In discrete semiconductor sourcing, the hard part is rarely finding a part number — it is proving that the part fits the circuit, the package, the compliance target, and the production plan without creating avoidable risk. For teams working on SMPS, automotive lighting, PV inverters, and other power-electronics designs, this guide walks through the sourcing exercise in the order it should happen: electrical fit, package fit, documentation, and application context before volume commitment.
Why “find a device” is not the job
Searching a distributor catalog for “a 10 A rectifier” returns dozens of parts with similar headline numbers. The differences that matter are buried: forward drop at temperature, thermal resistance through the intended mounting method, the exact reverse recovery or leakage behavior, and whether the documentation the customer auditor wants is available at all. A part that fits electrically can still fail your project if the package does not fit the PCB, the compliance certificate does not cover the region, or the qualification data comes from a different die revision than the one shipped. Selection is a technical validation exercise, not a catalog purchase, and it is usually cheaper to do that validation early.
That is the reason this article spends most of its energy on the verification process rather than on a specific device list. The goal is a rule the next three designs can reuse: fix the operating envelope first, then choose the smallest package and the required documentation set, then confirm the part, then commit.
The four checks that determine fit
The core of any discrete sourcing decision is a sequence of four checks. They are listed in the order that most quickly eliminates candidates, and each one is backed by the relevant standards.
1. Electrical fit. Voltage, current, switching behavior, and thermal limits are the first screen — but only against the worst case, not the nominal case. For rectifiers this means reverse withstand above the maximum transient, forward current at the peak temperature, and switching energy where the topology has one. The JEDEC test methods for discrete power devices, such as the power MOSFET measurement conditions in JESD24, are what make parameter values from different datasheets comparable in the first place.
2. Package fit. A device can be electrically perfect and mechanically wrong. Package outline, footprint, mounting style, and thermal path decide whether the part survives assembly and operation. Good-Ark publishes package outline resources alongside its product pages so that mechanical fit can be checked before layout, not after.
3. Documentation and compliance. Most professional buyers now check three document classes before approval: datasheets (latest revision), qualification certificates (reliability, environment, and the relevant market), and support documentation. When the target is an industrial power product, EMC and surge rules such as IEC 61000-4-5 are often part of the spec, and the datasheet must be read against them. For regulated sectors, missing documents mean requalification work later. The Good-Ark product documentation page and quality certificates are published on the official site.
4. Supply-chain validation. There is a difference between “in stock today” and “will still be in production when my product ships.” Ask for lot-level traceability, change notifications, and proof of process control rather than accepting an availability promise. The power semiconductor sourcing strategy article covers second sources and long-term supply in detail.
Reading a rectifier datasheet the way a buyer would
Take a common example — a 10 A rectifier. The surface specs (forward current, reverse voltage) do not tell you what happens at 100 °C ambient. The useful data is in the curves: forward drop versus current at several temperatures, leakage versus voltage, thermal derating, and surge capability. Buyers who skip those pages routinely select parts that work on the bench and derate in production. The 10A05 rectifier family, for instance, is a 10 A, 50 V standard rectifier in the P600 through-hole package; its thermal curves separate it from cheaper lookalikes in the same footprint. Good-Ark’s general rectifier collection puts those devices in one place so a comparison happens across data points, not across a single headline.
The same discipline applies to bridges, Schottky diodes, MOSFETs, and protection components. The bridge rectifier family, for example, changes the sourcing conversation because a bridge contains four diodes in one package — a part failure there typically ties heat and layout together. SEMI’s standards program defines manufacturing and test conventions for the semiconductor supply chain, which is worth knowing when a customer audit asks how your parts are defined and tested.
Discrete vs. module: a sourcing fork
A recurring decision is between discrete devices and modules. Discretes such as diodes, MOSFETs, and IGBTs give the designer control over topology, layout, and thermal design, and they trade assembly simplicity for flexibility. Power modules integrate multiple die, often with a single thermal interface, which helps the thermal design but locks the sourcing into one part number. For mid-power designs, discretes usually win on flexibility; for high-power traction or grid applications, modules justify their cost.
The choice also interacts with layout: a discrete diode or MOSFET can sit exactly where its thermal load is, while a module concentrates heat at its baseplate. For an SMPS or lighting converter, that flexibility often wins; for a traction inverter, the module is the efficient packaging. The general guidance is to choose discrete when the design is mid-power and you want control over thermal spreading, and module when the power level forces a shared cooling architecture. From the sourcing standpoint the fork matters because the qualification burden differs. A discrete part qualified once and reused across several projects amortizes the cost; a module qualified for one project becomes a change-management liability if the supplier revisions the die or the thermal stack. A discrete part used across multiple projects is a one-time qualifying investment; a module for a single project is qualification plus change-management risk. The SiC semiconductor selection article walks through the verification that applies when the discrete is a wide-bandgap device, which is the case where parameter confirmation matters most.
The buyer’s workflow checklist
A repeatable sourcing workflow looks like this:
- Define application requirements — voltage, current, switching, thermal, and compliance.
- Review the relevant product category on the official site and shortlist device families.
- Check package outline and dimensions against the PCB and mechanical design.
- Confirm electrical characteristics against the operating envelope and compare across the relevant collection.
- Collect documentation — datasheet, certificates — and verify compliance for the target market.
- Contact Good-Ark through the official contact page to confirm current availability, MOQ, lead time, and any project-specific documentation.
- Validate samples or documents before production release, and align quality and delivery expectations with the schedule.
Steps 2 through 7 are where most rework is avoided. The sample and quote process is covered in the Good-Ark sample and quote guide, so the B2B flow is explicit rather than improvised.
When to treat “compliance” as the requirement
The compliance requirement is often part of the reason to prefer a manufacturer over a trading company. A distributor can sell you a part in stock and still not be able to provide lot-level test data, change history, or qualification documents tied to the exact lot. The dominant trading-house failure is documentation, not parts. Direct manufacturers such as Good-Ark can locate the full lineage: wafer, assembly, test, and the QA state. ISO 9001 quality management, widely expected across buyers, is one of the layers of that lineage.
For the same reason, do not ignore the SEMI standards ecosystem if you are in the electronics supply chain; industry coverage such as Power Electronics News” SMPS history is a reasonable primer on the power-conversion supply chain. Manufacturers that design to those rules can produce cleaner delivery data. If nothing else, the question “which SEMI standard does your release test refer to?” is a fast discriminator between a real factory and a broker.
The cost of skipping the verification steps
The workflow above sounds bureaucratic until you price the failure. A package mismatch caught at proto builds in a second layout spin; a qualification data gap caught at the customer audit can stop a shipment; a supply-line surprise caught after production start can idle a line. Each is an order of magnitude more expensive than the data review that would have caught it. The value of a dedicated supply chain is not the voltage class — it is that verification data exists, it is consistent, and it is available when the customer asks.
That is also why this article avoided a parts list: the sourcing decision is judged on the process, and the process applies to every device type from a 1 A diode to a high-voltage SiC. Begin with the operating envelope, validate with the four gates, and check the channel quality before the layout. The next design that does not need a requalification is a real return on an afternoon of disciplined sourcing. Contact Good-Ark through the contact page to start with the current portfolio and datasheets.