EcoAffordable: Mass Timber Mid-rise Housing Co-op

Hi, I'm Lincoln Clarke, and I just graduated high school in Toronto, Canada, where both homelessness and affordable housing are significant issues. In cities, a huge driver of housing costs is the price of land. In the Toronto area, a tear-down house will easily sell for over a million dollars. With land prices that high, standalone houses are simply not an affordable housing solution. My aim for this project is to overcome the high cost of land by building a mid-rise housing complex. With many stories, construction becomes the main cost, not land. Fortunately, we can implement innovative solutions to reduce the costs of building, as opposed to land, which we can't do anything about.

Apartment style living also allows for more opportunities to create a cohesive community through shared spaces and proximity between residents.

1. Autodesk Fusion
2. (laser cutter)
3. (2mm basswood wood)

I'll choose 189 Dundas St West, Mississauga, Ontario, which is an empty lot near where I live. Although it may look like a park from above, it's not, it's just overgrown with weeds and shrubs. The main reason I chose this lot was because it's surrounded by apartment buildings on both sides, which makes zoning easier for a mid-rise residential building. It's also very close to multiple parks, grocery stores, a hospital, and a train station that commuters can use to get into downtown Toronto. This is also a large lot, allowing us to achieve scale, which can help reduce costs.

The ground floor of the building will be dedicated to retail space, which will be highly valuable because of the high density of this area.

In order to come up with ways to make construction cheaper, I researched the elements that contribute most to the cost of traditional concrete mid-rise housing projects. Then I further broke down the larger categories to help pinpoint the expensive aspects of traditional construction.

1. Structural frame (load bearing pillars, concrete floors...) => 30-40% of the building cost

Element of structure — % cost of the structure

1. Formwork (including labor for setup) — 35–45%
2. Concrete Material — 20–30%
3. Rebar Material — 15–20%
4. Rebar Labor (tying, placement) — 10–15%
5. Concrete Labor — 5–10%

2. MEP integration (Mechanical, Electrical, Plumbing) => 20-25% of the building cost, HVAC is the main cost driver

We'll focus specifically on HVAC.

Element of HVAC— % cost of HVAC

1. HVAC Equipment (Furnaces/ACs/heat pumps) — 30–40%
2. Ductwork Materials — 15–20%
3. Ductwork Installation Labor — 15–20%
4. Refrigerant piping — 10–15%
5. HVAC Controls + Thermostats — 5–10%

3. The building envelope (insulation, cladding...) => 10-15% of the building cost

Element of the envelope — % cost of the envelope

1. Cladding Material — 40–50%
2. Cladding Labor — 20–25%
3. Window Units — 20–30%
4. Flashing, Waterproofing — 5–10%
5. Thermal Insulation — 5–10%
6. Attachment Hardware — 2–5%

The structural frame is the largest driver of cost, with high potential for innovation. Formwork costs 40% of the structure, totaling around 15% of the building's overall cost. Rebar and concrete each cost around 10% of the building's cost.

Formwork is the single largest cost of the building, and yet it isn't even present in the final structure! We could also cut costs by replacing the complex 2 material structure (rebar + concrete) with a single material. 3D printing came to mind as a means to eliminate formwork and use only one material for the structure, but it poses structural challenges for mid-rise housing, and so far has only been used for shorter buildings. 3D printing also comes with high upfront costs for complex machinery. Mass timber, on the other hand, eliminates formwork and uses a single material for the structure. It's also a much simpler an better established construction technique, which minimizes the risk of cost overruns and makes skilled labor easier to find. Mass timber uses wooden beams for the structure and cross-laminated timber (CLT) panels for floors and outer walls.

The HVAC equipment will remain expensive, but it's worth spending on since it can deliver efficiency gains long term. We'll use heat pumps, which are a larger investment upfront but are highly efficient in most climates. They also eliminate the need for separate furnace and AC systems. Refrigerant piping and ductwork throughout the building are very expensive and can be mostly eliminated with a decentralized HVAC system, each unit having its own heat pump and simple ductwork. We'll use cylindrical pipes, which are generally cheaper than square ones. These will be left exposed on the ceiling to eliminate the cost of embedding ductwork in walls and ceilings.

Like the HVAC units, we can't cheap out on the cladding materials because we want this housing project to be affordable not only in its construction but its entire lifecycle. We'll use high quality insulation and waterproofing. Our main cost savings will come from reducing labor by prefabricating exterior wall panels which will be installed much faster on-site. These wall panels will include the CLT panel, waterproofing, insulation, cladding, and windows in one piece.

One of the main goals of this project is sustainability. This is supported by the use of mass timber, which has a much lower carbon footprint than concrete or steel. Using heat pumps, which run on electricity is also more sustainable for 2 main reasons:

1. Uses electricity instead of natural gas - In Ontario, electricity is mainly produced with renewables and nuclear, resulting in a much lower carbon footprint. (see pie chart above)
2. More than 100% efficient because heat is transferred, not produced - More efficient than electric baseboard heating.

Mass timber and decentralized HVAC with heat pumps are both cost saving and eco-friendly, hence the name EcoAffordable.

The building needs to be to code, which requires that all bedrooms have an exterior window. As a result, making one solid rectangular block that covers the whole lot is not an option because inner units wouldn't have windows (not to mention it would be poor quality housing).

There are 2 main building layouts we could chose, as illustrated above:

1. A rectangular donut around a central courtyard
2. Five standalone rectangular buildings

In both illustrations, the buildings are 20m wide/deep, allowing for a double loaded corridor with an 8-9m apartment on each side. We'll decide on the type of parking in the next step, but both these sketches provide 30 m of extra space behind the building(s) for parking. If we choose underground parking, the buildings can be made slightly larger.

