Chapter 14 converts the completed electrical, magnetic, winding, parasitic, and thermal design into a transformer that can be built repeatedly and verified objectively. The engineering task now shifts from predicting nominal performance to controlling variation: materials, turns, winding sequence, gap, insulation, leads, assembly, test methods, acceptance limits, traceability, and engineering changes all become part of the design.
← Previous Chapter
Next Chapter →
In This Chapter — Click to Expand
Table of Contents
1. A Flyback Transformer Design Is Not Complete Until It Is Manufacturable
A prototype that works once does not yet define a production transformer. A manufacturable design must specify the physical construction clearly enough that multiple operators, winding machines, suppliers, and production lots can reproduce the intended electrical and safety performance.
The manufacturing definition should preserve the engineering intent behind magnetizing inductance, leakage inductance, DCR, turns ratio, polarity, isolation, winding placement, thermal performance, and parasitic behavior.
The production test plan then verifies the characteristics that can reasonably be measured on every unit while qualification and periodic validation cover characteristics that are too destructive, time-consuming, or system-dependent for 100% testing.
SOLIDMAG ENGINEERING INSIGHT
Manufacturing Controls Are Part of the Engineering Model
If an important design variable is not controlled on the drawing, winding specification, bill of materials, process instruction, or test plan, it is not truly controlled in production.
The objective of Chapter 14 is to translate design variables into measurable and auditable manufacturing requirements.
2. Build the Production Transformer Specification
The production specification should identify the exact materials and construction used to create the validated transformer.
- Core manufacturer, geometry, material grade, and approved alternatives
- Bobbin part number and material
- Target gap or factory AL option
- Primary, secondary, and auxiliary turns
- Conductor type, size, strand count, foil dimensions, or triple-insulated wire
- Start and finish terminals
- Winding direction and polarity
- Turns per layer and layer count where controlled
- Winding order and split-winding connections
- Insulation tape, barriers, sleeving, margins, and approved insulation system
- Shield material and connection where used
- Lead routing, pin assignments, and termination requirements
- Core assembly method, adhesive, clip, or clamp
- Varnish, impregnation, potting, or encapsulation where used
- Electrical test requirements and acceptance limits
- Drawing and specification revision
ENGINEERING CAUTION
Avoid Unqualified ‘Equivalent’ Materials
Changing ferrite grade, bobbin polymer, tape, wire insulation, adhesive, or other safety-related material can alter loss, temperature, spacing, dielectric performance, and regulatory status. Define approved substitutions explicitly and validate them before release.
3. Control Turns, Winding Order, and Polarity
A single missing turn, reversed winding, or incorrect series connection can materially change reflected voltage, auxiliary output, current stress, and regulation.
The winding specification should make polarity visually unambiguous. Useful controls include:
- Start and finish labels for every winding
- Dot convention on the electrical drawing
- Pin numbers
- Winding direction
- Section order for split primaries
- Series-junction definition
- Layer-by-layer turn count where required
- In-process turn-count verification
For an ideal transformer, voltage ratio provides a useful turns-ratio check:
A low-voltage AC ratio test can detect wrong turns or wrong connections without operating the transformer at converter voltage.
SOLIDMAG ENGINEERING INSIGHT
Polarity Is a Safety and Reliability Variable
Incorrect polarity is not merely a regulation problem. It can prevent energy transfer, create abnormal switch stress, reverse an auxiliary supply, or defeat a control scheme.
Polarity and turns-ratio tests should be part of the production strategy whenever a winding error can produce a serious converter failure.
4. Magnetizing Inductance Is a Primary Production-Control Variable
Magnetizing inductance connects the physical gap and turns count to converter current ramp and stored energy. It is therefore one of the most important electrical acceptance measurements.
The production drawing should define the acceptable inductance range and the measurement conditions, including:
- Test frequency
- Test amplitude
- Primary winding terminals
- State of all other windings
- Core assembly condition
- Temperature or reference condition
- Instrument or equivalent method
A measured inductance below the intended minimum can raise peak current. A value above the intended maximum can alter operating mode, current offset, demagnetization timing, and control behavior.
