Power over Ethernet (PoE) is a convenient technology that sends electrical power along with data over a single Ethernet cable. It lets us plug in devices like security cameras or Wi-Fi access points without needing separate power outlets. However, many users have encountered situations where a device labeled "PoE" refuses to power up when connected. This can be puzzling and frustrating. The truth is, not all "PoE" is created equal. There are several technical reasons why a PoE-capable device (the Powered Device or PD) might not get power from a PoE source (the Power Sourcing Equipment or PSE). In this article, we'll break down the common causes in an accessible way, with clear explanations and simple analogies. By the end, you'll understand issues like PoE standards vs. non-standard implementations, Mode A vs. Mode B power modes, power requirements mismatches, cable length limitations, and PoE power budget limits.
Standard vs. Non-Standard PoE
One of the first things to check when PoE doesn't work is whether both the PSE (like a PoE switch or injector) and the PD (your device) follow the same PoE standards. The IEEE has established several PoE standards to ensure compatibility: 802.3af (often just called "PoE"), 802.3at ("PoE+"), and 802.3bt (sometimes called "PoE++", which includes higher-power Types 3 and 4). These standards define how devices should detect each other and how much power can be delivered safely. For example, 802.3af (2003) provides up to about 15.4 W per port (12.95 W usable by the device), 802.3at (2009) raises that to 30 W (25.5 W usable), and the newer 802.3bt (2018) standard can deliver 60 W (Type 3) or up to 90 W (Type 4) per port for high-power. All PoE equipment adhering to these standards perform a "handshake" (signaling process) to ensure the PD is PoE-compatible and to negotiate how much power is needed. This handshake keeps everything safe and interoperable – an IEEE-compliant PSE won't send power unless it detects a valid PoE device on the line, and it won't exceed the device's power.
Non-standard PoE implementations, on the other hand, don't follow the IEEE rules. You might also hear these called passive PoE or just PoE-capable (as opposed to compliant). In a passive PoE system, the source doesn't do any negotiation – it simply injects a fixed voltage onto the cable pins, whether or not a compatible device is there. Common examples are some proprietary systems that provide 24V DC over Ethernet for certain wireless antennas or camera. If you plug a passive PoE injector into a device that isn't expecting that voltage, it's a bit like plugging a 120 V appliance into a 240 V outlet – you could permanently damage the device. Conversely, if a passive PoE device expects 24 V but you only have a standard 48 V PoE switch (which won't power it due to no handshake), the device stays off because the switch never recognizes it as a valid PD. In short, active PoE (IEEE standard) is smart and safe – it only powers when communication confirms both sides support it – while passive PoE is an always-on, "dumb" power feed. Always double-check what type of PoE your equipment uses. Mixing standard and non-standard PoE is a frequent reason devices fail to power on.
Beyond passive PoE, there are other non-standard PoE variations you might encounter, often created before the current standards or to deliver extra power. Here are a few:
- "PoE Legacy" – This term usually refers to early pre-802.3af equipment (for example, some old Cisco IP phones or access points) that used proprietary PoE signalling. These legacy PoE PDs don't present the normal IEEE signature (the 25kΩ resistor) that modern PSEs look for. Unless a switch explicitly supports legacy detection, it may not power such a device. For instance, Cisco's older 7940G/7960G phones only supported Cisco's pre-standard inline power, not 802.3af. If you plug one into a standard PoE switch that doesn't recognize the legacy method, it won't turn on.
- "High PoE" – Before IEEE 802.3bt standardized 60 W and 90 W PoE, some vendors offered midspan injectors to power high-wattage devices like PTZ cameras. These often used all four wire pairs to send ~50–60 W and were proprietary. For example, Axis marketed a "High PoE" midspan for their heavy PTZ cameras. These injectors typically don't negotiate – they just output a higher voltage/current. A device not designed for that exact system won't work (or could be harmed) if connected. Always match such injectors with the specific device they are meant for.
