What this decision actually has to settle
A steam cleaner that works well on a machine exterior can still be the wrong choice for the fixtures and workholding that sit on that machine. A vise jaw that holds flat after cleaning, a soft-jaw set that does not blister, a pallet whose datum faces still read the same on the surface plate: these are the acceptance tests for workholding, not for the machine. So the questions below are not about whether hot steam lifts cutting oil and way grease. They are about whether this cleaning method can be trusted around parts whose geometry is the reference for the next operation.
That reframing changes who answers what. The shop owns the condition record, the datum list and the re-measurement. The supplier owns the written answer on the machine being offered, including the industrial steam cleaner class, its mode behaviour and its service terms. The two records meet at a single trial, and the trial is designed before the enquiry is sent.
Confirm the deposit first, because it sets everything else
Workholding does not collect one kind of soil. A pallet returning from a wet process carries cutting-oil film and dried coolant paste. A vise in a dry milling cell carries dry swarf and a thin mist residue with almost no oil. A fixture near the coolant tank carries the sticky residue left when mist evaporates on a horizontal face. Ask your own people, in writing, what the worst part in the range actually carries, then confirm it by looking at a part that has not been wiped.
Two questions follow directly from the deposit. If the soil is mostly dry chips with light oil, does the job need heat and dwell at all, or is the binding problem mechanical wedging in slots and screw threads? If the soil is a dried paste, is the difficulty the deposit itself or the fact that it has been re-worked by clamping and turned into a packed layer at the part contact? A shop that cannot answer these two questions cannot evaluate a supplier's answer, because there is nothing to compare it against.
Confirm the material and surface condition of the workholding
Write down what each item in the range is actually made of and how its working faces are finished. Ground and hardened steel plates and pins behave differently from cast bodies, and both behave differently from aluminium or polymer fixtures and soft jaws. A brass or bronze wear pad, a laminated soft jaw, a bonded shim, a plastic locating cap and an anodised face are all reasonable items in a real tooling store, and none of them is the same cleaning problem as a ground base plate or a hardened tooling insert. Ask, for each family, who in the shop is responsible for the condition of that tooling when it comes back from cleaning, because a tooling store with no named owner produces no usable record.
Then list the surface conditions in the same document: hardened faces, ground faces, honed or lapped faces, as-machined faces, plated faces, painted bodies, elastomer seals and any adhesive-bonded assembly. The purpose is not to alarm anyone. It is so that when a supplier proposes a mode or an accessory you can point at a named item and ask what happens to that item, rather than agreeing about the category.
Name the families of items in the range, one line each for the tooling in the store, with the working face and the hidden feature of every family: base plates, sub-plates and risers, with their ground faces, dowel holes and threaded holes; clamps, studs and their seats, with threads, undercut corners and contact wear; locating pins, diamonds and pads, which are the faces that define position together with the bores they sit in; vises, chucks and mandrels, with their slideways, jaws and the seals between moving and static parts; and soft jaws, grips and caps, usually polymer, laminated or bonded, often with a machined texture that holds soil.
The datum faces: which questions decide whether the method is allowed at all
For every datum face, record in one line what must not change after cleaning, and how you would notice if it did. A base plate must stay flat enough to sit on the table without rocking. A locating pin must stay at its size and still enter its bore without force. A clamp seat must not gain a raised edge that lifts the clamp out of plane. A pad must not gain a burr, a lifted corner or a residue ridge under the part. Those four statements are shop language, not engineering language, and that is deliberate: the person doing the check after cleaning is usually not the person who ground the face.
Then ask the two questions that decide whether cleaning happens at all. First: does anything in the range carry a face whose condition cannot be re-established once it is disturbed, which means that item belongs to a manual reconditioning route instead of any cleaning process? Second: for a face that can be disturbed, can the change be detected without taking the fixture out of service, which means a measurement someone can repeat in the aisle rather than a decision in the metrology room.
