Work platform

It's almost time to put up the roof framing. The rafter units are not too heavy to carry, but they're not light, either. And they're long (16'). And we have to put them on top of an 8' wall, which is a long way to reach.

We passed one up today to see how it felt.

I carried it from the storage area to the building, and passed it through a gap in the wall to Julie. Her job was to set one rafter tail on one wall, slide it up, then hook the other rafter tail on the other wall, then center it. She couldn't reach high enough, so I had to run around the building, come in through the door, and help her get it in to place. Too much running back and forth to get all rafter units up like that, and it was hard work. How to make it easier?

• Get a couple helpers for 1/2 an hour.
• Carry the rafter units to the building, as far as I can get them. Then run around, and together we drag them in & hoist them up.
• Rent scaffolding.
• Build scaffolding.

A combination of the above may be the best choice. I decided to build a raised work platform, as a start.

I mostly improvised. The one cool thing I did was to create 4 short, temporary jack studs to attach to existing studs. The building is 10' across, so a pair of extra 2x4 10' sticks rest on these jack studs. 

There are few more sticks here and there to finish the structure. The building, which we've worked hard to make strong and stable, helps the platform be strong and stable, too.

A sheet of plywood rests on top as a nice, large surface to stand on. It's about 3' up, mostly because I had some 3' cut-offs lying around already. Maybe 4' would have been better - 1/2 the height of the walls? Whatever, it works.

Most of the lumber is used full-length.  The plywood plywood deck is only 6' long, but I didn't cut it - it was like that already. Screws hold things together, making it easy to take apart, but the screws carry minimal load  - loads mostly go through wood resting on wood.

Top plates & porch beam

The porch beam holds up the roof over the porch. Over the enclosed building, the studs hold up the rafters, but around the porch, the roof load is carried along porch beams to supports at the end. Here's the plan, with the beam pockets marked.

And the beam pocket as seen in my last post:

First I climbed up the ladder to cut back the sheathing around the beam pockets. I used my grand-father-in-law's jigsaw, since it fit in that small space and was light enough to use up on the ladder.

Hoist the beam in position. One end in the pocket, one end clamped to the porch post. Level & tack in place with a deck screw (3 1/2" #10) at each end.

One end of the beam rests on a jack stud in the wall, which is pretty secure. The other end is next to (not on top of) a porch post. How to carry the load securely in to the post? One option is lag screws or through-bolting, which works pretty well but is bit tedious up on a ladder, with all the predrilling, etc.

I decided to add a "jack stud" to the side of the post. It is attached to the side of the post, running from the underside of the porch beam to the deck beam below. A series of screws through the jack stud in to the post help stabilize things. 

We plumbed the posts (in 2 dimensions) and braced them to each other (at the proper spacing) and diagonally to the deck. Then some 1/4" timber screws through the beam in to post. The timber screws will help with uplift, if the wind tries to steal my roof.

Once the beams were up, we added the double top plates all the way around, including over the porch beam. That ties the porch beam in to the main wall nicely. I'll be adding some tie plates at the joints, as an extra measure.

Fixing a window framing mistake

When framing a window under a load-bearing wall, you need to put in a header to carry the roof load. Typically it's built out of 2x lumber on edge, with plywood sandwiched between to make it the same thickness as the 2x4 wall. (2x6 walls are done a little differently.)

I screwed up and put the header in sideways, with the lumber laying flat. This is far, far weaker.

I tried taking it apart, but it would have meant tearing up the king stud to pull the nails out.

Will it still be strong enough? Here's the math:

The roof design load is 10psf dead load + 25psf snow load = 35psf total. Rafters/studs are 2' O.C., and rafters are 8' long (I designed to use them full length.) Tributary area is 2' x 8' = 16 sq. ft. Design load = 16 sq. ft. x 35 psf = 560 lbs.

The weight limit for a span is normally Fb * b * d * d / (9 * L), where:


Fb = Fiber bending stress of the wood you're using. My 2x4s are Douglas Fir "#1 or BTR". According to WWPA, that wood has a fiber bending stress of 1200. (I have no idea what unit that is!)

b = breadth of the beam, in inches. A 2x4 laid flat (oops) is 3.5" wide.

d = depth of the beam, in inches. A 2x4 laid flat is 1.5" deep. Notice it's squared - this is why lumber on edge is so much stronger.

