A bow saw cuts firewood faster than a folding pruning saw, with less effort, and anyone who has used both in the same afternoon knows it within about three strokes. The frame holds the blade in tension, so the blade can be thin. The thinner the blade, the thinner the kerf, and the less material removed per stroke, which makes for an easier cut. Yet the saw most people actually carry is the folding one, because it is more compact and more convenient to pack.
This is a dimensioned build for a bucksaw that collapses into a bundle the length of its own blade, sized around a standard 610 mm (24 inch) blade. It is aimed at anyone with a saw, a chisel and a drill.
Tension is the whole point
A panel saw resists buckling by being thick. Push a thin steel strip through a cut and it folds sideways on the return stroke, so a saw meant to be used unsupported needs a plate stiff enough to hold itself straight. That stiffness has to be paid for with metal, and every extra thousandth of plate thickness is more wood turned into dust on every stroke.
A bow saw cheats. Stretch a thin blade between two points hard enough and it cannot buckle, because it is never in compression. All the straightness comes from the frame. That is why a bow saw blade can be 0.7 mm thick when a comparable hand saw plate is nearer 0.9 mm, and why the same log takes noticeably fewer strokes.
So the frame has exactly one job: pull the two ends of the blade apart and keep pulling while you work. Everything else in the design is in service of that.
The mechanism is a windlass. A cord runs between the tops of the two arms. Push a stick through the cord, twist it, and the cord shortens as it winds on itself. The arms pivot about the stretcher between them, so pulling their tops together drives their bottoms apart, and the blade pinned across those bottoms comes taut. The stretcher is the fulcrum. The whole frame is a pair of levers sharing a beam.
Making it fold without losing anything
The failure mode of a takedown saw is not breakage. It is a small metal part in leaf litter at dusk. Wing nuts, cotter pins and separate blade pins all work beautifully on a bench and all eventually get dropped in a wood.
So the rule for this build is that nothing on the saw comes off in the field. The blade attaches with a captive pin: a 4 mm pin driven permanently through each arm near its toe, spanning a slot cut down the middle of the arm. The blade eyelet drops over the pin from inside that slot and stays there for the life of the blade. Slacken the cord and the blade goes loose but does not detach. There is no fastener to undo, nothing to unthread, and it works in gloves.
Which means the blade and the two arms are permanently one assembly. To pack it, the arms have to fold with the blade still attached, and the only thing they can rotate about is the pin the blade already sits on.
Swing each arm ninety degrees about its own pin and it lies flat along the blade. One arm covers the left half, the other covers the right. The blade ends up sheathed inside the slots in its own arms, which is exactly where you want a set of exposed teeth to be.
The dimension you do not get to choose
Here is the part that reorganises the rest of the build.
If both arms fold down onto the blade and lie end to end, then together they can be no longer than the blade span. Each arm gets half. A 610 mm blade has roughly 597 mm between eyelet centres, so each arm can be about 300 mm above its pin and no more. Go longer and the two arms collide in the middle before either has finished folding.
The arm length is therefore fixed by the blade, not chosen by you. And the arm length is what the stretcher position is measured against, and the stretcher position is what sets the throat, the clear gap between blade and stretcher that decides the fattest log the saw will pass through.
You do not pick the throat depth of a folding bucksaw. The blade picked it, at the moment you decided the saw should fold.
For comparison, the classic non-folding H-frame uses 400 mm arms with the stretcher at mid height, and gets around 200 mm of throat from the same blade. Folding costs roughly a third of that. Worth knowing before you cut anything, because if you mostly saw thick rounds, the right answer is a saw that does not fold.
Throat against leverage
Given a 315 mm arm, the only remaining freedom is where the stretcher sits on it, and that is a straight trade.
Put the stretcher at mid height and the lever arms above and below the fulcrum are about equal, so the cord carries roughly the same tension as the blade. Comfortable to tension, but the throat comes out around 115 mm.
Slide the stretcher up to about sixty per cent of the arm height and the throat grows to 146 mm, which is the difference between passing a 115 mm round and a 145 mm round. The cost is leverage: the cord now sits 111 mm above the fulcrum while the blade hangs 169 mm below it, so the cord has to carry about one and a half times the blade tension for the same result. In practice that means a couple more turns on the toggle and a slightly beefier cord.
This build takes the throat.
Where the slot and the stretcher fight
One detail nearly sinks the design, and it is worth understanding rather than discovering.
The blade slot has to run the full length of the arm, because the folded blade needs somewhere to go along its whole half. It also has to be deep, because the blade is 25 mm wide, so the slot goes 35 mm into a 45 mm arm and opens on the inner face. That is the same face the stretcher wants to seat against. A conventional mortise cut there would break straight into the slot.
