Skip to main content
  1. Home
  2. Technical Blog
  3. Industry Watch
  4. Testing NEV Components Is Not a Copy of the 12 V Standard - Three Document Layers and Three Extra Verdicts for High-Voltage Parts

Testing NEV Components Is Not a Copy of the 12 V Standard - Three Document Layers and Three Extra Verdicts for High-Voltage Parts

Key takeaways

The difference between high-voltage and 12 V components is not the voltage figure but the number of verdicts. Beyond functional status, insulation, withstand voltage and the timing of safety actions all have to be judged. The base layer offers a menu of conditions, the product layer selects from it, and the admission layer decides whether the result can leave the building - none of the three can stand in for another.

Copying the environmental and electrical-load clauses written for 12 V parts onto high-voltage components is a frequent and expensive misjudgement in NEV projects. What high-voltage parts add is not a higher voltage but three extra verdicts: insulation and withstand voltage, the timing of safety actions, and transients set by precharge and contactor switching. The national standard here is not one document that covers everything - it is three layers, each covering a different span. This article sets out the division of labour, four real differences from 12 V parts, and four items to align before samples are submitted.

1. Three layers: base, product, admission

A convenient shortcut appears often: run everything against GB/T 28046 and add "judged per OEM specification" to the report.

The problem is that this document only answers half the question. It defines the conditions under which a part operates, the methods used to apply them, and how functional status is classified. It does not tell you where the acceptance line sits for a specific product. That line lives in the product standard and the OEM specification.

LayerCoversDoes not cover
---------
Base: GB/T 28046, ISO 16750Environmental and electrical-load conditions, test methods, functional status classesAcceptance criteria for a specific product, safety action requirements
Product: product standard, OEM specification, technical agreementVerdict items, limits, monitoring points, sampling rulesHow generic environmental conditions are defined
Admission: mandatory standards, type approval, export regulationsWhether the product may be sold, which mark applies, which approval routeInternal verification depth and sampling plans

The order cannot be reversed. The base layer offers a menu of conditions, the product layer decides which dishes are ordered, and the admission layer decides whether the table may leave the kitchen.

Using an admission-layer certificate to answer a product-layer question is a standard reason for rejection in supplier audits - a certificate shows the product may be sold, not that insulation still holds after the latest structural change. The adoption relationship between the base layers is covered in ISO 16750 versus GB/T 28046, and where OEM specifications go beyond the national standard is in a separate piece. This article deals only with the NEV-specific misalignment.

2. Four real differences from 12 V parts

Placed side by side, the difference falls into four dimensions, and none of them is solved by raising a number.

Dimension12 V partHigh-voltage partWhere the difference lands
------------
Transient sourceGenerator load switching, stop-start, inductive disconnectionPrecharge circuit, contactor make and break, DC-DC switchingDifferent energy source, different pulse shape and repetition count
Verdict itemsMainly functional status classesFunctional status plus insulation, withstand voltage, safety action timingSeveral verdicts unrelated to function but mandatory
Failure consequenceLoss of function, reset, bus dropoutArcing, overheating, electric shock riskCriteria lean toward safety, monitoring density differs
Test organisationConventional live workLive high voltage, discharge confirmation, two-person verificationThe safety process itself belongs in the test outline

The third row decides a great deal. A 12 V part under vibration is watched for terminal wear and intermittent contact. A high-voltage part under the same vibration is watched for those, plus whether the insulation distance has changed and whether creepage still holds after a fastener loosens. Same shaker, same fixture, entirely different verdicts.

12 V and 24 V cannot be converted by a simple multiplier, and the high-voltage side is not a linear extrapolation of 12 V either. The 12 V versus 24 V comparison covers misalignment within one system; high-voltage parts belong to another system altogether.

One further risk dimension is often missed: partial discharge. It does not show up in continuity testing, yet frequently opens up to a measurable level only after environmental stress. The partial discharge and imaging article covers the sequencing problem in detail.

3. How the three extra verdicts are implemented

First, insulation and withstand voltage are measured in pairs, not added on later. Environmental and mechanical stress are themselves the means by which insulation degrades slowly; the insulation measurement is the ruler that reads the result. One reading describes only the moment it was taken, so one measurement before testing serves as a baseline, one after reads the drift, and long programmes add intermediate points. Test points - positive to case, negative to case, positive to negative - along with humidity state and recovery time must be in the outline: a sample just out of a damp heat chamber carries surface condensation that pulls the reading down, and measuring immediately versus after recovery are two different conclusions.