L = length of the span, in feet. My window rough opening is 2'.

Put it all together: 1200 * 3.5 * 1.5 * 1.5 / (9 * 2) = 525lbs. Not enough.

That formula is for an evenly distributed load across the span, which assumes that the strain on the wood varies linearly with the distance from the point of support. To put it another way, it behaves like a point load on the center of the beam that's 1/2 as large. So, divide my max load by 2.

However, I have 2 pieces of wood here (twice as strong), so multiply my max load by 2. So we're back to 525 lbs. Still not enough.

There's one more factor to consider: there's a double top plate that can carry some of that load, too. If you assume it's just strong as the misplaced header, then I'm fine, but I've never seen that described before, so I was wary of counting on it.

However, we're close. A little more wood could do the trick. And I had the perfect scrap of wood - a piece that was intended to be a window header, but it was 1/4" too short. If I lay it flat across the top, it increases the strength of the header by 50% more - enough to be certain we're strong enough. And I can end-nail the cripple on to this extra header *before* I install it, avoiding the hassle of toe-nailing.

That went together really nicely.

As I've been learning to build, I have asked myself the question: what is the difference between me (a novice) and an experienced pro? There are several obvious answers:

• Pros cost money, I just cost time.
• I'm way slower, because I spend a lot of time figuring out *how* to do this, and because I make mistakes I have to recover from. Also, I get tired because I'm not used to the work.
• I waste material due to mistakes.
• Pros are good at this, and know how to produce high-quality work.
• I care about the outcome far beyond my concerns for reputation and callbacks, so I'll take care when it matters to me.
• Pros cut corners to save money & time (which is also money to them) if they think they can get away with it.

But there's one more point that I have become aware of: pros are really good at hiding their mistakes. A novice can do work that's just good (or just as bad) as a pro, but a pro will make it look like they did a good job, even if they didn't.

Erecting walls

To anyone who has done platform framing before, building walls is the most obvious of activities. For novices, there is a lot of complexity to work through, a lot of learning to be done. I spent most of my time reading, drawing, pondering, conversating. Oh, and resting my hammering arm.

We could frame with either 2x4 or 2x6. 2x6s costs a little more, the lumber is usually a little better, the wall is stronger, and there's more space for in-wall insulation. Stick-built houses around here are usually built with 2x6s, but spaced 24" O.C. (instead of 16" O.C. for 2x4s), to use less wood. That's fine, but then you need stronger sheathing and drywall to span the larger gaps, so the savings are reduced.

Because our building is so small (12' x 10'), losing 4" in each dimension to thicker walls seemed like a lot to lose. Also, heavy insulation isn't important with the use patterns we're expecting in this building. So, we went with 2x4s.

Even though stud spacing for 2x4s is typically 16" in a house, in our little building I chose 24". The main reason is that the 10' dimensions don't work out well with 16" stud spacing. You end up with a couple studs really close together. If I spaced the rafters at 16" to match, I'd have 2 rafters really close together, too. Sheathing ends need to meet right on studs/rafters, and irregularly spaced studs/rafters make that a bit messy. We saw that on the floor, where the middle section of subfloor had to be 6' and 4', not 2' and 8', which would have let me use a full sheet.

It's not a bad idea to make studs and rafters line up. Then the roof load is transferred straight in to the studs, without relying on the top plates to carry that load. In fact, this is a key element of Advanced Framing, allowing a single top plate instead of the traditional double top plate. I'm keeping the double top plate, because it makes joining the walls at the corners easier (as a novice, I need easy).

The building is only a single story, so the wall can be a little weaker, than a 2-story building. So, we picked 2x4s 24" O.C.

When I bought the lumber to frame the walls, I ordered pre-cut studs. This way they're all the same length - I don't have to count on my ability to cut to length reliably.

My first time framing a window:

Once framed, we tilted up the wall to see how it felt.