Ultimately, we want to optimize the cost-per-unit and the quality of living for residents.

To start, the donut provides a slightly larger floor area, better utilizing the lot. Lot utilization isn't the only thing to consider though—we also need to think about the usable area for apartment units and the cost of stairs/elevators for each solution.

The Ontario building code mandates the following for a 12 story residential building:

1. 45m maximum egress distance from units to closest stairwell
2. At least 2 stairwell exits per floor
3. At least 1 elevator capable of accommodating firefighters and paramedics with a stretcher (This elevator can also be used as a regular passenger elevator)

Although the code only specifies a minimum number of elevators, it also states that buildings with fewer elevators than industry standard will be unlikely to get building permits. We will aim for 1 elevator per 80 units.

For rough calculations, let's assume 1000 square foot units (92 m²), which is a comfortable size for a family of 4. For this unit size, the rectangular buildings would have around 125 units requiring 2 elevators each, giving a total of 625 units and 10 elevators. The donut design would have around 650 units, requiring 8 elevators. The unit-to-elevator ratio of the rectangular buildings would be 62.5:1 while it would be 81.2:1 for the donut.

The rectangular buildings would have 2 stairwells each, for a total of 10 stairwells. A double-loaded corridor running along the donut shape would be 340 m long. The maximum distance from units to stairwells is 45m, giving an absolute maximum of 90 m between stairwells. Let's say 85 m for a margin of safety. 340/85=4 so we would need 4 stairwells, which is much better than the 10 total required for the rectangular design.

Overall, the donut design is slightly better for the cost of elevators, and significantly better (4 vs. 10) for stairwells.

The donut is also better for fostering community because of the large central courtyard which can be used as a park, to host community events like barbecues, and more.

We'll choose the donut.

Underground or above-ground parking? In both cases, we'd be looking at a multi-level garage.

The cost of building underground parking can vary significantly from $50,000 - $120,000 per stall. Let's assume it would cost $70,000 per stall. Above ground garages tend to cost around $30,000 per stall, but come with the drawback of occupying land.

We'll do a cost analysis, assuming we would leave 30 m of space at the back of the lot for parking, and that that space would be occupied by the building if we build underground parking. For both cases, we'll allocate 0.5 parking spaces for residents plus 0.15 for visitors, per unit. Although this doesn't sound like a lot of parking, our building site is located on a transit corridor and so building with this reduced parking ratio is approved by the city of Mississauga. In addition, we'll allocate 35 spots per 1000m² of retail space.

Above-ground parking

1. 650 apartment units => 650*(0.5+0.15)=472 parking spots
2. 5700 m² of retail space => 5700*35/1000=200 parking spots
3. Total of 472+200 parking spots=672
4. Total cost of parking = $30,000*672=$20,160,000

Underground parking

1. Additional apartment units from making the building 25m longer: 10 extra units per floor
2. 650+11*10 apartment units => 760*(0.5+0.15)=494 parking spots
3. 6600 m² of retail space => 6600*35/1000=231 parking spots
4. Total of 494+231 parking spots=725
5. Total cost of parking = $70,000*725=$50,750,000

In the underground parking scenario, the building above ground would also cost more because of the additional size. I'll do a rough calculation for the additional cost of the larger building. It will be very approximate, but that's okay as we'll see later on. Let's assume a total cost of $300 per square foot using mass timber construction. for 110 extra 1000 sq ft units, $300 * 1000 * 110= $33,000,000 additional cost.

Taking into account both the additional cost of building parking and the building above ground, it's around $63 million extra.

Are 110 * 1,000 sq ft units + 900 m² of extra retail space worth $63 million?

Not quite. Let's say each unit goes for $340,000, which although expensive, would still be considered rather affordable in the Toronto area (and that is with profits). Retail space would probably sell for around $4,000 per square meter.

In total, that would be an additional value of $340,000*110 + $4,000*900=$41 million.

Overall, above-ground parking has better economics: $41 million of extra value isn't worth $63 million of extra costs.

We had allocated 30 meters at the back for parking, but now that we've decided on a design, we'll do more detailed calculations about the necessary space for the above ground parking. It should not be taller than the main building, both for aesthetics, and for zoning, since parking will be adjacent to 2 story houses. It would be ideal if the parking garage was slightly shorter than the main building so that the height of buildings rises gradually from the houses to the apartment.

Here are the mandates of the building code for my chosen property:

1. Above-ground parking garages require a minimum 7.5m setback from all lot lines
2. 9+ story apartment buildings adjacent to other apartment buildings require a minimum 9m setback from side lot lines
3. No minimum requirement specified for setback from street, but 5m is the lowest reasonable amount.

We previously calculated that around 672 parking spots will be required, and let's assume that number stays the same. The number of apartment units probably won't change even if we make the building slightly smaller to optimize the parking garage. Instead, each unit may need to be slightly smaller. The 672 parking spots we calculated also isn't an exact number we need to stick to. Of those spots, 200 were for retail parking at 35 spots per 1000 m², which is industry standard, but slightly more than required by code. Some of those can be eliminated, especially given that part of the retail's customers will be residents who won't require additional parking. The reason it's so important that this number isn't exact is because reaching a specific number could require the addition of an entire floor to the parking garage, and being slightly under is fine.

The economics of a car elevator vs. ramps in the parking garage

Although a car elevator may sound high-tech and expensive, it may actually make economic sense because of the additional parking spaces created by eliminating ramps.

