Start by writing the surfaces down, because the deposit is not one thing
Before anybody holds a watch, somebody has to write a list of the surfaces that are actually in scope. The dried coating a buyer is trying to remove is not a single substance. There is the wet film left straight after a run, still mobile, mostly cutting or flushing fluid. There is the worked film that has been through several shifts, oxidised, dark, and holding fines against the panel. There is way oil that has crept out along a slide and dried into a ridge nobody waxes about because it is under a cover. There is hydraulic bleed from a fitting weep, which spreads into a band on the floor and up the plinth. There is the airborne mist film that lands on everything within reach of the enclosure, thin, even, and invisible until a hand passes over it. And there is the paste, the one that decides the whole job: fluid, swarf and fines bound together into a crust in a corner, on a chip conveyor hood, around a wiper, at the base of a column. That whole family of residue is what the trade calls machining fluid buildup, and it is why a single number never works.
Each of those behaves differently when heat and water meet it, and the differences are what a time estimate is made of. A fresh film is a pass. A crust is heat, then waiting, then a pass, then a second look. So the first page of the estimate is not a duration at all: it is a surface list with three columns, what is on it, how thick and how old, and whether a person can reach it without a ladder, a torch or a second pair of hands. Anyone who puts a figure on this duty from a machine photograph rather than a written list is guessing, and the guess will be low.
Draw the boundary first, because it sets the shape of the job
In a machine shop the boundary is set by the machine's geometry, not by how well it was built, and the buyer's time plan covers whatever falls inside that boundary. Write down, before any pass, what will not be touched: way covers and the slideways beneath them, wiper seals, the ball screw and its guard, bearing and shaft ends, the terminal box and where its cables enter, the encoder and the scale that reads position, the loom of sensor plugs, air and oil fittings, the filler neck and breather on any reservoir, the vents on the spindle and the servo, the cabinet cooling fan, and every nameplate, warning triangle and inspection stamp. Every line on that list needs an alternative method named beside it, not a shrug, because the shop still has to clean those places by hand and that hand time belongs in the estimate.
Two entries deserve to be written in the shop's own words rather than copied from a general list, because they carry the most risk and the most hidden time. Panel that has been repainted or re-coated at any point in the machine's working life, and vinyl or paper decals, date stickers and repair labels: heat and dwell do things to finishes with a history, and only a taped test patch with photographs before and after tells this shop what happens on this panel. Then the inside of the machine. The boundary this article works to stops at the exterior envelope, and it stops there deliberately. Directing a wand at an open guideway, a track or a chip conveyor strips the lubricant the slideway needs, and a machine that runs dry for a shift after being cleaned has not been maintained. Any time saved by going inside is borrowed from the next week's accuracy.
Break the job into blocks and time each block, not the whole job
A single stopwatch reading for "clean the machine" is not usable in a plan, because it mixes decisions with walking and waiting. Split it. Block one: making the machine safe and available, which is lockout, waiting for the fluid to sit, letting hot surfaces fall, moving pallets, tables and guards into the position the wand needs, and closing doors that have to be closed. Block two: shielding, which is the exclusion list made physical with sheeting and tape. Block three: the pass itself, surface by surface, orientation by orientation. Block four: dwell, the waiting between heat and removal, which the operator does not experience as work but which the clock does not stop for. Block five: drying and rechecking, wiping the crevices, looking under covers and at the shielding after the first pass. Block six: putting the machine back, removing shielding, replacing anything taped, re-lubricating what was removed, and re-checking that the safety devices do what they are supposed to do. Block seven: the run-off, which is where these projects end.
Time each block once on paper before the trial, then measure it, and compare. Two of them are almost always under-estimated and are the reason most shop figures fail. Shielding is a per-machine activity, not a per-shift one, and on a machine with many exclusion entries it is the largest single block after the pass itself. The blocks after the pass, dry, recheck and restore, are invisible while the machine is being washed and then appear in the schedule as an overrun at hand-back. Write the two blocks at the front of the sheet so nobody reports a cleaning time that is really just a washing time.
Decide the mode per surface, and let that decision drive the clock
The range is offered with a switch between dry and wet steam, which means the choice is real and somebody has to make it deliberately for each surface rather than at the start of the job for the whole machine. A thin mist film on a painted column does not want a soaked panel. A crust at the bottom of a column wants heat and time, and heat and time consume dwell, not water. A vertical panel behaves differently from an overhead one, because gravity moves the water and everything the water has picked up. Read the mode as a per-surface instruction on the list, with the dwell noted next to it, and the total falls out of the surface count instead of out of a habit.
The reason to insist on this in writing is arithmetic: heat and dwell do the cleaning work, while the quantity of water is a cost, a drying burden and a run-off problem. A shop that reads a tank capacity as though it were a delivery figure will plan badly twice, once on how long the machine can run and once on how much liquid has to be contained and disposed of. Whatever the buyer is eventually quoted, the water figure to ask for is an hourly one, at the hand tool, for the exact unit, with its measurement point stated. If a quotation cannot produce that, then no floor time built on it can be trusted.
