Recently, I was restoring a small spice rack made of very thin and very soft wood (pine-like) - I unglued it, restored pieces and was about to glue it back together. I saw that the best way to go was to glue it all in one go, since it would be much easier to clamp when fully assembled than to glue it in steps (see the sketch).

small spice rack from thin planks

So I decided to put a few small 2mm (1/16") dowels (i.e. bamboo toothpicks) in the pieces to make sure it won't slide under clamps.

What I did (and created "the problem"):

I drilled holes in one piece, made small "dowel pin locators" by cutting nails, marked the opposing pieces, and then started to drill them...

Unfortunately, even with a drill press, the drill wandered off by approx. 0.5mm (1/64") when it hit harder grain - that would be fine for larger hole for a larger dowel, but since it was so small, even this miniscule error was unacceptable (by which I mean the pieces would be slanted or wouldn't fit at all).

My "sort of solution"

I tried several times, but drilling just didn't work for me, so in the end I used my tiny 2mm (1/16") chisel and made square holes and used square dowels. This worked, but it took way too much time.

(In fact this was not the first time I was unable to drill a very small hole into a softwood with hard grain lines exactly where the mark was.)

The question:

Is there some fool-proof, simple and most importantly reliable way accurately drill very small holes 0.5 - 2mm (1/50" - 1/16") in softwood with hard grain lines?

  • 1
    I've done a minor edit since, as one of the tags already implied, your problem wasn't so much a soft wood, but instead a problem with softwood.
    – Graphus
    Jul 19, 2023 at 17:17
  • Considering the size, would a pinvice work? Jul 20, 2023 at 7:47
  • @JourneymanGeekOnStrike that was my tool of choice, then I thought it wanders off because I can't hold it well and/or its too slow, so I tried small electric drill (like a dremel), then i tried drill press, but it always moved on the grain a little - I thought I'm just not skilled/experienced enough, but it's apparently just hard to do - see Graphus answer.
    – Jan Spurny
    Jul 20, 2023 at 8:01
  • Re. your Comment below, not sure if Chris actually meant a bradawl — he refers to marking a hole, which is something a conventional awl is for (in an analogy with metalworking that would be the equivalent mark to the one made by a centre punch). Bradawls on the other hand were for actually 'drilling' small holes (chiefly in softwoods because they absolutely suck at doing deep LOL). They are basically obsolete, which is why you see them used so so infrequently; if you want to try one out for yourself that's easy, just grab a nail and clip the head off, then mount it a handle [contd]
    – Graphus
    Jul 21, 2023 at 17:57
  • ...of some kind. If you don't have any fat dowel handy just whittle a bit of stick down, it'll do for the experiment. Although common wire nails are just mild steel this will still allow you to see what the tool is like in practice once your epoxy has cured tomorrow. I'm all about hand tools and traditional working practices, and I own four or five bradawls now...... and never use them. I continue to buy them mainly to save them from oblivion and to practice restoring the handles :-)
    – Graphus
    Jul 21, 2023 at 18:02

5 Answers 5


Is there some fool-proof, simple and most importantly reliable way accurately drill very small holes 0.5 - 2mm (1/50" - 1/16") in softwood with hard grain lines?

Unfortunately, not really.

The problem with many softwoods is the huge difference in hardness between the dark and light bands in the grain/growth rings. The dark latewood is quite hard and the pale wood in between, the earlywood, is soft to very soft and presents a path of least resistance that drill bits are prone to following.

Even with a distinct starting mark from a sharp awl drill bits can wander into the soft part of the grain, even when the bit entered the wood in just the right spot initially. Remembering that grain is three dimensional, and just under the surface a growth ring can be vertical or veer off steeply to either side, so the tendency to drift is unpredictable to say the least.

Using brad-point bits (also called lip-and-spur bits or dowel bits) instead of the common twist bit can help, but we soon run out of sizes as the hole size goes smaller. And for very tiny holes the only bits commonly available are twist bits.

So, since we sometimes have to live with a hole that isn't quite where we wanted it workarounds are necessary.

Drill the secondary hole larger
You transfer across your hole positions as normal, and then deliberately drill an oversize hole so that you're sure the dowels will not be skewed to one side or not fit at all. Note that this requires a gap-filling adhesive is used.

This isn't ideal for a few reasons. For starters, someone might not want, or be able to, glue with epoxy or another gap-filling adhesive, but more importantly for this specific use-case it doesn't provide the exact positioning of the parts that was the initial requirement.