- Cisco UPOE – Cisco's Universal PoE (UPOE) was a proprietary extension introduced around 2014 that pushes about 60 W per port by using all four pairs (similar to later Type 3 PoE++). Cisco also coined UPOE+ for 90 W. While UPOE and UPOE+ were precursors to the 802.3bt standard, they initially worked only between Cisco's own switches and endpoints or with Cisco-approved intermediaries. In practice UPOE is very similar to standard Type 3 PoE++, but if you have a Cisco UPOE switch and a non-Cisco PD (or vice versa), there could be compatibility issues in how they negotiate power. Newer IEEE 802.3bt-compliant devices should also work on UPOE ports (Cisco made UPOE+ backward-compatible with 802.3bt Type 4), but it's something to be mindful of if one device is strictly proprietary.
- Other Proprietary PoE – A few other vendors had their own PoE-like implementations (e.g. "Passive PoE 48V" or enhanced power for specific systems). These are less common now but still exist in certain products. Essentially, if you see terms like Passive, Injector-required, or brand-specific PoE requirements, that's a sign of non-standard PoE. You must use the correct matching equipment. A non-standard PD may not even get recognized by a standard PSE and a non-standard PSE may overpower an incompatible PD. Incompatibility between different PoE types is a top reason for "mystery" PoE failures.
To summarize this section: Always verify that your PSE and PD share a common PoE standard or known compatible mode. If one is "active 802.3af/at/bt" and the other is "passive" or proprietary, they won't play nice. Using only IEEE 802.3af/at/bt on both sides is the safest way to ensure any PoE-labeled device actually gets power. Mixing standard and non-standard PoE is like mixing different charger types – the plug might fit (the Ethernet jack), but the electrical "language" is different.
(For reference, the table below shows the key IEEE PoE standards and their power limits. Notice how newer standards provide higher wattage to meet the needs of power-hungry devices. Any device claiming compliance with one of these should work with a PSE of the same or newer standard.)
| PoE Standard | Year | Common Name | Max Power from PSE | Power Available at PD |
| IEEE 802.3af | 2003 | PoE (Type 1) | 15.4 W | 12.95 W |
| IEEE 802.3at | 2009 | PoE+ (Type 2) | 30.0 W | 25.5 W |
| IEEE 802.3bt Type 3 | 2018 | PoE++ (4-pair) or "BT" Type 3 | 60 W | ~51 W |
| IEEE 802.3bt Type 4 | 2018 | PoE++ (4-pair) or "BT" Type 4 | 90 W | ~71 W |
(Note: Cisco's proprietary UPOE (2014) delivered up to 60 W similar to Type 3, and UPOE+ (2018) up to 90 W similar to Type 4. Devices labeled UPOE/UPOE+ should now be cross-compatible with the equivalent IEEE standards, but initially these were vendor-specific.)
PoE Powering Modes: Mode A vs. Mode B
Even if both your PSE and PD are standard PoE, there's another wrinkle that can cause issues: the powering mode used on the wire pairs. Fast Ethernet and Gigabit Ethernet cables contain four twisted pairs of wires (eight conductors total). PoE can send power over these pairs in two main ways, called Mode A and Mode B. The end result (delivering DC power) is the same, but the difference is which pairs carry the current:
- Mode A (Alternative A) – Power is delivered on the data pairs. In typical T568A/B wiring, these are pins 1–2 and 3–6, which are the pairs used for data transmission in 10/100 Mbps Ethernet. Mode A superimposes DC power onto those same pairs (a technique often called "phantom power"). Don't worry – the data signal and DC voltage coexist fine because they operate at different frequencies. A PSE using Mode A is usually an endspan (built into a network switch), so you might hear Mode A devices called endspan PSE. In Mode A, pins 1&2 carry the positive DC voltage and pins 3&6 carry the negative (or vice versa, since polarity can be automatically adjusted).
