The interval most machine shops run on the outside of their machine tools was never chosen. It arrived with the first cleaning round, survived every machine change, and is now enforced by nobody in particular: operators do it when the room looks wrong, a contractor does it when they are already on site, and the plant finds out how loose the arrangement is during a customer audit. If you are the person who buys the equipment and signs the service scope rather than the person holding the lance, that gap is your exposure, not the operators'. This article is written for that side of the transaction: how to put a frequency claim on paper that the plant can actually answer, and how to audit the answer.
How a shop ends up with an interval nobody owns
There are four ordinary routes into this state, and none of them looks like a decision while it is happening.
The first is the machine's own paperwork. The operating and maintenance instructions that ship with a machine tool describe contamination in the spindle, the guideways and the toolchanger. They rarely address the painted panel, the pendant housing, the cable carrier or the outside of the enclosure, and they say nothing about how often those parts should be cleaned in a shop running a wet process. A buyer who reads them carefully still leaves the room without an interval, because the document was never written for that question.
The second is the service visit. Cleaning of the exterior often rides along with a maintenance call, and it inherits that visit's rhythm. Where the machine is serviced under a contract, the outside of the machine is cleaned when the contract engineer happens to bring equipment. The interval is real, but it belongs to the maintenance schedule and nobody has written it down as a cleaning commitment.
The third is shared labour. Where one crew looks after several machines, the easiest one to reach is cleaned most, and the one tucked behind a wall is cleaned when the crew has time. The resulting pattern looks like a schedule from the shop floor but disappears as soon as anyone asks which machine was cleaned last.
The fourth is a person. In many shops the exterior of the machine tools is kept presentable by one individual who knows which panels stain, which corners collect chips and which machine the customer always looks at. When that person is missing, transferred or gone, the interval disappears with them, and the plant discovers that nothing was ever written down.
Put the interval's content on paper before you argue about its length
Buyers argue about how often. That argument is unwinnable without first agreeing what the cleaning is. A frequency claim that can be audited states five items, and a proposal that omits any of them is not a claim, it is a promise.
- The task list. Which surfaces are covered by name, not the phrase "machine exterior". A list that reads "panels, guarding, pendant, chip conveyor cover" can be audited; a list that reads "the machine" cannot.
- The boundary. Where the cleaning stops. The outside of the enclosure and the inside of it are different jobs with different risks, and the inside belongs to the machine maker's instructions.
- The excluded parts. Way covers, seals, optical scales, electrical enclosures, coolant tank openings, pneumatic fittings and anything with a rating label on it. Exclusions stated in advance are protection; exclusions discovered on the day are a dispute.
- The time to return to service. Not the time the cleaning takes, but the time between finishing and the machine cutting again. It is that number that competes with production, and it is the number a plant should be told in advance.
- The written machine instructions relied on. Which document the cleaning method follows for this particular machine. If nobody can name it, the method was invented on the day.
A supplier who will state all five of these in writing is describing a process. A supplier who states only a frequency is describing a habit, and habits do not survive a change of crew.
The access you buy decides the interval you can keep
Buyers set intervals as if the only constraint were willingness. In practice the constraint is physical. Ask how a person reaches each named surface with the machine in place, and the honest answer usually rules out the schedule you had in mind.
The back of a machine tool is the classic case. Machines are installed against walls and against each other, and the rear panel, the rear cable routing and the exhaust side are frequently unreachable without moving something. An interval that includes those surfaces is only real if the shop has committed to moving something, which means floor space, a route, and clearance agreed with whoever owns the layout. This is a purchasing decision, made years before anyone talks about cleaning.
The same logic applies to the small things that make an interval survivable: a door that opens far enough for an arm, a panel that stays out of the way once opened, a place to set a tool down, a floor that drains rather than ponds, lighting that shows a film rather than hiding it. None of these appear in a cleaning specification, and all of them determine whether the plant can keep what was signed.
A practical audit question for a buyer: choose the least accessible machine in the shop and ask when it was last cleaned on all sides. If nobody can answer, the shop does not have an interval; it has an intention.
Take your audit samples from the shop calendar, not from a rota
Checking on a fixed date tells you what the schedule looks like when nothing interferes. Checking at the moments when the shop's own events force a decision tells you what the schedule is worth. Five calendar moments are worth a buyer's attention.
- After the first batch on a freshly cleaned machine. The rate at which the outside returns to its previous state is the plant's own data, and it is only visible if the machine was actually cleaned first and someone looked afterwards.
