Multi-split diversity: why a System 4 shares its output

Diversity is the name for an assumption the designer makes. It says a household will not call on all of its rooms at the same time. Capacity is then chosen against ordinary use rather than against the busiest hour the flat could ever produce.

The architecture behind it is covered elsewhere. What a multi-split aircon is, and why one outdoor unit serves several rooms in most Singapore flats, is settled ground on that page. This one picks up at the capacity the shared unit was given.

The reasoning is sound engineering, not a corner cut. A condenser large enough to hold every head at its printed figure would sit well below that ceiling on nearly every day it runs. It would cost more, take more space on the ledge, and draw more current to do the same ordinary work.

The assumption never reaches the household, which is what turns it into a complaint later. It is not printed on the box. It does not come up at handover. A family learns the shape of it by living outside it, usually on the first hot evening when everybody is home.

Each head's rated figure is a laboratory measurement to begin with, and the guide on actual cooling capacity covers what a real room takes back from it. Diversity stacks on top of that rather than replacing it. Both are at work on the same hot evening.

Makers do publish the consequence, in a document buyers are rarely shown. A combination table lists what each head delivers for every mix of heads that can run together. One such table, covering a four-head outdoor unit rated at 7.1 kW, produces the figures below.

The symptom has a signature, and the signature is occupancy. Cooling that disappoints on a full evening and behaves on a quiet one is reporting how many heads are running. It is not reporting the condition of any part.

A fault that has developed does not behave that way. Fouling on a coil, gas sitting below where it was set, a blower turning slower than it should: each of those is present with two heads running as much as with four. Quiet afternoons show them too.

That difference is the whole diagnosis, and the household is better placed to see it than a visitor is. Nobody arriving for an hour knows which rooms were on last Sunday. The people living there do.

There is a check the household can run without any tools. On the next full evening, switch two heads off and leave the rest running. Rooms that recover have named the ceiling. Rooms that stay poor point somewhere else, and the occupancy pattern was a coincidence worth setting aside.

Note what that check produced before anybody is called out. Which heads were on, which room fell behind first, how the others responded once two were switched off. A visit built on those notes is aimed at something specific.

A spare port is an invitation, and it is where most households cross the line without noticing. The store room becomes a study. A System 3 quietly becomes a System 4 on an outdoor unit that has been reliable for years, and the cooling everywhere else changes character.

Nothing was added outside. The outdoor unit produces what it always produced, and that output now divides one more way. Three rooms that used to hold their setting on a hot night have a smaller share than they had last season, and none of the three was touched.

Port count and capacity allowance are different limits, and quotes routinely treat them as one. A four-port outdoor unit will physically accept a fourth head. Whether that particular set of heads sits inside the published allowance is a different matter, and the tables carry the answer.

Ask what the connected indoor capacity becomes once the new head is on, and ask it while the quotation is still open. The figure is published and it takes a minute to look up. Where a supplier cannot produce it, nobody has checked, and the three existing rooms are the ones who pay for that.

Households that genuinely fill every bedroom every night are a specification problem rather than a repair problem. Shift work, three generations under one roof, a room somebody works in late: each of those lifts simultaneous use above what the standard assumption covers.

State that pattern in plain terms before a model is chosen. How many rooms are occupied on a normal weeknight. Which hours overlap. Which room has to hold its setting while everything else is running. A supplier can size against answers like those. A floor plan on its own gives them nothing.

The lever with the most in it is usually how the load gets split, not how large the condenser is. Two outdoor units, where the ledge and the building rules allow a second one, remove the sharing question instead of negotiating with it. That trade belongs to one condenser vs two condensers, which works it through properly.

Where one outdoor unit is the only permitted arrangement, the mix of heads is what remains to be argued over. A head sized generously for the room that matters most claims a larger share of the output once every room is running. Published tables show that directly, and it is a cheaper lever than lifting the entire system.

Buying the outdoor unit a size up is a separate decision carrying costs of its own, and the guide on oversizing sets out what those are for a room that does not need the extra. The question here is narrower. It asks what the system has to hold when every room runs, not what any single room needs on its own.

A technician checking one room on a quiet afternoon will find a system in good order, because at that moment it is one. Pressures read where they should. The temperature split across the coil is normal. No part that can be tested is failing.

That is the honest limit of a single-room visit, and it deserves saying before one is booked. The condition being described only exists while several heads are running. Test one head and the condition is not there to be measured.

A visit that can settle it looks different. Every head on, the flat in its normal evening state, readings taken while the system carries the load the complaint describes. That needs cooperation from the household, because the flat has to be put into the state producing the problem.

Treat a clean result from a single-room check as incomplete rather than reassuring. It rules out very little. A report that the system tested fine, taken at the wrong hour, has confirmed something nobody was disputing.