The bottom floor would be dedicated to heavier SUVs and trucks while upper floors would be dedicated to lighter duty vehicles. This means the elevator would only need to handle lighter vehicles, which allows us to use a less expensive elevator. An industrial elevator capable of holding a car and servicing 10 floors would likely cost around $900,000 (including installation costs)

If each elevator trip takes 1 minute, and 50 cars want to leave in a 30-minute window, 20 cars are left in queue. We'll need 2 elevators, partly to speed up traffic flow and also to have a backup in case one elevator breaks.

So elevators cost $1.8 million. But ramps are expensive to build too.

There are actually only a few ramp/elevator parking garage layouts that work with the lot size, so I've created a blueprint of 4 possible layouts that we can then compare to optimize costs.

Here's why there are only a few logical layouts:

1. The width is fixed to the width of the lot minus buffer space on each side required by the building code
2. Straight ramps can only go in the width direction of the lot, not the length direction because they require over 45 meters of length. This would cause the parking garage to eat up too much area of the main building.
3. Each design should be as efficient as possible with its footprint, so aisles should be double loaded (cars on both sides)

The following building code requirements also apply:

1. Minimum 2 stairwell exits
2. Maximum egress distance of 45m from stairwells

This will require a stairwell at each end of the parking garage in the width direction, which is nearly 60m.

The layouts shown in the pictures above all assume 2.75 x 5.2 meter parking stalls and 7m aisles.

We can immediately eliminate layout 4 because it would need to be 15 levels, which is too tall to be adjacent to the houses behind.

While layout 3 does make efficient use of the building, the helical ramp poses challenges in terms of the building code. The ramp is still considered a part of the building, and so the edge of the ramp, not the garage itself, must be 7.5 m from the lot line. Similarly, the helical ramp protrudes in the direction of the main building, and we still need a space between the building and the ramp for cars to drive through a drop off zone. The helical design isn't a good use of land because it's like it occupies the area of the smallest rectangle it would fit in, which comprises the ramp.

Layouts 1 and 2 are the viable options for comparison.

First, the elevator layout (2) is 6 meters shorter in depth, allowing the main building to be 6 meters longer. This would result in an additional 240 m² per floor.

Let's compare the cost of building layout 1 and 2. This calculation will again be very approximate, but we'll see that doesn't matter.

We'll assume that the cost of building a precast concrete structure is $920 per square meter.

For layout 1

1. 7 floors
2. 40.8x65m area
3. Total cost of 7*40.8*65*920 = $17,079,000

For layout 2

1. 8 floors
2. 34.8x65m area
3. 2 car elevators each costing $900,000
4. Total cost of 8*34.8*65*920 + 2*900,000 = $18,449,000

Let's consider how the smaller size of layout 2 may justify the extra cost of $1,370,000.

Let's estimate the value of the additional space layout 2 provides for the main building:

1. 240 m² on the ground floor for retail space @ $4,000/m² = $960,000
2. 240 m² for the 11 floors above, creating 2 additional large units per floor (or 4 additional small units). Let's value the 2 large units at $400,000 each. 11*$400,000*2=$8,800,000

This is an additional $9.76 million of value. But, we need to account for the additional cost of making the building bigger.

Because of scale, the marginal cost of adding a few meters of length to the building is most likely less than the average cost of the building. In a previous section, we estimated a cost of $300 per square foot for mass timber construction. Let's discount that to $250 for this calculation. 240 m² * 12 = 31,000 sq ft.

The total additional cost of building would be $7.75 million.

The verdict

Layout 2 costs $1,370,000 more to build, and making the main building larger costs another $7.75 million. This combined additional cost of $9.12 million results in $9.76 million of additional value from the added retail space and units. The net value of layout 2 is $640,000 higher. This alone, however, doesn't make it a clear winner because these calculations are rough. But, layout 2 creates 40 additional parking spots, which I think provide a good margin of safety for the calculations given they could each be rented at around $3,000 per year. Over 10 years, this would give them a value of $1.2 million. Additionally, the car elevator provides a much better parking experience than having to circle around floors and up ramps, likely resulting in less accidents over time.

We'll go with the car elevator design (layout 2).

This step comes before we sort out the details of a floorplan because I want to decide on the overall look of the building to create some AI renderings for inspiration (the next step).

Foundation

The foundation will be expensive, but it's a necessary cost of constructing a structurally sound building. We won't have a podium (1st above-ground floor out of concrete), which reduces costs. The ground floor does need to be concrete though. There 2 main options:

1. Reinforced concrete mat: 0.5-1m thick slab of concrete spanning the building's footprint.
2. Pile foundation with thinner concrete slab: Deeply drilled piles under columns, with a thinner slab spanning the building's footprint. This is generally more expensive because of engineering costs and drilling. It's mainly used when soil is unstable.

The other option is spread footing, which is like a pile foundation but wider, shallower piles, but this can only be used with very solid ground, which is overly optimistic. Let's assume the soil has medium stability, in which case the cheapest solution is the concrete mat.

Structure

Mass timber - Glulam beams and CLT.