Where you can, use through-dowels
With through-dowels hole alignment is automatic because the entire hole (the holes in both pieces of wood being joined) is drilled in one go. So a minor miss-positioning of the initial hole from the desired position – the hole in the 'top' piece – automatically translates to the position for the second hole in the 'bottom' piece.

Bonus option
This is for the specific initial problem you outline, rather than a complete replacement for dowels as an alignment aid or reinforcement.

Essentially what you do is the following, but then stop:

I drilled holes in one piece, made small "dowel pin locators" by cutting nails

The clipped-off nails generally have little difficulty in embedding themselves into the next piece of wood and already do what you wanted to do, prevent sliding around after the glue is applied and the pieces are brought together under clamp pressure.

This is an old trick and used in both softwoods and hardwoods, but the harder the wood used the shorter the clipped-off nails need to be to ensure that clamp pressure can make them penetrate fully into the next piece of wood — if overlong they'll act as 'standoffs' and prevent the workpieces coming together.

  • Thanks! The "Bonus option" would definitely be the best solution for my problem - and I already had the nails in.. well, next time :). As for through dowels - that would be my choice if it wouldn't be a restoration - it would work for the back piece, but not for the rest.
    – Jan Spurny
    Jul 19, 2023 at 19:31
  • 1
    As for the "larger secondary hole" idea - I like it and I think I can make it even better - I think I could make a larger hole (in this case - barely - but I could) and then glue in some hardwood dowel, then flush cut it, plane, and I would have a nice hardwood endgrain material where precise small hole could be drilled easily! I'll try it next time!
    – Jan Spurny
    Jul 19, 2023 at 19:36
  • 1
    Drilling a larger hole for a dowel to allow for more precise drilling of the matching hole is a great idea! It's quite a bit of extra work but obviously worth it for some applications.
    – Graphus
    Jul 20, 2023 at 5:25

Is there some fool-proof, simple and most importantly reliable way accurately drill very small holes 0.5 - 2mm (1/50" - 1/16") in softwood with hard grain lines?

Nothing is foolproof, but there are some things that can help prevent deflection:

  • sharp bits: Some people never think to sharpen their drill bits, but keeping them sharp helps them cut through harder areas more easily, limiting deflection.
  • fairly high speed: Higher speed allows for more cutting action. It can also cause heat to build up, though, so withdraw the bit periodically to clear the hole and give the bit a moment to cool down.
  • feed slowly: Like the higher drill speed, feeding slowly gives the bit more time to cut straight instead of deflecting.
  • limit the bit length: The more the bit sticks out from the chuck, the further it can deflect. Small diameter dowels don't usually need to be very long, so the holes don't need to be deep. You don't want to tighten the chuck on the fluted part of the bit, of course, but there's no use in having more than a millimeter or two of the unfluted shank sticking out from the chuck. Also, it's not uncommon to snap small bits, and you can often sharpen the snapped bit to make a shorter but stiffer bit.
  • As a direct answer (rather than a workaround), this. Generally, "fine work" is often better with high speed (5000+ RPM) / low torque / slow feed. A Dremel or some such is a better tool for this. One should not even "feel" the bit cutting. In such conditions, the drill bit will not care much which material it's cutting, provided it's sharp.
    – Zeus
    Jul 21, 2023 at 0:44
  • 1
    @Zeus, in homogenous materials this is dead on, but many woods are very much not homogenous and the OP's Q is specifically about some of the worst-case stuff one can imagine for hard/soft contrast. I've drilled tons of holes using my Dremel-type drill over the years, but even at my drill's top speed ~20,000 RPM (with near-zero runout) deflection can be a huge problem, not necessarily due to the bit wandering but the whole hand holding the drill. You just can't control it, and it happens so quickly at these speeds. Once I broke my 5th or 6th sub-mm bit I declined to drill such holes using it.
    – Graphus
    Jul 21, 2023 at 5:36
  • @Graphus, well, hand-holding may be a problem. I have a mini drill press that holds the Dremel for such occasions. But generally speaking, this approach (high-speed cut) is simply less dependent on the material, including its homogeneity. But yes, in some sense it is less forgiving to mishaps and slips. And yes, sub-mm is often a challenge, but 2 mm should be relatively trivial.
    – Zeus
    Jul 21, 2023 at 7:11

Try this. Line a straightedge up with your point and use a sharp marking knife to draw a very short, deep, line. Turn the edge to a different direction and repeat to form a small X. Hammer in a fine nail, starting gently. It will not wander. This can be your guide hole or perhaps your final hole with a bigger nail. Practice this with variations on some scrap or on a hidden part of the piece.

