Tails First or Pins First?

I've cut dovetails using both the tails-first and pins-first methods. When I was taught to do this, it was using the pins-first method, and although I was a beginner, the joints were very tight and accurate. They just required a lot of correction along the way.

Soon afterward, I saw a video by Rob Cosman on making dovetails and he cut the joint tails first. It looked to have many advantages for speeding the process and making it more accurate so I started experimenting. I soon found that a pencil was inadequate for marking the pins, where it worked just fine for marking the tails on a pins-first joint. This caused me a lot of grief as I didn't have a marking knife. I tried a number of solutions, including a sheetrock knife (don't try this at home, kids), but they all compromised the tightness of the joints. Finally, Fine Woodworking did a test on marking knives and included a $3.37 solution: the Xacto Knife. Problem solved, and my tails-first joints got much tighter.

Yesterday, Christopher Schwartz posted a list of reasons why he has adopted the tails-first method, and a couple of them are destined for my shop: gang cuts should be a real time saver, and I'll be trying the rabbetting trick soon after the moving fillester arrives in my shop.

First Dovetail for an Eight-year-old

Here's a great post by Konrad Sauer about his son's first through dovetail. This is the kind of experience that blow me away about being a parent.

Box Assembly

Yesterday night I returned to the shop. The bottom and insides of the cherry dovetailed box had been finished with two coats of Tung oil finish and a coat of wax. The finished cherry glowed almost golden, practically begging me to glue it together.

Having come this far without incident, I wanted to be sure to get the glue up right. I checked my labels, still visible on the outside of the box, and laid out the parts in relative position with the bottom in the center. I placed the sides with the bottom slots of each piece laid alongside the bottom. I had clamps ready if I should need them; also a 12" rule, a mallet, four pine blocks cut with fingers slightly thinner than the tails, glue spreaders and glue cup cut from a small paper cup, a spray bottle of water, paper towels, and scraps of cherry created when I cut the dovetails.

Surveying the scene, I ran through the process in my head. It seemed that everything I needed lay in front of me. I used some poster putty to temporarily attach the paper glue cup to the bench top (have you ever chased a glue cup under the bench while your open time was ticking away? I have.). Then I poured enough wood glue into the cup and started spreading glue on the long grain of the pin boards. Perhaps I should have also spread glue on the tail boards, but I chose not to do so. I knew the tails had a fairly tight fit, and I wanted to minimize squeeze out.
After adding glue to all the pins, I lightly inserted one set of tails into the corresponding sets of pins, first on one side and then the other. Using one of the pine blocks to protect the cherry side from the mallet, I tapped the end home, working alternately from one edge to the other. When that end was set, I turned the box over and slid the bottom into its slot, then repeated the process of inserting the tails and driving the second end home.

I now had a box. Using the 12' rule, I tested for square. Then I looked at the inside for squeeze out. At this point I realized two sections were not driven all the way home, so I applied the mallet just a little bit harder to drive them into place. The box was still square, and there was only one spot of squeeze out. Since the inside was already finished, cleaning this was no big deal. I sprayed a paper towel lightly with water and wiped away the glue spot.

At the last, I looked at the dovetails critically. Only one spot looked like it needed significant help, so I made a small wedge from one of the cherry scraps. I dipped this in glue and lightly tapped it into the gap (making sure to line up the grain so it would appear to be part of the pin – I think end grain hides better than edge grain and its easier to make the wedge that way). So now the assembly is done. My only worry is that I may have driven this wedge a little too far and deformed the tail enough to be noticeable. We'll see tonight when I trim it flush.

Dovetailed Box - Progress Report

I continued work on the box as much as possible this weekend (amidst a two day camping trip, yard work, and a barbecue):

• All sides of the bottom, and the insides of the box sides have been finished with two coats of tung oil finish and a coat of lightly buffed wax.
• My maker's mark has been stamped in the bottom.
• The profile on the top sections of the box sides has been beveled. One of the bevels is slightly steeper than the other two, but trying to correct this would just cause problems. It looks fine as it is.
• I cut pine blocks that fit the tail fingers of the ends. These will be used to protect the box from the clamps I will use to set the box square and keep it that way until dry.
• I set out the clamps and set them to the right size for the project, cut a paper cup (both for a glue dish AND glue spreaders), and laid out the parts: ready to glue (possibly tomorrow night).

In all, I'm happy with the progress, but I'm anxious to finish so I can move on to the bench project. I want the benches for Mid-July for a test run, and I don't have much free time between now and then.

Pounce Wheel and Pounce Bag: Another Marking Technique

Christopher Schwartz wrote an article about pouncing a while ago, over at Woodworking Magazine's blog. This is a technique that seems like it could speed the process of transferring dovetails. Probably not significant in a basement shop, but definately relevant to makers who may be laying out and cutting multiple dovetail joints.

In the comments on the article, someone mentioned a pounce wheel:

I've used a pounce (pronounced pöns, I believe) bag before, but never for dovetails.