Exterior wall paneling

The exterior walls will be prefabricated panels with the following layers, from inside to out:

1. 2x "5/8 type X gypsum board as drywall and fire encapsulation
2. CLT panel for structural support
3. 2x "5/8 type X gypsum board for fire encapsulation
4. Air and vapor barrier membrane layer
5. Mineral wool board continuous insulation (This achieves a good balance between cost and thermal performance while also being non-combustible, which is a requirement)
6. Water resistive barrier
7. Furring strips (Screwed into the CLT, through the other layers
8. Cladding

The choice of cladding significantly affects cost. The choice, in order of increasing cost, is between

1. Vinyl
2. Corrugated aluminum
3. Engineered wood
4. Aluminum composite panels.

In terms of aesthetics, engineered wood and aluminum composite panels are nicest. Engineered wood however, is combustible, which would require additional fire safety layers in the wall paneling. This would drive up cost so it isn't a good option for affordable housing. Aluminum composite panels are generally triple the cost of corrugated aluminum, so they aren't a good choice either. Vinyl also faces fire safety issues because technically it is combustible even though it has a high ignition temperature. Corrugated aluminum is the best choice because it's both inexpensive and fire-resistant.

Corrugated aluminum, however, isn't as aesthetically appealing. I was inspired by the hotel in the first picture above. If you look closely at the picture on the left, it's actually quite an old brick building. But, they've refreshed it simply with paint, and the other side that's painted black with white lines actually looks very modern. You wouldn't know it's an old building at all. Painting wide, diagonal lines is a cheap fix to the appearance of corrugated aluminum, which will give the building a much nicer modern feel. The aluminum will be a light gray, so I plan to add blue, charcoal, and white lines. The paint can be applied in the factory, during pre-fabrication of side panels.

Electrical and pluming integration

Because of freezing during winter, the building code requires that plumbing be embedded in interior walls. While electrical can be integrated in exterior walls, it would require a gap between the finished drywall and the CLT panel, and then the drywall layer would no longer qualify to encapsulate the CLT, requiring a second layer of gypsum. All plumbing and electrical will be in interior walls.

HVAC

Apartments will all have individual heat pump units. The exterior units will be relatively compact, measuring around 0.4m * 0.8m * 1m. This size heat pump is sufficient for spaces around 90 m² (our average unit size). Heat pumps are more efficient than electric/gas heating in temperatures down to -15°C to -20°C, but in the Toronto area, it sometimes drops below that during winter. To ensure the heating system can always keep up, we'll use a heat pump with integrated electrical resistance backup. It's the same as a normal heat pump except air also blows over heating coils on the inside unit. Those coils can be turned on when it gets really cold, increasing efficiency and allowing it to keep up in frigid weather.

Parking garage

For this we'll just use precast concrete, which does normally look utilitarian. Again, we'll give it a modern vibe with painted lines on the exterior. To make it match the main building's aesthetic while making the parking garage a distinct building, we'll use purple, charcoal, and white instead of blue.

The first picture above shows what the lot layout looks like now.

I've created a rough, block-like design for the building and parking lot, as shown in the second picture. Then I had AI create a more detailed description, which I slightly modified.

Here is the final description I fed to Microsoft Copilot AI image generator:

---------------------------

Overall Scene: A realistic architectural rendering of a modern, 12-story mass timber apartment building situated on a lot measuring 80 meters by 175 meters. The scene should capture the building from a slightly elevated, oblique angle, showcasing its front and one side, with the parking garage visible at the rear. The lighting should be even and clear, highlighting textures and forms without harsh shadows, suggesting a bright, clear sky, daytime. The overall aesthetic should be clean, contemporary, and functional.

Main Apartment Building:

1. Shape: Rectangular "donut" shape, with sharp, rectilinear edges and corners, not rounded.
2. Height: 12 stories.
3. Outer Dimensions: 62 m width, 121 m length.
4. Courtyard Dimensions: The central open courtyard is 16 m by 80 m, giving the donut a thickness of 23 m. The ground of the courtyard will probably not be visible from the rendering angle I requested.
5. Exterior Siding: Corrugated aluminum siding, in a neutral gray tone. The corrugated texture should be subtly visible, adding depth. The cladding is painted with diagonal wide likes, blue, white, and charcoal, in somewhat random placement. Only include up to 7 lines on one face. These are meant to be just a couple-few lines, not busy.
6. Windows: Arranged in a precise, regular grid pattern across all visible facades. Each window is a standard rectangular size (not full height of the floor) and should appear slightly recessed into the facade.
7. No Balconies
8. Roof: Flat roof, with minimal visible mechanical elements.

Parking Garage:

1. Location: Located behind the main apartment building, above ground. It's a separate building.
2. Construction: Pre-cast concrete structure. The characteristic modular panel lines of pre-cast concrete should be discernible.
3. Color: Light, utilitarian gray. Include similar lines as with the main building, except make them purple, white, and charcoal.
4. Dimensions: Approximately 65 meters wide and 35 meters deep. It’s 8 levels.

Lot and Surroundings:

1. Lot size: 80m x 175m dimensions. (80 m width on the street)
2. Street (Front): A paved street in the foreground, gently curving to accommodate the building's frontage. There is a sidewalk between the street and building, with a strip of grass, then a cement pad/walkway and the entry to the building
3. Driveways (Sides): Wide, dark-paved driveways extend along both sides of the apartment building, providing access.
4. Green Space: Beyond the driveways and flanking the entire lot are well-maintained grassy areas, rendered with a natural, vibrant green.

Make sure to show that the building is twice as long as it is wide.

---------------------------

This generated the rendering shown in the third picture, which I'll use as inspiration. There are some errors, like the number of floors being slightly off, but it's good for the general concept.

Here are some of the features I'll add:

1. Retail space on the ground floor, with priority renting for residents who want to run small businesses
2. A daycare
3. A shared central courtyard with the following spaces
4. BBQs and outdoor picnic area
5. A playground with a play structure
6. A large open green space for outdoor games and community events
7. Shared WiFi throughout the building to reduce costs for residents

We'll have 2 floorplans: One for retail (next step) and one for residential floors. All the apartment floors will have the same layout, but there will be different sizes of apartments on each floor to allow for variability in family size.