- Mode B (Alternative B) – Power is delivered on the spare pairs. In 10/100 Ethernet, the pairs on pins 4–5 and 7–8 do not carry data, so Mode B uses those for DC supply. Pin 4&5 typically carry the positive side and 7&8 the negative side of the DC supply. Mode B PSEs are often midspan injectors – devices that sit between a non-PoE switch and the PD. Because they use spare pairs, they inject power without touching the data on the other pairs. Thus, Mode B PSEs are sometimes referred to as midspan PSE. (In gigabit or faster Ethernet, there technically aren't "unused" pairs since all four pairs carry data; but a Mode B approach will still put power primarily on what would be pins 4–5 and 7–8, while Mode A uses 1–2 and 3–6. In 4-pair PoE like 802.3bt, both sets of pairs can be used simultaneously for even more power.)
Now, according to the IEEE standards, a compliant PD (Powered Device) must be able to accept power from either Mode A or Mode B. The PD is supposed to be polarity-insensitive and wired internally to draw power from whichever pairs the PSE decides to energize. The PSE, on the other hand, is allowed to implement just one mode (A or B) or both. In practice, many PoE switches output Mode A, and many midspan injectors output Mode B, but this isn't a strict rule – some switches can do either or both. As long as the PD supports both modes, everything should interoperate. If both devices are truly 802.3af/at compliant, you typically don't need to worry about Mode A vs B at all – the PD will happily take power on whichever pairs it gets.
However, issues arise when a device is PoE-capable but not fully compliant. Some manufacturers (especially of early or inexpensive PoE devices) cut corners. For example, a device might be built with only two pairs wired internally (perhaps to save cost), effectively meaning it only accepts power on one specific mode. This would be a non-compliant PD, since the standard disallows implementing only one mode. But such devices exist – often labeled "PoE compatible" or "passive PoE". If you happen to pair one of these with a PSE that uses the opposite mode, the device won't power on despite both saying "PoE". It's an interoperability problem outside the official spec.
It's important to note that for 10/100 Mbps networks, Mode B requires all four pairs in the cable to be present (since it needs the spare pairs to carry power). If you had an old Ethernet cable or connector wired with only two pairs (which can still carry data at 10/100), a Mode B injector wouldn't be able to send power. This is uncommon today, but worth mentioning if dealing with older wiring.
In summary, Mode A vs. Mode B mismatches can cause PoE failures when using non-compliant gear. A fully standard PD adjusts to either mode automatically, but a "PoE-capable" (not certified) device might specify a required mode. For instance, some IP cameras or sensors from lesser-known brands might only list "PoE on pins 4,5 (+) and 7,8 (–)" – which means they expect Mode B. If you plug those into a switch that only provides Mode A, nothing will happen. Conversely, a device wired for Mode A only (power on data pins) would fail with a Mode B midspan. Ideally, stick to IEEE compliant devices which guarantee mode flexibility. If you must use a non-standard PD, find out which mode it needs and use a PSE/injector that provides power on those pins. Mode mismatches are less common than standard mismatches, but when they occur, the symptom is the same – a "PoE" device that frustratingly stays dark.
PD Power Demand Exceeding PSE Capability
Another common reason a PoE device fails to power on (or keeps resetting) is that the device needs more power than the PoE source can provide on that port. Even within the IEEE standards, there are multiple power Classes that a device might belong to. When a PoE PD connects, it indicates (during the negotiation handshake) which power class it falls under – essentially telling the PSE "I'm going to need up to X watts." If the PSE can't meet that requirement, a proper PoE switch will refuse to power the device at all. This is actually a safety feature: rather than brown-out or damage something, the PSE just won't turn on power if it knows it's inadequate. For example, imagine plugging a high-power PTZ camera (perhaps needing 25 W or more) into a small old 802.3af PoE port (which maxes out at ~15 W). The camera is likely a Class 4 device (PoE+ range) while the switch port can only support Class 0-3. The switch will detect the camera's class and won't energize it because it can't supply what's requested. To the user, it looks like nothing is happening – because the switch is essentially saying "I can't power this, so I won't even try."
