The Verdict, the Alternatives, and the Test That Decides It

Introduction

The question was whether the GL-X3000 and a good antenna are worth getting. The answer is conditional, and the condition is knowable — it just has not been measured yet. This volume gives the alternatives an honest hearing, states the recommendation, and describes the cheap test that turns a conditional answer into a decision.

5.1 The alternatives, fairly assessed

5.1.1 A cell signal booster

A booster — the weBoost Drive Reach is the standard recommendation for vehicles, chiefly for its uplink power — amplifies the outside signal and rebroadcasts it inside the vehicle, helping every device at once with no plan changes and no SIM juggling.

Its advantages are real and specific. It helps phones, not just a router, which means it improves voice calls — something no amount of router MIMO does. And in genuinely marginal conditions it can lift a signal that a passive antenna cannot, because it is adding gain rather than merely collecting better.

Its disadvantages are equally real. A booster operates on only a handful of the many bands in use. It inserts an amplification stage that adds noise and latency, and it cannot preserve MIMO — the rebroadcast signal reaching the router’s internal antennas is a single spatial stream, so a booster and a 4×4 antenna array are alternatives, not partners. One vendor’s summary is blunt: compared to a roof antenna feeding a router directly, a booster degrades data performance in the large majority of situations, while being better for voice and for the extreme fringe.

The rule that emerges is clean: a booster is a voice-and-extreme-fringe tool; a MIMO antenna is a data tool. For this build, where the inReach already covers emergency communication and the want is data, the antenna is the better fit.

5.1.2 Starlink Mini on a Roam plan

This is the alternative that actually addresses the stated problem, and it must be taken seriously rather than dismissed as expensive.

The Starlink Mini is roughly $199–249 for hardware, with Roam service starting around $55 per month for 100 GB, about $80 for 300 GB, and unlimited well north of that. It works in motion under Roam. It is the size of a laptop, draws modest DC power, and — decisively — it works where there are no towers.

Its limitations matter in the Keweenaw specifically: it needs sky, and the Keweenaw is heavily forested. Under a dense canopy at a dispersed site among mature hardwoods, the Mini will struggle, and a clearing may have to be part of choosing where to camp. Performance in motion on winding, tree-covered roads is interrupted.

But the comparison is stark. Where there is no cellular coverage, the entire GL-X3000 investment returns zero and the Starlink returns a working connection. No antenna decision changes that.

5.1.3 Doing nothing

Worth stating because it is a legitimate answer. The truck already has a phone with Android Auto, a tablet with offline Gaia maps, and an inReach Mini for emergencies. Downloading maps and forecasts before leaving pavement covers the actual functional need. Everything discussed here buys comfort, not capability.

5.2 The recommendation

Conditional, and the condition is which failure mode the Keweenaw week actually was.

If the Keweenaw had one flickering bar, or service that worked outside the truck and failed inside it — buy the GL-X3000 and a proper four-port antenna. That is a reception failure, this equipment is the correct fix, and the dual-SIM AT&T-plus-T-Mobile configuration addresses the coverage variability on top of it. Expect roughly $380 for the router, $200–400 for an antenna that genuinely covers 600 MHz, and $50–100 for feedline and connectors — call it $650–900 installed, plus service.

If the Keweenaw had zero bars, consistently, from a phone held outside the truck — do not buy the router expecting it to fix that trip. Buy a Starlink Mini instead. It is cheaper in hardware, it addresses the actual failure, and it is the only thing on this list that produces a connection where no tower exists.

The strongest case, and the one the equipment list is already drifting toward, is both — in that order. They solve different problems and neither substitutes for the other. Cellular is faster where it exists, uses less power, works under tree cover, and works while moving; satellite works where cellular does not. A rig that travels the Upper Peninsula seriously ends up with both, and the sequencing question is simply which failure is more common on the trips actually taken. On the evidence of one Keweenaw week, that is satellite.

Two things should be bought regardless of which way the decision goes:

A proper external antenna is worth having even if the router is deferred, because it is the component that will still be correct in five years when the router has been replaced. Antennas do not obsolete; modems do.
The prepaid or suspendable service strategy from volume four, because paying for twelve months of two SIMs to support six weeks of travel is the most common way this equipment becomes a bad investment.

Two things should not be bought:

The GPS element on the antenna — it costs receive diversity on a shared port, and the truck already carries four GNSS receivers. Volume three covers this.
A booster in addition to the MIMO antenna, expecting them to combine. They do not; the booster defeats the MIMO.

5.3 The test that decides it

This costs nothing and takes one trip.

Return to, or recall precisely, the places in the Keweenaw where service failed. Note them on the Gaia track — the answer is location-specific and a county-level average does not settle it.
At each, hold the phone outside the vehicle, in the open, elevated, and look at the actual signal reading rather than the bars. On Android this is Settings → About phone → SIM status, or a field-test app. The number wanted is RSRP in dBm.
Interpret it:
- Better than about −110 dBm — there is a usable signal and the cab is what is killing it. This is the reception failure. Buy the router and antenna; they will work.
- Between about −110 and −120 dBm — marginal. A roof antenna will probably convert this into a working connection, and this is precisely the case the equipment exists for.
- Worse than −120 dBm, or “no service” with no reading at all — there is nothing to receive. This is the coverage failure. No antenna fixes it. Buy the Starlink.
Do this on more than one carrier if possible — a borrowed phone on a different network at the same spot is the single most informative measurement available, because it directly tests the dual-SIM premise.
Cross-check against the FCC’s national broadband map rather than carrier marketing maps, which are consistently optimistic about rural coverage.

5.4 Recording the outcome

Whatever is measured and whatever is bought, it belongs in this dive as a new volume rather than in a text file. The measurements above are the most useful thing this project can produce about connectivity, because they are specific to the terrain actually travelled and no review can supply them.

References

GL.iNet — Spitz AX (GL-X3000) product specifications and router documentation.
GL.iNet community forum — external antenna threads, including staff clarification of the internal 5 GHz Wi-Fi antenna.
Quectel RM520N-GL module band tables.
Mobile Internet Resource Center — cellular antenna and mobile router reviews.
MobileMustHave — roof antenna versus booster technical comparison.
Waveform — cell phone signal booster guide and GL-X3000 antenna guide.
Coverage aggregations for Keweenaw County, Michigan; the FCC National Broadband Map as the authoritative cross-check.
Starlink Roam and Starlink Mini published plans and pricing.
Coaxial cable attenuation figures for RG-58 and LMR-240.