Being so big means I have plenty of muscle. Picking up the wall frame was easy. Easy enough that we decided to sheath the wall while it's flat on the ground. That forced some decisions about the sheathing at the gable end and letting the sheathing hang down to tie in to the floor system. To work that out, we had to think about the overhang at the gable end, and I decided we needed lookouts.

Julie really enjoyed the chiseling work.

Once that was taken care of, we proceeded to sheath up the wall. I used 10d galvanized box nails to fasten the sheathing. What a pain in the ass. They bend easily. A knot, or a bad hammer hit, and they'd buckle over. The galvanization meant they stuck in the wood really well, and were hard to pull out. The heads usually folded or pulled off in the process, so there was nothing to grab on to. Often I just flattened the nail in to the sheathing. It looks sloppy, but I had little choice.

After the first piece of sheathing was done, I stepped back to look at the work, and saw that the wall wasn't squared. Arggg! I forgot to square before sheathing! I spent a day figuring out what to do. Keep it as-is? Remove 50 difficult nails? Cut the lumber to square? Cut the wall apart and rebuild? After careful study, I decided that it was close enough to be workable, and continued. I did trim the sheathing on one side so the next piece of sheathing could fit properly.

Finally, the sheathing was done and I picked up the wall. It was very heavy, but I still was able to pick it up alone.

Once up, Julie held it steady while I shifted it back and forth until it was in just the right spot (in 2 dimensions) and plum (measured at both ends) and braced.

Bill helped me tilt up, align, and stabilize the 2nd wall, which will get the door.

The remaining walls continued in a similar vein. I got faster each time. I built the 4th wall and tilted it up in 1 day, which is like 8x slower than a pro. I'm improving!

The 3rd & 4th walls only got 1 piece of sheathing before going up, because the remaining sheathing has to go on the outside of the 1st & 2nd walls, too. Also, we'll use the gap to pass up the rafter units.

The next step is porch beams, then the roof. Here's the pocket where they'll rest on the wall (the plywood needs to be cut away) and the beams themselves. The beams are from trees on our site, full-dimensioned 2x6.

Rafter units

Once the floor was built, you'd think I'd build walls. No. Rafters first!

I started by drawing the roof lines on the deck. Then I used them to mark & cut 1 rafter, including a bird's mouth. The bird's mouth cut didn't go well, and getting them right makes a big difference in how the roof goes together. I decided on a different approach.

It's common to buy pre-made roof trusses, built in a factory. I could have done that, but I wanted to try doing it myself once. But I can do something similar: Rafter units, made of 2 rafters and a ceiling joist.

I built a template on the deck to align all the components, for uniform rafters. There's a placeholder for the ridge board:

This blocking represents the wall, and aligns the ceiling joist:

There's another one on the other side:

Used 1 good rafter as a guide for the plumb cuts on all the others:

I also cut some strips of plywood for later use:

Nailing is loud, especially on a deck, and shards of metal can fly at high speed, so some protection is a good idea:

How to build a rafter unit:

1. Slide rafters in to position:

2. Screw them down to stop them from shifting during assembly:

3. Attach the tops of the rafters to keep them from shifting. The nails were just what I had on hand. Note the plywood isn't really structural (it doesn't need to carry a load), it just holds the rafter unit together until the roof is done. 

4. Place the ceiling joist. There's a little wiggle room, but make sure it doesn't stick out above the rafter.

5. Nail things together. My nails are 3 1/4" long, and the wood is only 3" thick. Driving the nails at this angle actually makes things stronger, while keeping the points from sticking out.

6. 5 nails at each end. Note that they're not *quite* all the way in, so that if I change my mind, I can get a cat's paw in to pull the nails without tearing up the wood.

Completed rafter units:

I built one extra, just in case, but I used screws so I can easily take it apart and repurpose the lumber.

Now, to the walls. 

Building the floor

After completing the deck, we turned to the floor of the building. Joists continued 16" O.C., but the building is 10' wide, which doesn't divide by 16". Instead, the last 2 joists are spaced 12" O.C.

