Start from the table list, not from the machine
Walk every preparation area with a clipboard and write one row per table before choosing anything. A table is not one surface. The working top and its inserts, the seam where the top meets the splashback or the wall, the underside of the top and any undershelf, the legs and adjustable hollow feet, the trolley or castors if the unit moves, and the doors, handles and cut-outs of any refrigerated preparation unit belong on separate rows. Two preparation rooms with identical tables produce different plans when one is bolted to the floor and the other stands free with a gap behind it.
Give each row three readings. The first names the deposit you actually find at that spot: loose dry crumb, a fatty film that has set cold, a proteinaceous or starchy residue that has dried onto the surface, or a film that has been spread and re-spread until it no longer looks like dirt. The second records the material and its condition, because a scored, pitted, hazed or re-caulked surface is handled differently from a sound one. The third records reachability: can you get to it with the room in production, only after service, or only by lifting the unit.
Write the exclusions on the same sheet, in the same pass, before anyone proposes a method. A surface that will not be cleaned needs a named alternative and a note of what that alternative costs you, so that the standard stays complete rather than quietly shrinking every month.
Put the conditions in writing before the method
The rule that keeps a process from collapsing is simple: inspect the surface, then choose the method. On preparation tables that means checking the finish for scratches, etching and discoloration; checking whether seams and seal lines are sound, opened, or re-caulked; checking whether the top is laminated with an edge trim that can lift; checking for electrical penetrations, control cut-outs or adjacent services within reach of a wet pass; and checking whether the unit is sealed to a wall, bolted to the floor, or free-standing with a hidden back face.
Then record the room-owned conditions, because these decide most decisions that follow. Which service gap exists, and who authorises it. Where water can be drawn from and how far the hose actually reaches. Which drain each group of tables discharges to, and whether discharge is permitted there. How the room falls and where liquid will stand. How long the area needs to stay dry before food returns to the top. Which adjacent surfaces, finishes and fittings must survive the job. And what the kitchen does today, so the new process replaces something rather than being added on top of a routine nobody stopped.
Put an owner, a frequency and a defect route on every row
Repeatable means someone is answerable for each line on the sheet. Write one page per group of tables: the row itself, the owner named by role rather than by name, how often it runs, the method, and what happens when the row fails. The route for a failure is the part teams leave out. If a seam is found opened during a clean, the process has to say who is told, what is quarantined until it is signed off, and what the table may be used for in the meantime. Without it, the observation is reported once, remembered by one person, and gone by Friday.
Keep frequencies honest to the shift pattern. A daily top-of-table clean, a deeper pass on the underside, legs and feet on a longer cycle, and a documented strip-down of anything that can be safely moved are three different tasks with three different owners. Merging them into one line called "deep clean" is how preparation tables end up cleaned twice a day in the places everyone can see and never in the places nobody looks.
Choose the mode row by row: heat and dwell, or water volume
Steam decisions on a preparation table are usually a decision between heat held against the surface for a measured time and the volume of liquid you are prepared to introduce and then recover. They are not a decision between stronger and weaker. A set fatty film responds to heat and dwell; a loose dry deposit needs neither; a large floor-adjacent area may be limited purely by how much liquid the room can take and remove. What matters on the sheet is which mode a given row needs, because that is what makes a per-row plan workable at all: the tabletop and the band of legwork near the floor rarely want the same setting.
The published page for Fixed Steam Cleaning Machine PowerDe HW02 describes a fixed unit rather than a portable one, and it is that page which shows a selectable dry and wet steam output exists in this class at all. Their own voltage, heating-power, temperature, pressure and water-inlet rows belong to that single page: those figures do not rate a food-service duty at all, they do not carry across to another model or variant, and they should never be treated as the delivered unit's numbers.
Two cautions belong in the standard itself. Two cautions belong in the standard itself. The temperature printed for a boiler is not the temperature measured at the lance tip, so the useful question is what reaches the surface, not what leaves the machine. And a tank capacity is not a working time; it says how much is carried, not how long a task can run before a refill interrupts the cycle.