- Before an audit or a customer visit. If the shop cleans because someone is coming, the interval is being driven by optics rather than by the machine. That is not automatically wrong, but it should be an acknowledged policy with a stated reason, not an emergency.
- After a cutting-fluid or tooling change. New chemistry, new additives, a new coolant concentration, a switch to a different tool coating: the soil leaving the machine changes with it, and the previous interval was measured against the old soil.
- After a process or workload change. More shifts, longer cycles, heavier cuts, a different material. A shop that added a night shift did not add a cleaning shift, and the interval inherited from single-shift working was calibrated for a room that no longer exists.
- At the contract renewal. The moment the plant has leverage over a service provider, and the moment to test whether the frequency in the paperwork matches what the crew actually did.
A machine tool steam cleaning scope should be written so that each of these five moments produces a finding someone can act on. If the record cannot answer a question at those moments, the cleaning programme has no evidence base, only a calendar.
Two boundary considerations set the range; the shop decides inside it
A buyer who wants a defensible interval should pin down two considerations that bracket it, then let the plant choose within the bracket rather than exporting a number from another plant.
The first is about the room, not the machine: how long can the shop stay presentable to the people who work in it and the people who visit it? A shop that must look orderly for customers, for recruitment, or for a contract auditor has a short upper bound on the outside of its machine tools, because the exterior of the machines is the largest visible surface in the room. A workshop whose customers never enter the floor has a much longer one. Neither answer is a fault; they justify different intervals and different money.
The second is about the machine maker: how long can a deposit sit on a surface before the machine's own documentation stops treating that surface as maintained? That is the lower bound. Accumulated chips at a way cover edge, fluid tracking down a panel into a seam, grime at a cooling fan strainer, residue at a door seal: these are the places where exterior soil turns into a machine condition rather than a housekeeping matter. Where those locations exist on a given machine, the interval that keeps them clear is set by the machine, and no negotiation with a contractor changes it.
Between the room's upper bound and the machine's lower bound, the shop is free to choose daily, weekly, monthly or quarterly work, and to mix them: the panel faces daily, the conveyor cover weekly, the rear access monthly. That mix, written down, is a cleaning frequency. A single number applied to every machine in the shop is not.
What each rate actually costs, priced from both sides
Cleaning the outside of machine tools costs money in two places, and buyers only ever see one of them.
The visible cost is the labour and the equipment: the person doing it, the machine doing the work, the consumables, and the containment and recovery of what comes off the machine. The less visible cost is production. A machine that is stopped, cooled, cleared, cleaned and dried has lost cutting time whether or not anyone charged for it, and in a shop running multi-shift the loss is what makes an aggressive interval fail in practice.
The far side of the trade is also real and equally unpriced. Deposits that stay on an exterior make leaks and weeps invisible until they are expensive. Grime on a strainer or a fan intake changes how a cabinet breathes. Chips accumulated at a cover edge interfere with how the cover does its job. Wear particles carried into a door gap abrade the sealing surface. None of this is a cleaning cost until it becomes a downtime invoice, which is precisely why leaving it out of the comparison biases the decision toward less cleaning.
A buyer evaluating a CNC machine cleaning equipment proposal should therefore ask the supplier to price both sides of their own recommendation: what the interval costs to run, and what the plant is being spared. A proposal that presents only one side is not making an argument, it is making a sale. Where the scope covers a whole production area, the same comparison is worth running against industrial steam cleaning equipment options at quotation stage, so the plant is choosing between stated numbers rather than impressions.
What to refuse in a proposal, and what the finding should read like
Refuse an interval offered without the task list, the boundary and the exclusions behind it. Refuse a price quoted per visit when the question is whether the plant can live with the interval. Refuse a recommendation lifted from another shop, since the soil, the shift pattern, the access and the machine documentation are all different. Refuse a scope that includes interior machine parts without naming the machine maker's document it follows.
And decide in advance what an audit finding is allowed to say. "Cleaning is satisfactory" is not a finding. A finding names the surface, the condition, the machine, the date of the last recorded cleaning of that surface, and the last one is the machine documentation relied on. Five named fields turn an inspection into evidence; adjectives turn it back into an opinion.
The question the operators are actually being asked is not whether they clean often enough. It is whether anyone can demonstrate what "often enough" means for their machines, in their room, against their machine documentation. A buyer who installs that proof (the stated task, the agreed boundary, the sampled moments, the priced trade, the finding format) gets a frequency that survives the crew change, the machine replacement and the audit. A buyer who only ever sets a number gets a rota that quietly stops on the day the person who understood it leaves.
