All guides

informational / troubleshooting

Starlink Obstruction Map: How to Read It and Fix Warnings

Learn how to run and read the Starlink obstruction map, judge direction and repeatability, fix placement, and handle false obstructed warnings on one page.

Starlink Obstruction Map: How to Read It and Fix Warnings

The Starlink obstruction map is a sky diagram, not a verdict that your service is doomed. It shows where the terminal’s view is blocked, or where a phone-based scan thinks it will be blocked. Read direction and repeatability first. A one-off red patch while a truck is in the driveway is not the same as a tree that clips the same sector every evening.

This guide also covers the common case where the app says obstructed, but the sky looks clear. Use the official app for the scan, then compare the result with the real installation and repeated behavior.

What the map is for

A Starlink dish does not stare at one geostationary point. It works a moving field of satellites. Anything that occupies part of that field—trees, a steep roof, a neighboring wall, a mast, a parked RV—can interrupt a handover. The obstruction view exists to show where in the sky those hits occur so you can move the dish, raise it, or accept a known compromise.

It is not:

  • a live weather radar;
  • a guarantee of speed;
  • a substitute for a speed test;
  • proof that Starlink is “down” in your county.

If the dish cannot see enough sky, you will feel brief drops, slowdowns, or a long optimizing wait. If the map is clean and those symptoms remain, look elsewhere: disconnect classification, packet loss, Wi-Fi, or load.

How to run a useful scan

Use the current Starlink app on the kit you actually own. Menu labels move between app versions; the job does not. You want a view that plots obstructions around the dish, and—if the app offers it on your hardware—a phone scan that estimates the sky from a candidate mounting spot before you drill.

Do this:

  1. Stand where the dish is, or will be, not in the kitchen. A phone scan from the driveway measures the driveway.
  2. Give the dish time to settle after a move or reboot. A scan during boot is a portrait of confusion.
  3. Hold the phone still for a phone-based check. Wrist wobble paints fake wedges.
  4. Note the time, weather, and anything temporary in the scene: vehicles, people, a ladder, a seasonal decoration, snow on the dish.
  5. Run more than one scan on different hours if the first result will justify a saw or a new mount. LEO geometry changes through the day.

Save or screenshot the plot if the app does not keep history you like. You are going to compare plots, not admire a single image.

How to read direction, not just color

Most obstruction visuals are a circle or dome: center is overhead, rim is toward the horizon, and clock positions are compass directions. Your app’s legend is the authority for which way is north. If you assume “up on the phone is north” without checking, you will prune the wrong tree.

Read the plot in this order:

Where is the mark? A blob in one sector is a physical object in that direction. A thin ring near the horizon may be trees or buildings at the edge of usefulness. A spray of specks all around can be a bad scan, snow, heavy rain, or a dish that was not ready.

How high is it? Obstructions near the horizon are often tolerable if the rest of the dome is open. A block near zenith is a much bigger deal—overhangs, a too-close roof, or a dish sitting under an eave.

Does it match what you see? Walk to the dish and look. A chimney to the southwest should land in the southwest sector. If the map screams east and the only tree is west, you are not looking at a trustworthy plot yet.

Does it repeat? The useful question is not “is there red?” It is “is the same sector red at 8 a.m. and 8 p.m., on two weekdays, with no truck in the way?” Repeatable geometry is a placement problem. A unique smear is a candidate for a false warning.

Percent-style scores in the app, when shown, are summaries. They can hide a nasty narrow wedge that only hurts at certain hours. Prefer the picture.

Placement fixes that actually change the map

Move the dish to buy sky. That is the whole game.

Height. A few feet of mast can clear a fence line or a neighboring roof edge that a patio table never will. Height also increases wind load—choose a mount that can take it rather than a taller wobbly pole.

Setback from the house. Dishes tucked against a wall see a wall. Slide toward the yard, a ridge, or a railing with a wider dome.

Foliage. Leaf-on season is harsher than winter. A scan in March can lie about July. If the offending sector is a deciduous tree, decide whether you will prune, relocate, or live with seasonal drops.

Vehicles and seasonal clutter. Boats, trailers, and RVs migrate. A “perfect” winter install can fail when the camper comes home.

