A cleaning trial on machined workpieces is rarely a test of the machine. It is a test of whether the factory can describe its own oil residue precisely enough that a result can be repeated. Most trials fail because both sides arrive with a vague idea: one batch of parts, a hope that heat will help, and no agreed definition of clean. The information below is what should be settled on paper before any trial day is scheduled.
Name the oil and name its condition, not just its source
"Cutting oil" describes several very different soils. A wet film of neat cutting fluid that has just left the nozzle behaves nothing like an emulsion that has dried into a milky film, and neither behaves like a baked residue that has been driven through a warm dryer or left on parts stored near a heat source. Each one needs a different combination of heat, dwell and mechanical action, and each leaves a different kind of failure behind.
Before the trial, record the fluid by its product data sheet name, whether it is neat or a water-mix emulsion, whether it carries additives that leave a film after the carrier evaporates, and how long the residue has been on the part. Then classify the residue by state: still wet, dried but rewettable, or hardened. That single classification usually matters more to the outcome than the choice of equipment from the commercial steam cleaning equipment range.
Describe the workpiece geometry and what it traps
Machined parts hold oil in places that are not visible from the outside: cross-drilled holes, blind bores, threaded seats, under flanges, in the grooves of knurled surfaces, and inside the gaps between a pressed-in bearing and its housing. A trial that reports success on flat faces says nothing about those locations.
Supply a drawing or a photograph from the same angle every time, and mark every cavity, bore and seam that must be clean. State the material and surface condition as well, because an untreated machined surface, a passivated surface and a painted or coated one respond differently to heat, and residue can be pushed into a porous or rough finish rather than lifted from it.
Say what "clean" means in a form someone can check
A pass or fail judgement has to be written down before the trial starts, otherwise the result becomes an opinion. Practical checks that a shop floor can actually run include: no visible film under a work lamp, a water-break test where the sheeting behavior of water is compared to a known clean reference panel, wipe tests with a weighed cloth, and, where the parts feed a coating, bonding or painting step downstream, an actual coating trial on the cleaned parts.
Also state what must not remain. Residual water in a bore, detergent film, and displaced oil that has run into an adjacent cavity are all failures even when the visible surfaces look good. If the parts go on to assembly, say so, because a film that is harmless to a machinist can be fatal to an adhesive joint.
Fix the process variables, one at a time
A trial that changes everything at once cannot be repeated. Set and record the mode in use, the nozzle or lance distance, the travel speed of the hand or fixture, the number of passes, the dwell time allowed before wiping or vacuuming, and whether any detergent is used and at what dilution. Change only one of these between runs and keep a written line for each.
Dwell is the variable most shops forget. A hardened film usually will not respond to a fast pass at any distance; it responds to time at temperature. Where a fixed unit such as the PowerDe HW02 steam cleaner is the candidate, the figures on its product page are the reason dwell and duty cycle need testing rather than assuming: a nominal 6 kW heating element, an 80 L SUS304 tank, and a 1/2-inch four-point internal-tooth tap-water inlet. Those numbers describe what the installation must supply and how long the unit can run before it needs attention; they are not a claim that any given oil on any given alloy will lift.
Give the trial site the same attention as the parts
Oil trials produce an oily run-off, and that run-off has to go somewhere that is not the shop floor drain. Record where water is drawn from, whether the trial position has a protected outlet on a suitable circuit, how the run-off will be collected, and whether the floor area can be closed, dried and released afterwards. A trial that cannot be cleaned up safely will not turn into a routine.
Water quality belongs in the record as well. A tap-water inlet is a real plumbing requirement, and hard supply water leaves scale inside the machine. Scale is a maintenance variable that shows up later as weaker output and is frequently mistaken for a cleaning problem, so it should be logged with the trial parameters rather than discovered afterwards.
Run the trial on your own worst case
A supplier's demonstration panel is not your worst part. Bring parts taken from your own line at the end of a shift, with the residue state you actually see, and let your own operators run the trial with their own tools and their own working rhythm. Then repeat the winning settings three times on fresh parts from the same source. A setting that works once is a coincidence; a setting that works three times is a process.
Judge the outcome against your current baseline rather than against a brochure. Record the time per batch, the number of parts handled, the water consumed, the number of parts that need a second pass, and the time until the parts can go into the next operation. Those figures are the ones a production meeting will act on, and where several routes look viable, customised steam cleaning equipment may be worth discussing only once those measurements exist.
Turn the result into a one-page controlled method
The output of a successful trial is not a machine purchase; it is a short written method. One page carrying the part list, the residue class, the mode and nozzle distance, travel speed, passes and dwell, the dilution if any, the drying route, the pass or fail checks, the personal protective equipment, and the signature of whoever owns the process. Every future shift then follows the same method instead of re-inventing it.
Keep one line per batch as evidence, and put the cleaning machine into the same preventive maintenance list as the production assets, with a named owner for descaling and for the checks on nozzles, hoses and inlet filter. A machine that drifts out of condition quietly changes your cleaning result.
Send the questions before asking for a model
If you want a supplier to run the trial on your residue instead of their own, send the fluid data sheet name, the residue classification, the marked-up part drawing, the pass criteria you intend to apply, the trial window you can release, and the utility facts for the position. That is what makes a meaningful test possible; a model number on its own is not enough to design one.
Send that same pack to PowerDe through the contact page and state which trial question you want answered first: removal of the residue class you photographed, or the practicality of the route on your own line.
