congestion data), seasonal foliage models for VHF/UHF Fresnel zone fill-in.
- Fleet-wide patterns. If 12 sites on the same repeater are degrading simultaneously, the repeater is the problem — not the sites. Aevus's correlation engine catches this across the fleet; individual site monitoring doesn't.
What Aevus never does: reconfigure your radios. The IL-9000 boundary denies any write to field equipment, including radio management. We surface the degradation, recommend the dispatch, and let the operator make the call. Full architectural brief →
That's industrial radio.
If your fleet runs hundreds of remote sites on licensed UHF, 900 MHz ISM, cellular, or any combination — and you want to know which radios are failing before they fail — that's the conversation we're built for.
Related Articles
How industrial telemetry works from sensor to historian.
8 radios degraded for 2 months unnoticed.
MRP, DLR, PRP, HSR — failover times and where each fits.
Ready to see how this applies to your operation? Start a pilot conversation — no commitment, no field changes.