When installing hardware, it is important to have precise control of the location of the pilot holes for the screws. The options are to start with a mark made by a scribing point or awl, then use that to locate the drill bit, or drill the hole directly without a preceding mark. Here is my approach.

I mark first, and most of the time, I try to center the mark in the countersunk hole of the hardware. Sometimes, however, I will intentionally very slightly decenter the mark to laterally “pull” the hardware tightly into the mortise, or to one side of a mortise that was made a bit too large. In any case, I do not want a mark that will result in the screw pulling the hardware in the wrong direction.

The key is that you want that control. It’s that one-sided tolerance thing again. I love that concept.

Here is my low-tech, reliable method.

I use a scriber point or pyramidal awl to prick a starting mark for the drill bit. It is surprisingly easy and accurate to judge just by eye if the mark is centered in the hole in the hardware. Good lighting is essential, as in the first photo below. A crescentic shadow on the side of the hole will cause you to misjudge the center of it.

The mark in the photo below is actually centered but the shadow interferes with judging it.

Sometimes I use a fine point scribe (the two tools to the left in the photo below) to make a tiny mark to seat the point of a brad point bit. Other times I use my wonderful Czeck Edge awl (at right in photo) to start and enlarge a hole suitable to seat a twist bit. Twist bits are commonly available in more sizes than brad point bits, which is particularly helpful for using tapped holes and machine screws to mount hardware in wood.

This simple approach makes it easy to reliably make a deliberately off-centered mark. Just as important, it is also easy to move a mark made with a scriber or awl. Just stab the sharp point into the sidewall of the original mark and push or swish the tool to make a slightly bigger mark with a new center location.

How about self-centering bits (one brand is Vix bits)? In all the brands I have tried, there is always a bit of wobble of the drill bit within its housing. Even though I would drill squarely to the work surface, there is some random error in the location of the hole. The errantly drilled hole is then nearly impossible to “move.” Furthermore, there is no reliably controlled way to produce a deliberately off centered pilot hole. Depth control is also difficult or the depth is inadequate. Such bits have been banished from my shop for some time.

How about center punches – those that you tap and those that are spring-loaded in various ways? They avoid some of the disadvantages of self-centering drill bits, but all of the ones I have tried, even Starrett’s, also have some wobble of the scriber within its housing. They are faster than the low-tech approach I use but less consistent and, again, there is no good way to make deliberately off-center marks. Such tools also have been banished from my shop.

In summary, to accurately locate pilot holes for hardware screws, go low tech, use your skills, keep in mind the one-sided tolerance concept, and use the skillful adjustability of craftsmanship.

A reader recently emailed me to ask the best kind of question – a simple one with broad implications. Using only hand tools, he asked, how do you “cut several pieces of wood exactly to length.” He stated further that he wanted to make several pieces “exactly the same length.”

There are really two issues here. In furniture making, it is not often that you need to saw a piece to an exact absolute length, such as 18 5/16″. By exact, I mean to a tolerance well under 1/16″. Usually, a component must be cut to precisely match another part or an opening, after the general dimensions of the piece have been established. What’s more, most of the time, sawing to length does not demand high precision.

Let’s examine the matter by looking at the fundamental types of construction in furniture making.

Post and rail

In a table frame, the critical matter is for the lengths of the aprons on opposite sides to be equal. The precision comes from knifing the tenon shoulder lines with the pieces ganged together. The absolute distance between the shoulder lines is not critical; it just has to be the same for the pair of aprons.

The sawn crosscut at the end of the apron just defines the length of the tenon, and since there should be extra depth in the mortise, it does not have to be a precision cut. As for the legs, gang them together, mark out the ends of the mortises, and cut the legs to length with adequate precision by hand.

Frame and panel

For a cabinet door or back panel, the precision in the frame components comes from knifing pairs of shoulder lines together. Later, you will plane the slightly oversized frame to fit the door opening or case. As for sawing the panel to length, those crosscuts end up inside the grooves, which give some margin for error. So again, the crosscut hand sawing just has to be good, not dead-on precise.

Board construction

For a case or box, the absolute dimension is again much less important than making the opposite sides match. This is where shooting is an essential technique of woodworking. Saw the two boards to length as consistently as you can, shoot three ends, then incrementally shoot the fourth and final end to make two boards of exactly the same length with clean, square ends.

Oops, you overshot that fourth end? Shoot the end of the other board down to match. The absolute length is not important.

By contrast, fitting a drawer front to an opening is a matter of meeting an absolute measurement, and there is no going back. For this, shooting allows you to sneak up on a precision fit in a way that sawing cannot.

