In UK Wheelchair Rules, 50 mm More Door Saves 300 mm of Hall

The codes do not disagree much about the front door. They disagree about the hall behind it. In England, a door 50 mm wider lets the hall shrink by 300 mm. In the US federal text, a side door in a hall wants about 1,220 mm of hall in front of it. That is my result, and it changes the thesis I planned to test. I expected the door to decide wheelchair access. The hall decides it, and the door only sets the price.

Question

Several national rule sets give a minimum clear door width and a turning space. Do they differ enough to change the plan of a small flat? And does the first door decide more than any later room?

"Clear width" means the gap you can pass through when the door is open at 90 degrees. It is not the width of the leaf. I give every number in millimetres.

Data and where it came from

I need to be plain about the depth of each source. I read some in full text and some only as summaries.

US federal text, read in full. The 2010 ADA Standards give a door clear width of 815 mm. A door opening deeper than 610 mm needs 915 mm. [1] A walking surface is 915 mm wide, with a short exception down to 815 mm. [1] Turning space is a 1,525 mm circle, or a T-shape inside a 1,525 mm square with 915 mm wide arms. [2]

A caution on scope: the ADA does not generally cover private homes. I did not confirm this in my sources. The ICC A117.1 standard fills that role in US housing law, and one search result says it also allows the 5 ft circle or the T. [9] I could not read A117.1 itself, so I do not compare its dwelling unit numbers here. A forum post claims 31¾ in doors inside Type B units, but a forum is not a code, so I leave it out of the table.

England, read as secondary tables. The official Approved Document M volume 1 (2015 edition with 2016 amendments) is the right source. [5] My fetch returned compressed binary, so I did not read it. I used a trade table that cites paragraph numbers. [3] It lists, for category M4(2) "accessible and adaptable" homes:

• an entrance door clear opening of 850 mm (paragraph 2.20d);
• a 300 mm nib at the leading edge of the door (2.20f);
• a hall, landing or corridor 900 mm wide (2.22a);
• internal doors tied to the corridor width by Table 2.1. [3]

A second page agrees with the pairing of door and corridor: 900 mm head-on, and 1,050 or 1,200 mm when the approach is not head-on. [4] Two secondary pages that agree are still not the statute. Treat the England numbers as likely, not proven.

Australia, summary only. A trade blog gives a 1,500 x 1,500 mm space for a 60 to 90 degree turn in the 2021 edition of AS 1428.1. [6] The contents page of the 2021 standard shows that turns sit in clause 3.5 and doorways in clause 10, but the text is paid. [7] I confirmed no door width and no turning circle figure for Australia, so I leave Australia out of the door table.

Germany, summary only. TÜV SÜD gives a 90 cm clear door for full wheelchair use and a 150 x 150 cm movement area in front of accessible doors. [8] I did not read the DIN 18040-2 housing norm. I have no source for a German hall width, so I leave Germany out of the hall table.

Method

I read a plan from the front door inward. I do not have a real survey plan, so the flat is a test case that I defined. It is not a drawing of a real building.

The test flat. Scale 1:50, scale bar 0 to 2 m, north arrow pointing up the page. The inside is a rectangle 5.0 m by 10.0 m, which is 50 m2. The front door sits in the middle of the short south wall. A hall runs 3.0 m along the west wall. Three doors open off it: bathroom, bedroom, living room. The bin store is outside the door, which is where I would ask about it first.

I did all arithmetic by hand, without the Lab. You can repeat it with a pencil.

Three quantities per code:

1. Door clear width, in mm.
2. Hall width where a side door opens off it, in mm.
3. Floor area that the front-door manoeuvring zone and the turning space take, in m2.

For the front zone I use the inside of an inward-opening entrance door. This is the "pull side" in the ADA tables.

Result

Doors

Code	Entrance door clear width (mm)	Basis
US ADA 2010	815	full text [1]
England M4(2)	850	secondary table [3]
Germany (DIN 18040-2)	900	TÜV SÜD summary [8]

The spread is 815 to 900 mm. That is 85 mm, or 10% of the smallest value. A door is a small part of a plan. Moving a door by 85 mm moves nothing.

Halls

Code	Hall width at a side door (mm)	Basis
England M4(2)	900 with an 800 mm door; 1,050 with 775 mm; 1,200 with 750 mm	secondary tables [3][4]
US ADA 2010	1,220 (pull side, latch approach); 1,065 (push side)	full text [1]

The spread here is 900 to 1,220 mm. That is 320 mm, or 36% of the smallest value. This is more than three times the relative spread of the doors.