ENGINEERING CAUTION
An Inductance Number Without Test Conditions Is Incomplete
Ferrite permeability, fixture parasitics, instrument settings, and winding connections can affect measured inductance. Specify the test method well enough that supplier and customer measurements can be compared.
5. Leakage Inductance Must Be Measured with a Defined Short-Circuit Method
Primary-referred leakage inductance is commonly measured with the secondary and relevant auxiliary windings shorted according to the defined test method:
The exact shorting configuration matters, especially for multi-output transformers. The production specification should state which windings are shorted, how they are shorted, and the measurement frequency.
Leakage-inductance acceptance limits should be derived from the actual converter tolerance for:
- MOSFET drain overshoot
- Clamp or snubber loss
- Efficiency
- Regulation
- Ringing
- EMI
SOLIDMAG ENGINEERING INSIGHT
Leakage Limits Should Come from Converter Performance
Do not choose an arbitrary percentage of magnetizing inductance as the specification simply because it is convenient. Use the converter’s acceptable voltage stress, clamp loss, and EMI behavior to define the manufacturing limit.
6. DCR Verifies Copper, Length, and Termination Quality
Winding DCR is a useful production indicator because it responds to conductor size, turn count, copper length, termination resistance, and some classes of manufacturing defects.
Specify DCR limits at a known reference temperature or define a temperature-correction method.
A high DCR result may indicate incorrect wire size, too many turns, excessive lead length, poor termination, damaged strands, or unusually high measurement contact resistance.
For milliohm-level secondaries, four-wire Kelvin measurement may be necessary to separate winding resistance from fixture and lead resistance.
A production transformer requires a repeatable test sequence that verifies the electrical, insulation, mechanical, and traceability characteristics that protect converter performance and manufacturing quality.
The production test flow should verify turns and polarity, magnetizing inductance, leakage inductance, winding resistance, dielectric integrity, workmanship, and traceability before shipment. Each test should have defined conditions and acceptance limits tied back to the transformer design and converter-level requirements.
Figure 14-1. Production transformer test flow showing turns/polarity, magnetizing inductance, leakage inductance, DCR, dielectric tests, visual inspection, and final traceability.
7. Dielectric Withstand and Insulation-Resistance Testing
Isolation testing verifies the insulation system between circuits that must remain electrically separated. Common production or qualification methods can include dielectric-withstand testing and insulation-resistance testing.
The required test voltage, waveform, duration, ramp, leakage-current limit, and test points must come from the applicable product standard, approved insulation system, safety-agency requirements, and released transformer specification.
Typical test boundaries may include:
- Primary to secondary
- Primary to accessible core or shield where required
- Secondary to core or shield where required
- Separate isolated outputs where the product requires isolation between them
ENGINEERING CAUTION
Do Not Invent a Generic Hi-Pot Voltage
A production dielectric test is safety-critical. The correct voltage and duration depend on insulation classification, working voltage, product standard, construction, and certification plan. Use the released compliance requirement rather than a generic internet value.
High-voltage test procedures also require suitable guarded equipment, fixtures, discharge provisions, operator protection, and documented work instructions.
8. Visual and Mechanical Inspection
Electrical tests cannot detect every manufacturing defect. Visual and mechanical inspection should verify construction features that preserve safety and repeatability.
- Correct core and bobbin
- Correct winding and tape sequence
- Required margins
- Sleeving placement
- Lead routing
- Pin assignments
- Solder quality
- No damaged wire insulation
- No exposed conductor across an isolation boundary
- Core seating and gap construction
- Core clip, adhesive, or clamp installation
- Potting or varnish quality where applicable
- Marking and lot traceability
- Mechanical dimensions and PCB fit
For safety-critical features that cannot be inspected after completion, use in-process inspection or documented process controls before the feature becomes hidden.