By the way, to answer another question that came up, one cannot use a stop block on a shooting board set up to trim endgrain because you are not cutting to a line. Rather, you are removing a certain amount with each pass, based on the blade depth. You must advance the board after each pass, so stop block set-ups are not used.

One more thing – and it’s important

Even if you use a table saw and miter gauge/sled to make your crosscuts, as I do, you still want to be able to nuance them with hand techniques. Very often in fine woodworking, we want to intentionally tweak a component slightly out of square, make something not quite straight, or correct a fit.

Theoretical exactness is often not the goal, and understanding the variance can elevate your technique and results. That is why incremental hand-tool techniques are so valuable. Some examples of this follow.

The top and bottom of a solid wood case may intentionally have ends that are not quite square because you want the case a trace wider at the back than at the front to allow well-fitting inset drawers.

Even if those table aprons were supposed to come out just right, you find during dry assembly that you need to tweak one shoulder with a shoulder plane. Now that apron does not quite theoretically match the opposite one, yet the whole piece fits together just right.

You can adjust the side snugness of a drawer front in its opening with the last shooting pass or two.

After fitting the hinges for a cabinet door, even though everything was supposedly dead square to start with, you plane the edges of the door to produce surrounding gaps that are consistent to the eye.

In the craft of woodworking, sometimes what you thought was exact is not right.

Handplaning thin boards, those less than about 1/2″ thick for most species, sometimes produces surprising frustration. The solution to many problems is to consider what is happening on the underside of the work piece.

I usually plane such wood, for example, a thin drawer bottom, on a very flat planing board with front and side stops only about 7/32″ high, or on a nice flat section of my workbench against a low-profile planing stop that inserts directly into the bench. This avoids distortion from the pressure of bench dogs and gives me a better feel for how the plane is meeting the wood.

The handplane sole acts somewhat like the feed rollers in a thickness planing machine. It presses on the flexible board and forces it downward to close gaps under the board. This effect varies, sometimes unpredictably, with the length of the plane and the skew angle of planing.

Suppose now we are trying to smooth plane the surface of a drawer bottom or panel and the underside is a bit concave. (Here I do not mean an even-thickness board that is simply cupped a bit.) A finely set plane will miss areas on the topside that are over the concavities underneath.

This is especially annoying if you are trying to clean up a few spots of tearout and the plane keeps missing them. The tolerances involved are quite small because in these situations the plane is trying to take a very thin shaving, say .001″.

The solution is to be cognizant of the interaction of work piece and the work surface, and resort to good old shims. Sometimes I’ll use blue tape but more often just a few shavings stuffed under the board. or even some paper.

In fact, this little trick is handy even if the thin board is perfectly flat and you want to raise a small area to plane away a defect without having to plane down the whole surface just to hit the defect.

When using a shim, it usually is helpful to dog the board to keep the ends down and thus push up the shimmed area.

Note that if the work piece is of even thickness and just a bit cupped, this will usually not be an issue. The board will flatten against the bench as it is being planed, all areas will be hit with the plane, and the board will simply spring back to its cupped state when you are done planing. Within limits, that is not a problem because the panel will usually be easily coaxed into flatness by the frame or drawer grooves.

So, the real key in all of this is to not take your handplaning setup for granted but rather to be aware of the actual interaction between a finely set handplane blade, the plane sole, the work piece, and the work surface. With this awareness, it’s just a matter of common sense to diagnose and fix what otherwise could be vexing problems.

Wendell Castle once commented that often not enough time is spent designing a piece. The same can be said of researching a piece.

Unless you have previously executed something very similar, neglecting adequate research can lead to a lot of wasted effort building a disappointment. The research phase of a project should be enjoyable as the possibilities unfold and your woodworking knowledge expands.

Here are four categories that require attention. To illustrate the breadth of research sometimes required, I have used the example of a wedding wine box that I recently completed. This piece had very special significance and I wanted to approach it fully prepared.

1. Function

Almost all woodwork is functional. You do not want the beauty of a piece to belie an inability to do its job. Think of it this way: making a bat requires more than understanding wood and turning; you have to understand baseball.

I researched all the dimensions of the wide variety of standard Bordeaux and burgundy bottles to design a versatile cradle that would accommodate a range of bottles. I also had to learn about how wine should be stored long term.

2. Materials

This is not an area for guessing or shortcuts. Processes that are routine in one wood can be fraught with surprises in another species. What’s more, nearly every project involves several non-wood materials that woodworkers have to understand.

