Celtic Knot – #4 Final Inlay Assembly & Pen Build

After letting the inlay’s glue dry for three days it gets unclamped and the edges of the layers all get cut flush.  The top two photos show what that looks like.  The bottom photo shows the layout of the inlay blank which will give me ten pieces.  Four of them are needed for each blank plus one in case there is a problem like the chip in the lower right corner of the piece.  To minimize waste the pieces get cut out with the bandsaw.

Once the inlay pieces are cut the same process for adding them is followed.  Here the first one is dry clamped for a test fit.  As a check I made a diagonal line across the piece before cutting it.  If the inlay thickness matches the saw kerf, then the line should line up across the joint which it does so that’s encouraging.  Next is gluing and clamping this inlay piece in place.

The same process of cutting and gluing the inlay is done three more time ending up with a blank whose four sides look like this with everything aligned.

Next the blank’s excess length is cut down to a couple of inches or so longer than the piece needed for the pen.  The end centers of the piece get located using the same method shown earlier and the piece gets turned round.  Next, a hole is drilled down the center of the blank just like the screwdriver was.  Last is to part off the blank so it’s just a little longer than the brass tube which gets glued inside the blank.  Once all that’s done this is what the blank looks like along with the tube that gets glued in.  The tube can be glued in using CA glue, epoxy or polyurethane.  I like to use polyurethane because as it cures it expands filling any gaps between the brass tube and the wood blank.

After the polyurethane cures overnight, the excess is cleaned off and the blank is put in a mandrel with bushings whose small diameter is inserted inside the brass tube.  The top photo shows all those pieces laid out.  The bottom photo shows the assembly mounted in the lathe ready to turn the blank to a true cylinder whose axis is parallel to the centerline of the lathe and the brass tube.  This step is needed to take care of any out of alignment or drifting of the drill bit while the hole was being drilled.

Once the blank is trued it is removed from the lathe and the disk sander is used to bring the wood down to just a tiny bit longer than the brass tube.  The inset shows how close the wood is to being flush with the brass tube.

Next a trimming tool is used to cut away the wood so it’s flush and square with the brass tube.  The top photos are of the blank and the trimming tool.  The tool can be put in a hand drill and spun to trim but I like to do it by hand so I can feel if there is a problem to avoid chipping out of the end.  The bottom left photo is right off the sander before trimming and the right photo is after trimming.  When I can see the shiny brass edge of the tube it and wood are flush.

Now that the blank is trimmed to its final length it gets mounted back on the mandrel in the lathe so the final turning can be done.  The photo below shows that with the inset photo showing how much wood needs to be removed to get down to the steel bushing.  The bushing matches the diameter of the finished pen so when the wood matches its diameter the finished blank will be flush with the pens exposed metal pen parts.

For a finish I decided to spray on three coats of lacquer.  The top photo shows the mandrel taped to keep the overspray off of it.  After giving the lacquer a few days to cure the tape comes off and the mandrel goes back into the lathe for a little work on the finish to remove any dust that settled in the lacquer before it dries.  That’s done using a 12,000-grit wet sanding pad and the lathe on a slow speed.  The bottom photo shows the piece all done.

Assembly is next, here are the hardware parts that go into the pen.

The pieces are pressed into the brass tube inside the just finished wood piece.  I use a bar clamp (top photo) with plastic covers so to not scratch or mar the finished metal parts.  First is to press the top and its clip into the tube. The middle photo shows it almost in place.  Once that’s done the lower assembly which includes the retracting mechanism for the ink refill is pressed into place shown in the bottom photo.  

Once that’s in place the ink refill is put in and the pen is done.  Well, at least I thought it was done.  When I tried to twist the tip end on the left to bring out the ball-point end it would barely budge.  Since it was so hard to turn, I used a rubber gripper and twisted the mechanism back and forth probably a hundred times to see if it would loosen up with minimal success.  This really surprised me as I have never had a problem like this with all the pens I have made.  Checking an identical mechanism, it worked just fine so the problem is limited to the one I used.  Unfortunately, once the parts of this pen are assembled, they can’t be disassembled without destroying the pen so swapping out the problem mechanism is not an option.