9. First-Article Inspection and Prototype Qualification
Before normal production release, first-article or qualification units should be evaluated more deeply than routine production units.
A qualification plan can include:
- Complete dimensional inspection
- Material and supplier verification
- Turns and polarity
- Magnetizing inductance
- Leakage inductance
- DCR
- Interwinding capacitance where controlled
- Dielectric withstand and insulation resistance
- Converter waveforms
- MOSFET and rectifier voltage stress
- Full-load efficiency
- Thermal rise and winding/core temperature
- EMI performance
- Load and line regulation
- Startup and transient behavior
- Overload and fault behavior
- Environmental or reliability tests required by the product
SOLIDMAG ENGINEERING INSIGHT
Qualification Proves the Design; Production Test Controls the Process
Not every qualification measurement belongs on every production unit. The test plan should distinguish design-validation tests, first-article tests, periodic audits, and fast 100% production tests.
The goal is enough production coverage to detect meaningful process escapes without turning every unit into a full engineering laboratory exercise.
10. Converter-Level Validation Is Essential
A transformer can pass bench LCR measurements and still behave poorly in the actual converter. Final validation must therefore place representative transformers in the intended circuit.
Verify across the required line and load range:
- Primary current waveform
- Peak current and current-limit margin
- MOSFET drain voltage and leakage spike
- Secondary rectifier voltage stress
- Output regulation
- Auxiliary-winding voltage
- Demagnetization timing
- DCM/BCM/CCM behavior where applicable
- Clamp or snubber temperature
- Transformer winding and core temperature
- Efficiency
- Conducted and radiated EMI
The transformer and converter should be validated as one system because parasitics, layout, clamp topology, switching speed, and thermal environment interact.
Bench measurements of the transformer alone are not sufficient to prove converter performance. The finished transformer should be validated in the intended flyback circuit so its electrical waveforms, semiconductor stress, thermal behavior, regulation, and EMI performance can be evaluated together.
Converter-level validation should include primary-current shape and peak value, MOSFET drain-voltage overshoot and ringing, secondary-current behavior, output-voltage regulation and ripple, transformer and semiconductor temperatures, and conducted or radiated EMI margin. These measurements confirm whether the physical transformer and the surrounding converter operate consistently with the assumptions used in the design.
Figure 14-2. Converter-level flyback validation showing oscilloscope checks for primary current, MOSFET drain voltage, secondary current, output voltage, thermal measurements, and EMI verification.
11. Define Acceptance Limits from Design Tolerances
Production limits should map directly to converter-level design limits. A useful structure is to trace each measurable transformer parameter to the failure mechanism it controls.
| Production Parameter | Too Low Can Cause | Too High Can Cause |
|---|---|---|
| Magnetizing inductance | Higher peak current; reduced current-limit margin | Mode/timing shift; slower current ramp |
| Leakage inductance | Usually not a functional problem if within lower bound | Higher drain spike, clamp loss, EMI |
| Primary DCR | May indicate wrong turns/wire or measurement issue | Higher copper loss and temperature |
| Secondary DCR | May indicate wrong construction | Higher loss, regulation error, termination heating |
| Turns ratio | Incorrect reflected voltage and duty behavior | Incorrect reflected voltage and semiconductor stress |
| Interwinding capacitance | Usually lower CM coupling | Higher common-mode current and EMI |
| Dielectric leakage | Possible test artifact at very low values | Insulation defect or contamination |
Acceptance limits should include measurement uncertainty and the actual tolerance budget. If the production specification consumes all of the design margin, the design is not robust enough.
12. Tolerance Stack, Capability, and Statistical Process Control
A released transformer should be designed so normal manufacturing variation remains comfortably inside the electrical specification.
For a measured parameter x, a simple process-capability metric is:
When process centering is also considered:
These statistics are useful only after the process is stable and the measurement system is capable. The appropriate capability target depends on company quality policy, risk, and customer requirements.