Whatever the mode, the sequence stays fixed: loose soil off first, then heat and dwell, then mechanical removal, then recovery, then a genuinely dry surface. Skipping mechanical removal is the commonest failure; it converts a thick deposit into a thin film over a wider area and the table looks worse under raking light the next morning. Skipping the dry finish leaves moisture in seams, feet and hollow sections, which is where the next problem starts.
Decide the water balance before the first pass, not during it
Food-preparation rooms fail on water, not on heat. Before any wet work, fix and record four things: how water reaches the work area and through what fitting; which drain the run-off from each table group goes to and whether that discharge is permitted; how the liquid will be recovered while it is still on the surface rather than after it has travelled; and how the room is closed, dried and re-opened. If the recovery step depends on a mop that also carries deposit into the next area, the process has no recovery step.
Write the run-off route into the cycle time, as a task with an owner. A plan that quietly assumes the floor will dry by itself is a plan that will be broken the first time the room is busy.
Prove it on one table with a bounded trial
One table, marked with tape at the edges of the test area, is enough to decide. Photograph it bare, from a low side angle that keeps the grain of the surface visible rather than flattening it, and keep the photographs. Write the trial conditions while the machine is still cold: which surface, which mode, how long the dwell was held, what was used to remove the deposit, how it was dried. Write down the circumstances that halt the trial before steam is applied, so that a bad result ends the trial rather than extending it.
Then run it twice and take the timing from the second run, because the first run is a learning run. Inspect again the next day after production has been through the room, under the same raking light as the before photographs. Three checks need no laboratory at all: does the deposit remain where the tape was, does the surface look the same under raking light as it did before, and do seams, feet and hollow sections come back dry within the window you set. If the deposit is merely redistributed, the trial has failed even if the table looks cleaner that evening.
Keep a written exclusion list with a replacement for every entry
Exclusions are part of the standard, not an exception to it. On preparation areas the list usually includes cutting boards and end-grain blocks, timber and any dressed timber surface, printed or applied markings and labels, decals and branding, lacquered or coated fascias, opened or re-caulked joints until they are repaired, and any hollow foot or leg that cannot be dried within the window. Each entry needs a replacement method and what that substitute costs you in return: a board is replaced or re-finished on a cycle rather than cleaned on the line; a marking is cleaned by hand with a damp cloth and accepted as wearing; a joint is repaired before any heated pass touches it.
Ask the supplier for their exclusion list in writing as part of the offer, and ask what they recommend instead. A supplier who cannot produce one is describing a machine, not a process.
What the supplier should be asked, in row labels
By this point the questions write themselves, and every one of them belongs to a specific unit on a specific quotation. Ask for: the steam temperature and the vapour pressure measured at the working point, not at the boiler; the heating power and the supply voltage matched to the outlet and the cable protection at that location; the tank capacity and the water use per hour as two separate answers; the water inlet fitting and whether it matches the site tap; the footprint and travel route including handles, thresholds, kerbs and any doorway the unit must pass; for a battery-fed variant, the working voltage and the run time for the machine actually quoted, plus who owns the packs and the charging place; the protective features for that unit named item by item, without assuming one unit's protection carries across to a variant; and whether the same unit also performs suction, since that function is not listed on every page in the range.
Ask the same list of each unit you compare, and refuse answers written as prose. The side-by-side is only fair when both suppliers have filled in the same rows. Where the duty is unusual, where the room is tight, or where the existing routine has to be rebuilt rather than topped up, that is the point to bring it to customized steam cleaning equipment rather than hoping a catalogue configuration happens to fit.
A single sheet per table group, trained at the table, reviewed on a date
The finished process is not a manual. It is a single page for each group of tables: the rows, the deposit type, the mode and dwell, the drying condition, the exclusions with their replacements, the owner, the frequency, and the defect route. Train it at the table with the person who will use it each shift, on the surface they will actually face, and hold the page to a review date rather than to a revision history nobody reads.
Two things this approach does not prove. A successful trial on one table does not vouch for the whole room, and a supplier's figures for one machine do not answer for another model or variant. Treat the trial as evidence about your own surfaces and the trial room's conditions, and treat any published page as a summary of one machine rather than a contract for the unit that arrives. That distinction, held consistently, is what makes the process repeatable rather than merely repeatable-looking.