First we need to optimize the width of the donut. For our previous calculations, we assumed 20m, which is a good rough estimate, but let's be more precise now.

We'll have a combination of 1, 2 and 3 bedroom apartments. For 2-3 bedroom apartments around 90 square meters, 10-12 meter deep layouts are generally most efficient. This allows putting more units on the same exterior perimeter while still allowing enough natural daylight. Assuming 2m double loaded corridors, a 20m wide donut would give only 9 meters of depth for units, minus structural thickness. It would be more efficient to make it slightly wider. I want the courtyard to be at least 16 meters wide. We have 62 meters of width to work with, resulting in a 23 m wide donut. Minus the 2m hallway, this leaves 10.5 meters for units on each side, which is a pretty efficient depth.

Designing for mass timber

Our goal is to optimize the mass timber grid for cost while meeting vibration requirements and fire ratings.

We'll start with fire ratings required by the Ontario building code because that's what will determine the thickness of CLT floor panels, which will influence the grid layout. For tall mass timber buildings, Ontario requires encapsulation to achieve fire requirements. Although they do allow for buildings to request "alternate solutions" and demonstrate performance-based fire safety, this requires jumping through more regulatory hurdles, and it can already take upwards of 3 years to get buildings approved in the Toronto area. Additional regulatory challenges will increase costs. To keep costs down, we'll follow the encapsulation guidelines and cover 5-ply CLT wood ceilings with 2 layers of "5/8 type X gypsum. The floor on top doesn't need to be modified because it will already have an acoustic mat and flooring, which is considered sufficient for encapsulation.

We're using 5-ply CLT — Now we need to decide on a grid layout.

Vibration is one of the main considerations when choosing a layout. The larger CLT floor panel span, the more vibration occurs. This is in large part because of resonance. The longer the span, the lower the fundamental frequency of the panel. We want the fundamental frequency of our panels to be above 8Hz because if it's below that, human activity like walking/running/jumping (2-6Hz) will cause resonance and vibrations will be very perceptible. Thinner CLT panels are also more susceptible to vibrations.

Here's what we need for the grid:

1. It covers the donut's 23 m width
2. 5-ply CLT
3. Above 8Hz frequency

We can actually do some simple calculations to find the maximum span distance of our CLT. Here's the general formula for calculating the fundamental frequency of simply supported beams/panels.

We can solve for the maximum length given the minimum 8 Hz frequency.

For 175V 5-ply CLT, which is a widely produced CLT panel,

1. EI=2.54*10^6 N.m²/m
2. m is approximately 150 kg/m²

So, our maximum span will be 5 m, which will provide a comfortable floor. I know 5 m seems really close to the maximum of 5.05, but the 8Hz frequency is already conservative enough. We want to maximize this span because it reduces costs by requiring less columns and beams.

We have a 23m wide donut, and we need to come up with a grid with 5 m spans.

We have a couple options:

1. Run the CLT panels in the 23m width direction of the donut with 23/5=4.6m spans
2. Run the CLT panels in the direction opposite to width with 5m spans

I've created these layouts in the pictures above, which I used to decide on the cheaper option. Both drawings above show a sample section of the building, 23 m wide, and of different lengths to make it an integral number of each grid. Note that these are sample sections of the long rectangular parts of the buildings, and that the corners, and distances between, do complicate things. The 9.2x9.2m grid works perfectly for the corners, while the 7.67 x 10m grid would need to be shortened to a square 7.67 by 7.67 grid at the corners. Between corners, distances also won't be perfectly divisible by the grid size. This is okay. In both cases, it will result in a slightly rectangular grid, for example 9.2 x 9 or 7.67 x 9.7. You might be thinking it would be best to adjust the length of the building to make an integral number of bays fit, but these slight variations don't require redesigning connectors, and so won't affect cost much. Making the building smaller would reduce value created more than any efficiency gains would reduce cost.

The beams that aren't connected to columns are intermediate beams to support CLT spans. The CLT panels will span their length direction in the direction perpendicular to the intermediate beams.

The material cost of the volume of wood accounts for around two thirds the structural cost, so the solution with a lower volume of wood per area is likely the better option. I'll calculate the length of beams that are required for each of these sample structural frames, taking into account the length of the individual beams. The reason length of individual beams is important is because shorter beams will require a smaller depth, resulting in less volume of wood. We'll assume all beams are 0.175 m wide.

9.2 x 9.2 m grid

1. 23 x 55.2 m area = 1269.6 m²
2. 28 total columns
3. 460 m of 9.2 m long glulam beams - Beam depth of 0.75m
4. 32.2 m of 4.6 m long glulam beams - Also beam depth of 0.75 m because there are a relatively small number of these beams and smaller depth would require designing separate connectors, reducing the efficiency of scale and repetition.
5. Total beam volume of (460+32.2) * 0.75 * 0.175 = 64.6 m³
6. Beam volume per area = 64.6 m³ / 1269.6 m² = 0.0509 m³ / m²

7.67 x 7.67-10 m grid

1. 23 x 60 m area = 1380 m²
2. 28 total columns
3. 240 m of 10 m long glulam beams - Beam depth of 0.80 m
4. 299 m of 7.67 m long glulam beams - Beam depth of 0.65 m
5. Total beam volume of 240 * 0.80 * 0.175 + 299 * 0.65 * 0.175 = 67.6 m^3
6. Beam volume per area = 67.6 m³ / 1380 m² = 0.0490 m³ / m²

The 7.67 x 7.67-10 m grid is more efficient. Although the 9.2 x 9.2 m grid is square which has advantages in terms of repetitiveness and scale, bulk lumber discounts and economies of scale still exist with the rectangular 7.67 x 10 m grid because this building is so big. The number of each type of connection and beam we'd need to manufacture is high enough that economies of scale are achieved either way.

