Start with the standard, not with the test chamber
Before anything is scheduled, write down which standard or customer specification the prototype is being qualified against. A safety requirement that arrives as “must pass IP testing” is not testable. The usable version names the test method, the severity, the exposure time and the pass criterion. Once those four items are written down, the requirement turns into questions the design can answer: which enclosure boundary is being challenged, which ports are open during the test, and what counts as a failure — water ingress, a loss of dielectric strength, a temperature rise, a surface that becomes too hot to touch.
Where the equipment is intended to be cleaned with water or steam in service, the review should also cover the cleaning method the end user will actually apply. Industrial industrial steam cleaning equipment works at a higher temperature and pressure than a hose, and a prototype enclosure that tolerates a cold rinse can still fail a hot, pressurised one. Recording the intended cleaning method in the requirements file gives the test something concrete to simulate, and gives the quotation something concrete to be judged against.
Sort the requirements into three columns
A practical way to review the list in one sitting is to place each requirement in one of three columns: what the prototype must do, what the test rig must provide, and what has to be true of the sample itself.
- Device requirements. Electrical isolation, protective conductor continuity, touch temperature, ingress protection, stability, moving-part guarding, emergency stop behaviour. These are the items the prototype is judged on.
- Test-equipment requirements. Test voltage and available fault current, spray pressure and flow, chamber volume, the ability to hold a load condition, calibrated instruments, and the interlocks that stop a test safely. A requirement with no matching capability on the bench is an assumption, not a test.
- Sample requirements. How many units, at what build stage, with which fasteners and seals, wired in which configuration, in which operating state. Prototype samples often differ from the production design in exactly the places a safety test is sensitive to.
The third column is where reviews most often go thin. If the ingress test is run on a unit whose cover fasteners are the temporary ones from the machine shop, or on a sealing arrangement that will not exist in the shipping build, the result says nothing about the product. Mark every sample deviation next to the requirement it affects, and decide in advance which deviations are acceptable for that round and which make the round meaningless.
Check the interface between the design and the rig
Half of the findings in a good review come from the boundary between the unit under test and the test setup. Work through it item by item:
- Cables, hoses and fittings that pass through an enclosure during the test — these are the usual leak paths, and they are frequently not the ones the design drawings show.
- The state of the equipment during the test: powered, unpowered, running a duty cycle, or cold-soaked. A requirement written against one state does not transfer to another.
- What is instrumented and what is merely observed. A measured value can be argued about later; an operator's note that “it looked dry” cannot.
- Who holds the sample. A fixture built for one prototype revision often quietly changes the load path or blocks a vent for the next one.
This is also the point at which the supplier of any ancillary equipment should be asked to state what their machine provides, rather than what the brochure claims. On a fixed steam cleaning machine intended for a production bay, the items worth confirming in writing are the supply voltage available on site, the heating power the local circuit can carry, the water inlet arrangement, and how the machine's own protection functions are proved. Those same parameters are what a test protocol needs if cleaning is part of the qualification, and a buyer who collects them once can use them in both conversations.
Run a short pre-test review meeting with the people who build the sample
The review works best as a one-hour meeting between the engineer who owns the requirement, the technician who will run the bench, and whoever assembles the prototype. Three questions carry most of the value:
- Where does the sample deviate from the production design, and which requirements does that invalidate?
- Which requirement has no written pass criterion, only an expectation?
- Which test can the bench actually run this month, and which one is waiting on an instrument or a fixture?
Write the answers into a table with one row per requirement: requirement, test method, rig and instrument used, sample condition, planned round, criterion, result. A table of that shape is also what a customer, a notified body or a purchasing department will ask for later, and it is far cheaper to build once than to reconstruct after the prototype has been cut up.
Decide what the prototype round is allowed to prove
Not every safety requirement can close on a first prototype, and pretending otherwise creates a file full of green boxes with no evidence behind them. Say plainly which items are being screened, which are being verified, and which are deferred to a production-intent build. A screening result supports a design decision; a verified result supports a declaration. Recording the difference protects the schedule when a later round reopens a question.
Where the product line spans more than one application, the same review is what stops a supplier from quoting a machine for a duty it was never meant for. PowerDe's equipment range covers cleaning machines for vehicle, kitchen, medical and industrial duties alongside process and automation equipment, and a buyer asking for a customised cleaning solution should expect the engineering conversation to start with the site conditions and the duty, not with a model number.
A practical checklist before the first test
- Each requirement has a named test method and a written pass criterion.
- Each test method is matched to an instrument or fixture that exists, with calibration records.
- Sample deviations from the production design are listed, with the requirements they affect.
- Service conditions — including any cleaning method the product will face in use — are recorded as test conditions.
- Screening and verification results are labelled as such, with the responsible engineer named.
- Failures have a route to a corrective action and a retest, not only a comment.
A prototype that has been through this kind of review tends to come out of the test with a decision attached to it. That is the difference between a build that shortens the path to a saleable machine and a build that only produced photographs.
