We have cut on Western backsaws for fifteen years. A customer left a dozuki behind after a class and it has been on the bench ever since, so here is an honest account of six months with a saw that works the other way round.
0.3mm kerf. The line you knife is very nearly the line you get.
The kerf
This is the whole story. A dozuki cuts on the pull stroke, so the plate is in tension and can be far thinner than a plate that has to be stiff enough to push. A third of a millimetre against nearly a full millimetre changes what a joint feels like: there is simply less material to be wrong about.
What took adjusting
No weight on the cut. Every instinct from a Western saw is wrong — it wants to be drawn, not driven.
Starting is different. You lift into the first stroke rather than dropping into it.
The spine limits depth absolutely. On anything over about 60mm it is the wrong saw and there is no arguing with it.
Sharpening
You do not. The teeth are impulse-hardened, which is why they stay sharp so long and why a file will not touch them. When it is done you buy a new plate. That is a real cost and worth knowing before you start: a Western saw you can maintain forever is a different proposition from a consumable.
It has not replaced anything. It has taken the work where a thin kerf matters, and left the rest alone.
It lives on the bench now for dovetails and small tenons. The tenon saw still does everything over 50mm deep, and the pair of them together covers more ground than either did alone.
We bought the jig in March to see whether it would earn a place in a workshop that cuts dovetails by hand. Six months and roughly forty drawers later, the answer is a qualified yes — with a longer list of caveats than the marketing would suggest.
What it is
It is an aluminium template jig with a fixed 8mm comb, clamped to the bench and used with a 14mm dovetail cutter and a guide bush. Half-blind and through dovetails both come off the same template by flipping the comb. Setup is a matter of squaring the stock against the stop, tightening two cam levers, and taking a test cut in scrap.
The comb reversed for through dovetails. The stop on the left is the only setting that matters.
Where it pays
For drawer runs of more than about eight, the setup time is repaid. Below that, the hand tools are out and finished before the jig is squared up. The crossover is not a fixed number — it moves with the timber. In dry oak the jig wins earlier, because paring end grain by hand is slow and the router does not care. In soft pine it wins later, because hand-cut tails go quickly and the router tears the exit face unless you back it up carefully.
Drawers
By hand
With the jig
2
50 min
70 min
6
2h 30
2h 10
12
5h
3h 20
24
10h
5h 40
Timed on 12mm ash sides, single operator, setup included.
Where it does not
The spacing is fixed. Every joint comes out with the same rhythm, and there is no way to narrow a tail at the edge of a wide board.
It wants stock of consistent thickness. Anything more than about 0.3mm out and the shoulder line wanders.
Half-blind depth is set by the cutter projection, so a change of drawer front thickness means a fresh test cut.
It will not do anything at all on a curved or tapered side, which is most of what we build.
The tails it cuts are unmistakably machine tails — even spacing, uniform slope, a small radius in every corner where the cutter could not reach. That is a look, not a fault, but it is not what a customer paying for hand work expects to see. We have taken to using it for the carcass drawers on a kitchen, where nobody will ever pull one right out, and cutting the show drawers by hand.
A jig does not make you faster at dovetails. It makes you faster at the fortieth dovetail.
Setting it up
Note: the cutter projection below is measured from the base of the guide bush, not the router table. Getting this wrong is the single most common way to ruin a drawer front.
Through dovetails, 12mm stock — projection 12.5mm
Half-blind, 18mm front — projection 9mm
Half-blind, 22mm front — projection 11mm
Write the numbers on a piece of tape stuck to the jig. You will not remember them in six weeks, and the test cut that reminds you costs a piece of the drawer front you were about to use.
Verdict
It stays. It lives under the bench rather than on it, which is about right: a tool for the batch work that pays for the interesting work, rather than a replacement for knowing how to cut a joint. If you make one blanket chest a year, save your money and buy a better saw — unless you are a beginner who wants repeatable results while learning what a good dovetail should feel like. If you have a kitchen to fit out, it will pay for itself on the first job.
Sharpening is the skill that unlocks every other skill, and it is the one people put off longest. It is also much simpler than the volume of writing about it suggests. Here is what we actually do at the bench, on an ordinary Tuesday.
1000 grit, slurry raised, working the whole surface of the stone rather than one track.
Flatten the back once
A new chisel needs its back flattened, and it needs it done properly one time. An edge is where two flat surfaces meet; if the back is hollow the edge is a line of scratches no matter how well you hone the bevel. Do the first 30mm and no more — nothing beyond that ever touches the work.
Angles
Tool
Primary
Secondary
Bench chisel, softwood
25°
30°
Bench chisel, hardwood
25°
32°
Paring chisel
20°
none
Mortise chisel
25°
35°
A degree or two either way makes no practical difference. Consistency does.
The routine
1000 grit until a burr forms along the whole edge. Check by running a thumbnail up the back.
6000 grit, same bevel, until the burr flips.
Back flat on the 6000, four or five strokes, to take the burr off.