For the long sections of the building, the grid will be very close to 7.67 x 10m, so we'll use 5m span, 3m wide CLT floor panels The 5m span will go in the 10m direction of the grid, supported by the intermediate beam. The 7.67 dimension of the grid isn't divisible by the width of the CLT panels, but that's okay since CLT panels are designed to be supported from their short ends, and they don't need to be supported by beams on their long sides. There will be a fractional number of CLTs per bay, but that's okay.

Now we need to apply the 7.67m x 7.67-10m grid to the entire rectangular donut.

We'll start with the corners, which will be a 7.67 x 7.67m grid. Between corners, we'll use a rectangular, longer grid to reduce the number of columns, beams, and CLT panels.

On the short edges, there's 16 m between the corner grids, as shown in the second picture. So, we'll have a rectangular grid of 7.67 x 8 m.

On the long edges, there's 75.1 m between corner grids, as shown in the third photo. Our goal is to make the rectangular grid as large as possible while remaining under 10 m.

75.1/10 = 7.51, so we round up and split the distance into 8 sections

This results in a 7.67 x 9.39 m grid.

The first picture shows the final grid. One great thing about this layout is that we can use the same width CLT panels for all floor space, only length of CLT changes. On the short ends, CLT panels run in the direction of the short ends, and on the long sides, the CLT runs in the direction of the long edges. The short ends and long sides have the same thickness from exterior to courtyard: 23m. This the CLT covers the 23m in the direction of its width. The largest standard CLT panel width is 3m.

23/3=7.67, so we split this into the width of 8 CLT panels, each 2.875 m wide. (In reality, they would be just slightly wider — still less than 3m, because they need to overlap with the columns on the edges, which this drawing doesn't account for)

There will be 8 total elevators, as calculated earlier, split into 2 banks of 4 elevators, each at one of the short ends of the building (Facing the street and the parking lot)

The building will have 2 floorplans for apartment floors. Floors 2-6, will be serviced by all of the 8 elevators, and floors 7-12, which will be serviced by 4 elevators. 4 of the elevators will only go to floor 6. This cuts costs/area in 2 ways:

1. Reduces the cost of elevators + installation by having 4 of the elevators only go half the height of the building
2. Utilizes the footprint of 4 of those elevators on the upper floors. Because of the mechanical equipment, the elevator does still take up its footprint on the floor above the top floor it services. (It services up to 6, so will also occupy 7)

So, we'll have one floorplan for floors 2-7, and a separate one for floors 8-12.

We need to adhere to the Ontario building code and ensure a maximum 45 m egress distance to the nearest stairwell.

The perimeter of the outside of the hallway shown above is 282 meters. The units are 10.5 m deep, which is a part of that 45m travel distance. So, the maximum hallway travel distance is 30m because there are probably some units with 15 m inner travel distance to the door. So, we can have a maximum of 60 m between stairwells.

282/60=4.7. So we'll need 5 stairwells.

One approach would be to optimize stairwell placement to make all the distances between stairwells equal. However, they don't need to be equal because we have some flex. We just can't have above 60m spacing. We want flex so we can just place stairwells wherever is convenient according to the floorplan - This way we don't need to alter unit layout as much (although we do still need space for stairwells).

We'll optimize the distance and then move them a bit from there. The ideal spacing is 282/5=56.4. We'll align the first stairwell with the elevator bank at the back of the building, and go from there. The second picture shows the final stair layout along with both banks of elevators.

Now we need to draw the lines for the interior walls to divide each floor into units, and rooms within them.

Bedrooms

The ratio of different types of units (1BR:2BR:3BR:4BR) is a crucial choice to ensure the building serves the community as best as possible. We'll mainly follow the City of Toronto's affordable housing guidelines, which recommends

1. 40% 1-bedroom units (no less than 525 sq ft, average 590 sq ft)
2. 40% 2-bedroom units (no less than 650 sq ft, average 725 sq ft)
3. 15% 3-bedroom units (no less than 900 sq ft, average 1000 sq ft)
4. 5% 4-bedroom units (no less than 1100 sq ft, average 1175 sq ft)

However, Mississauga has one of the largest average household sizes in the Greater Toronto Area, with 14.3% of households having 5 or more people. Around 10% of households are multigenerational, also one of the highest in the region. This makes larger units a better match to the demographics of the area. Additionally, larger units are typically underserved by the rental market because smaller, 1BR units tend to command higher rent per square foot than larger ones. For this reason, we'll change the ratio slightly to the following:

1. 25% 1-bedroom
2. 45% 2-bedroom
3. 20% 3-bedroom
4. 10% 4-bedroom

In terms of the size of units, we'll make them larger than the minimum set by building codes/Toronto's guidelines, which result in cramped spaces. Affordable housing should still provide good livability.

All units will have one slightly larger master bedroom of the same dimensions, designed for a couple. The 2, 3, 4BR units will have secondary bedrooms, which will all have the same dimensions. For the secondary bedroom, we'll aim for the 9.3 m² City of Toronto guideline for bedroom floor space, excluding a closet.

Initially, here are the dimensions I decided on.

