Start with the conclusion: responsibility follows the object under test, not the invoice
When a device manufacturer and a component supplier start arguing, the opening question is almost always "who pays for this test". What actually decides ownership is never the cost, it is the state the object under test is in:
- The object under test is the component itself, separated from the device - the supplier obtains the evidence, and the device manufacturer verifies and retains it.
- The object under test is a system fitted to the device and operating under the device's real conditions of use - the evidence can only be obtained by the device manufacturer on a fully assembled sample. No supplier can produce it on your behalf.
That single rule resolves most of the argument. What is left usually concentrates on one thing: whether the scope of the component-level evidence the supplier handed you actually covers your installed configuration.
Four interfaces where the dispute keeps happening
Batteries and chargers. What the supplier can provide is a safety evaluation at cell or battery pack level, for example a report issued to IEC 62133-2, where the object under test is the battery itself. What the device manufacturer owns is the behaviour of battery, charger and device in combination - the battery and charger requirements addressed by ISO 7176-25 as fitted to the wheelchair. Passing the former does not mean passing the latter, because the latter's object under test includes the routing of the wiring, the mounting position, nearby heat-generating parts, ventilation conditions and the matching of the charging interface. Submitting a supplier report as if it were a whole-device conclusion is one of the most frequent errors on this chain.
Controllers and drive systems. The electrical characteristics of a controller can be evaluated at component level, but the way the control logic behaves on the assembled device - start and stop behaviour, response under fault conditions, coordination with the brakes - has the whole device as its object under test by definition. The split here: component-level evidence from the supplier, device-level behaviour verified by the device manufacturer in the assembled state, and the two pieces of evidence traceable to the same component version in the technical file.
Structural parts. For frames, seat supports and footrests, suppliers can rarely provide usable evidence, because strength items have the assembled device as their object under test in the first place. Take the same length of tubing: change the assembly method, change the weld positions, and the load path changes with it. There is essentially nothing to hand off here. What the device manufacturer can do is require consistency evidence for materials and processes from the supplier, so that material variation is at least pinned down as a variable.
Other bought-in electrical and connection parts. There is a further group of items, for example the part corresponding to ISO 7176-31, where the two sides regularly push responsibility back and forth. The method does not change: first establish whether the object under test for that item is a standalone component or the assembled device, then assign ownership. Confirm the content and conditions of application of the specific items against the current valid version of the standard text.
Responsibility split at a glance
| Type of evidence | Object under test | Who normally obtains it | What the device manufacturer must still do | Consequence if it is missing or mismatched |
|---|---|---|---|---|
| Cell and battery pack safety evaluation (IEC 62133-2) | The component itself | Supplier | Check that the model on the report matches the model actually fitted, word for word, and that the report is current | The device-level electrical assessment has no foundation and may be halted pending component-level evidence |
| Battery and charger as fitted to the device (scope addressed by ISO 7176-25) | Assembled device | Device manufacturer | Provide an assembled sample matching series production, with the charger it ships with | Substituting a component report is a mismatch of the object under test, and the conclusion does not stand |
| Structural strength | Assembled device | Device manufacturer | Lock down material grade, welding process and assembly method | The supplier changes material without notice and completed strength data stops being representative |
| Control system behaviour | Both component and device level | Split between the two parties | Confirm the controller version matches the one verified | Controller firmware is revised, behaviour changes, and nobody re-verifies |
| Material and surface treatment properties | Material or component | Supplier | Check grade, supply batch and change records | After a material substitution, contact-related and durability conclusions are left hanging |
| Quality system and process evidence (ISO 13485) | Organisation and processes | Each party maintains its own | Retain supplier evaluation, purchasing control and change control records | The evidence chain breaks during a system audit and cannot be traced back to source |
The battery chain is where most projects get caught
Sort through the cases and battery-related mismatches come in three forms.
One: the model on the report does not match the model fitted. One character different in the suffix, a different capacity grade, a revised internal protection board - the report is genuine, it just does not correspond to the sample in front of you. This gets flagged directly during assessment, and what has to be reworked is not the test, it is the evidence chain of the entire technical file.
Two: battery and charger come from different suppliers, and "the plug fits" is treated as a match. A plug that fits only shows the mechanical interface is compatible. Charge termination logic, the trigger conditions for protective action, behaviour under abnormal conditions - none of that has been verified by anyone. When a problem surfaces on the assembled device, both suppliers will say their own component is conforming, and both are right, because nobody owned the combination.
