HVCH Sample Acceptance: Linking Test Results to the Production Configuration

"The sample passed" is not a complete sentence

"Sample test results" apply only to specific items, specific test methods, and specific dates; they do not represent the overall performance of the product. If the approving authority mistakenly interprets them as proof of the product's overall performance, problems will inevitably arise.

We have repeatedly observed this pattern on the supplier side: a sample passes approval, mass-produced products are subsequently delivered, but failures occur during vehicle testing. Investigations have revealed that the issue was not a single component failure, but rather a change in configuration—for example, a software version one update newer than the sample submitted for testing, the use of connectors from a different supplier, a change in sealing material due to supply issues, or the use of a component batch not included in the sample. In reality, no component "failed"; rather, the product configuration itself changed while the approval status remained unupdated.

We have a clear principle, which we communicate well in advance of any sample discussion: if the "version" field in the sample record is left blank, it cannot be considered "approved" but must be treated as a "pending issue." Even if approval is forcibly signed off on in such a situation, the risk is not transferred to the supplier; rather, the problem is merely deferred until the vehicle project phase, at which point the cost of resolving it will be far higher than if it had been addressed during the approval meeting.

Five Key Fields That Give Test Results Clear Meaning

Test results are only meaningful when accompanied by the following five fields: model number, hardware version, software or firmware version, control interface type, and test conditions. If any of these fields is missing, the test result loses its "context"—it may still appear to be evidence, but it can no longer be linked to any specific product configuration.

For heaters, the "control interface type" field is particularly critical. If test results are based on a "Series B" unit equipped with a CAN bus and an enable interface, they are of absolutely no reference value for the "Series E" unit you actually ordered—which is controlled by a thermostat—even if both units have the same power rating. Because the interfaces differ, the vehicle's logic and fault behavior are entirely different. Identical power ratings do not mean the products themselves are identical. Similarly, differences in power and voltage specifications follow this logic: performance data of 20–24 kW measured within the 400–800 V voltage range applies only to the specific product model that generated that data, not to the general "series name" listed in the product manual.

If these five fields are explicitly listed in the sample record, then the test results have an accurate "pinpoint"; if only three of them are listed, then the results fall, at best, within a rough "range." The approval process should require the provision of this precise "pinpoint."

Bench results and vehicle results are two different statements

Bench test results are real data specific to the bench environment: they are based on controlled circuits, stable supply conditions, specific cooling media, and are free from installation constraints. In-vehicle test results, on the other hand, reflect the system's performance under actual installation conditions, encompassing supply characteristics under load, environmental fluctuations, operating cycles, and the actual installation locations of components. Both types of results are critical and cannot be used as substitutes for one another.

The most commonly overlooked issue during the approval process is the question of "translation": Under what vehicle operating conditions do bench test results remain valid? And which vehicle test should be used to fill the remaining validation gaps? It is essential to clarify this issue in writing prior to approval. If a failure occurs during the vehicle testing phase, the same issue will escalate into a dispute over "who should have raised this question in the first place."

There is a trade-off here, and we do not shy away from acknowledging it. Requiring validation evidence under full-vehicle operating conditions prior to approval will slow down the project schedule; accepting only bench test evidence will accelerate progress but shift the risk to the subsequent full-vehicle development phase. We prefer to have these "time-consuming" discussions early on—not only is this less costly than investigating failures during the vehicle assembly phase, but it is also the only opportunity to have such discussions before the hardware is finalized.

There is already established terminology in this field to define this process.

The industry already has a name for this discipline

In the automotive supply chain, this process is known as the "Production Part Approval Process" (PPAP). It involves samples and related supporting documentation and is closely linked to the production process; re-approval is required whenever a new part is introduced, an existing part is modified, or changes are made to the production process. This procedure is part of the industry's "Advanced Product Quality Planning" (APQP) system, whose core principle can be summarized in one sentence: Every result must be based on a clearly defined configuration; every change must undergo re-approval.

You do not need to adopt the entire system verbatim; simply draw on its core principles. The remainder of this guide essentially applies this principle to the specific context of heater products: sample records correspond to "configuration"; deviation lists correspond to "change control"; and production release conditions correspond to "approval"—ensuring that both parties can verify the details through clear documentation.

Items to Inquire About in Writing from the Supplier

Please include the following questions when shipping samples, and file the supplier's responses along with the purchase order. Specific values are not listed here because each value depends on the specific model and test items:

• Complete configuration records for the tested samples: including model, hardware version, software or firmware version, and control variant.
• Specific conditions underlying each test result: Cover circuit configurations, power supply, coolant, environmental conditions, and load/duty cycle, rather than simply providing the final test conclusion.
• Explanation of the transition from bench testing to vehicle application: Clearly state which vehicle operating conditions the bench test results apply to, and specify which vehicle tests are required to validate the remaining conditions.
• List of deviations: List all differences between the tested sample and the final production configuration; if there are no differences, a written statement must be provided.
• Trigger conditions for re-approval: Specify the types of changes to hardware, software, interfaces, materials, or production processes that require sample re-approval prior to shipment.
• Production release conditions: Specify the documents and their version numbers that production units must comply with (both parties should retain this document).

Information Required from You

If you wish us to confirm sample approval prior to product release, please submit a sample record containing the five elements listed above, along with the test conditions and the vehicle operating conditions you expect the test results to cover. We will compare this information with the production configuration; if any inconsistencies are found, we will point them out promptly before the approval process is completed (i.e., while the process is still in the document review stage). We will not directly apply the test results of a specific sample to configurations not built by us, nor will we speculate on relevant parameters for models we have not designed.

Related reading

• IP67 and EMC in High-Voltage Heaters: What to Check Before Installation
• Products

Frequently asked questions

Does a sample passing the test mean that the mass-produced product is automatically approved?

No. Test results apply only to the specific configuration used under specific test conditions. Mass-produced products are considered within the scope of approval only when they are fully consistent with that configuration in all parameters. If any discrepancies exist, they must be treated as a new application, and the corresponding validation evidence must be resubmitted.

Can a sample that has passed CAN bus testing serve as the basis for approval of a temperature control device?

No. The interfaces, vehicle logic, and fault manifestations differ between the two. Even if the power rating is the same, they cannot be considered products of the same category; no test results can be directly applied across these boundaries.

Which changes require re-approval?

Any change involving the five key parameters (fields), or any process change that causes the product to deviate from the validated configuration, requires re-approval. These trigger conditions must be clearly defined in writing prior to approval, rather than being addressed retroactively after a failure occurs—at which point any change may appear "obvious."

Who is responsible for the deviation list?

The party designated in writing in the qualification documentation—that is, the supplier or integrator—is responsible. If the deviation list does not explicitly identify the responsible party, it is merely a "to-do list" rather than a formal plan; until a responsible party is identified and a document number is assigned, the list should serve as a barrier to product release.