In some cases, a high-power PD might be backward-compatible to lower power but with reduced functionality; in other cases, it's not. For instance, some wireless access points or smart LEDs might operate in a limited mode if only 15 W is available (e.g. radios at lower capacity), whereas others simply will not boot without their full 30+ W. These are what we call non-backward-compatible PDs – they don't gracefully scale down. If you connect such a device to an underspecced PSE, it either never turns on or continuously reboots because it browns out when attempting to draw full power. A PoE LED lighting panel that needs 60 W, for example, won't light at all on a 15 W port – it's not like it'll just be dim; it's off. Always ensure your switch/injector's per-port power rating meets or exceeds the PD's requirement. If you accidentally plug a Type 4 (90 W) device into a Type 1 (15 W) port, it will not come alive (and thank goodness, because otherwise the port would overload).
It's also worth mentioning that even during operation, power needs can change. Some devices draw more power in certain modes of operation than in others. A great example is a PoE PTZ camera with a heater or IR illuminator. During normal idle operation, it might only draw, say, 10 W. But when you start panning/tilting and the internal motors engage, or when the internal heater kicks in on a cold night, the power draw can spike significantly (maybe up to 20–25 W). If the PSE budgeted only for a lower draw, these spikes could exceed what the port can supply, leading the camera to shut down or reboot mid-action. This can happen if the device was negotiated as a lower class initially or if the PSE is at its limit. The result is intermittent operation: the device might power on, but then resets when it draws that extra power for a feature. If you observe a PoE device power-cycling when in heavy use, it's a red flag that power demand exceeds what the port provides. The solution is to use a higher-capability PoE source for that device or disable some high-power features.
In summary, always check the PoE class or wattage of your powered device. Match it with a PSE port that can deliver that class. Here's a quick reference:
- 802.3af ports supply up to 15.4 W (sufficient for Class 0-3 devices, like basic phones, small WAPs, simple cameras).
- 802.3at (PoE+) ports supply up to 30 W (needed for Class 4 devices, like advanced APs, PTZ cameras, video intercoms, etc.).
- 802.3bt Type 3 ports go ~60 W, and Type 4 up to ~90 W (for very high-power devices – think LED lighting arrays, all-in-one touch computers, etc.).
If you connect a device requiring PoE+ to a PoE-only switch, it won't power on by design. And if a device is right at the edge of the budget, it might flicker on/off under load. The cure is simple: use a PoE injector or switch port that meets the required standard of that device. When planning a PoE deployment, budget a little headroom – don't expect a 15.4 W port to reliably run a device that claims "12.95 W" usage if that device may sometimes need the full 15 W. It's like trying to run a high-end gaming PC on a tiny UPS – under full load it's going to trip power. Better to have a margin or the ability to negotiate a lower-power mode if supported.
(Technical note: The IEEE PoE standards have a concept of power classes (0 to 8) and a mechanism for negotiation. In fact, 802.3bt introduced dynamic power allocation and even "power demotion" – a PSE can agree to a lower class if it can't meet the PD's request, rather than simply refusing. However, not all PDs can operate in a demoted power mode. For simplicity, ensure your PD and PSE are appropriately matched in capability.)
Excessive Cable Length
Sometimes the issue isn't with the devices at all, but with the cable in between. Ethernet cables have a recommended maximum length of 100 meters (328 feet) for data transmission. When using PoE, that distance guideline still applies. One reason is that longer cables introduce more electrical resistance, which causes a voltage drop over the cable's length. Essentially, the further electricity has to travel, the more pressure it loses by the time it reaches the other end (like water losing pressure through a very long hose). The PoE standards account for some voltage drop (they assume up to 100 m of Cat5e cable in worst case). For instance, a PSE might send 48 V knowing that the PD at the far end might only see about 44–46 V by the time it arrives, which is still fine. But if you significantly exceed the 100 m limit, the drop may become too large – the PD could end up with a voltage below what it needs to run, especially under load.