A few boards of gorgeous curly ovangkol (shedua) caught my eye. I had not worked with this species before, so I needed to explore the range of figure it had to offer to be able to choose top quality stock. I looked at objective data on its physical properties and movement characteristics. Most important, I did some practice sawing, chiseling, and planing to appreciate its working properties. It was surprisingly incompressible and somewhat brittle so there was very little margin for error in the joinery.

I considered lots of options for secondary woods, settling on wenge and a billet of killer figured redheart big-leaf maple. I trialed finishes, tested glues for special situations, and also researched leathers.

Then there’s hardware. Ugh. There are always oodles of options here, though often I am not happy with any of them and end up modifying or at least fine-tuning the best available materials.

3. Constructions

Almost every piece I make involves at least one modified or non-standard construction technique. It really helps to consider solutions that other woodworkers have used, though it is important to use sound principles and experience to distinguish good information from bad.

If you never venture from the conventional, you miss out on a lot of fun in woodworking, but you must build right, so do your research.

In this project, I could not find a satisfactory solution for a cradle that would snugly hold a range of wine bottle sizes. What I worked out is no marvel of engineering but I did have to sit at the drawing board for a long time scratching my head, and make mock ups, before settling on a solution.

4. Techniques and Tools

Similarly, every project is an opportunity to develop as a woodworker by learning new techniques and reinforcing skills that you have used before. A good woodworker should never be too proud to practice even those skills that were acquired some time ago.

And yes, there is also the excuse – oops I mean perfectly valid reason – to buy a new tool, which, by the way, has to be studied and tuned. In this project, because I did not do enough research, I needed the excellent Lie-Nielsen drawer lock chisels to bail me out, as mentioned in an earlier post.

In summary, researching function, materials, constructions, and techniques/tools is smart woodworking. Note to self: don’t cheat on your homework.

Here is most of the power sanding gear in my shop. Once again, I hope you will find a useful tip or two in here.

The gentler Bosch ROS20VS (5″ disc, 2.2 amps, 3/32″ diameter orbit) random orbit sander gets much more use than the larger Bosch 3725DVS (5″ disc, 3.3 amps, 3/16″ diameter orbit) because I am mostly using these tools to smooth surfaces, not to form them. As mentioned in the first post of this series, there are much better options, with and without sanding, for the later task.

In any case, these Bosch sanders perform very well at reasonable cost. The 6″ size has 44% more area, helpful for bigger jobs, but the 5″ is handier and suits the scale of most of my work. I always use these sanders hooked up to my Fein shop vac with the auto-start feature.

A crepe rubber stick (above, left) is a must for cleaning power sanding discs but also works well on hand sandpaper. Get a big one.

Norton’s new ProSand Multi-Air Cyclonic hook-and-loop discs (above, left and right) have superseded their 3X discs (center). Amazingly, the 246 laser-cut holes (I read that; I didn’t count them) give better dust extraction than the big holes in 3X and other brand discs. The ProSand discs also have more sanding area, an efficient ceramic abrasive, and are easier to apply because you do not have to align holes – just put the big one in the center.

This storage tree keeps the discs organized. It is nothing more than dowels set at a slight angle into a wood strip.

The Ridgid EB4424 combination oscillating spindle-belt sander, shown above in belt mode and below in spindle mode, may be the best thing Home Depot sells in their “tool corrals.” The designers thought of just about everything. This certainly is a sanding tool for shaping wood.

The oscillating action, which runs true, makes shaping smoother and more controlled, as grabbing is minimized, and leaves a smoother surface on the wood. Changeover between modes is fast and convenient. Belt tracking is easy to adjust and the setting is retained very well.

The most significant downside of this machine is the cheap table but it is not a deal breaker. Though it adjusts from 0° (shown) to 45° and has a serviceable miter slot (3/4″? Yea, sure.), it should be flatter and firmer. Dust collection with the Fein vac is surprisingly good for this type of sander.

I highly recommend this machine.

By the way, those purple 3M flexible mesh sanding sheets found their way into the two photos above, at the left edges. I forgot I had them since 3M’s much better Ultra Flexible Sanding Sheets became available. (See the previous post.)

Norton’s blue Norzon belts are incredible wood eaters. Most of the extra sleeves and belts get stored in this boot box.

The Singley drum (above, left) uses ordinary sandpaper, cut to size and easily wrapped around and tucked into the drum. With this tool, drum sanding can be done economically in a wide range of grits. Many sizes are available.

The Robo sander, chucked in the drill press, works like a flush trim/end-bearing router bit but more gently and without the risk of tearout.

