The question "should we test to ISO 16750 or to GB/T 28046" is not really about two standards, but about two routes to the same one. The GB/T 28046 series is the national standard that adopts the ISO 16750 series with modifications (MOD); its five parts map one-to-one and share the same clause skeleton. What actually gets a report returned during review is a mismatched version year: the published national standard stopped at the 2011 edition, while Part 4 of ISO has moved on to the 2023 edition, where the temperature-class chapter uses a completely different code set.
1. The adoption relationship: MOD, not a translation
GB/T 28046 and ISO 16750 stand in a modified adoption (MOD) relationship. The foreword of the standard carries an adoption statement naming the corresponding international standard, the edition it corresponds to, and the modifications that were made. That statement is the only authoritative basis for judging the relationship and is more reliable than any second-hand interpretation.
Modified adoption means three things:
- The framework follows: chapter structure, test items and pass/fail logic track the international edition;
- Changes are permitted: editorial changes may be made for domestic industry conditions, individual clauses may be deleted, or certain requirements may be stated more specifically;
- Clause-by-clause equivalence is not promised: so finding different wording, or a missing clause, when you compare the national text against the ISO original is a normal outcome of the adoption mechanism, not a printing error.
In one sentence: one methodology, two texts. Project results are usually close, but a report must never blur which one it used.
2. How the parts map
| GB/T 28046 part | Scope | Corresponding ISO 16750 part |
|---|---|---|
| Part 1 General requirements | Functional status classification, general test requirements, mounting-location categories | 16750-1 |
| Part 2 Electrical loads | DC supply voltage, overvoltage, superimposed AC voltage, load dump, voltage drop and start-up characteristics | 16750-2 |
| Part 3 Mechanical loads | Sinusoidal vibration, random vibration, mechanical shock, free fall | 16750-3 |
| Part 4 Climatic loads | Temperature classes, temperature cycling, damp heat, salt spray, solar radiation | 16750-4 |
| Part 5 Chemical loads | Reagent list and test methods | 16750-5 |
A note on sourcing: the part division above follows the table of contents of both standards. The currently published national edition is the 2011 edition, corresponding to the 2003–2007 parts of ISO 16750. For the exact edition mapping, defer to the adoption statement in the foreword and to the national standards full-text public system.
The most practical use of this table is catching missing items. If the customer protocol says "test to ISO 16750-3", the mechanical loads you have to run sit in that single part; do not casually fold climatic items in to round out the scope. Conversely, when a protocol says only "to ISO 16750" without naming parts, you must go back and ask which parts — the test severity of the five parts differs enormously.
3. Version-year mismatch: three practical problems
1. Temperature classes: the code set has already been replaced
The 2023 edition of ISO 16750-4 changed the old letter-based temperature classes to four levels, U, V, W and X, each with its own high and low temperature endpoints. The 2011 national edition does not contain this code set at all.
The consequence is immediate: when the customer's technical protocol says "to class U" and all you hold is the national edition, the clauses do not line up and picking the class becomes guesswork. This is not a job for guessing — pick the wrong class and both the test severity and the conclusion are wrong. The division of the four classes and the selection logic have been covered separately (what changed in the ISO 16750-4:2023 temperature classes); the point here is simply that without a clear answer on the edition, the conditions cannot be set.
2. Electrical loads: fine print in the condition definitions is the easiest to miss
Electrical loads is the most frequently revised part; clauses on supply voltage range, voltage drop and reset behaviour, and start-up characteristics may have their condition wording adjusted in every revision. Part 2 of the national standard still follows the 2011 wording.
For foreign-invested customers and export projects, protocols often cite the ISO part directly. The right move is not to "approximate it with the national standard" but to open the protocol, confirm which edition it cites, and configure to that edition's condition table. If the two editions disagree, run the stricter side and say so in the report.
3. Mechanical loads: the mounting-location scheme decides the class you pick
Vibration and shock severity classes are tied to mounting location — engine bay, passenger compartment, luggage compartment, exterior attachments — each with a different severity level. How mounting locations are divided and which class each takes differ subtly between editions, so the same part tested to two editions may end up at different severities.
The detail behind those condition differences (what PSD, Grms, peak acceleration and pulse width actually mean) is broken down in the condition differences for mechanical loads and is not repeated here.
4. Which standard number goes on the report
| Project type | How the report is written | Note |
|---|---|---|
| Direct commission from a domestic OEM or Tier 1 | Cite the corresponding GB/T 28046 part plus year | Most domestic technical protocols cite the national standard directly |
| Foreign-invested customer or export project | Cite the corresponding ISO 16750 part plus year | Protocols usually cite the international edition; the edition year is not optional |
| Both required | One test run, two reports | Only if the conditions can be satisfied simultaneously; where they differ, run the stricter side and note it |
| Only a verbal statement from the customer | Confirm the standard number and edition in writing first | A verbal edition cannot serve as the basis for a report |
The most common mistake: the report cover says ISO 16750 while the condition table is copied from the 2011 national edition, with not a word about the edition year. The reviewer reaches the edition line and one question stops the report.
One further layer is easy to overlook — functional status must resolve to a specific class whichever text you use. Whether a reset, a communication dropout or a display flicker during the test counts as a pass, and how the A-to-E scale is applied, must go into the test plan together with the standard number (the A–E classification in GB/T 28046.1). Get the standard number right but leave the criteria vague and the report still fails review.
5. A few straight answers
To judge whether a report is usable, look at three places first: the standard number, the edition year, and the test condition table. With all three present the report stands; miss any one and it only looks right.
An adoption relationship is not clause equivalence. Copying the standard number and edition cited in the customer protocol into the test plan at project kick-off is the cheapest step available — when TMC LAB takes on an environmental reliability commission, we check the edition cited in the protocol first and build the condition table from it, so that nobody discovers after the fact that the class was picked on the wrong basis.
German OEM projects add one more layer: LV 124 and ISO 16750 overlap on some test items and diverge on others, and the customer's corporate standard governs.
For clause-by-clause breakdowns, start with the ISO 16750 and GB/T 28046 standard pages; electrical loads detail is under ISO 16750-2 and German alignment under LV 124.
Turnaround time is arranged according to the actual project. You are welcome to state your target date when submitting samples; we will check it against the current schedule, confirm a feasible plan with you, and update you at each milestone.