Upon close inspection it looked like the wood to chrome tip joint might be the problem.  Working under that assumption I took an X-Acto knife and tried to take a little off the layer of wood that’s glued to the end of the brass tube.  This layer is only about 3 hundredths of an inch thick so I had to be very careful.  Well, while removing the last little bit the knife slipped making a scratch all the way through the finish about ¼” long down the face of the pen’s body.  If that wasn’t bad enough when I screwed the tip back on the mechanism didn’t work any better.  The mechanism was as stiff as ever and now the body of the pen had a big scratch in it.  As a last-ditch effort, I squirted WD-40 down inside the mechanism and let it soak for a couple of days to see if that might help.  Fortunately, it did improve some so I added more and let it soak for another couple of days which loosened the mechanism up enough to make the pen usable. 

Now all I have to do is fix the scratch in the pen’s body.  Since the pen can’t be taken apart it’s a little harder to fix the finish.  The solution is to take a scrap piece of soft wood, turn it round then drill a hole so the cap end of the pen fits snuggly in it and mount the pen in the lathe.  That’s shown in the two photos below.

Once in the lathe I could very carefully spin it at a low speed and sand the finish off back just beyond the scratch feathering a smooth transition to the undamaged lacquer finish.  With that done the threaded tip end gets masked off and three coats of lacquer get sprayed on to the end without getting any on the pen’s clip.  After the final coat of lacquer is applied and cures for a few days the pen gets mounted back into the lathe and wet sanded with 6,000, 8,000 and 12,000 grit pads.  After that some very fine polish compound is used to finish blending the lacquer layers together.  Below is the finished pen which looks just like it did before the scratch.

While waiting for the lacquer to cure I went ahead and glued up another blank for future use using cherry for the main body this time.  Here is what the blank looks like with the four pieces in place and turned round.  It’s a little bigger in diameter than the one used to make the pen which is why the knot is a little more open.

Once all the problems had been resolved the Celtic Knot looks a lot more complex to make than it actually is. 

Celtic Knot – # 3 Turning 1st & 2nd Test, Gluing Inlay Blank

After the cherry and maple glue joints for the fourth cut have cured as before the maple is cut down so it is flush with the cherry.  Looking at the four faces of the blank the good news is the left edge of the maple is aligned like it’s supposed to.  The bad news is the right side is nowhere close to what was planned.

Since I have finished the four-stage glue up for the blank I might as well go ahead and turn it round to see what it looks like.  This starts by cutting off a 3 ½” long piece, finding the center with the center finder and making a small dimple (red arrow) at the center.  This will be used to locate the live center on the lathe.

After turning the blank round to ¾” diameter the finished faces look just as weird as the square blank did.  The photos below show the blank rotated a quarter of a turn in each photo.  The top photo looks the best as that’s the one from the result of the 2nd cut.  From there each one gets more and more distorted.

In looking at the ¾” turned blank I wondered what it would look like if it got turned thinner and here are the results.  There are two columns of photos below.  The left column shows the same face as it is turned down in 1/8” increments from 5/8” diameter down to ¼” diameter.  Each row is the same size with the right photo showing the face on the opposite side as the left photo.

Next is to make another test piece using the piece of inlay maple that exactly matches the width of the saw kerf explained at the end of the last post.   The process of cutting and gluing is the same as before so no need to go through that again.  Here is after the fourth and last cut where three maple pieces have been glued in place.  It looks pretty good with the maple pieces lining up like I had expected.

