Changing a PoE Device's Output from 12V to 5V: Can You Keep the Same Power Transformer? 2.4× Current, Auxiliary Winding and Same-Series 5V Parts | VOOHU

This situation comes up often when a PoE camera or wireless AP gets a board revision: in the old design, the isolated supply on the PoE powered device (PD) side outputs 12V, but the new main board runs directly from 5V; or the team wants the isolated supply to output 5V directly to save a buck stage. The hardware team's first reaction is often: the power has not changed, so just change the feedback resistors, keep the original PoE power transformer, and no board re-spin is needed.

Done that way, even if the prototype powers up, the design cannot go straight into production. With the same power at a lower voltage, the current goes up: at 13W, a 12V output delivers about 1.08A, while a 5V output must deliver 2.6A, 2.4 times as much. The transformer's turns ratio, secondary wire gauge and auxiliary winding were all designed around the original output voltage. That is why, in thePoE power transformerselection table on the VOOHU website, WH13P-12L, WH13P-24L and WH13P-50L, all 13W parts in the EP13 series, are three separate parts for 12V, 24V and 5V respectively. Using data from that selection table, this article explains why the transformer has to change when the output voltage changes, which part to switch to, and what else on the board has to change with it.

1. The Short Answer: 12V to 5V Needs a Different Part, but No Larger Core Is Needed When a Matching Version Exists

First, for a large output voltage change such as 12V to 5V, switch to a transformer made for the new output voltage. Changing only the feedback resistors may give a working prototype, but the secondary current is already beyond the original design, and the voltage of the auxiliary winding that powers the controller changes as well. Problems such as auxiliary winding undervoltage often show up right at power-up as a hiccuping output and repeated reboots; problems such as secondary winding heating and higher primary peak current only appear at high temperature, full load and low input voltage.

Second, some power levels on the VOOHU website come in both 12V and 5V (or 24V) versions. At the 13W level of the EP13 series, for example, the WH13P group alone has 12V, 24V and 5V versions (see the comparison table in Section 5). If the original part's power level has a version for the new voltage, you can switch within the same core series without moving to a larger core.

Third, not every part has a version with another output voltage in the same series and at the same power. If the original part's power level only has a 12V version, follow Step 2 in Section 9 and look one power level up or in another core series.

Fourth, “same core series” does not mean “pin-compatible drop-in replacement”. The VOOHU category page states this clearly: EP and EFD are core series labels, the actual dimensions and pinout follow each part's own drawing, and the series name does not imply interchangeability. Before switching parts, always put the old and new datasheets side by side and check the pin definitions and pads pin by pin.

2. Lower Voltage, Harder Work for the Transformer: at 13W, a 5V Output Delivers 2.4× the Current of 12V

Output power equals output voltage times output current. At the same power, going from 12V to 5V raises the current from about 1.08A to 2.6A. The listed output currents of the three WH13P parts in the 13W level of VOOHU's EP13 series reflect exactly this relationship:

The “output current” column on the VOOHU website uses two formats. A range (such as 1.08A~1.99A) or “≥ 2A” is a current band, not the range of current the part can deliver. WH13P-12L, for example, is listed as 1.08A~1.99A, but its listed 13W at 12V works out to about 1.08A, so take 1.08A as its current limit, not 1.99A. In general, take the current limit as the part's power value divided by its output voltage, and keep your calculated current below that. Where the column gives a specific value (such as 3A or 5A), keep your calculated current below that value.

Once the current goes up, at least two places work harder. The first is the secondary winding: copper loss is proportional to the square of the secondary RMS current. The output current is 2.4 times the original; the secondary now conducts for a larger share of each cycle (from about 0.49 to about 0.69), which spreads the current out somewhat, but in the example in Section 3 the secondary RMS current still ends up about twice the original (about 1.5A to 3.1A), and copper loss rises to about 4 times, so a secondary wire gauge designed for a 12V, 1.08A output is no longer enough. The second is the output rectifier: assuming a 0.5V rectifier drop, rectifier loss is about 0.54W at 12V and 1.08A, and about 1.3W at 5V and 2.6A, rising from about 4% to about 10% of the output power.

