PD will not power up on the port
very common
“Plugged the camera in and nothing. No light, no boot, dead. The switch acts like there's nothing on the port at all.”
Likely causes
-
PoE disabled on the port, or the port sitting in a denied or latched error state
20%The switch shows the port with power off, denied or in a fault state, and a PD you know is good does nothing there either, but the same PD powers immediately on the port beside it. Denied for budget is Switch out of PoE budget; a fault state that survives a link bounce is Port stays dead after you fixed it.
-
Open, short, or conductors crossed between pairsets on the pairs the PSE is using
20%Wiremap shows an open or short on the pairset the PSE feeds, or conductors crossed between pairsets so 4/5 and 7/8 are no longer intact pairsets. A 4-pair patch cord run straight from switch to PD powers it fine. Common where a jack was terminated data-only by another trade. Note what is NOT here: a polarity reversal within a pair is not a PoE fault.
-
Type mismatch: PD needs 4-pair Type 3/4 and the PSE or a midspan in the path only delivers 2-pair
20%Switch detects a class then reports insufficient power, or the PD blinks once and dies. PD datasheet calls for 802.3bt. Check for an old midspan quietly sitting in the path, and check how many pairsets the port reports.
-
Passive or non-standard PD that presents no valid detection signature
15%PD is labelled passive PoE or ships with a barrel-jack injector. The switch logs a detection or invalid-signature failure, never an overload, and the manufacturer's own injector powers it with no argument. A compliant PSE refusing this device is correct behaviour, not a fault to configure around.
-
Legacy capacitive-signature PD on a PSE with no pre-standard detection
5%Pre-802.3af era gear: the old Cisco inline-power population. Detection fails, and the platform either has a legacy or pre-standard detection setting or it does not. This is the only case where that setting is the answer.
-
Moisture or a damaged sheath putting leakage across the pairs: outdoor and OSP drops
10%Resistance to ground reads in the kilohms instead of megohms. Detection succeeds intermittently then the port reports a short or overload. Water in the enclosure, the jack or the low point of the run.
-
Dead or damaged PSE port
10%A known-good PD fails on that one port and powers on every other port, and the suspect PD powers everywhere else. If the port is dead in both directions (no link either, not just no power), suspect damage from non-compliant hardware in the path and go to Injector or splitter killed the port.
What to bring
- Switch CLI or vendor app for inline power status
- Two known-good 4-pair patch cords
- Certifier with wiremap and DC loop resistance
- PoE loaded tester that draws real load, not just voltage
- Manufacturer injector for the PD, or a known-good compliant midspan
- Multimeter
Steps
-
Step 1: Swap the patch cords at both ends for known-good 4-pair cords
Power and link come up: the old cord was 2-pair, 28 AWG or damaged. Half the cords loose in a TR are one of those. Costs nothing and takes a minute.
If that doesn’t do it
Move the PD to a port you know delivers power
-
Step 2: Move the PD to a port that is currently powering something of the same class, and look at which ports work
PD powers there: the fault is the port, its config or the switch budget, not the device. Still dead: the PD or the horizontal run. Before you move on, check whether every powered device on that switch sits inside one numeric block, because that is a capability boundary rather than a fault.
If that doesn’t do it
Read the port's inline power state, or go to That port was never a PoE port if the working ports form a block
-
Step 3: Read the port's inline power detail and write down exactly what it says
Admin off, power denied for budget, detection failed (signature or wiring), overload or short (wiring or water), or a latched fault state that a link bounce will not clear. Note class detected against power allocated: they tell you different stories.
If that doesn’t do it
Confirm the port config for PoE, Type and, only for a legacy PD, pre-standard detection
-
Step 4: Confirm the port is enabled for PoE at the highest Type the switch supports and clear any per-port maximum
Legacy or pre-standard detection is for capacitive-signature pre-af PDs only. It will never power a passive PoE device and no other PSE setting will either: that device needs its own manufacturer injector, and that injector must never be connected to a switch port in either direction. Syntax, Type support and whether legacy detection exists at all vary by platform, so look up the vendor's inline power reference for that model.
If that doesn’t do it
De-energise the channel, then wiremap the permanent link
-
Step 5: Disable PoE on the port, confirm the channel is dead, then wiremap the permanent link end to end
All four pairs continuous, no shorts, no splits, and no conductors crossed between pairsets. Pair polarity reversal is not a PoE fault (the PD rectifies each pairset and takes either polarity), so do not burn a trip re-terminating a reversal expecting power. An open on 4/5 and 7/8 kills power outright only with an Alternative B PSE, a midspan, or a PD that demands 4-pair; on a typical Alternative A endspan, 2-pair power survives and it presents as low power or a Type 3/4 refusal instead.
If that doesn’t do it
Go to PD powers but runs in low power mode if 2-pair power survives, otherwise measure resistance
-
Step 6: With the channel still de-energised, measure DC loop resistance and resistance to ground
Compare against what length and gauge predict rather than against a remembered limit: 24 AWG solid runs about 8.4 ohm per 100 m per conductor, so a full-length pairset (two pairs paralleled each way) lands near 8 to 9 ohm. Know whether your certifier is reporting per-pair or per-pairset resistance before you compare it to anything, because a per-pair figure reads roughly double. Anything to ground below megohms is moisture or a nicked conductor.
If that doesn’t do it
Re-terminate both ends, then replace the run
References
- IEEE 802.3-2022 Clause 33
- IEEE 802.3-2022 Clause 145
- ANSI/TIA-568.2-D Clause 6 (check for a later revision)