Master bedroom dimensions: 3.2 m along the exterior wall, 3.3 m room + 0.6 m closet depth = 10.56 m² floor space

Secondary bedroom dimensions: 3.0 m along the exterior wall, 3.1 m room + 0.6 m closet depth = 9.3 m² floor space.

However, these aren't practical for construction, as shown by the drawing of this layout in the second picture. They're almost the same total depth, but because they're slightly different, we need to build a zigzag into the wall. To allow for one straight piece of wall, we'll make the master bedroom 0.1m shorter, and the secondary bedroom 0.1 m longer. We'll also make the secondary bedroom 0.1 m narrower because we're making it deeper.

Final master bedroom dimensions: 3.2 m along the exterior wall, 3.2 m room + 0.6 m closet depth = 10.24 m² floor space

Final secondary bedroom dimensions: 2.9 m along the exterior wall, 3.2 m room + 0.6 m closet depth = 9.28 m² floor space.

Both bedrooms actually have slightly more floorspace because the closet won't go the entire width.

The first picture shows the final layout, with closets and beds. A queen bed fits comfortably in the master bedroom, and a twin bed for secondary room, which can be a bunk bed if 2 people are to sleep there.

Living space + Kitchen + Dining room + Bathroom + (office/flex)

The living room should have a window, so it will also have an exterior wall.

Let's start with the total unit size to determine the length of the living room's exterior wall. I'll explain my calculations for a 2BR, and at the end I'll include the final designs for all types of units. Our 2 BR size will be around 950 sq ft or 88 m². This is comfortable for a family with up to 2 children.

The units have a 10.5 m depth. So their width should be 88/10.5= 8.4 m. The bedrooms take up 6.1 m along the exterior wall, leaving 2.3 m along the exterior wall for the living space. This is wide enough for a dining room.

The third picture shows the apartment without any living space subdivisions yet.

The bathroom doesn't need light, so we'll put it on the left side near the hallway. The living room can go right outside the bedrooms, and the kitchen will be directly behind the living room and dining room, near the hallway.

The second picture shows a general layout, without precise dimensions yet.

The dining room, living room, and kitchen will be open concept, which is both a modern look and it's cheaper because it requires less internal walls.

Communal laundry room or in-unit laundry closets?

The main advantage of a communal laundry room is that it's cheaper. However, there are also many arguments for in-unit laundry machines.

1. More convenient for tenants, resulting in higher quality of life and lower tenant turnover.
2. Large families, which our building is geared towards, generate lots of laundry, so the burden of carrying it downstairs is much larger, and a task that must be done more often.
3. Reduced maintenance of common areas
4. Less with conflicts between tenants like when laundry machines are left full
5. Modern builds typically include in-unit laundry, which can make a communal laundry room feel especially dated and undesirable

We will go with in-unit laundry machines. The fact that modern builds have moved towards this strongly suggests that the value tenants get from this is worth the additional cost.

We'll save money on dryer ductwork by using ventless dryers, which condense the evaporated moisture into either a tray that you empty after, or directly into a drain.

So we need to add space for a stacked washer/dryer. We'll also need a coat closet, and ideally a linen closet.

The office/flex room can house the washer/dryer, and the coat and linen closet can take up the flex room's area, as it's conveniently positioned by the door.

Next, I decided on the bathroom dimensions. The layout of sink/toilet/bathtub I chose is ideal both for maximizing space and because it keeps all the plumbing connections on a single wall, reducing cost. In fact, this is one of the reasons I chose to put the washing machine where it is: It's on the same wall, allowing for easy plumbing connection.

The width of the bathroom is simply the length of the bathtub — 1.60m. The depth of the bathroom is the combined widths of the bathtub, toilet, and vanity, plus a little bit of space to be used for a garbage can. This small gap at the other side of the vanity will also make the bathroom feel more spacious.

Then I designed the office/flex room layout, which was fairly straightforward given the need to integrate both closets and the washer. Because the rest of the apartment is open concept, there was very little to design in terms of the floorplan — it can be furnished how residents want it. The last step was making the kitchen counters, which would include the oven, a dishwasher, and the sink. You might be wondering about where dry food is to be stored because there's no pantry closet, but there will be ample cupboard space below the whole L-shaped counter and above the counter on the one wall. Additionally, there is space to add an island between the kitchen and living room, which could add even more storage space. I've chosen not to add such an island because it would increase costs unnecessarily.

1 Bedroom modifications

Since the 1-bedroom is narrower, there's no space for an office/flex room.

3 Bedroom modifications

For a 3 bedroom apartment, which could easily occupy 5 people, there should be an additional half bathroom with an extra toilet and sink. Notice in the design that both bathrooms and the spot for the washing machine are all on the same wall, again to keep down plumbing costs. The extra bedroom on the exterior wall makes the unit wider, allowing for a both a flex room and an office. Because of the bathroom placement with the doors partly inside the flex room, I decided to make it somewhat of an open room with no door.

Here will be the layout from one end to the other. The goal is to make the space family friendly and to facilitate community. Play, eat, gather.

The courtyard has dimensions 75m x 16m.

Playground Area (16m x 19m): Mulch with a play structure and benches on the side for caregivers

Low planting buffer between play area and BBQ.

BBQ + Picnic Zone (16m x 19m): 4 gas BBQ grills with many picnic tables around. Overhead lighting for evening use

Multi-use Lawn / Event Space (16m x 37m): Grass area. Families can use it for outdoor games, and it can be used for community events like markets, movie nights, live entertainment, or yoga/fitness classes.

I'll make a plan for a laser-cut architectural model. Unfortunately this will remain a plan because I wasn't able to laser-cut the prototype. I've instead created a rendering with AI. It conveys the concept well, although the scale isn't accurate.

