VOOHU push-pull transformer selection: duty cycle asymmetry, flux bias and startup waveform verification

The push-pull isolated power supply seems to be stable during the prototype stage. However, after entering cold start, low input voltage, heavy load or high temperature working conditions, the current side of the switch tube may increase cycle by cycle, the output may restart repeatedly, the peak may suddenly increase, and the temperature rise of the transformer may even be inconsistent with expectations. Such problems are often simply attributed to "insufficient power of the transformer", but what really needs to be answered first is: whether the volt-seconds applied to the core in the two half-cycles are equal, whether reliable reset is completed during startup, and whether the drive, winding and PCB circuits maintain sufficient symmetry. If this chain of evidence is not established, simply replacing the material number with a larger core or a higher inductance may not eliminate the bias magnetism.

VOOHU official websitePush-pull transformer product catalogCurrently, WHST060, WHST06E, WHST06K, WHST06Q and other series can be checked.VOOHU push-pull transformer solutionThe application background of isolated power supplies is also given. The public details page lists fields such as inductance, turns ratio, DCR1, DCR2, isolation withstand voltage and operating temperature. This article uses real material numbers that can be accessed from four official Chinese and English websites to create a preliminary screening table, but does not directly write the web page fields as controller compatibility, output power or safety conclusion. Formal confirmation of materials still requires obtaining the specifications, schematics and pin definitions corresponding to the complete suffix, and completing startup and steady-state waveform verification on the target controller, switching frequency, input range, load and PCB.

1. The risk of push-pull transformer is not only the turns ratio, but also the volt-second balance of two and a half cycles.

The duty cycle is close to equal, which does not mean that the volt-seconds received by the core are exactly equal.

The ideal push-pull circuit consists of two switching tubes driving the center tap primary alternately, and the two half cycles establish magnetic flux in opposite directions in the magnetic core. The actual volt-second is determined by the winding terminal voltage and the effective conduction time. The driving pulse width, dead zone, switching tube voltage drop, power supply path impedance, copper foil voltage drop and winding resistance will change the effective value of each half cycle. Just because the two-channel PWM looks close to 50% on the oscilloscope is not enough to rule out bias magnetism; the actual conduction range of the two switching tubes, the waveforms at both ends of the primary and the current slope should be compared at the same time, and attention should be paid to whether the error accumulates with input, load and temperature.

Start-up, power-off restart and protection recovery are more likely to expose bias magnetism than steady state

The steady-state controller may have entered a closed loop and limited the duty cycle, and the first few cycles of power-up are affected by soft-start, undervoltage lockout, driver supply settling speed, reset state, and output capacitor charging current. If one of the circuits activates first, if the pulse widths are significantly different, or if protection exits and then resumes in an asymmetric sequence, the core may gain net DC volt-seconds during the initial phase. During debugging, a complete record from before the driver is valid to after the output is stable should be saved, instead of just intercepting hundreds of stable cycles; cold start, hot start, rapid power-off and power-on, and fault recovery should be covered separately.

DCR1 and DCR2 on the webpage cannot be regarded as proof of two semi-primary symmetries

The DCR1 and DCR2 disclosed on the product page are device-level winding fields, which are usually used to understand the DC resistance magnitude of different windings. Even if a certain part number shows that DCR1 and DCR2 have the same values, it cannot be inferred that the number of turns, leakage inductance, DC resistance and lead parasitics on both sides of the center tap are exactly the same; nor can it replace the pin diagram, winding structure and physical measurement. When verifying two and a half primary, the corresponding pins should be confirmed according to the complete specification, and then compared using consistent fixtures, test frequencies and connection methods to avoid misjudgment of differences in probes, contact resistance or cables as device asymmetry.

2. First use the public fields on the official website for preliminary screening, and then complete the boundaries between the controller and magnetic parts.

Inductance must be understood along with test conditions, input voltage and switching frequency

The inductance values ​​disclosed on the four candidate details pages range from 86μH to 475μH. These values ​​can help narrow down the sample range, but cannot alone determine the transmittable power, peak flux or light load stability. The web page does not fully display the test frequency, test voltage, tolerance, core parameters and saturation boundaries required for this article, so the current specifications should be obtained from technical support and reviewed based on the controller operating frequency, maximum effective on-time, most adverse input and allowable flux swing. If the project modulates the output by changing frequency or duty cycle, the verification points also cover the actual limits that the control loop may reach.