My usual use for one is to transfer curved patterns from paper to full size templates. The technique is to lay the paper plan onto the template stock, and trace over the line with a pounce wheel. A pounce wheel is a spiked wheel on the end of a stylus. It sort of resembles a Texas cowboy's spur, in miniature. Anyway, the pounce bag is bounced on the row of tiny holes, and a clear dotted line shows up on the template stock, ready to bandsaw to shape.

This is a very cool way to transfer curves from a full-sized paper pattern to a template or workpiece. The pounce wheel is something I've seen when browsing the tools section of the fabric store while my wife improved her stockpile of fabric, so if you want a pounce wheel and don't know where to find it, try calling your local fabric store.

Patching a Poorly Cut Dovetail

I've been practicing by making boxes with half-blind dovetails and frame and panel bottoms. Although I was taught to cut near the line and approach slowly and carefully, I know it is faster to cut perfectly with the saw. So I've been cutting right to the line and dealing with the consequences, knowing I will spend more time now for reduced time in the future.

The first two joints came out quite nicely, but on the third joint my hope for the best didn't work so well. I wrote last week about cutting on the wrong side of the waste line. The good news is that my practice of cutting to the line meant I was only one saw's width too wide instead of more. But there was a saw's width gap.


I didn't think that could be fixed with a simple wedge (as I plan with the bottom side of the same tail), so I decided to patch it with some of the waste generated while cutting the pins.

The first step was to identify a waste piece that would fit, and make sure it fit snugly. I chose to patch the pin rather than the tail, because end grain is less likely to show noticeably. Then I made sure the side and corner of the tail were square and flat.

Even though my chosen waste piece had a flat section, there was a notch where I had started the chop. I needed a safe way to flatten this, and a chisel was definitely not that way. While puzzling over this, I remembered a miniature plane that might work. It looked to be just right for this application, so I tuned gave the blade a quick flattening and sharpening. Then I clamped it in the vice and ran the intended patch over it until it was flat.

Note: Looking at the picture, I realize this probably is not safe for the ends of my fingers. I was lucky and did not make finger shavings. Let me know if you have a safer idea for how to do this in the future.

Then I used a chisel to cut a matching angle for the bottom edge and test fit the patch to the tail.

Glue up was next. I used Tightbond II and a small c-clamp to attach it.
After letting that sit for a day, I trimmed the patch flush with the pin.

Then marked it for a new cut.

And trimmed to the line with a chisel. Actually, I trimmed past the line on one end (insert a favorite string of curses here) but the result is much better than I started with. The remaining gap can be wedged during glue up or hidden by peening the end grain.

Thoughts on Dovetails

Since I've been making dovetails, I've thought a lot about why they are made the way they are. The first time I tried a dovetail, I automatically laid out the joint with half tails at the top and bottom. I soon learned that this was not recommended.

Tage Frid, in the first book of his excellent series Tage Frid Teaches Wooodworking stressed that dovetails should start and end with half pins and not half tails. I believe that the reason for pinning top and bottom is to provide a mechanical restriction should either board try to cup away from the joint.

Using half pins at top and bottom theoretically fixes the pinboard and tail board from cupping. The half tail offers no such restriction. Because the angle of the tail runs with the grain rather than across it. there is nothing preventing the edges of either the pinboard or tailboard from cupping away from the joint. The first such restriction will be the first pin. So the rule of starting and ending with half pins is a good one because it improves the chances of the joint staying tight.

Today I discovered that this good-sense rule was not always practiced in the past.
If you look at the section on Antique Dovetails in this Popular Woodworking blog entry, you'll see that only one of these follows Tage Frid's admonition. Instead, most started with a half pin at the top, and ended with a half tail at the bottom. As evidenced by the half pins at the top, these craftsman obviously knew that half pins offered an advantage. But why didn't they pin the bottom? At first I guessed that they wanted the top to stay absolutely flush with the drawer because the user was most likely to notice gaps there. But if they were willing to pin the top, was it so much more work to create a half pin at the bottom?


On further thought, I suspect there was some planned advantage to this layout: there must be a reason why leaving out the bottom half pin was faster and easier. I suspect that reason is the groove holding in the bottom of the drawer. I obviously can't see the interior of any of the drawers Glenn Huey showed us, but I suspect that both the sides and front have a channel for the bottom that runs right through area of the half tail. Laying out with a half tail at the bottom allows the maker to work without fear of accidentally exposing these channels with a misplaced saw cut or chisel (something that can easily happen with a half pin). This advantage probably justifies a gamble on the stability of bottom end of drawer.

Dovetail Error #2

No, this is not a mock up for next year's jack-o-lantern teeth, but a real life example of cutting the tail instead of the waste. This demonstrates why you should always mark the waste, on both visible edges, before starting to cut and chop.

It almost makes me feel better that this error is over a year old. And it was a mock-up using scrap wood. And I was hurrying.

My New Year's resolution? Work smarter, not faster. The speed will come, and going slowly once is still faster than going quickly twice.

Dovetail Error #1

Notice the chisel (my thinnest) and how it compares the the waste area next to it. If your thinnest chisel is a 1/4", you cannot cut pins that run to the width of the saw kerf: your chisel will be too wide to remove the waste.