Yew "fence post" longbow, part 8 - filling time

Reducing and heat treating the belly has revealed the weakness inherent in the knot in the upper limb. The knot was pretty punky, and adding heat to it only made it crumble worse. While I've compensated for it some (probably not enough?!) by adjusting the width of the limb at that point, it makes me highly uncomfortable having this open "wound," if you will. As I started digging out the punky parts, the knot has revealed itself to be somewhat conical in nature, with the tip of the cone just protruding in a flattened point on the back of the stave, though it was not punky fully though. This results in the blister in the back being somewhat filled -- more crater like than an actual tunnel.

I've decided to try and fill this with a combination of yew wood dust (collected in the dust collector by grinding some yew offcuts on the belt sander) and super thin cyanoacrylate (CA) glue (a.k.a. "super glue"). What I've read on-line is that the wood dust should be reasonably fine and that it should be compacted in small layers, then soaked through with the CA glue.  Doing this on the back is pretty straightforward. It will also be covered by rawhide, so I'm mostly trying to fill the little crater here so that I don't end up with an air pocket after the rawhide goes on.

Filled and glued.

Leveled and sanded smooth.

The hole on the belly side takes considerably longer. I add just a bit of wood dust, then compact it down with the stem of a q-tip that has had its head removed. Glue is dribbled in a bit at a time to soak through, after which I pack down the wood dust again and use a paper towel to wick up any puddling glue before it dries. The whole process gets repeated again.

Partly filled...

...fully filled...

...being filed...

...and finished!

With luck, the filling process will provide some sort of structural support -- but I'm not going to count on it. I am learning that with bows, as each new problem reveals itself, although we may take measures to remedy it, we are constantly flirting with disaster. I feel like Odysseus, trying to navigate between Scylla on one side and Charybdis on the other. Let's hope things turn out better for this bow than for Odysseus's men...

Yew "fence post" longbow, part 7 - resplicing

I decided to bite the bullet and cut the splice apart. I've outlined the steps below, just in case anyone is crazy enough to want to do so too!

Step one was to mark out the areas to be cut, on either side of (and parallel to) the current glue lines. By marking the new lines out a mere half inch on either side of the end of the splice along the center line, I should only end up losing about an inch in total from the length. One can note from the images that this results in a narrow (but reasonable) area in which to cut the slanted parts of the splice. By my reckoning, I should be able to cut exactly along the glue line, then pare back to the penciled lines and have a new splice joint pretty close to fitting spot on...

...but first, I'll need to drill a hole just big enough for me to snake a coping saw blade through! The hole is drilled not at the very end of the splice, but in my closest approximation to what will be a waste area (based on the penciled lines). This means I will need to saw in both directions, but I want to have as clean a spliced line as possible.

Starter hole from the back.

Starter hole where it exited the belly.

Fortunately, I was able to drill a pretty straight hole (having a wood bit and marking the start with an awl helps). Below you can see the coping saw "in action." (Actually, I had to switch hands and am not really sawing here, as the other hand is holding the camera -- but you get the idea!) This is followed by an image of the finished "inner cut."

Things got a little jagged running the other way, as I was using my left hand. It got a little detracked on the belly side, so I had to recut a section there. No worries -- I'll fill that tiny section with a wood offcut, and the glue should patch it right up. I finished up the last bit with a Japanese pull saw, as it had no back and just fit in the slot. This worked great, as it was a lot sharper than the coping saw blade, despite the latter being new. It also did wonders on cutting the slants quickly but accurately.

Below is the cut-apart splice, waiting to be trimmed up. (The blue on the interior of the right piece is paint or something from the writing on the Japanese saw. It will come off in the cleaning up.) One nice thing is that by following the previous glue lines, the resulting splice angles should line up pretty accurately and it should only be a matter of needing to refine the surfaces slightly. You'll notice from the following image that I biased my cuts to one side of the glue line so that I could just focus on cleaning up the outside edge on each joint instead of working hard to reduce the interior angles.

I apparently got caught up again in the doing rather than the recording and forgot to take additional pictures of the finished splice during fitting and the glue-up process. Suffice to say, it took a little bit of finessing and I had to use some yew offcuts to fill the gaps (most significantly, the saw kerf from the "inner cut" on the center line, as failing to do so would have shifted the centerlines of the two limbs fractionally). Mostly I think it turned out okay:  I only lost an inch (the stave is now 75 inches overall), but I did decide to add a tiny bit of reflex by joining the limbs together at a very small angle. It is probably not enough to be noticeable, but every little bit counts. I daresay that the joint looks a little more solid than it did the first time, so maybe being forced to redo it was providence?!