The fix is to stop treating it as a mortise. Cut a pocket instead: 20 mm along the arm, 10 mm deep, 19 mm across, opening on the inner face. Because the slot is only 1.5 mm wide and the pocket is far wider, the pocket does not cross the slot, it simply widens its mouth for 20 mm. The stretcher end drops in, tension holds it there, and the folded blade passes through the same space with room to spare. Leaving 3 mm of wood on each side of the pocket stops the stretcher wandering sideways.
Two things about that slot are worth stating because they look alarming and are not.
It costs almost nothing in strength. The arm bends in the plane of the frame, and a 1.5 mm cut straddling the neutral axis removes about three and a half per cent of the second moment of area. The arm is essentially as stiff as it was.
But it must stay full section at the pocket. The pocket sits at the fulcrum, which is where the bending moment peaks. Taper the arms above and below it if you want to lose weight, and leave the wood alone within about 40 mm either side.
Cut list
Ash is the right wood: tough, springy, and it takes the repeated tension cycles without setting. Beech and hickory both work. Avoid softwood, which crushes at the pocket.
| Part | Qty | Length | Width | Thickness |
|---|---|---|---|---|
| Arm | 2 (mirror pair) | 315 mm | 45 mm | 25 mm |
| Stretcher | 1 | 572 mm | 38 mm | 19 mm |
| Toggle | 1 | 200 mm | 19 mm | 6 mm |
| Pin | 2 | 25 mm | 4 mm diameter | - |
Plus a 610 mm bow saw blade and about 1.4 m of 550 paracord.
The arms are a mirror pair, not two identical parts. The slot opens on the inner face of each, so they are handed. Mark them before you cut.
Measure your own blade before committing to the stretcher length. Eyelet spacing varies between makers and 597 mm is typical, not universal. Better still, cut the stretcher last and fit it to the frame you actually built.
The stretcher looks slender for the load. It is not: at 572 mm between shoulders, 38 by 19 mm in ash, Euler buckling comes out around 7.9 kN assuming pinned ends and a modulus near 12 GPa. That is an order of magnitude check rather than a certification, but it is a comfortable order of magnitude.
Building it
Shape everything before the pins go in. Once a pin is driven the blade is captive, and every subsequent operation has a sharp 610 mm blade swinging off the workpiece.
- Dress both arm blanks to 315 by 45 by 25 mm. Mark which face is inner on each, remembering they are handed.
- Rip the blade slot down the inner face of each arm, 1.5 mm wide and 35 mm deep, running the full length. A bandsaw or a thin-plate rip saw both work. Test that the blade slides in freely before going further.
- Chisel the stretcher pocket in the inner face, 20 mm along the arm and 10 mm deep, centred 189 mm up from the toe, leaving 3 mm of wood on each cheek.
- Round the arm tops and cut a shallow notch on each outer edge, about 15 mm down from the top, to stop the cord riding off.
- Shape the lower ends into handles. Chamfer everything. Sand now, because you will not get clean access later.
- Drill the 4 mm pin holes 20 mm up from each toe, square through both cheeks.
- Drop the blade eyelets into the slots, line them up with the holes, and drive the pins. Peen or epoxy the ends flush.
- Assemble the frame and cut the stretcher to the length that gives you a light interference fit in both pockets.
- Fit the cord as a loop of four strands between the arm tops, above the notches, and slip the toggle through it.
- Oil the wood. Boiled linseed or a hardwax oil, several thin coats.
Tensioning and packing
Wind the toggle four to six turns and stop when a flick of the blade gives a clear note rather than a dull slap. Rest the toggle end against the stretcher so it cannot unwind.
Slacken it when you finish. Cord left wound overnight creeps, and wood left under load takes a set, so a saw stored tensioned is a saw that will not tension properly next time.
To pack: unwind fully, lift the stretcher out of its pockets, swing both arms down onto the blade, and lash the stretcher and toggle alongside with the cord. The bundle comes out 637 mm long and about 45 by 50 mm in section, which lashes to the outside of a pack and will not go inside one. That is the honest cost of a 610 mm blade, and it is the same for every folding design.
What it will not do
The throat is 146 mm. Roll a bigger log and cut from two sides and you will manage 290 mm, but slowly and with a wandering kerf.
Seven millimetres of teeth stay exposed at the middle of the folded bundle where the two arms meet. It is not enough to cut a pack, but it is enough to find a thumb. A scrap of leather over the join solves it.
And replacing the blade means punching out a pin, which is a workshop job rather than a field one. That is the deliberate trade behind never losing a fastener in the dark. Carry a spare cord, because that is the part that actually fails, and leave the spare blade at home.