- Home
- The shutdown window
What actually fits into an empty week, and what does not
Every building in this town has one or two windows a year when it is empty. They are known twelve months in advance, they are short, and the jobs that go wrong are almost always the ones that were squeezed into one that was never long enough.
The window is the last step of a job, not the whole job
A replacement is not a week of work in an empty building. It is a survey, a load calculation, a specification, an order with a lead time attached, possibly a permit, possibly a crane and a road closure, then the physical work, then commissioning, then a set of readings taken with the building at something like normal load. Only the middle part needs the building empty.
When a job fails it is nearly always because that sequence was started at the wrong end — the window was booked first and the lead times discovered afterwards. Equipment that arrives on the Thursday of a one-week shutdown cannot be commissioned by Friday, and a building that reopens with a system half-installed is in a worse position than one that waited.
So the useful questions are: what are you actually trying to fit in, and how much time do you genuinely have?
What you are trying to fit in, and how long you have
Nothing is sent anywhere and nothing is recorded. This is what we would ask you at a planning meeting.
A short window is genuinely useful, but it is useful for a particular category of work. The things that do not fit are not a matter of effort or of how many people are on site, and no amount of goodwill changes them.
What cannot be compressed: equipment lead times, which are what they are on the day you order; a permit or an inspection, which runs at the authority’s pace and not yours; crane scheduling and any road or car park closure that goes with it; and commissioning, which needs the system running against something like a real load to mean anything at all. A system started on Thursday and switched on Friday has not been commissioned, it has been turned on.
What genuinely does fit in a few days: almost all maintenance, coil cleaning, drains, filters, most repairs, balancing, controls and thermostat work, and a full survey. That is a great deal of value and it is frequently the better use of a short window than a replacement that will overrun.
And our position, plainly: we would rather tell you a job does not fit and put it in the next window than begin it and hand you a building that reopens with a system half-installed. A deferred replacement is an inconvenience; a half-finished one is a closure. If somebody else is willing to promise the compressed version, ask them specifically when the equipment is arriving and when commissioning happens.
Work backwards from the reopening date, not forwards from the shutdown
Fix the date the building must be usable, then subtract: commissioning and snagging, the physical work, delivery, order lead time, permit and inspection, specification and load calculation, survey. Whatever date that lands on is when the job actually starts, and it is usually several months before the window.
What is worth settling early: whether the existing curb, electrical supply and duct connections can be reused, because each of those is either a small item or a large one and nobody knows which until somebody looks; and whether a crane is needed, because that brings scheduling, access and sometimes a closure with it.
What we would insist on: commissioning readings taken and handed over — airflow, static pressure, charge against the manufacturer’s target, temperature split, and amp draw. On a plant replacement they are the only evidence you will ever have that the thing was set up rather than merely connected.
And a plain warning about sequencing: if the ductwork needs work, it happens with the old unit out and before the new one is in. Discovering the duct problem after the new equipment is connected costs the window.
The best possible use of an empty building, and the most under-booked
Duct repair, sealing, re-supporting and balancing all need the space empty, quiet and dusty-tolerant, and none of them needs long lead times or a permit. That combination makes them close to ideal shutdown work, and they are almost always deferred in favour of something that would have been better done later.
What we would do first: measure. Static pressure against the equipment’s rated limit, airflow at every register, return capacity and duct leakage. In most buildings that produces a short list of specific, bounded jobs rather than a vague duct programme.
Why balancing needs the window: it involves being in ceilings, adjusting dampers and re-measuring repeatedly, and it cannot be done properly around occupants. It is also the single cheapest way to fix most complaints about a room being too hot or too cold, and it does not involve buying anything.
One caution: balancing is only meaningful once the system is otherwise sound. Balancing around a failed damper, a disconnected run or a starved return produces a compromise rather than a result, so the repairs come first and both have to fit.
Small work, large consequences, and the one thing to insist on
Thermostats, sensors, schedules, zone dampers and staging are quick to change and slow to verify. The physical work usually takes hours; proving the building behaves as intended takes longer and needs the system running.
What we would do: record what the settings are now before changing anything — schedules, setbacks, balance points, lockouts, and which sensor controls which zone. That record is what lets you undo a change that turns out to be wrong, and almost nobody takes it.