Indoor and window myths. Glass, window tint, and frames are obstructions. If you are testing on a sill, the map is allowed to look ugly. That is not a false warning; it is an honest reading of a bad site.

After a move, wait for the dish to finish adjusting, scan again, then measure in the hours that used to fail. A prettier map with the same evening drops means you fixed a picture, not the path—or the remaining issue is weather, load, or Wi-Fi.

“Obstructed, but it’s not”: false and misleading warnings

An apparently false warning belongs in the same diagnostic process as a visible obstruction.

The app can report obstruction when you are staring at blue sky. That does not always mean the software is broken. It usually means the terminal or the phone scan saw a blockage you are not counting, or a different fault is wearing an obstruction label.

Work through these before you argue with the plot.

Temporary objects you discounted

Delivery vans, your own truck, a neighbor’s dumpster day, scaffolding, and even a person standing beside the dish during a phone scan will print on the map. Rescan with the scene empty. If the mark vanishes, it was never a tree.

The phone and the dish are not the same instrument

A handheld scan estimates a future mount. The dish reports what it is experiencing now. They disagree when:

  • you scanned from a slightly different spot than the actual mount;
  • the phone lens saw a roof that the raised dish now clears, or the reverse;
  • the dish has snow, rain, or dirt on the radome that the phone camera did not treat as sky blockage.

Trust a settled dish plot over a stale phone screenshot from last weekend’s site walk.

Snow, ice, rain, and wet debris

A white sky can still be a blocked aperture. Snow on the face of the dish is an obstruction to the radio even when the horizon is photogenic. Clear it safely; use Snow Melt as described in the Snow Melt guide. Heavy rain can degrade the path and produce ugly statistics that users read as “obstructed.” Separate that from geometry using rain, storms, and the bad-weather pillar.

If the map is dirty only while precipitation is on the terminal, you do not have a false obstruction. You have weather on the antenna.

Alignment versus obstruction

People mix “it hasn’t finished pointing” with “something is in the way.” A dish that is still optimizing, recently moved, or fighting a loose mount can show poor sky while the real issue is that the terminal is not yet looking where it should. See stuck on optimizing connection. Stiffen the mount if the dish nods; a moving dish writes moving obstruction.

Stale maps and skipped rescans

Some users keep an old screenshot as gospel. Firmware, a physical nudge, a new vehicle, and leaf-out all age a plot. Reset or refresh in the app if it offers that control, then run a new scan. Do not prune from last autumn’s image.

Network and Wi-Fi failures mislabeled by the user

The app says online. A phone on weak mesh says the internet is gone. That is not dish obstruction. Classify it with Starlink keeps disconnecting. Ethernet will tell the truth faster than another obstruction scan.

Reflections, metal, and tight spaces

Nearby sheet metal, railings, and another dish can confuse a scan or add multipath that looks like a messy map. A too-small courtyard with sky only in a slot may be “clear” to the eye and blocked to the terminal. The eye is not a radio.

When it really does look like a bug

After empty-scene rescans, a dry radome, a finished optimization, and a wired test that still shows obstruction with a wide-open dome, you have a support case: document timestamps, screenshots, hardware generation, and what you already ruled out. Do not skip those steps and assume the map is haunted. Most “false” reports in this cluster fail one of the checks above.

A practical sequence

  1. Photograph the real sky from dish height, noting compass directions.
  2. Run a settled scan; read sector and elevation, not only a score.
  3. Remove temporary objects; rescan.
  4. If the same sector repeats, move or raise the dish, then scan and test again.
  5. If the sky is open and the map is dirty only in weather, treat weather—not a saw.
  6. If Wi-Fi is the only casualty, stop using the obstruction map as your tool.

Obstruction is a placement problem when the picture repeats. It is a language problem when “obstructed” is standing in for rain, snow, a reboot, or a phone that never had signal. Keep those meanings apart and the map becomes useful.

Primary next step: After a placement change or a “false obstructed” check, run a Starlink speed test on a wired client in the hours that used to fail. A cleaner map should show up as fewer incomplete runs and stabler latency, not only as a prettier circle.