Cost the run-off as a scheduled task, or the estimate will not survive the first week
Ask, before the trial rather than after it: where is this water allowed to go, and who signs that answer. A machine shop usually has a fall to a gully and an interceptor, and frequently has nothing else. The mixture coming off a machine is not clean water; it carries fluid, fines and whatever the shop has previously put on those panels. Containment, the recovery of swarf and separated oil, the clearing of a gully, the paperwork if someone has to sign for the waste: each of these is a person with a bucket and a duration, and all of them belong in the estimate as their own line. The slip hazard while the floor is wet is a schedule item too, because it decides whether work can continue next to the bay or has to wait.
Choose the unit of work the shop will actually measure
A machine is the wrong unit for planning, because a shop does not clean one machine at a time. Decide what is being counted and stick to it: a machine, a station inside a cell, a group of panels, a per-shift wipe task, a weekly deeper task, or the exterior envelope of one machine type. Then run one lot. In the lot, treat the first machine as set-up, not as data, and write down what changed between it and the second. Do the second as the real run. Do two more to check repeatability, because a number produced once on a machine nobody has cleaned this way before is a curiosity rather than a standard. One lot per machine family: a press, a turning centre and a machining centre are not one family and do not share a number.
Run one bounded trial with the acceptance rule written first
A trial without a written start condition produces a story, not a figure. Before any water moves, write two things down and put them where the operator can see them. The acceptance rule: what a clean surface looks like, judged on taped test patches at the worst corners, with close-range photographs before and after, so the decision is a comparison rather than an opinion. The stop rule: what ends the trial, and it should name the three failures that end these trials in machine shops, moisture anywhere it should not be, damage to a decal, label or finish, and any sign that material has been pushed into a gap, a slideway or a seal. One row per patch, one signature per row, and the person who signs not the person who held the wand.
Reach is a question this trial has to answer as well. Where does the unit stand, how far does it have to be moved around this machine, and can a person get to the far side of the envelope without repositioning twice. Reach and dwell, not effort, are what cap how much of a shop a buyer can plan for in a shift, and a floor figure that ignores them describes an afternoon, not a week.
What a machine changes in the number, and what it never settles
A shop comparing options should read specification rows with the same scepticism they bring to a vendor claim. Published rows are not written alike between suppliers, and notation in this product family is not always consistent in how temperature and pressure are labelled, so a figure taken from one page and moved to another machine is not a comparison; it is a guess with a decimal point. When a buyer asks a supplier to put the relevant rows in writing for the exact unit, the list is short: whether the unit carries its own power and water supply or has to be plugged and plumbed at the point of work; where it can physically get to, and how far; the delivered figure at the hand tool rather than a holding capacity; the accessory set supplied as standard, since the accessory set decides reach; and the model designation on the document matching the machine that will arrive.
The machine also answers a question the buyer will be asked by a planner: whether it is a fixed position the shop builds around, or a unit on wheels that follows the machines. On the fixed unit the page carries a voltage inside the model name and states dry and wet steam as the modes; it publishes no machining figure, and it does not present itself as a solution for cleaning inside a machine. Anyone comparing options should read that positioning as a constraint to plan around, not as a defect, and should refuse to carry an application written for a fuelling island across to a machining centre.
And the limits of the exercise, stated plainly. A time estimate settles nothing about whether this shop is allowed to discharge what comes off the machine, nothing about whether heat near a finished panel is acceptable here, nothing about whether the machine's own supplier would object, and nothing about whether a person is competent to work around live panels. Where a shop needs the duty worked out with the supplier rather than after the fact, the site describes customised customized steam cleaning equipment, and the steam cleaning equipment testing pages describe hardware testing as a service stage rather than a promise about a machine.
Turn two measured runs into a one-page standard the shop can keep
Two honest measured runs are worth more than any catalogue. Buyers who want the figure to survive an audit keep the sheet that produced it. From those runs, write one page per machine family and keep it short enough that it will actually be read at the machine: the family name, the surface list with its deposit grade for each line, the mode and dwell for each line, the exclusion list with its named alternative, the seven blocks with their measured figures for that family, the run-off route with the name of whoever signs it, the acceptance photographs as the reference, and one line saying when the page will next be reviewed and by whom. Train against it in place, on the machine it belongs to, and correct it when reality disagrees with it rather than defending it.
Then treat the page as the estimate. The buyer who asks a supplier how long this will take is really asking whether the shop has a measured figure it can defend to a production planner, and the answer is either a sheet with two runs and signatures, or a number somebody repeated from a brochure. The equipment a shop ends up comparing those figures against sits inside this manufacturer's PowerDe cleaning equipment range, which is presented around vehicle and medical work, and that positioning is worth carrying into the conversation early rather than discovering late.