Yew "fence post" longbow, part 6 - handle "set back"

Normally when bowyers speak of "handle set back," they mean the shaping of the bow in such a way that the handle is closer to the archer, so "set back" from the back of the bow. In this case I mean "set back" in the sense that ordinary folks use: a failure followed by a temporary delay.

I was feeling pretty good about heat treating the belly and worked from both tips toward the handle. The tips seem to have taken a fair amount of twist correction, and some small amount of reflex was materializing. I thought the glue (Unibond 800) was not heat sensitive--but it turns out I thought wrong. The glue had darkened and bubbled in spots, reminding me of the crunchy foam structure of chocolate honeycomb candy my grandmother sends from the East Coast. On inspection, it looks like I've cooked the glue and weakened the joint. A quick search on-line shows that yes, Unibond can be degraded and weakened by heat. Unibond is a great glue, and frankly I should know better: just about any glue will weaken under any temperatures high enough to scorch wood.

Cooked glue joint!

First rule: Don't panic! All is not lost: I can cut the handle apart and splice it back together over again, but it is going to be work. For one, I could do a neat cut in half and totally recut the splices, but it means losing 4 inches in overall length, going from a 76" stave with 74" nock to nock length to a 72" stave with 70" nock to nock length. That's a considerable loss, especially considering that I fought the initial stave to retain every bit of length I could. I could try to recook the glue and hope it fails then clean up the edges and reglue, but I'm afraid that in doing so I will either a) ruin the wood by overheating or b) not be able to weaken the glue enough to simply pull it apart with my mere mortal strength. Even if I could, I'd still have to clean up the edges and get rid of any trace of the original glue, which could be problematic.

There is a third way between these two extremes, though I'm not happy at the prospect: I can cut the splice apart along the glue lines, refit the joint, and reglue. It will be time consuming and would actually mean cutting a new splice in fresh wood just outside the glue line, then cutting a matching splice in the other limb in such a way that I pare out all of the old glued sections. The advantages are being able to glue a brand new joint while retaining most of the current length (if thoughtfully done, I might in theory only lose a half inch). On the other hand, cutting the center spliced section means having to drill a small hole in the middle of the stave in order to snake a coping saw blade through. There will also be even less room for error, since the stave has already been taken down to final width at the handle. (I might end up losing some there...)

This is not the happiest thing that could happen, but I will say this: I have learned an enormous lesson about glue and heat, and it has already changed my mind about how I want to approach my next bamboo backed osage bow project. This will be a lesson I won't forget -- and I'm sure I'll learn something new in the cutting and re-splicing process.

Yew "fence post" longbow, part 5 - heat treating

Not a lot of pictures for this one... I guess I got caught up in actually doing rather than documenting!

At this point I'm heat treating the limbs for a few reasons:

1. To pull the twist out of the tips of the stave.

By using clamps and weights to create a lever to twist the limb in the desired direction, it is possible to take out some of the twist at the tips. I use the approach of not forcing the limbs against a form. Rather, by using the clamp and weights levered out at the end, the tips are gently coaxed into position as the wood is heated and relaxes. The trick is to try and coax them a little beyond where one wants them eventually, as some of the wood has memory and springs back a little bit. One can see the clamp and the weights in the picture below.

2. To try and add a little bit of reflex into the stave, as it is fairly straight limbed and slightly deflexed in areas.

I know that adding reflex can make a bow more difficult to tiller, but I'm not anticipating adding much: just enough to help make up for some of the inevitable string follow that will happen. As the bow takes set from being worked in, the limbs will start to bend toward the archer. If I can give the bow a little reflexing, the degree of string follow will be diminished in the finished bow (though the amount of string follow may be the same -- it just starts from a different place). Consequently, performance should go up, albeit only slightly.

3. To try and create a little more performance by altering the belly wood. 

Heating the wood on the belly, tempers it and makes it more resistant to compressive forces, which can also slightly increase performance by two means: 1) increasing the draw weight per overall mass or allowing a lower mass for a given draw weight, and 2) forcing the belly to do more work. For the latter, wood is usually stronger in tension than in compression, so having the belly do a little more work can help balance out the load bearing. One downside, though: I recognize that some of the effects of heat treating the belly will be diminished by the fact that the limbs are not at their final thickness, so some of the tempered parts of the belly will disappear under the rasp and scraper. Why not wait until later? I plan to back the bow with rawhide using hide glue, so I don't want to soften the glue with heat and have the backing come off. As with many things, bow building is a matter of compromises!