What we would insist on: a written note of every setting as left, handed over. Controls are where buildings drift, and a building whose settings nobody wrote down drifts invisibly for years.
And the boundary: where a site has a building management system, the programming belongs to whoever owns it. We will read it, work with it and tell you what it is doing to the equipment, but we will not reconfigure somebody else’s controls strategy in a shutdown week and leave you with it.
Sort the list by consequence before anybody prices it
A backlog is usually a mixture of four different things: safety items, things that stop a building being used, things that cost money quietly, and things that are merely untidy. Working through it in the order it accumulated means the fourth category gets done and the second does not.
What we would do first: walk the list with you and rank it, then split it into what needs the window and what does not. In most backlogs the majority does not, which means the window can go to something that genuinely needs it.
What tends to be at the top: anything gas, anything with a carbon monoxide implication, condensate drains that have failed, and anything where a single failure takes a building out of use. Those are rarely the items that have been on the list longest.
And the honest observation: a backlog that keeps growing is usually a maintenance-interval problem rather than a budget problem. Clearing it once without changing the interval produces the same list in three years.
The best thing you can do with a short window, and the cheapest
An empty building is the only time every unit, ceiling void, plant room and roof is reachable at once without an appointment. A full survey in that week is worth more than almost any physical job, because everything afterwards depends on it.
What it produces: every unit found and listed with model, serial, manufacture date, what it serves, filter size and the location of its isolator and condensate run; measured readings; photographs; and a replacement horizon per unit as a range with a reason attached.
Why the range matters: “four to seven years, because of the refrigerant it holds and the state of the coil” is a budget line. A single year for a machine that currently works is invented, and any contractor offering one is guessing on your behalf.
What it changes immediately: next year’s window stops being a guess. You will know what needs it, what does not, and what has to be ordered in February for a job in August.
About a replacement
Ask for the lead time on the specific equipment in writing before the window is booked. It is the number the whole plan hangs on and it is usually assumed rather than confirmed.
About ductwork and balancing
Ask for static pressure and airflow measurements before and after. Balancing without a before figure cannot be shown to have achieved anything.
About controls
Ask for every setting to be recorded as found and as left. It is ten minutes of work and it is the difference between a change you can undo and one you cannot.
About a backlog
Ask for it to be ranked by consequence rather than by cost. A list sorted by price gets the cheap things done and leaves the expensive risk in place.
About a survey
Ask for the document in a form you can keep and forward. A survey that lives in a contractor’s system is not yours, and you will want it when you change contractor.
With a few days
Read the note above before anything else on this page. In short: maintenance, repairs, balancing, controls and a full survey all fit comfortably; equipment replacement usually does not, because lead times, permits, cranes and commissioning are not compressible. We would rather defer than hand you a half-finished building.
With about a week
A genuinely useful window, and enough for one significant job plus the routine work — provided everything that is not physical work has already happened. Order placed, equipment delivered and on site, permit issued, crane booked. If any of those is still outstanding when the window opens, choose a different job for it.
With a month or more
Enough for a plant replacement done properly, including the duct work that usually gets discovered halfway through. Two things worth doing with the extra time rather than adding scope: commission against a real load rather than an empty building, which may mean returning once it is occupied; and finish the record while everything is still open and reachable.
With the window not fixed yet
The best position on this page, and the one to use. Fix the reopening date first, then work backwards through commissioning, physical work, delivery, lead time and permit to find the date the job actually starts. That calculation frequently shows that this year’s window is already too close and next year’s is comfortable — which is a much better thing to know in February than in July.
Four things to settle before a window is booked
-
The reopening date, working backwards
Commissioning, physical work, delivery, lead time, permit, specification, survey. The date that produces is when the job starts.
-
Written lead times for anything being ordered
For the specific equipment, not for equipment in general. Everything else in the plan depends on this one number.
-
What the window is actually for
The part that needs the building empty, and nothing else. Work that could have been done occupied should not be consuming the window.
-
What happens if something arrives late
Agreed in advance and in writing. Every window has a point after which starting is worse than deferring, and it is better to name that date than to discover it.