Three: the scope of the supplier report does not cover the device's conditions of use. The component was evaluated in an open environment; on the device it sits inside an enclosed compartment right next to a heat-generating part. The conditions of use have changed. This mismatch is hard to spot and is usually traced backwards only after something anomalous shows up in device-level testing.
There is only one way to handle it: check the supplier report at the moment you receive it, not the week before samples ship. For powered products, run your overall test package and sample requirements against the scope of powered wheelchair testing first.
Intake check for supplier reports
| Check | What to look at | What to do if it fails |
|---|---|---|
| Model consistency | Compare the model on the report cover against the purchased part number and the marking on the installed part, character by character | Ask the supplier for a written model correspondence statement, or obtain evidence for the correct model |
| Scope of application | Whether the configurations, structures and conditions of use covered by the report include your installed state | Anything outside scope gets verified by the device manufacturer at device level |
| Report status | Who issued it, whether it is current, and whether it is the full report rather than a summary page | A summary page alone is never accepted - ask for the full report |
| Version and changes | Whether the component has been revised and whether the report corresponds to the current version | If there is a revision but no new evidence, treat it as a new component |
| Associated parts | Whether the charger, cables, protection board and similar parts share the same source and are covered | Obtain separate evidence for associated parts, or bring them into device-level verification |
| Contractual duty | Whether the purchasing contract states that the supplier maintains this evidence over time | Add the clause, with an explicit obligation to notify changes |
The system view: turn verbal agreements into purchasing documents
Under an ISO 13485 framework, this sits across purchasing control, supplier evaluation and change control. In practice there is one action that matters: write into the purchasing documents which evidence the supplier provides, under what circumstances that evidence ceases to be valid, and who obtains it again once it does. An email promising cooperation is not a substitute.
"The supplier shall cooperate in providing relevant documentation" is the same as writing nothing. When evidence actually has to be topped up, the scope, the deadline and the cost all have to be renegotiated - and by then the product is usually sitting there waiting to ship.
Changing supplier or part number: when re-verification is mandatory
Four questions. If the answer to any one of them is yes, take the scope of retesting seriously.
- Does the change alter the physical make-up of the object under test - material, structure, internal devices?
- Does the change alter the device's interface, mounting method or conditions of use?
- Does the scope of the existing evidence still cover the post-change state?
- Does the change fall on something flagged as a key characteristic in the design output?
| Type of change | Usual conclusion | Exceptions to watch |
|---|---|---|
| Same supplier, same model, new batch | Normally does not trigger retesting | The supplier changed a process or a sub-part number internally without telling you |
| Same supplier, different model | Component-level evidence must be obtained again | If the interface or mounting method changed too, device level needs review as well |
| New supplier, parameters claimed identical | Component-level evidence must be obtained again | Identical parameters do not mean identical implementation, and device-level behaviour may change |
| Material or wall thickness of a structural part adjusted | Strength conclusions need reassessment | Local reinforcement is still a change - do not assume a change in the safe direction is exempt |
| Controller firmware upgrade | Depends on what was changed | Anything touching fault response or brake coordination must be re-verified at device level |
Three things the contract has to say plainly
- The evidence list and who obtains each item. Itemised, with the form of delivery for each. No blanket wording.
- A statement of the scope of each piece of evidence. Require the supplier to state the models, configurations and conditions of use the evidence covers, and who owns whatever falls outside.
- Change notification duties and their triggers. Which changes must be notified in advance, who decides whether re-verification is needed, and how the cost of re-verification is shared.
Negotiated at signing, these three points take a few sentences. Negotiated while the product waits to ship, they cost you schedule. For the test packages and common handling approaches used on comparable products, see testing services and case studies. How the risk management file maps onto individual tests is a separate topic and is not opened up here.
When the interface will not divide cleanly, let a third party draw the line
SUNGO Mobility Testing Lab is the dedicated wheelchair and mobility aid testing lab within our group, accredited by CNAS, CMA and IAS (USA), with laboratories in Shanghai and Hefei. We handle third-party testing of powered wheelchairs, mobility scooters and their components on an ongoing basis, and during the planning stage we can turn the device-level and component-level evidence interfaces into a list: what has to be done on an assembled sample, what can be accepted directly from supplier evidence, and what needs a supplementary statement. To discuss a test plan or pricing, call +86 132 4819 8029, or send your bill of materials and existing reports through request a quote for an initial review. Please note that accreditation marks only demonstrate that the laboratory has the corresponding technical competence within its accredited scope; they do not constitute a commitment regarding market access outcomes.