Useful parameters for statistical monitoring may include:
- Magnetizing inductance
- Leakage inductance
- Primary and secondary DCR
- Turns ratio
- Interwinding capacitance if important to EMI
- Critical mechanical dimensions
ENGINEERING CAUTION
Do Not Confuse Test Limits with Process Capability
Passing specification limits does not prove the manufacturing process is well centered or statistically capable. Track distributions and trends so drift is detected before units begin failing.
13. Measurement-System Repeatability Matters
When a tolerance is tight, the measurement system can become a significant fraction of the allowable range.
Evaluate:
- Instrument accuracy
- Fixture resistance and inductance
- Lead compensation
- Kelvin connections
- Test frequency and amplitude
- Operator variation
- Contact repeatability
- Temperature
- Calibration interval
A production limit should not be tighter than the test system can resolve reliably. Measurement-system analysis should be used where appropriate.
14. Engineering Change Control and Approved Alternatives
Transformer performance can shift when seemingly minor materials or processes change. The engineering change process should identify which substitutions require review or requalification.
| Change | Potential Impact |
|---|---|
| Ferrite material | Core loss, saturation, permeability, temperature |
| Core supplier / geometry | Ae, AL, gap behavior, loss, fit |
| Wire or foil | DCR, AC loss, fill, insulation |
| Tape / insulation | Dielectric strength, thickness, creepage, temperature class |
| Bobbin | Window dimensions, pin spacing, safety approvals |
| Gap process | Inductance, fringing, tolerance |
| Winding order / layer count | Leakage, capacitance, AC loss, EMI |
| Adhesive / varnish / potting | Thermal path, mechanical stress, safety status |
| Winding machine program | Turns, placement, tension, repeatability |
SOLIDMAG ENGINEERING INSIGHT
A Transformer Is a Controlled Process, Not Just a BOM
Two transformers built from nominally similar parts can behave differently if winding placement, tension, insulation thickness, gap assembly, lead routing, or termination changes.
Revision control should include the construction process as well as material part numbers.
15. Traceability and Production Data
Traceability allows field failures, drift, and supplier changes to be tied back to the actual build history.
- Transformer part number and revision
- Date or lot code
- Core lot / material where required
- Bobbin and insulation material lot where required
- Winding machine program revision
- Operator or line
- Test-station identification
- Measured production data
- Rework status
- Deviation or concession records
Data retention depth should reflect product risk, customer requirements, regulatory needs, and expected product life.
16. Worked Example — Turning the Chapter 13 Design into Production Limits
DESIGN ASSUMPTION
Educational Production-Control Example
The acceptance limits in this example are illustrative. Final production limits must come from the actual converter tolerance analysis, measurement capability, safety requirements, and validated supplier process.
| Design Quantity | Nominal / Illustrative Target |
|---|---|
| Primary turns | 32 |
| Secondary turns | 4 |
| Turns ratio | 8:1 |
| Target magnetizing inductance | 147.3 µH |
| Illustrative AL tolerance | ±10% |
| Illustrative leakage target | 1.5 µH |
| Primary DCR at 25°C | 98.7 mΩ |
| Secondary DCR at 25°C | 1.54 mΩ |
Step 1 — Convert AL Tolerance into Inductance Limits
These values are consistent with the tolerance example developed earlier, but a production release would verify that both limits remain safe across controller and frequency tolerances.
Step 2 — Verify the Realized Turns Ratio
The production test should detect incorrect turns or polarity rather than relying only on visual inspection.
Step 3 — Establish DCR Reference Conditions
The nominal DCR values should be referenced to 25°C or corrected to the drawing reference temperature before comparing measured units.
