Before adding capacitance to a DC-DC converter input, check the source network, its resonant frequency, and the converter’s input behavior. This review covers the DC-side input LC network; selecting an AC mains EMI module or an output filter is a separate task. A filter that reduces switching noise can also create an impedance peak that interacts with the converter’s control loop. More capacitance is therefore not a general cure for an unstable or noisy input.
विषयसूची
An SMPS input filter can reduce conducted switching noise, but its interaction with the converter matters. For an OEM hardware engineer, the immediate decision is whether a proposed filter can move into controlled prototype testing or needs further analysis first. A useful review combines the exact converter’s requirements, the source and cable model, component limits, and frequency-dependent evidence. A static impedance calculation provides a screening flag, not a stability certificate.
Identify the complete input network
Treat the filter, converter and surrounding circuit as one controlled assembly, as described in OEM buck-module qualification.
Draw the upstream supply output, distribution protection, connectors, positive and return wiring, added inductance, external capacitance, and converter input. State where a filter model begins and ends. An ideal laboratory voltage source connected with short leads does not reproduce a long harness powered by a current-limited upstream supply.
List the converter’s existing input capacitance where documentation makes it available. Include capacitor equivalent series resistance, inductor winding resistance, layout parasitics, and source behavior in the applicable model. Leaving these out can change the predicted damping and the frequency at which an interaction appears.
The external application report SNVA801, Analysis and Design of Input Filter for DC-DC Circuit shows why an LC filter can affect a converter control loop and create resonant amplification. Its worked converter is a specific example; its component choices are not a transferable WEHO filter recipe.
Understand negative incremental input resistance
Within a regulated operating region, assume temporarily that output power and efficiency remain constant. The input then approximately behaves as a constant-power load: Iin = Pin ÷ Vin. If input voltage falls, input current rises to maintain that power.
Differentiating this simplified relationship gives rin = dV/dI = −Vin2 ÷ Pin. The negative sign describes the slope around an operating point; it does not mean the converter generates power or has a negative DC ratio V/I. The DC voltage and current can both remain positive.
The referenced paper on input source impedance explains this mechanism and its interaction with a non-ideal source. Real input impedance varies with frequency, control behavior, and operating state. The simple expression does not apply unchanged during startup, current limiting, or undervoltage shutdown.
Screen a hypothetical LC proposal
This original calculation is a hypothetical review exercise, not a measured response, a selected WEHO converter, or a component recommendation. Assume a regulated 120 W output, 90% conversion efficiency, and a proposed 10 µH / 100 µF input LC network. Treat efficiency as constant only for this screening calculation; use effective capacitance and inductance in a real design.
| Quantity | Calculation | Result |
|---|---|---|
| Converter input power | 120 W ÷ 0.90 | 133.33 W |
| Input current at 24 V | 133.33 W ÷ 24 V | 5.56 A |
| Incremental resistance at 24 V | −24² ÷ 133.33 | −4.32 Ω |
| Input current at 18 V | 133.33 W ÷ 18 V | 7.41 A |
| Incremental resistance at 18 V | −18² ÷ 133.33 | −2.43 Ω |
| Ideal LC natural frequency | 1 ÷ (2π√(10 µH × 100 µF)) | Approximately 5.03 kHz |
The lower input voltage reduces the magnitude of the negative incremental resistance while increasing the steady input current. That makes the low-voltage, high-power condition an important review corner. The 5.03 kHz calculation identifies an ideal natural frequency; it says nothing by itself about the actual impedance peak or damping.
Suppose, additionally, an unvalidated source-network model predicts a 3.0 Ω impedance peak near that band. Dividing by the static magnitudes gives 3.0 ÷ 4.32 = 0.69 at 24 V and 3.0 ÷ 2.43 = 1.23 at 18 V. These are screening comparisons against a low-frequency approximation. They are not the actual frequency-dependent impedance ratio and cannot establish stability or instability. Both cases still require the converter’s dynamic input information and the complete source model.
Request the dynamic evidence
Define the source envelope at the converter terminals first. A battery-fed DC-DC input-range review helps define its low-input operating corners before assessing filter interaction.
Then confirm that the source network and converter are each stable in their defined standalone conditions. Where suitable models or measurements are available, compare the magnitude and phase of the source network’s output impedance with the converter’s input impedance over the relevant frequency range. Define the same physical port for both. The resulting frequency-dependent impedance ratio is more informative than a DC resistance reading or one resonant frequency.
Ask the converter manufacturer what external input network is permitted, which operating modes its model covers, and how an added filter changes the integration requirements. If a recommended network exists, retain its topology, component characteristics, layout assumptions, and operating conditions together. Copying only a capacitor value strips away essential context.
A well-separated impedance relationship can be a useful design criterion under its stated assumptions. An overlapping magnitude alone does not fully describe the closed-loop system. Where the review identifies a possible interaction, an experienced power designer should choose the applicable analysis and confirm it with appropriate testing.
Check component and startup limits
| वस्तु | Check under actual conditions | Why the value alone is insufficient |
|---|---|---|
| Input capacitor | Voltage rating, effective capacitance, ripple current, ESR, temperature | Bias and temperature can change capacitance and damping |
| Series inductor | Saturation current, thermal current, winding loss, inductance versus bias | Startup and load peaks can exceed the steady current |
| Source and harness | Current limit, cable inductance, resistance, connectors | The filter operates with this source network, not an ideal source |
| Converter input | Permitted capacitance, voltage limits, startup and protection behavior | A running-state calculation does not cover energization |
| Any damping network | Manufacturer design method, component stress and dissipation | Changing a network can trade resonance against heat and efficiency |
Do not insert an arbitrary series resistor or replace capacitors with unspecified low-ESR parts as a field fix. A damping solution needs a defined topology and verified stress limits. It also needs an electrical and thermal review for the assembled equipment.
Use a controlled validation plan
Keep the electrical noise measurement method explicit; use the ripple acceptance and measurement matrix for that separate evidence record.
First record a repeatable baseline with the intended source, harness, input voltage range, and output load states. Identify the disturbance being investigated: conducted noise, startup ringing, sustained oscillation, or load-step recovery. Those observations call for different evidence and should not all be called “ripple.”
Then test the proposed network using suitably rated equipment and defined measurement points. Capture converter input voltage and relevant current during energization, load changes, and restart. Cover the low-input/high-power corner and component tolerances that the design review identifies. Set acceptance limits from the exact converter and system requirements before judging the traces.
A clean time trace in one test does not replace frequency-domain analysis where interaction remains a concern. Conversely, a screening flag should trigger investigation, not an unsupported claim that the converter is defective. Keep model assumptions and test coverage in the same review record.
Send a specific filter inquiry
For a specific non-isolated assembly, see WH-C241210 product details and the matching WH-C241210 datasheet (PDF). Its input and output share a negative reference; that does not remove filter interaction. These documents identify the item, but do not qualify the hypothetical LC filter above. Request the exact input-network limits and applicable dynamic evidence. Contact WEHO about input-filter integration through the contact form with:
- The converter model or required input/output ranges and load states.
- A source, harness, and filter diagram with component part numbers.
- The input-voltage and current traces, including startup conditions.
- The noise or stability symptom and the intended acceptance criterion.
This makes the next step specific: obtain a documented integration recommendation or identify the additional analysis required before committing the network to production.