I'd create it in 9 panels of 2mm thick wood.

1. 4 side wall panels facing the exterior of the building
2. 4 side wall panels facing the interior of the building
3. 1 panel for the roof

The side panels will have small holes cut for windows, and laser engraved lines to mimic the design that would be painted on the corrugated aluminum.

Once created, I'd super-glue these panels together to make the building, and then glue them onto a wood base.

For the parking garage, I'd follow a similar process. On the roof, I'd engrave lines to denote parking spots. The side walls would have long holes to mimic the ventilation openings in the parking garage facade.

In this project, I've presented a design for a 650 unit, 12 story mass timber apartment complex. My goal throughout was to create good quality housing as affordably as possible. I ensured the building was to code, and I made tradeoffs that increase initial construction cost but provide more long term value. For example, I could have opted for the bare minimum number of parking spots, which would have lower total construction costs, and probably would have resulted in slightly lower construction costs per unit. But the value of the additional units and retail space is higher than construction costs, and so the decision to go with the car elevator parking garage design with a smaller foot print but more levels was logical. I also integrated cost savings measures that may slightly reduce quality, but that I deemed reasonable for affordable housing. For example, individual HVAC units for each apartment with minimal and exposed ductwork inside. Exposed ductwork is simply aesthetic and doesn't decrease the functional value of the apartment. With proper styling where the industrial look feels intentional, this probably wouldn't decrease the overall value of apartment units.

This disconnect between value creation and cost of construction is where the economics of affordable housing become complicated.

Construction cost ≠ Value created

This is how real estate developers think. The value created is higher, and that's their profit. You can implement all these cost savings measures like exposed ductwork, but it remains questionable whether it actually makes housing more affordable to consumers. Exposed ductwork decreases construction costs, but doesn't really decrease the value of the apartments, at least not by much. From a developers perspective, that just means more profit, because the apartments will still be sold/rented at market price. Affordable housing isn't only an issue of construction costs, it's an issue of supply and demand. Demand for housing is very high in cities, which drives up prices, and building housing with cheaper construction methods does little to bring down prices. It would allow developers to sell units more cheaply if they wanted to, but realistically they'll sell them at market price.

Then who builds it?

One obvious option is that the government builds it and offers it at cost, or as subsidized housing. Without seeking profit, units would be much cheaper. The cost savings measures truly would be passed along to consumers. But governments are known for being less efficient than the private sector, so there would likely be more cost overruns and the government may need to sell units at a loss for them to be any cheaper than market price.

Another more interesting approach is a housing cooperative, where residents come together to build while cutting out developers. The goal is no longer profit, but affordable housing. This eliminates the disconnect between construction cost and the cost of housing. Although not a common development approach, this has been done before. Most recent examples include Baugruppe in Germany, which has build a few housing complexes this way. There are also many earlier examples in the US from the first half of the 20th century, mostly initiated by labor unions.

The first challenge with housing cooperatives is finding someone to start it. You'd need a few ambitious people to come together and get the project started by creating a general building plan, advertising it and getting some new members on board to finally form a legal entity with a board of directors. Then they could start getting financing and securing the land. The other option would be for a union to start the project. The union would help set up the cooperative and secure the land. This is probably more likely than a few ambitious people starting this on their own. The main disadvantage is that the building would have ties to the union, which may overly prioritize union members over non-union residents of the co-op.

Funding is the other major challenge with this kind of format, because banks have lower trust in cooperatives than well-established developers. Future residents would likely need to invest some of their own money, potentially partly from loans. Government loans at favorable rates may also be possible, but likely require lobbying. This can be supplemented by issuing bonds, which can be bought by members of the community (and anyone around the world). Crowdfunding is becoming a more popular and viable option with social media. Members of the community are more likely to support a cooperative because there's a certain trust created through the fact that it's local, and the amount lent by each person is much less. This is a particularly effective solution during times of low interest rates, where people are more likely to be enticed by rates that are higher than general mortgage rates, but still affordable to the cooperative. In the case of a union backing the cooperative, bonds for the building could be sold to the union pension fund.

The mass timber design works well with this model because it speeds up construction, shortening the amount of time funds need to be borrowed.

Once the building is finished, residents can buy their condominiums with regular mortgages, funding the bonds' maturity. For people who opt to rent, the cooperative's entity could purchase and mortgage those units, using the rent to pay the mortgage.

Retail space is the one area where the co-op would seek a profit. These funds would be used to cover maintenance fees, reducing what residents who own units would have to pay. If rent from the retail space was higher than maintenance costs (which is unlikely), apartment owners could benefit from profits.

Here are some of the advantages of a housing co-op:

1. Long-term affordability
2. Strong community and social cohesion
3. Protection from speculative real estate pressures

One of the main disadvantages would be union pressures in the case that it's union backed. The building is unlikely to be occupied by only union members, and the cooperative's entity and control would need to be structured carefully to ensure that all residents are treated equally. The union would need some incentive to lead such a project, so I would propose that the union members get first opportunity to buy/rent units, but once the building is done, the cooperative become entirely independent from the union.

This project presents a design that provides good quality housing while making construction more affordable. Ultimately, the affordability to residents depends on how the building is developed. I think the best way would be a cooperative approach.

I hope you liked this project - I certainly learned lots about mass timber construction, and construction in general. I especially enjoyed planning the layouts of the parking garage and the mass timber grid to be as efficient as possible. I also liked how I was able to use rough calculations to come to clear decisions.

Thanks for reading!