Shown beneath the Robo is the Veritas Drum-Sander Support System bearing. It is essentially a live center that sits on the drill press table, and whose point engages a dimple in a retrofit modified drum shaft. This trues and stabilizes the rotation of the drum.

This concludes the four-part series on sanding. As always, there is more than one good way to do just about everything in woodworking, but I hope this material has been helpful to you.

This is just about all of the hand sanding gear I use in my shop. I present this in the hope that you will find a useful tip or two.

A cork block, 1″ x 2″ x 3 7/8″, is the best tool I have found for hand sanding. It has just the right compressibility and resilience to produce a consistent and true sanded surface, and it feels comfortable in the hand. A piece of cork glued to a wooden block is a decent substitute but not as good. Almost all of the area of one-third of a standard 9″ x 11″ sandpaper sheet can be utilized by re-wrapping it just once.

The Norton pad gets less use but is handy for fine sanding large areas. It uses the same one-third sheet with little waste.

Convex and concave rubber sanding grips are handy, though I more often use ad hoc blocks made from combinations of wood, cork or rubber sheet, and pink foam insulation. Some of these are worth preserving but some live for only a single fleeting employment.

The sheet cutter makes it almost fun to size sandpaper. A hacksaw blade is screwed to a piece of plywood with enough slack to permit the sheet to easily fit under. A piece of thin plywood is glued to the blade to make it easy to press the blade firmly onto the back of the paper prior to tearing it. A slat is placed in one of the table-sawn kerfs that have been placed to yield the desired sizes of paper. By far most common are one-third sheets made by cutting across the narrow width of the full sheet to yield strips about 3 2/3″ wide.

This little plastic bucket screwed to the wall is a good place to store partially used strips. Contrary to comments by some shop visitors, it is neither a garbage can nor a urinal.

I’ve used Norton 3X paper for years with excellent results. Norton has superseded it with ProSand, which they claim has more durable abrasive and backing paper, so I will gradually restock with that.

Once you get these 2 1/2″-wide, PSA-backed (sticky) rolls in the shop, all sorts of uses arise. They are great to quickly make impromptu sanding blocks and tools, such as those below. I also reload the Veritas shooting sander with them.

These little shop-made sanding sticks solve vexing detail issues in almost every piece I make. Apply an oversized piece of PSA sandpaper to the squared end of a wooden tongue depressor, and then trim the excess with a utility knife. Of course, as needed, you can also create a V-point or other shape on the end. You’ll wonder how you ever managed without them.

From left to right in the photo above:

The silicon carbide “wet-dry” paper is there for sanding between coats and sometimes for wet sanding of oil-varnish. I have 600, 1000, 1500, and 2000 in stock but rarely use the finer grits.

3M Ultra Flexible Sanding Sheets, available in 100, 150, 220, and 320 grits, really live up to their name. The grit stays on and the backing does not crease or tear. These are a far better option for contour work than sanding sponges, which I have always found to be useless. You can back up these sheets with whatever you want – contoured rubber or foam, a sponge, or just your hand.

3M has also recently introduced Sandblaster Pro sheets with a grippy back. So far, I have not found an advantage from them for woodworking but they are very handy to flatten tools when simply placed on a granite surface plate. They stay put without spray adhesive or water.

I use the MicroMesh set of 3″ x 6″ sheets for tool and hardware alterations, not on wood. The grit ranges from 1500 to 12,000! The sheets are cushioned and thus not a good choice for sharpening.

Next: power sanding gear.

Let’s consider the finished surface qualities produced by sanding compared with handplaning. Here there will be no blanket declarations, including none extolling the superiority of handplaning, and no simple catalog of species with recommendations to sand or plane.

The real answer is to experiment as part of planning your project. In each case, consider the look you want and the practicalities of the building process.

You have to assess the surface of the wood (the particular boards at hand, not just as a species) in combination with the applied finish. Just as it pays off to plan the applied finish at the start of a project, so too should the method of finishing surfacing the wood be planned. Winging it is usually not smart.

Here are examples from the shavings and dust of my shop. YMMV. This discussion pertains to the final surfacing of wood that has already been trued or shaped; we’re only dealing with the last few thousandths of an inch of wood.

By the way, for exposed parts in fine woodwork, I never finish with the surface from a random orbit sander, no matter how fine the disc paper. When sanding, I always finish with hand sanding. In fact, most of the time, I don’t use the ROS at all.

For figured big leaf maple, one of my favorites, with satin gel varnish, I can see no difference in the final look whether the wood surface is finished off with hand planing or fine hand sanding. Therefore, I do whichever is easier and that is usually sanding.

