Outdoor PoE voltage drop decision tree — cameras reboot in rain or at night

Q: How do I confirm it is a voltage-drop problem with a multimeter?

Measure the actual voltage at the camera end (at the PoE splitter or cable termination) against the standard: 802.3af ~48V, 802.3at ~50-57V. If you read only the 30s of volts or lower, the line from switch to camera is dropping too much. Then measure the same port at the switch end; if the switch end is normal but the camera end is low, the drop is confirmed. Measure during the fault, not after it has recovered.

There is a headache outdoor-surveillance failure: a few bullet/domed cameras behave fine on dry sunny days, but reboot frequently in rain or at night, with heavy snow-noise on the image. The first reflex drifts toward "the camera got water in it" or "the network is unstable", so people swap cameras and tweak APs — money spent, problem remains.

The signature of this fault is "reboots/snow track the weather and time of day, and the root is power margin". A PoE-powered camera on an aged or over-distance line drops voltage until it falls below threshold and the camera power-cycles. The decision tree below handles the "outdoor PoE insufficient power" class.

60-second check — is it insufficient power?

Self-test with one sentence:

Camera reboots/snow concentrate in rainy / low-temperature periods; fine on sunny days; and the reboot times are suspiciously "on schedule" → strongly suspect PoE end voltage insufficient (excessive drop).

If reboots are random all day and unrelated to weather, look at the device or network itself — outside this page.

Troubleshooting decision tree

User reports: outdoor camera reboots in rain/night, snow-noise image
  │
  ├─ Look at the pattern
  │    ├─ Reboots concentrate in damp/cold periods? → power-margin suspect (check power)
  │    └─ Random all day, unrelated to weather?    → check device or network itself
  │
  ├─ Measure camera-end voltage (key)
  │    ├─ Measured far below 48V (af) / 50-57V (at)? → excessive drop ★most common
  │    └─ End voltage normal?                        → not power, move to network/device
  │
  ├─ Switch-end same-port output vs camera-end
  │    ├─ Switch end normal, camera end low? → line drop confirmed
  │    └─ Both low?                  → switch port power / PSU itself
  │
  ├─ ping -t the camera IP
  │    ├─ Loss lines up with reboot times? → reboot causes outage, corroborates power
  │    └─ Loss unrelated to reboot?     → real link problem
  │
  └─ Root-cause locating
       ├─ Cable over 100m / aged damp? → replace, add PoE extender/injector
       ├─ Adapter aged damp?         → industrial PSU / weatherproof box
       └─ Switch port power insufficient? → industrial PoE switch

The hinge of the whole tree is one line: first confirm "does the reboot track power margin", then decide whether to check voltage or network. That step "measure camera-end voltage during the fault" is the pivot — the drop will not wait for you to be ready; you must catch it while it happens.

Three-step locating procedure

Step 1 — catch "during the fault" end-voltage

:: Not a terminal command — field measurement notes (use a multimeter / PoE tester)
:: 1. Measure actual voltage at the camera end (PoE splitter or cable termination)
:: 2. Standard: 802.3af ~48V, 802.3at ~50-57V
:: 3. Reading drops by ten-odd volts or more → excessive drop
:: 4. Measure DURING reboot/snow, not after recovery (voltage returns to normal then)

Note: drop is set by load + line together; fault-state and recovered-state voltage can differ a lot. Catch a voltage frame at the moment the symptom appears.

Step 2 — switch-end vs camera-end comparison

1. On the PoE switch port see "output power/voltage" (commands differ by vendor)
2. Switch end normal, camera end clearly lower → line drop confirmed
3. Both low → problem is switch port power or chassis PSU, not the cable

This step separates "cable problem" from "switch problem" and decides whether to replace cable or switch.

Step 3 — ping -t to see if loss syncs with reboot

ping -t <camera IP>        :: watch whether timeouts cluster around reboot periods

If the timeout times line up with "device reboot" and reboots track the weather, power is locked in. A pure link problem rarely "follows the weather" this precisely.