One of the joists is also doubled. The lumber we bought was kinda crappy, so doubling the worst sticks was a way to compensate. The subfloor will have a seam there, so it adds more nailing surface, which we wanted because we're such framing newbies. Also, we're expecting an extra load in that area, so we're glad to have the extra strength. With better lumber and without the expected point load, I could have skipped the penultimate joist and left a 24" space instead.

Originally, we planned to lay the 4'x8' sheets of plywood subfloor along the joists. With that in mind, we started to install blocking to support the seams of the plywood. The 8' edges are tongue-and-groove, but the 4' edges are not, so they need the help. It's a little stronger to run the plywood the other way, so we changed  the plans to lay the plywood across the joists. Still, we installed the blocking anyway, to help stabilize the 12' joists.

Here's the plan:

Once the floor system was framed up, we added insulation (before the subfloor). We chose 2" polyisocyanurate foam panels, in part because that's exactly what is under the yurt, so it's familiar. Also, rodents won't nest in it, like they might in fiberglass, and when I work in the crawlspace I won't be breathing fiberglass. We have a lot of wooden stakes (Buffy would be proud) so we attached those to the sides of the joists, 2" below the top, to support the foam. Then we cut the foam to fit each joist bay. A few cans of Great Stuff expanding foam insulation filled in the gaps.

Looking back, I wish I had set the insulation a little lower, by maybe 1/4" or 1/2", so it wouldn't have a chance to interfere with the subfloor if it stuck out somewhere.  Also, I could have cut the panels 1/2" smaller and filled in with Great Stuff for a tighter seal than trying to make panels fit tightly, and ending up with a gap that's too small to get the foam in to. Still, I think it will be fine. After the Great Stuff cured I went back with a saw and cut off the extra.

 I bought 4 sheets of that insulation at $50 each + unknown environmental impact. We have most of a sheet left over, plus a bunch of scraps, and I've been wondering how I might have saved a sheet. You see, each sheet is 4x8 = 32 sq. ft., and the insulated floor is 10x12 = 120 sq. ft., so you might think I'd use 3.75 sheets. But the framing takes up a lot of space. Could 3 sheets be enough?

We could have waited wait until after the subfloor was in place before installing the insulation. Cut it to be 1/2" undersized, then glue it to the underside of the subfloor. Random-sized pieces can be fitted together to fill the bays. Great Stuff fills in gaps. You can even wait postpone this work until the building is complete, so you can see how the building works without insulation to decide how much to put in later.


Another alternative is to put the foam panels on the underside of the joists. You can use the sheets whole, with only a few cutouts for posts or whatever. The almost continuous insulation would be more effective. And you could fill the joist bays with batts or loose fill insulation, and get a super-insulated floor. That's out of place in this building, however.


Once the insulation was done, it was time to install the subfloor. I used 3/4" T&G Sturd-I-Floor, which is rated for 24" joist spacing, for extra stiffness. The yurt floor is 1 1/8" on 4" joist spacing, which is within spec but very springy. (If I could go back, I would install 2x4 joists under the yurt floor, across the existing joists. Even if they were only 24" O.C., it would be much stiffer.)

We applied subfloor adhesive to the joists, then dropped on a sheet of plywood, then wiggled it in to position. It's easy to move when it's floating on a bed of wet glue. When it was in position we started screwing it down. The board sheet wasn't completely flat, so I used my weight to press down the plywood as each screw went in. Here's the first board:

Because our joist spacing is a little inconsistent (16" O.C., except the last 2 that are 12" O.C.) we had to be careful about how we set out the plywood to get the edges right over a joist. The layout is 8' + 2', 4' + 6', and 8' + 2'.

Getting the tongues in to the groove was harder than I expected. We used a sledgehammer, but the insulation panels made it hard to get a good hit. It might have helped to spend some time with a water spray on one side and the sun on the other to take out the warp before installing.

Once all the subflooring was down, we went back and cut the panels again. They were a little long. I wanted them to be slightly short, so I could be sure they wouldn't interfere with the wall sheathing that connects to the joists. We measured carefully, cut, measured again, cut again. Like the old saying.

The resulting floor is really nice to walk on. Very firm and quite level, except one corner where a beam had a bend up at the end.