What to confirm about the equipment before any trial is booked
The published description of a given machine belongs to that machine, not to the model family or the class. When you request the figures for the exact unit being quoted, ask for them as rows you can keep, and expect the same rows to be filled in for any alternative: what the machine is positioned for, which cleaning modes it offers and how they are switched, what media it takes and through what connection, the utilities it needs, and what its own pages say about where it is used. For a fixed unit on the shop's list, the supplier's page for the Fixed Steam Cleaning Machine PowerDe HW02 is the record for that machine, and those page rows are what to ask about line by line rather than assuming the same values apply to another model or variant.
Mode behaviour matters more than headline numbers here. Workholding is full of gaps, threads, bores, seals and small cavities, and some of them must stay dry or stay empty. So the practical questions are: does the machine offer a drier mode and a wetter mode, how is that changed in the middle of a job, and does anyone at the supplier or in your own plant know how the drier mode behaves on a vertical face with a slot in it. Ask for answers in writing, tied to the unit you are being quoted.
How the trial should be built so a decision can be taken from it
Choose one worst-case item from your own store, not a sacrificial part bought for the demonstration, plus one identical item that is cleaned by the existing method and kept as a reference. Photograph both before anything is done, and take the measurements your own process already takes on them: base to table, setup reading on a dummy part, or the go condition of a pin in its bore. Keep the same person responsible for both sets of readings so a difference is a difference of the process, not of the hands.
Write the stop conditions before the first pass. Typical ones for workholding are: any material visibly displaced at an edge; any change in how a pin enters its bore; any change in how a part seats; any moisture found in a bore, thread or sealed cavity after the drying step; any discoloration or texture change on a non-metallic item. A trial without written stop conditions turns into an argument afterwards.
Run the trial twice with the same part list, and time the whole cycle rather than the pass itself: collection, transport to the cleaning point, shielding, the cleaning, the drying, the re-measurement, the drying check and the return to the store. A method that finishes a fixture quickly but leaves it damp, or that needs a long quiet period before re-measurement, has simply moved the cost somewhere your schedule will feel it.
Where the boundary is drawn and what replaces it
A boundary that only says do not is useless in a shop, because someone will still need the part clean. For every excluded item, name the alternative in the same row: hand wipe and a dry brush for a bonded or elastomer item; a solvent or aqueous bath route for a small part with internal passages; a reconditioning operation for a face that is already borderline; a dry air blow for a cavity that must not get wet. Where an item is excluded because of a coating, a bond or a material you cannot identify, the correct action is to identify it, not to guess at the cleaning.
Two habits keep most of the risk on this object manageable without any specialist knowledge. Nothing is aimed into a gap, a hole, a seal or a bore. And no mid-trial solvent wipe is used on an elastomer or a bonded part just because the steam left something behind; that decision belongs in the written process, made in advance.
What to send the supplier so the answer is comparable
Send a short package rather than a description in an email: the item list with materials and surface conditions; the deposit as you actually observe it; the datum faces with the one-line requirement each; the exclusion list with the fallback named; the utilities you can offer at the cleaning point and where run-off can go; and the measurements you will use as the acceptance test. If the cleaning duty of your range is not fully covered by an existing machine, that same document is the starting point for a customized steam cleaning equipment discussion, because it already states the constraint, the environment and the acceptance test in the supplier's terms.
Ask each supplier to answer the same eight questions in writing, in this order: which of the named items would they not clean and why; which mode they would use on ground steel and on the non-metallic items, and what they would avoid; how the datum faces are to be protected during the pass; how the part leaves the process dry enough to be measured; what they expect to change at edges, threads and bores over repeated cycles; what the whole cycle takes including drying and re-measurement; what they will say when a datum check fails; and what service, parts and support exist locally for the unit they are quoting. Comparing eight answers side by side settles the suitability question faster than another demonstration.
Sign-off and the record that stays behind
Close the trial with a decision rather than an impression. A named person signs the item list as accepted for steam cleaning, as excluded with the alternative named, or as provisionally accepted with a check that repeats on a stated cadence. The signed page carries the two photographs, the before and after readings, the observed cycle time and the stop conditions that were actually triggered.
That page is what the next buyer conversation and the next shift both rely on. Workholding is only worth cleaning when the shop can prove, part by part, that the reference geometry survived it; the questions above exist so that proof is produced once, in a controlled trial, instead of being discovered on a production machine at the start of a run.