Yew "fence post" longbow, part 4 - profiling

Now that the limbs have been spliced together at the handle, it's time to survey what lies in store and plan the bow accordingly.

Right now the overall length is 76", so will likely yield a bow around 74" nock to nock. This is good, as I'd like to tiller this as an English "warbow" of lighter weight -- maybe in the 60-70# at 30 (or 31? 32?) inches. The extra length will help here. Although it is meant to be a bend-through-the-handle bow, I'll want to stiffen up the handle section a little so that no undue tension is put on the spliced joint.

First, the back. One can see several of the flaws in the sapwood. Because of the narrowness of the resulting stave, there really isn't much room for avoiding some of these problems. The sapwood is also not very thick, so reducing it in thickness would not leave enough to make a decent selfbow, so I'll plan to back this with light colored rawhide, which will allow the appearance of a yew bow with the cream/brown contrast between the back and belly.

The twist at the end of the limbs can be seen fairly clearly now. The two photos below show the orientation of the crown at the tips compared to the center and middle part of the limbs (oriented correctly, with the crown centered).

The twist is not that great, but enough that I don't want to be wrestling with it while I'm tillering the bow. It is also worse at one end (the good end) than at the other. I'll plan to heat treat the belly some, and in the process I'll use some clamps and weights to gently coerce it to falling in line with the rest of the bow's crown.

Problems await on the other side, as the two knots I mentioned before could not be avoided. One is smaller than the other, and it may quite likely come out in the reducing of the belly thickness. The bigger one goes through to the sapwood side (but just a touch) and looks pretty solid so far. Hopefully I'll just be able to leave it along (or shoot it through with a spot of thin super glue) and get along nicely. The third image below shows the two knots in relation to each other; the smaller one is closer to the tip.

After lots of fussing, I managed to capture the center and crown of the stave. I mark it out using a line with chalk. The chalk is bluish-purple but comes off readily with water (or a little bit of sweat from one's palm). I use that as a rough but straight line, then follow up by marking over that in pencil, which does note rub off as readily. Then I marked the full width of the stave, 30mm at the center 4" section of the handle and 20mm at the tips. (Like many others, I switch between units of measure as a matter of convenience, marking out in mm when more precision is needed.)

You'll notice some dotted lines near that big knot. The dots actually follow the grain, and I'll respect that when I cut out the shape, leaving a little more wood around either side of the knot to help compensate for any weakness this section of the bow has.

When cutting out on the bandsaw, I did my best to keep the top of the crown uppermost, rotating the stave as it went through. In this way, the wood at the sides is always oriented perpendicular to the back, with the crown centered. It makes for some glaring twist, now that the excess wood has been taken away, but again, it will (hopefully!) come out when I heat treat and twist the tips into alignment. (In reality, they get twisted a little farther, as wood has memory and they will want to drift back to their previous position a bit.)

The true extent of the rot damage is now showing itself. It travels the limb a fair degree, but that will all be waste wood at the end of the day as the tips finally thin out, as can be seen in the second photo where I've marked the intended starting depth of 20mm at the tips.

Other sections are going to cause some challenges. Here's a section where at one time there may have been a limb. Now it is just an area where the grain runs in a divot and has cut into the working part of the limb. Some will be removed in initial thicknessing (see the pencil lines), and with luck the rest will as the belly is being shaped and tillered.

Similar challenges abound in the spliced handle section. The bow will be thickest here, but it is also where there's the least amount of wiggle room with the remaining wood. Pencil lines again show what will be removed, but there are still some significant sections of beveled wood remaining. 

[...time elapse...]

Finally the stave is brought closer to working depth by sawing off the waste wood outside the pencil lines. The dimensions I'm starting with are 20mm by 20mm at the tips and 30mm by 30mm at the handle section. We can't build a bow by numbers, but this is a starting place so that I can try and get the bow bending a bit in floor tillering. A quick weigh-in finds the stave at around 36 oz. -- which is about twice as much as it will eventually be if aiming from 70# @ 30", according to Steve Gardner's Mass Principle.

In working down the limbs, one can see that the small knot is close to disappearing. Right now it is a tiny hollow bit, which will likely come out altogether as the belly is reduced. On the other hand, that large knot has gotten slightly bigger -- and a whole lot more rotten, as the solid part came off with the waste! Just goes to show that one never really knows what lurks beneath.... I'm not sure how far the punky bits go, but I'll no need to plan to scrape out the bad bits and either a) fill the hole with a glue and sawdust mixture or b) make a wooden plug. In either case, since the knot is slightly visible on the sapwood, I'll have to assume that it is nasty all the way through.