The same goes for figured bubinga with oil-varnish mix. Bubinga responds exceptionally well to scraping, so little sanding is required thereafter.

Claro walnut, another favorite, with oil-varnish, seems to look more clear and lively when handplaned. With brush-on varnish however, I cannot tell the difference between planing and sanding. For highly figured Claro, its visual impact often seems to override subtle differences between planing and sanding.

Curly cherry with gel varnish, the finish I like best for it, is finicky. It looses some of its pizzazz when sanded. Pearwood similarly looks exquisite straight from the smoothing plane and can well be left unfinished, but after two coats of water-base acrylic, it is hard to tell if the wood was planed or sanded.

Oak, red or white, flatsawn or quartered, plain or figured, with wiping varnish and the grain unfilled, usually looks about the same to me, sanded or planed. Oil-varnish is different.

Again I emphasize that I always experiment at the outset of a project with the actual wood and finishes that I am using for that project, and try to anticipate the practical issues that I will encounter in building the piece.

We ought to be practical. A curvy table leg creates most of its visual impact from its form, while it is the surface and figure of a cabinet door panel that we appreciate. Again, choose planing or sanding based on the overall look that you are after and the practicalities of building. Maybe there are fine facets on the leg that sanding would obscure and a spokeshave is the right tool to use, or maybe there are gradual curves that look good sanded.

Work with a smoothing plane is usually more pleasant than with sandpaper, but sometimes planing is just awkward, such as when finishing off a dovetailed case. And let’s face it, sometimes we just don’t want to spend more time at the sharpening bench.

There are also some special situations. For example, when fitting a drawer, a hand plane is the only tool to use on the sides. Choose the wood for the sides to allow easy planing and usually leave it unfinished.

One more thing: when finish planing difficult wood, there are almost always a few spots of tearout that just seem unavoidable, or maybe the blade developed a nick (especially some A2 blades) and left a little row of raised wood. I touch up these areas with a 0.020″-thick sharp scraper rather than with sandpaper. Nothing is perfect.

Next: sanding tools

Sanding does not get much respect among woodworkers. Hand planing uses more gratifying tools in a more pleasing process, and when suitably employed, leaves lively surfaces and is more efficient. This series of posts will attempt to put sanding in perspective in the world of fine woodworking and present practical information on tools and techniques.

James Krenov wrote in The Fine Art of Cabinetmaking, “To me, sanding is not a way to express sensitivity with wood, even less a sign of true skill,” yet he recognized a legitimate role for sanding, recommending, for example, scraping and sanding as the best way to finish rowed woods like padauk.

The first thing to clear up is whether you are sanding to shape wood or to produce a nice surface on wood. Are you forming a curve or smoothing the surface of an existing curve? Are you flattening a surface or just smoothing an already flat surface? Sanding is hardly ever the best way to shape or dimension wood unless you are using jigged machines such as a spindle sander or thicknessing sander.

True, sometimes you are both shaping and smoothing, such as when forming a light chamfer with a hand sanding block or running a figured panel through a drum sander. But generally speaking, it is important first to be clear about just what you are doing when sanding. In fact, most problems with sanding come from inadvertently mixing shaping and finishing.

For example, you have a fantastically figured board in the rough, or perhaps is flat from the planer but full of tearout. You are afraid to touch it with a handplane so you take out the random orbit sander, start with a 60-grit disc and work through to 320. Unfortunately, despite all efforts to evenly distribute the sanding, the final smooth surface is wavy, the outer edges are dipped, and there is no hope of using this as a reference surface for further work, such as for a drawer front.

There were lots of better options that would have produced and/or retained a true surface that would then require only fine sanding, which would not squander the flatness. Among them, for various stages, are: a spiral cutterhead on the planer, a thicknessing sander, a jack plane worked across the board, a toothed blade in the jack plane, a Veritas scraper plane, a micro-toothed blade in the scraper plane, and a hand scraper.

An equally unpromising plan is to take a curved table leg rough sawn off the bandsaw and hope to use a curved sanding block with coarse paper as the primary final shaping tool. You will not get the proper feedback to produce a true curve that comes instead from high quality rasps, spokeshaves, and curved planes. When the curves have been trued with some of those tools, then you can use the curved sanding block to just finish smooth.

The point is that sanding – by hand and with small and large machines – has its place but it pays to be mentally clear about exactly what you are trying to accomplish with it, and restrict it to that task.

In the next installment, let’s consider the finished surface qualities produced by sanding versus handplaning. Be prepared for some surprises.