Step 4 — Define a Production Test Matrix
| Test | Example Frequency | Purpose |
|---|---|---|
| Turns ratio / polarity | 100% | Detect wrong turns, winding connection, or polarity |
| Magnetizing inductance | 100% | Control gap and current-ramp behavior |
| Leakage inductance | 100% or process-dependent | Control coupling, clamp loss, and switch stress |
| Primary / secondary DCR | 100% or sampling per risk | Detect wire, turn, and termination problems |
| Dielectric withstand | Per released safety/manufacturing plan | Verify isolation integrity |
| Visual / mechanical | 100% / in-process as defined | Verify safety construction and workmanship |
| Converter thermal / EMI validation | Qualification / periodic audit | Verify system-level performance |
Step 5 — Link the Test Plan Back to Failure Modes
Every production test should have a reason. Inductance protects peak-current behavior, leakage protects switch stress and clamp loss, DCR protects winding loss and temperature, turns ratio protects reflected voltage and output behavior, and dielectric testing protects the isolation system.
SOLIDMAG ENGINEERING INSIGHT
A Good Production Test Plan Is Traceable to the Design Equations
The most defensible acceptance limits are those that can be traced from converter requirements through the transformer design calculations to a measurable production parameter.
That traceability is one of the strongest advantages of an automated magnetic-design workflow.
A production-control plan should translate each important transformer design parameter into a measurable manufacturing or validation requirement. The 32:4 worked example shows how the original electrical and magnetic design can be linked directly to production tests and pass/fail criteria.
The completed control plan verifies turns ratio and polarity, magnetizing inductance, leakage inductance, winding resistance, dielectric integrity, construction quality, and converter-level performance. This traceability from design equations to production measurements helps ensure that manufactured transformers continue to meet the electrical, thermal, safety, and EMI requirements established during development.
Figure 14-3. Worked production-control example mapping the 32:4 transformer design into turns-ratio, magnetizing-inductance, leakage, DCR, dielectric, and validation tests.
17. Manufacturing, Testing, and Validation Checklist
- Core, material, bobbin, conductor, insulation, gap, and assembly requirements are released.
- Winding order, starts, finishes, polarity, and pin assignments are unambiguous.
- Magnetizing-inductance limits and measurement conditions are documented.
- Leakage-inductance test configuration and limit are documented.
- DCR limits use defined reference temperature and measurement method.
- Turns-ratio and polarity verification are defined.
- Dielectric and insulation tests come from the applicable safety/compliance plan.
- Hidden safety features have in-process controls.
- First-article qualification includes converter electrical, thermal, and EMI validation.
- Production limits are derived from converter tolerance analysis.
- Measurement equipment and fixtures are capable of resolving the limits.
- Process capability is monitored where appropriate.
- Approved alternative materials are controlled.
- Engineering changes trigger appropriate review and requalification.
- Lot and revision traceability are maintained.
18. Design Handoff to Practical Examples and Automation
By the end of Chapter 14, the transformer is no longer only an analytical design. It has a manufacturing definition, measurable acceptance limits, and a validation strategy.
Chapter 15 will bring the complete guide together through:
- End-to-end design examples
- Candidate comparison
- Iteration between electrical, magnetic, winding, and thermal variables
- Automated feasibility checks
- Core and material database selection
- CAD-ready geometry
- Manufacturing drawing outputs
- BOM and test-data generation
- Where engineering judgment still remains essential
SOLIDMAG ENGINEERING INSIGHT
Automation Is Most Valuable When the Design Is Traceable
A useful automated transformer designer should not merely output turns and a core number. It should preserve the chain from requirements to equations, candidate selection, geometry, tolerances, manufacturing controls, and validation outputs.
Chapter 15 will show how that end-to-end structure can become a repeatable engineering workflow.
Technical References for This Chapter
Power Integrations — InnoSwitch3 Family Design Guide (AN-72)
Power Integrations — Power Supply Design Techniques for EMI and Safety
Texas Instruments — How to Design a Flyback Converter
Final dielectric, creepage, clearance, production-test, and qualification requirements must be taken from the applicable end-product safety standards, approved insulation system, customer specification, and certification plan.
Related SolidMagnetics Resources
Ready to Design a Flyback Transformer?
Move from design theory to an automated magnetic design. Enter the converter requirements and evaluate a coordinated transformer candidate, winding geometry, loss estimates, thermal performance, and CAD output.
← Previous Chapter
Next Chapter →