If you have a very long run of cable and your PoE device won't power up or is behaving erratically, this could be the cause. Symptoms of insufficient voltage at the PD include devices that reset when drawing higher current (similar to the previous section's symptoms) or devices that won't quite start up at all even though the PSE port says it's providing power. In fact, one troubleshooting step for PoE problems is to measure the actual voltage at the device end or try a shorter cable temporarily. Long or poor-quality cables can introduce enough line resistance that the PD is effectively starved of power.
Ethernet cabling also has quality categories (Cat5e, Cat6, etc.), which affect resistance and performance. A higher-grade cable may have slightly lower resistance over the same length, meaning less voltage drop. On marginal runs, this can make a difference. But if you go far beyond spec (say, a 200 m cable run), no cable grade will save you – you need to use a solution like a PoE extender or an active switch in the middle to repeat the signal and inject fresh power. Standard PoE is limited to 100 m per segment not because the power magically stops at 100 m, but because the Ethernet signaling for data will start to degrade beyond that distance, and the power loss becomes significant too.
A practical example: Suppose you have an outdoor security camera at the far end of a parking lot, connected by a 150 m cable back to the building. Even if your switch port is PoE+, by the time that 48 V travels 150 m, the camera might only receive (for example) 36–38 V under load, which might be below its required input range. The camera could boot up when idle (drawing little current, thus less voltage drop) but then reboot when it tries to draw more power for the IR LEDs at night. Solution: shorten the run or install a PoE extender (which is a device that repeats the network signal and adds a fresh injection of PoE power for the next leg, effectively splitting the long run into two compliant segments). Always plan PoE installs with that 100 m guideline in mind. If you must exceed 100 m, use the proper equipment to do so, rather than pushing the cable beyond its limits. Also, ensure your cable is copper Ethernet of good quality; poorly made cables or CCA (copper-clad aluminum) cables have higher resistance and can cause more voltage drops even under 100 m.
To summarize: Distance can kill PoE. Long cable runs = higher resistance = voltage sag. Follow the 100 m rule for Ethernet runs. If your device is far away, don't daisy-chain lots of passive couplers or use substandard wire. Either keep it under 100 m or use PoE repeaters/switches for hops. This ensures your PD gets the juice it needs. Cable length issues often manifest as "it kind of powers on, then off" or "works when we test on a short cable, but not on the installed long cable" – classic clues pointing to this problem.
Overloaded PSE Power Budget
The last major reason for PoE failures we'll cover is a bit less obvious until you encounter it: the power budget of the PoE source. Every PoE switch (or injector) has a maximum amount of total power it can supply across all ports, often called the PoE budget. For example, a small 8-port PoE switch might have a total budget of 60 W or 120 W, a larger 24-port might have 370 W, etc. This limit is usually lower than the sum of the theoretical max of each port, because we rarely expect all ports to pull maximum power simultaneously. However, if you do connect enough high-draw devices such that their combined demand exceeds the switch's budget, something has to give.
Most modern PoE switches have intelligent power management: they will refuse to power any devices beyond the budget, or they'll drop power on the lowest-priority ports, etc. For instance, suppose you have a 8-port 802.3at switch with a 130 W total budget. You might assume you can put 8 PoE+ (30 W) devices on it (which would require 240 W), but you cannot – the budget only covers about four 30 W devices in this case. In fact, with 130 W total, if you try to attach a fifth PoE+ device, the switch will likely not power it on because doing so would exceed its capacity. If the switch firmware is smart, it might indicate a power denial for that port. Or it might power it and then quickly shut it off as the budget trips, resulting in no power for that device. The end user simply sees that some of the connected devices aren't turning on. On unmanaged switches, the symptom is just that the last devices plugged in don't get power.