Second, the timing of safety actions has to be monitored while stress is applied. How fast the circuit is cut after the interlock opens, how long active discharge takes to reach a safe voltage, which degradation path is followed after a fault is injected - none of this sits inside the functional status classes. Class A only says the unit still works, not that it cut off when it should have. Testing afterwards is equivalent to not testing, because the object of verification is the action taken at the moment stress is applied.

Third, read before and after, not once. Insulation resistance, contact resistance and sealing mean little as single points; the drift and its direction are what matter. Sealing can be scheduled alongside the IP rating work, but mind the order: sealing before vibration and sealing after vibration are two different questions.

4. Four items to align before submitting samples

  1. Voltage domain and bus architecture: 400 V or 800 V, dual-voltage platform or not, contactor or semiconductor precharge, integrated or discrete DC-DC. This decides how transient items are scheduled, not which standard applies.
  2. Document layer and revision: which revision of the base layer, the document number and revision of the product layer, whether the admission layer is settled. Three lines beat one sentence saying "per the national standard".
  3. Verdict item list: beyond functional status, whether insulation, withstand voltage and safety actions are in scope, at which points, and who reads them during the test. If this column is empty, no later data can be reviewed.
  4. Agreement on handling failure: high-voltage parts cannot be dismantled on site. The evidence-preservation order and who confirms discharge must be agreed in advance. Negotiating on the spot usually means the sample is lost.

Once these four are aligned the outline normally clears review in one pass. Miss any one of them and it comes back at report review, by which time the sample has usually been through a full round of stress.

A few plain words

Most disputes over NEV components are not about whether the test was run correctly, but about which layer of document owns a given item. Assigned to the wrong layer, even clean data cannot support a verdict.

When TMC LAB takes on this kind of commission, the voltage domain and bus architecture, together with the revisions of the base-layer and product-layer documents, are written into the outline first, and the monitoring points for insulation, withstand voltage and safety actions are marked separately before testing is scheduled. Conditions and limits follow the standard text, the OEM specification and the technical agreement; no values are preset at the outline stage.

Bringing the bus architecture drawing, the product-layer document number and the previous report along with the samples saves far more time than a single sentence saying "test it to the national standard". Scheduling is arranged per project; you can state your expected delivery date when submitting samples, and we will confirm a feasible plan against the current schedule and report progress at each milestone.

Related Q&A

Related questions

Frequently asked questions on the same topic. The first paragraph of each answer is the conclusion.

Can an NEV component be qualified with GB/T 28046 alone?
No. GB/T 28046 is a base-layer document: it defines environmental and electrical-load conditions, test methods and functional status definitions, but it does not supply acceptance criteria for a specific product. High-voltage parts are also judged on insulation resistance, withstand voltage and the timing of safety actions, which come from product standards, OEM specifications and the technical agreement. The workable arrangement is base layer for conditions, product layer for criteria, admission layer for market access, with the revision of each written into the test outline.
Why can load-dump clauses written for 12 V systems not stand in for high-voltage transient testing?
Because the energy source and the pulse formation mechanism differ. A 12 V load dump comes from generator load switching, while transients on the high-voltage side come mainly from the precharge circuit, contactor make and break, and DC-DC switching, each with a different pulse shape, source impedance and repetition count. Applying 12 V clauses to a high-voltage bus misses the real coupling path and can underestimate the overshoot produced by contactor operation. Actual waveforms and counts follow the product standard, the OEM specification and the technical agreement.
Why measure insulation resistance twice instead of once?
A single reading only describes the moment it was taken. Environmental and mechanical stress act precisely by degrading insulation slowly, so the measurement has to be paired: one before testing as a baseline, one after to read the drift, with intermediate points added where needed. The measurement also needs agreed test points, humidity state and recovery time - a sample just out of a damp heat chamber carries surface condensation that pulls the reading down, so measuring immediately and measuring after recovery give different conclusions. Criteria and thresholds come from the OEM specification or the technical agreement.

These answers share the same data source as the FAQ page.

Have a testing requirement? Send us your product details

Tell us the product model and test requirements. Our engineers will map out the test plan, the applicable standards and the item combination, and provide a quotation reference. The more specific the request, the sharper the answer.

  • Free scoping of the test plan and standards
  • Pre-scan testing during R&D to cut rework cost
  • On-site visits and witnessed testing supported
  • Scheduling per project, with progress updates at each milestone

Received — we will contact you shortly and follow up with a test plan and quotation.

By submitting you agree to be contacted by phone or email. Scheduling follows the project’s actual availability and is subject to mutual confirmation.

13510907826