Yew "fence post" longbow, part 3 - splitting and splicing

Laying out lines takes a lot of careful thinking and planning, checking and rechecking -- especially when there is so little room for error as there is here. On the other hand, bandsawing is relatively quick but requires an intense amount of focus when the stakes are high. One momentary lapse of attention, and the stave could be rendered into firewood -- and my planning for naught.

Well, the moment came and went, and here are the two much thinner staves. One can see the rotted section now in a good section of the stave for coming off in the profiling. From the long view, one can also see the slight jog I had to take to stay within the center of the stave.

Splicing would be a great deal easier if the wood was square or rectangular in cross section, in which case they could be stacked and cut in one go. Instead, I've marked out the waste areas on the sapwood and will need to cut each part carefully, keeping my saw lines totally perpendicular to a theoretical flat back (which is, of course, convex!). You can see the general steps below.

Once cut out, the two sections fit together fairly decently, though not perfectly. Any gaps can be filled with wooden shims so that I'm not depending solely on glue to fulfill any functional aspects.  Here are the sections dry fit with an offcut shim of yew in the background.

And here's the finished glued joint after cleaning up any of the squeeze out!

It all looks rather quick and easy here, but it was probably an hour in the marking and cutting, another hour or two in shimming, cursing, and finessing (not all necessarily in that order),  a half hour to glue and clamp, 24 to dry, and 10 minutes to unclamp and clean it up. Whew!

Now that the splice is complete, I can see that there is a bit of propellor twist at the very ends and a very slight amount of deflex. I'll plan to heat treat the belly to see if I can't get a bit of reflex into it while correcting some of the twist at the tips.

Yew "fence post" longbow, part 2 - laying out

After cutting off 4-inch section of rotted wood, it was still possible to see some rot penetrating into the stave. It looks, however, as if most of this will come off in the initial shaping and subsequent tillering, as it is fairly far from the sapwood. It also means that this section will have to be one of the outer tips if I am going to split and splice the stave back together -- which is a bit of a shame, since that end of the unsplit stave is slightly wider and would leave me more room for error in creating the splice at the handle.

Rotted end, before removal.

Remaining rot. How deep does it go?

Width of good end.

Width of rotted end. 

In these two images, one is able to compare the width of the two ends. The good end is almost a half inch narrower than the rotted end, which does not leave much room for error if I'm going to rip this into two sections to be spliced back together. Even though that makes me slightly uncomfortable, I'm going to do it anyway. How else will I learn something?

You can see from the photo at left what the ring count looks like. I mentioned in my last post that it was fairly low. Here I count between about 10-12 rings per inch. The wood, however, feels pretty heavy compared to other yew I've handled. It's kind of hard to tell from the photos, but the heartwood is very tan -- almost with a gray-green in some spots as opposed to the very reddish or orangeish yew I am used to working.

Since the wood has a lower ring count, and because the stave has pretty much no reflex (one part is slightly deflexed?), I'm planning to give the stave some heat treatment (probably after it is spliced and roughed out but before floor tillering), at which point I can induced a little relfex.

One complication is the fact that the stave, as it stands right now, jogs a little to one side at about 20 inches. In addition, there is a slight amount of propeller twist. It looks like I won't be able to lay this out with a straight cutting line along the back. Instead, I've decide to cut it straight until the stave starts to show the job and twist, at which point I'll change my cut so that it stays down the middle. (I could split this, but I worry about the lack of wiggle room. Even doing this on the bandsaw will only give me just the slightest amount of room for goofing up at the handle end. I figure I need to end with 1.25" at least to make this work.) You can just make out the jog and twist in the photos below.

Jog and twist.

Jog and twist.

One advantage to splitting the stave is that some of the bigger gouges in the sapwood are going to end up falling to one side or the other of the new thinner splits, and will eventually be worked off the entire limb as the front profile is brought in. You can see one such example in the circled section in the photo below. I think part of good bowerie is simply taking the time to really think through what the wood has to offer and how can can intelligently work around problems. An additional set of problems surrounds the two knots I identified in the last post. After splitting this stave, both should fall squarely within the middle of one of the limbs; the other limb will be knot free. I'll plan to make the knotty (naughty?) limb the upper one, since it should be slightly weaker and under a little less strain, anyhow.