Emergency stopgap — restore power temporarily

This is a field-handling checklist (not a terminal script; PoE power must be fixed on the cable/device):

1. Power this camera from a nearby independent PoE injector (shorten the path, reduce drop)
2. Or move the camera to a closer, higher-output switch port
3. Temporarily bridge the over-long cable with a better-quality segment to cut resistance
4. After recovery, watch one or two damp/cold cycles to confirm no more reboots

Honest note: fixing PoE power means "replace cable / replace switch / change power topology" — there is no copy-run script. The above is a checklist for field troubleshooting and temporary recovery, not code; cable spec and gear depend on your site.

Root cause — make weather-driven failures rarer

Recovering one camera is only a stopgap. Outdoor low-voltage sits in the fastest-aging environment; close it off in three layers:

Layer	Action	Prevents
Cable	Replace over-long/aged outdoor cable with better-shielded, thicker-conductor outdoor cable; add PoE extender/injector beyond 100m	long-line + aging drop breaking threshold
Device	Industrial PoE switch outdoors, surge-protected; connectors in a weatherproof box	sun/rain accelerated aging, surge burnout
Ops	"Outdoor weak-current box rainy-season inspection plan": pre-rain end-voltage check, damp-connector check, standing-water clearing	turn reactive repair into proactive fewer incidents

One often-missed point: PoE is more sensitive to drop than data. The same 100m-plus line may still pass Fast Ethernet, but PoE end voltage has already dropped below threshold. So "ping works" does not mean "power is enough" — the voltage step cannot be skipped.

Quick decision table (use it directly)

Symptom	First reflex (misjudged)	Real root cause	Where to look
Reboots only in rain/night	camera water damage	PoE end voltage insufficient (drop)	measure camera-end voltage during fault
Snow-noise image	network unstable	under-voltage disturbs image module	end voltage + switch port power
Fine sunny, dead in rain	signal interference	damp worsens aged-line drop	damp connectors + line resistance
Intermittent ping timeouts	link broken	reboot causes outage	does loss sync with reboot times
Swapping camera did nothing	camera quality issue	root is line / power	switch-end vs end voltage compare

This table is the landing version of "what it is most easily misjudged as" — eight out of ten tickets of this class open as "camera broken" or "bad network", and checking against this table skips most wasted effort.

FAQ

Q: Why does it only reboot in rain and at night, but work fine on dry sunny days?
A: Rain/night stack two factors: high humidity and damp connectors lower the aged cable's insulation so voltage drop worsens; and lower night temperatures make some adapters/connectors contract and lose contact, dropping voltage further. On a dry sunny day the drop is just barely acceptable; once it gets damp and cold it falls below the camera's under-voltage threshold, so the camera reboots repeatedly. That is classic insufficient power margin, not coincidence.

Q: How do I confirm it is a voltage-drop problem with a multimeter?
A: Measure the actual voltage at the camera end against the standard: 802.3af ~48V, 802.3at ~50-57V. If you read only the 30s of volts or lower, the line from switch to camera is dropping too much. Then measure the same port at the switch end; if the switch end is normal but the camera end is low, the drop is confirmed. Measure during the fault, not after it has recovered.

Q: Does ping -t packet loss relate to insufficient power?
A: Yes. An under-voltage PoE camera drops power and reboots first; the network goes down during reboot, showing as intermittent ping -t timeouts. If the loss lines up with the reboot times and the reboots track damp/cold periods, it is power, not the link itself. A pure link problem rarely changes "on schedule" with the weather.

Q: Must I replace cable over 100 meters?
A: PoE is more sensitive to drop than data. Fast Ethernet may still pass beyond 100m, but PoE over a long aged line drops end voltage below threshold directly. The spec caps a single Ethernet segment at 100m, and PoE needs more margin in practice. For over-long or aged outdoor cable, prefer a better-shielded, thicker-conductor outdoor cable, or add a PoE extender/injector mid-span.

Q: Are industrial PoE switches and weatherproof boxes worth it?
A: Very much so outdoors. Ordinary switches and connectors age fast under constant sun and rain; industrial gear has wider temperature range and surge protection, and a weatherproof box markedly extends life and cuts weather-driven failures. It turns "reactive repair" into "proactively fewer incidents", and works best with a rainy-season inspection plan.