[jimnotgym]

There is another important factor. Wood expands and contracts according to the moisture content of its environment and screws expand and contract according to temperature. The combination of these means that screws work themselves loose over time, often wearing the threads out of the wood.

In the UK, stair handrails (for instance) used to be mortice and tennoned with a drawbored peg. This is where an offset holde is drilled through the tennon and a peg driven in to pull it tight. A fashion for cheap construction has seen this replaced by screws in many cases. The problem is that uk houses tend to have wet plaster (or at least a skim on drywall) and high moisture levels at the time of installation. This means that the wood will shrink once installed... and your handrails come loose, the screws are hidden by this time!

In a very damp situation, say a wooden gate, this gets a cyclical drying and wetting, yet properly made oak gates can last many years due to their pegged tennon joints. Edit: clarity

[exDM69]

In addition to wood movement, there's another important factor when it comes to hobbyist woodworkers: accuracy.

With modern adhesives and precision machining like Mathias' work will make mighty strong joints even out of the simplest joinery method. A precision made box joint glued with wood glue is stronger than a finely made dovetail joint.

But when it comes to me bodging away in my garage with hand tools, that kind of precision is not going to happen. But historical "strong" joints like dovetails and mortises and tenons with wedges and drawbores are very forgiving joints and are quite strong even when they're not made to sub-millimeter accuracy. I can join a decent box with dovetails in about 30 minutes, put it together with a few drops hide glue and fill the gaps with sawdust and it'll work great.

But someone with a table saw and a simple box joint jig could join 10 boxes in the same 30 minutes it took me to make one.

[jdietrich]

Modern hand tool woodworking can match or exceed the accuracy of machine woodworking, albeit with a higher skill threshold. You can achieve very fine results with very basic tools. Machine woodworking is undoubtedly faster, however.

Tight tolerances are often completely counterproductive because of wood movement. Even with an impermeable polyurethane finish, you'll get significant seasonal movement. Effective fine woodworking is about designing to account for that movement - building drawer runners slightly undersized so they don't bind up, allowing door panels to float in the groove so they don't bow or split etc. One of the key skills of cabinet-making is developing an intuition for wood movement.

https://www.canadianwoodworking.com/tipstechniques/dealing-w...

[nemo44x]

What about torrefying the wood first? Wouldn’t this reduce the amount of movement in the wood?

[exDM69]

You can't really stop wood movement or even reduce the amount it moves. It can be slowed down a little with some finishes (then you have a "wood-plastic composite"). Kiln drying (do you mean this by "torrefying"?) will make the wood dry for a while, but when the temperature and (absolute) humidity will rise on the summer, the wood will absorb moisture from the air and expand and shrink again next winter.

[jdietrich]

Torrefaction (more properly, thermal modification) involves higher temperatures than kiln drying, causing chemical changes to the wood that result in permanently lower equilibrium moisture content and hygroscopicity. It improves the stability of wood and drastically improves the resistance of softwoods to decay, but it isn't a complete solution to wood movement. There is a significant tradeoff between stability and strength.

https://en.wikipedia.org/wiki/Thermally_modified_wood

[abakker]

It also changes the color, which in many cases is inappropriate for the project. However, if it were being painted I would definitely consider it. I bought some Torrefied ash recently - it's weird stuff. Kind of like working with toast instead of bread. the sawdust smells odd, too.

[kazinator]

> A precision made box joint glued with wood glue is stronger than a finely made dovetail joint.

To clarify, do you mean "than a finely made dovetail joint without glue"?

[exDM69]

No, I do not. According to the strength tests I've seen (and my experience with joinery reflects this), there is hardly any difference in strength between a box joint and a dovetail when done with equal precision and glued with modern glue, given that they have roughly equal amount of glue surface area. But a box joint is much easier to cut with precision on a table saw with a simple jig.

Dovetails aren't usually used without glue (or wedges). A tightly cut dovetail will stay together just fine without glue, but it has relatively little resistance against shear stress. A small bump can turn a square box into a trapezoid even with good joinery.

When I do dovetails, I use a small drop of glue on the long grain of the pins and tails. That's enough to keep it from trapezoiding and allows me to plane the joint flush.

[jimnotgym]

Plus the dovetail, mortice and tennon with wedges etc are still strong even after wood shrinkage has cracked the glue line. A simple glued joint just falls apart.

[jabart]

This is similar to current nails versus cut nails. Round nails help for sheer force, but after a while are easy to pull out. Cut nails are shaped different and work much better. I built two blanket chests using cut nails to hold it together, glue only to make a panel wide enough. Five years later it still is doing great, even with 5+ kids using it as a toy box. Screws would have worked themselves out by now.

Also built a workbench with pegged tenons, guess it weighs 100+ lbs, and if you bump it with enough force the whole things moves, but the tenons are still not loose.

[MatthiasWandel]

Yes, for outdoor applications, the glue usually doesn't hold, though sometimes, for some joints, if its tight enough, the glue can keep the moisture out. See here: https://woodgears.ca/deck/repair.html

[sandworm101]

>> The combination of these means that screws work themselves loose over time, often wearing the threads out of the wood.

There are also dramatic differences between woods. Even if you use the same wood from the same tree, glues can impact moisture absorption rates. The only totally stable joints are those that don't use either fasteners or glues, the sort of stuff used to build Japanese temples, but these are complex and bulky. There is no perfect joint short of growing a tree in the shape you want.

[aivisol]

Moisture expansion cycle creates incredible forces, especially when the construction is used outdoors with large seasonal temperature/humidity changes. Even a wood as soft as pine can easily rip of heads of the screws after a couple of years of use. (personal experience)

[newnewpdro]

I find this interesting and somewhat in conflict with my experience.

I've been doing restoration work on an old home in the desert, and many of the nails in wood exposed to the elements had become quite loose and worked their way out enough to just grab and pull the rest of the way out without any tools.

But the nails with with a screw pattern on them which twist as they're hammered in have done much better than the smooth ones. None of these have backed out, and they're all still very firmly set, in the same wood, in the same environment, for the same duration. The builder had used these screw-style nails for hangers, but smooth nails for everything else.

I've been using SS deck screws of various length and gauge to do all my restoration work, in part because the reused wood is so old and brittle it's impossible to put new nails into it without splitting, and based on the screw-nails doing so much better than the smooth ones, I assumed the deck screws would be similarly more lasting.

If what you're saying is accurate, I would have expected the screw-nails to have backed themselves out. But it was the opposite situation.

My interpretation of what happens is the wood swells and contracts from both the temp and moisture changes. The metal fasteners don't change dimension much at all in these relatively insignificant (for metal) temp differences. When the wood swells up the first time, it compresses itself around the firmly held nail, and the nail doesn't budge. When the wood shrinks back down, the compressed wood around the nail stays compressed and now the bore is bigger than before. This cycle repeats itself over the years, and since the nail has a pointy tip, every time the wood swells up and closes in on the nail, it nudges the nail out a little bit before tightening until things shrink again.

Based on this model, I think the screw is superior because the depth of the threads is greater than the dimensional difference between swollen and shrunken. The wood is free to go about its seasonal cycles, and the screw doesn't move because the threads are always engaged enough to resist that little pressing force at the pointy tip.

I'm a noob when it comes to all this, but that's my impression and only time will tell if I've made a horrible mistake using screws for all these joints. Though it's not like I had many options.