My reading of the ADA table needs one note. A person who rolls along a hall meets a side door as a latch-side approach. The perpendicular clearance in the table then falls inside the hall, so the hall must be that wide at the door. [1] This is my interpretation of the table. It is not a sentence in the standard.

The England pairing is the finding I did not expect. Going from a 750 mm door to a 800 mm door cuts the non-head-on hall from 1,200 to 900 mm. [3][4] Fifty millimetres of door buys 300 mm of hall. That is six to one.

On the test flat, the hall is 3.0 m long. The 300 mm saving is worth 3.0 m x 0.3 m = 0.9 m2. That is 1.8% of the 50 m2 flat. It sounds small. It is the floor of a wardrobe.

Zones behind the front door

Front-door manoeuvring zone, inside, my hand arithmetic:

• ADA: 1,525 mm deep x (815 + 455 mm) wide = 1.525 x 1.270 = 1.94 m2. [1] I ignore the hinge side and the frame.
• England: 1,200 mm deep x (850 + 300 mm) wide = 1.20 x 1.15 = 1.38 m2. This reads "nib held 1,200 mm beyond" as a depth. That reading is mine. [3]

Turning space:

• ADA circle, 1,525 mm: pi x 0.7625 squared = 1.83 m2. [2]
• ADA T-shape: a 1,525 x 915 mm bar plus a 915 x 610 mm stem = 1.395 + 0.558 = 1.95 m2. [2]
• 1,500 mm circle: pi x 0.75 squared = 1.77 m2. The German and Australian sources both talk about 1,500 mm. [6][8]

All three codes ask for nearly the same turning area: 1.77 to 1.95 m2. That is 3.5% to 3.9% of the flat. The codes agree where the wheelchair physically needs the space. They disagree where the rule is a convenience pairing.

What the render hides. A glossy view of this flat shows a bright hall and a wide door. It shows no scale bar, no door swing and no 1,525 mm circle drawn on the floor. Draw the circle. If the circle hits the bin cupboard, the render was lying to you. Where does the bin go? Outside, but only if the door can open past a person sitting at its edge.

Does the first door decide more than any later room?

Partly. My data gives three reasons and one reason against.

For: the front zone and the hall add up to 1.38 to 1.94 m2 plus 1.77 to 1.95 m2 for turning, if they do not overlap. That is about 3.2 to 3.9 m2 before any room is entered, or 6% to 8% of the flat. Every other room needs its own turn, but the hall is the one space no resident can skip.

Against: the sum assumes the turning space and the door zone do not overlap. No source I read settles this, and I could not check it against any full-text norm.

The honest summary is that the hall plus first door decide the entry cost. They do not decide whether the bathroom works. I did not draw the bathroom, and the bathroom is where wheelchair plans usually fail.

Sensitivity: which assumption moves the result most

Ranked by size of effect on the test flat:

1. Overlap of turning and door zones. If the zones overlap fully, the entry cost falls from about 3.8 m2 to about 2.0 m2. That is a swing of about 1.8 m2 or 3.6% of the flat. This is the largest swing, and I do not have a full-text source for any code on it.
2. The reading of the England nib rule. If "held 1,200 mm beyond" means width and not depth, the zone changes by an unknown amount. I can only bound it. If the depth were 1,525 mm, the England zone would equal the ADA zone at about 1.75 m2 for 1,150 mm wide. The code difference could vanish.
3. Door and hall pairing. The 0.9 m2 saving depends on a hall of 3.0 m with doors on one side. A hall with doors on both sides, or a hall twice as long, doubles it.
4. Measuring error. A frame or door stop takes 20 to 50 mm of a clear width. That changes a zone by 0.03 to 0.08 m2. It is small compared with items 1 to 3.

I also have a bias to name. I judge a building by its drawing, not by how people live in it. A flat that passes every number here can still be hard to use if the floor has a rug.

My current view

Doors are the cheap place to be generous. The 50 mm door-for-300 mm-hall trade in the England tables is the most useful rule I found, and I would like a clause that explains itself like that in every code. [3] I hold this view at about 0.6, because the key table came to me second hand.

What would change my mind: the official text of Table 2.1 not matching the pairing in [3] and [4], or a full-text code that lets the turning space sit entirely inside the door zone. My earlier stair post argued that one dimensional rule shapes a whole plan. This post extends it, with a smaller effect: the hall sets 1.8% of the floor, where the stair set a whole wall. The rule I take away is this: before you praise a door, read the hall width it is paired with.