This leads to an easily overlooked result: with the isolated supply delivering the same output power, the higher rectification loss at 5V means the input power the PD takes from the Ethernet cable may actually go up. Devices whose PoE power budget was already tight should be rechecked. If the move to 5V also removes a downstream buck stage, the loss saved by eliminating that stage is usually larger, so overall input power will generally drop, but it is still advisable to confirm this by measurement.

3. What Happens If You Only Change the Feedback Resistors: Same Turns Ratio, but Different Duty Cycle and Currents

One caveat first: not every PoE supply uses a flyback topology. The VOOHU website also cautions against treating the whole EP and EFD catalog as validated flyback designs; whether a given part is used in a flyback must be confirmed from the controller, the circuit and that part's documentation. Flyback is the most common topology in PD isolated supplies, so a simplified flyback example (continuous conduction mode) is used below to show how things change.

In a flyback, the secondary voltage is “reflected” to the primary through the turns ratio. This reflected voltage is roughly the turns ratio times (output voltage + rectifier drop). The duty cycle is set by the input voltage and the reflected voltage. Assume a transformer designed for 12V output, with a primary-to-secondary turns ratio of 4:1 as an example (an illustrative value, not taken from any specific VOOHU part), 48V input and 13W output. Efficiency is taken as 85% in both cases; in practice, rectification loss is higher at 5V, efficiency is lower, and the primary current will be even higher than shown in the table:

The table shows three things. First, the secondary RMS current doubles, copper loss is about 4 times the original, and the secondary winding temperature rise goes up noticeably. Second, the duty cycle drops from about 0.51 to about 0.31, so the same input power has to be delivered in a shorter on-time; the primary current during conduction rises by about 60%, the peak current rises with it, and the core runs closer to saturation. Third, as the input voltage varies between 37V and 57V, the whole duty-cycle range shifts downward, and the controller operates at a different point from the original design.

The VOOHU website notes that the turns ratio is not a guaranteed input-to-output voltage ratio, and that the output voltage is determined by the complete DC-DC circuit. Changing the feedback resistors can move the output setpoint to 5V, but the turns ratio, wire gauge and auxiliary winding are still matched to 12V, so the problems above, along with those in Section 4, will all appear.

4. The First Thing Likely to Fail: the Auxiliary Winding Voltage Drops Too

Many PoE power transformers have an auxiliary winding in addition to the main output winding, and some also have a synchronous rectifier (SR) drive winding. In a flyback, the auxiliary winding is wound on the same core as the secondary, so its voltage scales with the output voltage according to the turns ratio. Many controllers rely on the auxiliary winding for their supply once they have started up.

For example, if the original design has the auxiliary winding supplying about 12V to the controller at 12V output, changing the output to 5V drops the auxiliary winding voltage to around 5V, below the undervoltage lockout (UVLO) threshold of many controllers. The typical symptom: the supply starts at power-up, but the controller supply cannot hold up, the output voltage hiccups, and the device keeps rebooting. Some controllers instead fall back to running continuously from their internal startup circuit, which shows up as a hot controller.

In designs with an SR drive winding, the drive voltage drops in the same proportion, so the SR MOSFETs are not fully turned on and run hotter. When switching to a 5V version, re-confirm the connections and voltages of the auxiliary and drive windings against the new part's winding diagram. The VOOHU website also points out that the auxiliary winding's load capability is usually far lower than the main output's, and that pin definitions follow that part's drawing. So the auxiliary winding cannot be loaded like the main output.

5. VOOHU Parts with the Same Power and Different Output Voltages

The table below is compiled from the VOOHU PoE power transformer selection table. It lists parts that come in both 12V and 5V versions (some also in 24V) within the same core series and power level, with the output current listed on the VOOHU website in parentheses; 13V to 14V versions are in Section 6. When changing the output voltage, first check whether the original part is in this table; if it is, find the matching voltage in the same row:

The -L parts in the WH13P, WH13F and WH300F groups in the table follow a pattern: the number after the hyphen indicates the output voltage, with 12 for 12V, 24 for 24V and 50 for 5.0V, and changing voltage within a group keeps the same core series and power level. Other parts with the same prefix (such as WH13P-12L2, WH13P-12H and WH13P-12H5) have no matching -24 or -50 part on the VOOHU website, so do not use this pattern to infer part numbers. In the 25W EP13 level, where WH13P-12H5 sits, the 5V-output part on the VOOHU website is WHEPD13002; it is named differently, so check its pinout and windings against its own drawing.