Looking back, I wish we had used a 3rd beam down the middle of the floor. We could have dropped the joist size down and gotten a stiffer floor. No joist hangers required. I am considering adding it later.

I know it's a good floor because it attracts beautiful women:

More subpanels!

I haven't done any building for 2 weeks, so I thought I'd post some more abstract thoughts today, instead.

Every house I've lived in had all the breakers in one place, a load center (aka breaker panel) right next to the service entrance (aka electric meter). All circuits in the house are home-run to this location. (One house had a small load center, and a small subpanel next to it, but otherwise it was the same.)

Lots of cables run right next to each other, from the load center to other parts of the house. The worst case is usually the kitchen. You may have a separate circuit for each of:

• fridge
• appliance receps (often two circuits or an MWBC)
• disposer
• dishwasher
• range
• wall clock

That's a lot of copper all running to one place. 

As a long-time computer programmer, I'm used to breaking down bigger problems in to smaller ones, making things uniform/repeatable, etc. So I'm often tempted to apply this in other domains, mostly because it's familiar.

Instead of running a bundle of 12AWG cables to the kitchen, you could run one larger cable to a kitchen subpanel, and then fork off short local circuits from there.

You can use a lot less copper this way. For example, consider a typical installation of 6 circuits, 20A each. There's a hot, neutral, and ground - 12 conductors total. Even though total ampacity of all 6 circuits is 60A @ 240V, you don't need to feed that much to the subpanel, since the copper is shared across all the circuits. The subpanel could be 50A or less. And you only have 4 conductors, not 12. (Of course, if you guess too low, the subpanel's main breaker will trip, which is annoying. So don't take it too far.)

If you have an electric range, the subpanel circuit will be pretty hefty. You may decide to take the extra trouble to use an aluminum conductor, which is often much cheaper than copper. I usually avoid aluminum inside a dwelling because of increased fire risk, but in this case it may be worth it. Just take the time to make sure it's done right.

This approach will make motors (like a blender) last longer, and reduce power losses to heat. If you size the copper for the sum of the loads, but the loads are intermittent, each load will have a lot more headroom almost all of the time.

Another advantage is that adding a new circuit to your kitchen is much easier - the panel is right there already. Also, if you ever run out of room in that panel, it's easier to upgrade to a larger one, since you don't have to get the power company to disconnect/reconnect the main feeder - just shut off the upstream breaker. (This is also a reason I think the main disconnect for a service should be separate from the main load center. Put it outside in the same box as the meter.)

Also, if you overload a circuit, the breaker that trips is right there in the same room with you. You don't have to go far to reset it.

There are some disadvantages, of course. The subpanel itself and the upstream breaker cost money. Not much in the grand scheme of a house, but still. The electrician has a little more wiring work to do (but less work pulling cables from one place to another). Some people find a subpanel in a kitchen to be ugly (but not me!).

If you short a circuit, it could trip any of several breakers - the breaker on this circuit, the upstream breaker for the subpanel, or even the main breaker for the house. This is different from when you merely overload a circuit by plugging in too many appliances, say trying to pull 25A on a 20A circuit. A short is effectively pulling thousands of amps on the circuit, and whichever breaker happens to trip fastest is the one that you will have to reset.

Finally, it's weird. Most people like their homes to be conventional. Building trades are steeped in tradition.

While I'm at it, why not run a small subpanel to *every* room? A small load center is $15 or so, so the extra cost is not huge. If you're wiring 2 circuits to a room, using a subpanel saves you a couple conductors (hot-hot-neutral-ground instead of hot-neutral-ground x 2), much like using a MWBC (multi-wire branch circuit). Now, if you ever want a 3rd circuit in the room, it's easy to add.

This whole idea is derived from MWBCs. Some electricians swear by them (twice the capacity with just 30% more copper) while others abhor them (risk of 240V exposure if neutral is broken; bonded 2-pole breaker required). A subpanel to split the MWBC in to single-leg circuits, close to point of use, can work out really nicely.

Maybe electricians already thought of all this and don't put subpanels all over the place for good reason. I haven't heard that reason yet, but I'll keep my ears open.