Let's put this into an example to make it concrete. Consider that 8-port, 130 W PoE switch:
- It can supply 15.4 W on all 8 ports simultaneously (15.4 W × 8 = 123.2 W, which is within 130 W) – so 8 regular PoE (802.3af) devices like phones or small cameras are fine.
- But it cannot supply 30 W on all 8 ports (30 W × 8 = 240 W, way over budget). In fact, with 130 W, you could do at most 4 ports at 30 W (4 × 30 = 120 W, leaving a tiny buffer). If you plug in a 5th PoE+ camera, the switch has to say "sorry, no power available" on that one.
- Any mix of devices has to be managed so the sum stays ≤ 130 W. For example, 4 PoE+ cameras (30 W each = 120 W) and 4 regular PoE devices (15 W each = 60 W) would total 180 W – too high. The switch in that scenario might only actually power the first few and not the rest, or cap some ports to lower power classes.
Power budget overload situations can cause inconsistent behavior: you might notice that some devices power up if they are the only ones connected, but when you connect many devices, a few won't power. Or perhaps all devices power up but when they all draw peak power together, the switch drops one or two of them. Some higher-end PoE switches allow setting port priorities (so maybe lower priority ports get shut off first when budget is exceeded). In any case, the fix is to ensure the total demand matches the PSE's capabilities. Check your switch's PoE budget specification – it's usually listed in watts. Add up the consumption of your PDs (remembering that PoE+ ones count for ~25–30 W, etc.) and see if you're over. If yes, you have a few options: (1) limit how many high-power devices are on that single PSE, (2) upgrade to a switch with a larger PoE budget, or (3) use an additional PoE injector or midspan for some ports so the power draw is distributed.
Power budgeting issues are especially likely in scenarios like: a fully loaded switch in a CCTV system (many cameras), a switch powering numerous wireless APs with heaters or advanced features, or daisy-chaining multiple PoE powered repeaters. It's also worth noting that ambient temperature and PSU limits can slightly affect available power – some switches derate their PoE budget in high temperatures to avoid overload. But the key point for users: if some of your PoE ports are not powering devices and you know the devices are good, consider the possibility that the switch has run out of power budget. This is often indicated by an error message or LED status on managed switches. On an unmanaged switch, you might only deduce it by calculation or by testing devices one at a time.
| PoE Switch Total Budget | Device Load Scenario | Combined Power Draw | Result |
| 130 W total (8 ports) | 8× IP cameras @ 15.4 W each (802.3af) | ~123 W total | Within budget: all 8 cameras can run. |
| 130 W total (8 ports) | 8× PTZ cameras @ 30 W each (802.3at) | 240 W total | Over budget: only ~4 cameras would get power, others shut off. |
| 130 W total (8 ports) | 4× 30 W (PoE+) + 4× 15 W (PoE) devices | 430 + 415 = 180 W total | Over budget: reduce number of PoE+ devices or increase budget. |
As a closing tip: many PoE switches will have a spec sheet that clearly states something like "Max 30 W per port, 130 W total across all ports." Use those numbers to plan your deployment. If you need more devices than the budget allows, you might split them across two switches or use midspan injectors for some fraction of them. And if you suspect an overload, try unplugging one or two devices and see if the others come on – if they do, that's a sign you hit the ceiling of the power budget.
Conclusion
PoE is a fantastic technology that simplifies installations, but it has a few nuances. Devices labeled "PoE" can fail to work together if any of the above mismatches occur. To troubleshoot a PoE device that isn't powering on, you can run through this mental checklist: (1) Are both PSE and PD using the same standard (and not a passive or proprietary PoE)? (2) If standard, could there be a Mode A vs B issue with a non-compliant device? (3) Does the device require more power (higher PoE class) than the source provides? (4) Is the cable running too long or the cable quality causing excessive voltage drop? (5) Is the switch's total PoE power budget exhausted by too many devices? By systematically checking these, you can usually pinpoint why a given "PoE" connection isn't behaving.