The turns ratio only gives the conversion direction and cannot directly guarantee the target output voltage.

WHST06001E0Publicly 1:1,WHST06E14A1is 1:1.1,WHST06K14A0is 1:3,WHST06Q02E0is 1:2.88. The turns ratio can support early judgment of the buck-boost direction, but the output is also affected by the rectifier structure, diode or synchronous rectification voltage drop, duty cycle, winding voltage drop, load and controller adjustment method. There may also be differences in pins, packages, winding definitions and recommended operating conditions between different material numbers; the close ratio does not mean that it can be directly replaced, nor can the mass production output be derived solely based on the turns ratio on the web page.

Isolation withstand voltage is a device field and is not automatically equivalent to the system operating voltage or reinforced insulation.

The four candidates in this article have public isolation withstand voltages covering 2500VAC, 3000VAC and 4000VAC respectively. This field is suitable for excluding candidates that clearly do not meet the objectives, but whether a system complies with basic insulation, reinforced insulation or specific operating voltage requirements also depends on the insulation construction, creepage distances, clearances, pollution level, material group, test method and applicable standards. PCB pads and slots will also change the overall boundaries of the machine. Design reviews should place transformer specifications, PCB spacing, controller isolation strategies, and overall machine certification requirements on the same checklist.

3. Use four real material numbers to create a push-pull transformer sample branch

The fields in the table below are from the VOOHU Chinese and English product details page accessible on August 13, 2026. "On shelf" in this article only means that there are public product records accessible on the official website, and does not represent real-time inventory, price, delivery or adaptation commitment to a certain controller. The four material numbers are used to show the preliminary screening paths for different inductances, turns ratios and isolation voltages, and are not a list of mutual replacements; the complete suffix, package, pinout and current specification must be checked before sample application.

If the project requires close to 1:1 isolation transmission, you can firstWHST06001E0withWHST06E14A1are listed as two different verification branches; when a clear boost direction is needed, evaluate them separately.WHST06K14A0withWHST06Q02E0. But this grouping is just a way to organize experiments, not application recommendations. Each candidate must record startup current, steady-state current slope, switching node spike, output ripple, temperature rise and protection behavior under the same input, frequency, load, PCB and measurement settings, and finally determine the materials based on project limits.

4. From schematic to PCB, first eliminate system variables that will create asymmetry

The two drives and two primary circuits should be evaluated for electrical symmetry, not just for appearance.

The driving resistance, gate or base circuit, driving return current, device model and heat dissipation conditions of the two switching tubes should be consistent; the paths from the center tap to the power supply decoupling, as well as the copper foil, current vias and return paths from the two primary ends to the switching tubes should also match as much as possible. Layout mirroring does not necessarily mean that the parasitic parameters are the same. In particular, check whether one side passes through more vias, test points, fuses or longer power branches. The transformer pin definition must be checked one by one with the schematic network, package number and assembly direction to avoid misconnection of the center tap or terminal with the same name.

Clamping, absorption and measurement loops cannot be optimized on just one side

Switch node spikes can come from leakage inductance, junction capacitance, loop inductance, and commutation. If the values, positions, and recirculation of the clamping or absorbing devices on both sides are significantly different, it will directly change the voltage area of ​​each half cycle, which will not only affect the stress, but also may amplify the original slight imbalance. When debugging, you should first compare the passive components and wiring on both sides, and then compare the waveforms under the same probe, bandwidth and grounding method. Don't immediately increase the absorption on just one side just because that side has a higher peak; first confirm whether the difference comes from the switches, windings, layout, or measurement connections.

Probe safety and measurement consistency are part of the credibility of the conclusions

When measuring floating switch node and primary current, use a differential probe, current probe, or project-approved isolation measurement scheme that meets voltage, common-mode range, and bandwidth requirements, and follow power-down, discharge, and protection procedures. Comparison of two-way waveforms must use equivalent connection positions and probe settings, and record bandwidth limits, sampling rates, probe magnifications, and current clamp directions. An unevaluated common oscilloscope ground clamp may cause short circuit or personal risk, and also make the so-called "waveform difference" lose its reference value.

5. Use the four stages of startup, steady state, limit, and regression to verify the closed loop to eliminate magnetic flux bias.