The mistake, I think, is to get involved with rotation, either by using a drill press or by twisting an awl. Rotation brings you into contact with harder grain. Your fine chisel method avoided that. And if you are worried about those little X-marks, I think that at the end of the day they will be invisible after sanding or under whatever finish you use.

  • 1
    Starting with an X is an intriguing suggestion, and it could be done with fair speed once you got practised, with straightedge & square. But sadly deflection could still be an issue for one of the two cuts if the angle were unfavourable — at around 45° this should work great, because the opposing cut is also at a steep angle to the grain. But around 90° for the first cut the crossing cut would be around 0°. Visibility of the X shouldn't be a problem since they'll be hidden in the joint. If in a visible spot however, just like gauge marks the cross-grain portion would be highlighted by finish.
    – Graphus
    Jul 21, 2023 at 5:12
  • Yes, that may work.. I'm considering buing bradawl (i.e. the one that looks like a flat screwdriver) as Chris H suggested - this would work for the X too..
    – Jan Spurny
    Jul 21, 2023 at 8:03
  • Presumably the grain direction is visible, so getting roughly plus and minus 45$^o$ would be easy. If the marking knife is really sharp (Xacto?) the X-marks will close back up on their own as soon as they get any moisture, and any sanding will produce filler. These things can be encouraged.
    – Philip Roe
    Jul 21, 2023 at 16:18

Drill bits that small are rather flexible, as you've found. For the unspecified softwood and "pine" commonly sold here, unless you're on the edge of a knot, the following method works.

They're also short. This means that you can get much less flex by pushing the bit as far into the chuck as it will go. You can even grind off the chuck end of the shank.

That will allow you to drill a much more precisely-located hole, but a shallow one. If you need deeper, you'll have to loosen the chuck and expose a bit more bit.

In metalworking centre drills are common. These have a very thick shank compared to the size of the hole, and again can't go deep. But there's no danger of them wandering on a lathe. You could use one to spot softwood, but the imitation approach is pretty much as good.

However do note that a lot of small benchtop drill presses flex a bit (or don't run true), so even a stiff drill bit will move. Again, taking a tip from metalworking, marking the hole with a bradawl can help. You can also lower the press with it off, looking very closely and indent the workpiece with the non-spinning bit - but only if the bit is prevented from flexing.

  • Shorter drill bits seems like a good idea - maybe not "the solution", but definitely an improvement, thanks! But I'm afraid metalworking analogy stops here, because metal won't really compress and "move under the bit", whereas softwood unfortunatelly will. Iimagine cork layered with steel foil - the "cork" will compress and "move" even if the whole piece is held in vice, even with "ok" drill press (mine is not perfect, but it's solid). The centre drills again won't work - I had perfectly positioned "guiding hole" made with an awl, but underneath was hidden grain and it moved the bit..
    – Jan Spurny
    Jul 20, 2023 at 9:34
  • @JanSpurny when it comes to solutions to the real problem, Graphus has that covered (as usual). I don't use metalworking as an exact analogy (but thin aluminium does flex under the bit). Hidden grain is a pain, if it's bad enough for such big differences in compression across the width of a small hole that the workpiece itself is deformed
    – Chris H
    Jul 20, 2023 at 9:51

Im not sure if this idea would work in this scenario, since the surface area for the glue is so small, but in general a handy solution for a sliding glue up can be salt. Just salt the glued area before clamping the pieces together, the salt kernels work as minute dowels at the beginning and then presumably "melt away" in the glue while its dries. Sorry if everyone already knew of that and just figured it is not applicable in this case, but maybe it helps ^^

  • The salt trick gained some traction a few years ago, and it works but there's a but. In case you don't know the tip comes to us from the era of hide glue, where salt isn't an issue in a chemical sense, but Franklin (the makers of Titebond) recommend against it (and presumably this extends to all other PVAs, both white and yellow). There are some common strategies to help avoid issues as outlined in Wood Magazine's answer here but they're not a magic bullet sadly.
    – Graphus
    Jul 21, 2023 at 5:45
  • I use sand not salt.
    – jesse_b
    Jul 21, 2023 at 23:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.