Where one cell lists two parts (for example WH13P-12L and ST1339WH for 12V, 13W), their selection-table parameters are the same, but their pinout and winding arrangement are not necessarily the same. Once you pick one, lay out the pads to its own drawing; do not let the two share one PCB footprint until both drawings have been checked pin by pin and match.

6. How to Search for 3.3V, 13V to 15V, or a Higher Power Level

If the new design is not a simple 12V-to-5V change but moves to a different voltage, or changes both voltage and power, use the table below as a guide to the VOOHU selection table:

Two points to keep in mind. First, the power field on the VOOHU website is a reference value for initial comparison; it is not a PoE class and not a guarantee under all conditions. For a 5V, 25W requirement, for example, do not settle on WHEPD13002 just because it is marked 25W; also check it against the actual input voltage range, peak load and thermal conditions. Second, the website currently lists only one 3.3V part, rated 10W. When you need 3.3V at higher current, a common approach is to choose a 5V version of the transformer and add a buck stage after it.

7. Once the Transformer Changes, These Parts of the Board Must Change Too

Feedback divider and compensation

With the common TL431-plus-optocoupler feedback, the output voltage is about 2.5V × (1 + upper resistor ÷ lower resistor). The upper-to-lower resistor ratio is about 3.8 for a 12V output and about 1 for a 5V output. For a 3.3V output, the TL431's 2.5V reference plus the optocoupler drop already exceeds 3.3V, so a shunt regulator with a 1.24V reference (such as a TLV431 or similar) is commonly used instead. The optocoupler current-limiting resistor and the loop compensation also depend on the output voltage, so retune them together, then measure the output voltage at no load, full load and during load steps.

Output rectifier and output capacitors

Reselect the rectifier for the new output current. For 5V outputs of 2A or more, the usual choice is a low-Vf Schottky diode, or moving to synchronous rectification. Going from 12V to 5V, the output capacitors' voltage rating is still sufficient, but the ripple current rises with the output current, so check the capacitors' ripple current rating and temperature rise.

Output filter inductor

If the isolated output has an LC filter inductor or a ferrite bead, it carries the output current, which rises to about 2.4 times the original. Recheck saturation current and temperature rise, and if needed switch to a higher-currentintegrated molded inductor. If a buck stage follows, its inductor current is set by the downstream load and does not change; what needs checking is that stage's input current (about 2.4 times), the input capacitor ripple, and whether the downstream IC still has enough voltage headroom with a 5V input.

Primary clamp, pinout and pads

In a flyback design, the new part's turns ratio is different and its leakage inductance may differ too, so the reflected voltage and leakage energy may both change; recheck the clamp voltage and loss of the primary clamp circuit (RCD or TVS) for the new part. Follow the new part's own drawing for pin definitions, pads and height. The VOOHU website specifically notes that some parts require designated pins to be paralleled on the PCB; if the old board does not have those connections, the new part cannot simply be dropped in.

8. Going the Other Way, from 5V to 12V: the Problems Flip

When a 5V design is changed to 12V, the secondary current goes down and the wire gauge is adequate, but voltage-related problems appear. Using a simplified flyback example again: a transformer designed for 5V output, with a turns ratio of 9:1 as an example, has a reflected voltage of about 50V at 5V output; at 12V output, the reflected voltage rises to about 113V. To bring the output up to 12V at 57V input, the primary switch would have to withstand about 170V during the off-time from these two terms alone, which already exceeds the 150V MOSFETs used in many PD flyback circuits, and the leakage-inductance spike pushes it even higher.