The first paragraph: low energy startup, first confirm the sequence and polarity

When powering on for the first time under controlled low-energy conditions allowed by the project, first check the complementary relationship between the two drives, dead zone, transformer polarity, output rectification direction and protection threshold. Capture the entire process from when the drive power supply is established, the first set of valid pulses, and the output begins to rise to closed-loop stability, and observe whether there is continuous conduction on one side, loss of pulses, repeated starts and stops, or sudden bending of the current. If the startup logic has not been confirmed, it should not go directly to high input or full load.

Second paragraph: Steady-state comparison of current slope and switching stress for each half cycle

Superimpose and compare adjacent odd and even half cycles at multiple inputs and load points. A healthy system does not require every instantaneous point to be mathematically identical, but there should be interpretable and consistent trends in on-time, current starting point, rise slope, peak value, and turn-off spike on both sides. If the peak value on one side accumulates cycle by cycle, the latter part of the waveform becomes significantly steeper, the temperature rise of the switch tube continues to bifurcate, or the protection is triggered in a fixed half cycle, the power increase should be suspended and the drive, supply voltage drop, winding connection, clamping and core reset should be investigated one by one.

The third paragraph: combine low input, heavy load, high temperature and fast restart into the most unfavorable operating conditions

Bias problems tend not to occur at a single room temperature rated point. The verification matrix should cover the upper and lower limits of input, light load and heavy load, cold machine and warm machine, soft start, rapid power-off and power-on, short-term output overload or protection recovery specified by the project. Low input may push the controller to increase duty cycle, heavy loads increase peak current, and high temperatures change switch voltage drop and winding resistance; the sum of these factors makes it more likely to expose volt-second imbalance. The specific boundaries must follow the specifications of the controller, switching device, transformer and complete machine, and cannot be replaced by the general values ​​given in this article.

The fourth paragraph: Lock the material number, board number and test settings and do the temperature rise and complete machine return

The final report should be bound with the complete part number suffix, transformer batch, PCB version, BOM, input power, controller configuration, switching frequency, load, ambient temperature, probe and oscilloscope settings. In addition to the temperature rise of the transformer itself, the temperature of the switching tube, rectifier and adjacent devices must also be recorded, and project requirements such as output ripple, efficiency, protection recovery, conduction or radiation interference, etc. must be retested. Any changes to transformers, switching tubes, drive resistors, absorption networks, packaging, laminations, or copper foil should trigger an impact assessment, and the conclusions of the old board cannot be used.

6. Frequently Asked Questions (FAQ)

Can push-pull transformers with the same turns ratio be directly replaced?

It cannot be replaced based on the turns ratio alone. Also check the inductance and test conditions, pins and polarity, winding definition, DCR, isolation structure, package and pad, operating temperature, controller frequency and allowable flux range, and verify it through startup, steady state, limit and temperature rise.

DCR1 is the same as DCR2. Does it mean that the two semi-primaries are completely balanced?

You can't judge like this. The public fields describe the magnitude of the DC resistance of the device windings, which is not sufficient to prove that the two half-primaries are exactly the same in terms of turns, leakage inductance, lead parasitics, and measured DC resistance. The complete pin diagram and specification must be combined, and the same fixture must be used for actual measurement.

If a spike appears during startup, can it be determined that the transformer is saturated?

cannot. Spikes can also come from leakage inductance, switch loop inductance, commutation, probe connections, or snubber networks. Determining magnetic flux bias or saturation requires a combination of driving volt-seconds, whether the current slope becomes steeper cycle by cycle, whether odd and even half cycles continue to bifurcate, temperature rise and protection behavior, etc.

Should the higher the isolation withstand voltage be given priority?

uncertain. Isolation withstand voltage must meet system requirements, but higher device fields do not automatically bring correct turns ratio, appropriate inductance, compatible pins or complete machine safety conclusions. Applicable standards, operating voltage and insulation targets should be determined first, and then the selection should be based on electrical, mechanical, thermal and supply conditions.

Conclusion

The key to selecting a push-pull transformer is not to pick the one with the largest inductance or the highest isolation voltage from the web page, but to put the candidate material number back into the two-and-a-half-cycle volt-second balance for verification.WHST06001E0、WHST06E14A1、WHST06K14A0withWHST06Q02E0Four sets of verifiable public parameter branches are provided; only through symmetrical schematic diagram and PCB, complete startup capture, odd and even half-cycle current comparison, most unfavorable operating conditions and complete machine regression can the "possibly applicable" be converged into a traceable mass production conclusion.