Strop if you like. It is not necessary, and it will round the edge if you push.
Ninety seconds a chisel once the stones are wet. The reason it feels slower than that is almost always that the edge was allowed to go too far past sharp before anyone did anything about it.
How sharp is sharp
The paper test tells you nothing useful. Stand the chisel on your thumbnail: if it bites without pressure it will cut end grain. If it skates, go back to the 1000. Looking at the edge under a light is better than any test — a dull edge reflects a thin bright line, a sharp one reflects nothing at all.
Sharpen when the tool stops being a pleasure to use, not when it stops working. By the time it stops working you have already spent an hour fighting it.
Keep the stones flat. A dished stone puts a curve on the bevel and undoes everything above. A cheap diamond plate and thirty seconds a week is the whole of stone maintenance.
A secondhand Bailey-pattern plane usually needs an afternoon before it earns its place. None of it is difficult, and most of the advice online has you doing far more than is necessary.
Everything off before anything is checked. Half of what looks like a sole problem is a frog problem.
Sole and Frog
Flatten the sole with the iron in place and tensioned, but retracted. A plane body twists measurably once the lever cap is done up.
The frog only needs to be flat where the iron actually beds — the top two thirds. Chasing the whole face is wasted filing.
Set the chipbreaker somewhere near four tenths of a millimetre back from the edge for smoothing. Closer than that and the shavings jam.
What not to bother with
Lapping the sole to a mirror. It needs to be flat at the toe, the mouth and the heel. The rest is decoration.
Replacing the iron before you have sharpened the original properly. Most “bad” irons are just blunt.
Stripping the japanning. It does nothing for the cut and it removes most of the value.
A No. 4 that has been through this takes a full-width shaving you can read newsprint through, and it will hold that for a morning’s work. That is the whole test.
This is the long version of the dovetail class we run four times a year, written down so people have something to take home. It assumes you can sharpen a chisel and saw to a line. If you cannot yet do either, those are the two things worth practising before you cut a single tail.
Marking gauge set once, used on every piece. This is where accuracy is won or lost.
Stock preparation
Four faces flat, square and the same thickness. There is no recovering from stock that is out, because the shoulder line is referenced from the face and the baseline from the thickness. Ten minutes with a plane here saves an hour of fitting later, and it is the step everybody wants to skip.
Mark the faces and edges before you do anything else. Every subsequent measurement comes off the face side, and once the pieces are in a pile you will not be able to tell which was which.
Setting the gauge
Set the marking gauge to the thickness of the mating piece and do not touch it again until the joint is done. Gauge the baseline on both faces and both edges of every piece, tails and pins alike. A joint where the baselines disagree by half a millimetre will show a gap on one face no matter how well you saw.
Some people set the gauge a hair proud so the joint stands slightly above the surface and is planed flush afterwards. We do this on anything that will be finished with oil, and not on anything that will be painted.
Angle and spacing
The angle matters less than people think, and the spacing matters more. A 1:6 slope in softwood and 1:8 in hardwood is the usual advice and it is fine advice, but a joint cut at a consistent 1:7 in everything will look better than one where the slope wanders.
Board width
Tails
Half pins
Up to 100mm
2
2
100–180mm
3
2
180–260mm
4
2
Over 260mm
5 or more
2
A starting point, not a rule. Wider tails read as handmade; narrow pins read as fine work.
Sawing
Tails first. Saw on the waste side of the line, leaving the line itself intact — you should be able to read it afterwards. Keep the saw plate vertical in one plane and let the slope come from the angle you have marked; trying to control both at once is what produces the wandering cut.
Start the cut on the far corner and drop the handle as the kerf establishes.
Watch the line on the face, not the saw.
Stop at the baseline. Sawing past it is not recoverable and shows as a visible overcut inside the joint.
Cut every tail on every piece before you change tools. Repetition is what makes them consistent.
Chopping the waste
Fret saw out the bulk if you have one, then pare down to the baseline. Work from both faces and meet in the middle — chopping right through from one side blows out the far face every time. Undercut very slightly so the middle of the socket is lower than the edges; nobody sees the middle, and the joint closes cleanly.
Fit is decided at the marking gauge. Everything after that is just carefully not making it worse.
Transferring and fitting
Stand the tail board on the pin board, line up the baseline exactly, and knife around the tails. A pencil is two tenths of a millimetre of slop you do not need. Saw the pins on the waste side of the knife line and the joint will go together off the saw with a mallet tap.
If it will not close, do not force it. Find the tight spot with a piece of paper and pare that one place. A dovetail is only ever tight in one or two spots; taking a shaving off the whole joint to fix one of them is how a tight joint becomes a loose one.
Glue-up
Dry fit fully, then take it apart and lay out everything you need before you open the glue. Hide glue if you want a long open time and a joint you can take apart in thirty years; PVA if you want it done. Glue the pins, not the tails, and only the side walls — end grain drinks glue and gives nothing back.