Three other places run into trouble. First, the primary clamp. With a TVS or Zener clamp, if the clamp voltage is below the new reflected voltage (about 113V), the voltage across the primary winding is held at the clamp voltage, so the output can rise only to about (clamp voltage ÷ 9 − 0.5V) and cannot reach 12V. The feedback loop then drives the controller to its current limit, the energy the load cannot use goes into the clamp, and the clamp quickly overheats. The switch voltage stays within 57V plus the clamp voltage, but the circuit no longer works properly. With an RCD clamp, the clamp capacitor voltage is pushed above 113V, the resistor loss rises roughly with the square of that voltage, and the primary switch sees the roughly 170V calculated above plus the leakage spike. Second, the auxiliary winding voltage rises to more than twice the original and may exceed the maximum rating of the controller supply pin; the drive voltage of a self-driven synchronous rectifier rises the same way and may exceed the gate voltage rating of the SR MOSFETs. Third, the output capacitors need to move up from the 6.3V or 10V rating class. So a 5V-to-12V change also calls for the matching 12V version, for example going from WH13P-50L to WH13P-12L in the 13W EP13 level, then go through the surrounding circuits as described in Section 7.

9. A Four-Step Check and a Prototype Checklist

Step 1: Define the new output. Determine the output voltage, continuous current and peak current, and calculate the required output power.

Step 2: Search the same core series. In the VOOHU selection table, look for a part in the same core series and power level as the original part, with the new output voltage. If there is none, move up one power level or look in another core series.

Step 3: Check the drawings. With the old and new datasheets side by side, check the pin definitions of the main output, auxiliary and synchronous drive windings pin by pin, along with the pads, height and any pins that must be paralleled.

Step 4: Update the surrounding circuit, then build prototypes. Change the feedback divider, rectifier, output capacitors, filter inductor and primary clamp together, then test each item in the table below.

10. FAQ

Q1: If only the feedback resistors are changed and the prototype powers up, is everything fine?

No. Powering up only shows that the loop can regulate the output to 5V, not that the transformer is fine at this operating point. Secondary current, auxiliary winding voltage and primary peak current have all changed, and the result only counts once it has been tested at minimum and maximum input voltage, full load and high temperature. For production, switch to the matching 5V part number as described in Sections 5 and 6.

Q2: When the output voltage changes, do the RJ45 jack and LAN transformer at the port need to change too?

Generally no. The output voltage change by itself is no reason to replace them. The PoE power transformer is used in the isolated DC-DC power stage inside the powered device, while the LAN transformer in the data path is a different kind of component; the port side only sees the PD's input power. However, rectification loss increases after moving to 5V, so the input power may rise slightly; recheck it against the PoE power budget.

Q3: Can the output be changed straight to 3.3V?

The VOOHU website has one 3.3V output part:WHTS15EF1665Uin the EFD15 series, rated 10W, with a listed output current of 3A. For 3.3V at higher current, a common approach is to use a 5V version of the transformer and add a buck stage.

Q4: WH10P-50L and WH13P-50L are both 5V output parts. What is the difference?

The core series and power level differ:WH10P-50Lis in the EP10 series at 10W, andWH13P-50Lis in the EP13 series at 13W; both list an output current of ≥ 2A. Choose by actual load power. Pinout and dimensions follow each part's own drawing, and the two cannot be swapped directly just because they look similar.

Q5: When one voltage has several parts, such as WH13P-12L and ST1339WH for 12V, 13W, are they interchangeable?

Identical selection-table parameters do not mean identical pinout, winding arrangement or auxiliary winding voltage. The VOOHU website also states that the series name does not imply interchangeability. Once you choose one, lay out the board to its drawing; to switch to the other, compare both datasheets first and verify with prototypes if needed.

Q6: What if I need both 5V and 12V outputs?

A common approach is to choose the transformer for the higher of the two voltages, size the power for the sum of both rails plus a margin for the downstream conversion loss, and then use a buck stage to derive 5V from 12V. For example, if the two rails add up to about 10W, the 13WWH13P-12Lin the EP13 series is an option; if they add up to around 20W, look at the 25WWH13P-12H5. If you need multiple isolated outputs, send your requirements to the VOOHU FAE team for evaluation.

Need samples or technical support? If you are changing the output voltage of a PoE powered device, send us the original part number, the new output voltage and current, the input voltage range, the controller part number and the available board space. VOOHU FAEs can help you find the matching voltage version in the EP7, EP10, EP13, EFD15, EFD20 and EFD25 series, compare the pinouts and winding diagrams of the old and new parts, and provide samples and datasheets.

——— VOOHU Engineering Team

Technical support: fae.thorne@voohu.cn | Tel: 400-1048-018

Suzhou VOOHU Electronic Technology Co., Ltd. | www.voohuele.com