When the fourth and last maple piece glued in this is what the four sides of this test piece looks like.  All of the vertical joints line up and all of the diagonal pieces cross in the center of each face which is what it’s supposed to look like as compared with the first test piece shown at the top of this posting.  My original plan was to use this test piece on a slimline pen but the diameter of that style of pen is pretty small at about .3”.  In looking back at the first test piece when it was turned down to about that diameter the majority of the cherry shapes within the knot were lost.  Given that this test piece has thicker maple pieces I think that all of the cherry would be missing from the knot’s interior as would much of its detail.

This meant that I needed to come up with another use for the blank.  Looking for something to use the test piece on my wife mentioned she had a small screwdriver that’s frequently used in the kitchen not as a screwdriver but for odd jobs of prying things open and the like.  It looks similar to this.

Now that I knew what the test piece is going to be used for the blank gets cut down to the handle length plus some for chucking in the lathe.  Sequence for turning is to locate the blank’s center, mount opposite end in the scroll chuck then bring the live center to the marked center of the blank and tighten everything down.  Once that’s done the blank is turned down to ½” in diameter with a slight taper added to the blade end.

Next, a hole is drilled in the end where the screwdriver blade’s shaft will get epoxied into place.

The last bit of turning is to round the head of the handle.  This is done with a light touch so as to not break off the handle.  When done a saw is used to cut the handle free.  I don’t want to try and turn the handle completely off as it’s likely that when it breaks free there will be a little divot in the handle that would have to be addressed.

After the handle is cut off the screwdriver blade is inserted in it and the assembly is put in the drill chuck where the end of the handle gets sanded smooth.  From here all that has to be done is to spray a lacquer finish on the handle followed by epoxying the screwdriver blade in place but that’s going to be put in hold for a bit while I work on gluing the maple-katalox-maple inlay glued up. 

Getting the maple and katalox thinned down so the total of the three layers are really close to totaling .132” is next.  That’s done by setting the pieces on a carrier and running then through the thickness sander taking a little bit off with each pass.  The maple pieces are first since they are the thinnest pieces.  Below the top photo shows one of the two pieces needed brought down so they are only three hundredths of an inch thick.  That’s about as thin as I am comfortable making them at this point.  The center piece is katalox which is shown in the bottom photo.  Together the three layers total up to being about three thousandths of an inch more than the target thickness. 

Gluing the three pieces together to see what I actually end up is a little bit of a challenge because as soon as the glue is applied to the thin pieces, they start absorbing moisture from the glue and start to curl.  It’s a little like herding cats to get them all aligned, flattened and clamped in place.  The top photo is of how they are clamped together and the bottom photo is a close view of the left end.  Now it’s a matter of letting them set clamped to cure.  That’s going to take a few days as the cauls used in clamping have an impervious surface which will slow down the curing.

Next Up – Final Inlay Assembly & Pen Build

Celtic Knot – #2 Proof of Concept & a Problem

The proof of concept starts by taking piece of cherry for the body out of my “save-for-later” box and milling it to just under ¾” square and about a foot long.  A piece of maple from the same box is milled to the same thickness as the planned total thickness of the three-layer stack. 

The first step is to set a jig up on the table saw to make the 40-degree cut through the body.  To set the angle an adjustable triangle from my drafting days and a Sliding T-Bevel are used.  The photos show how the triangle is set and held against the table saw rip fence to set the Sliding T-Bevel to 40-degrees.

Using the Sliding T-Bevel the fence on the jig is set to the 40-degree angle making sure the T-Bevel’s edge is tight to the blade without touching any teeth and the handle is up tight to the jig’s fence. 

Next the jig’s stop block is set so about 2” of the blank’s body will get cut off.  It’s important that this setup is not changed from here on out so all the pieces that make up the knot are properly aligned.  A sliding clamp is used to hold the blank securely in place while the cut is being made.

Gluing up is done using this jig to keep the cherry parts aligned.  The jig’s base has the top layer covered with packing tape and the top backstop is covered with a vinyl surface neither of which the glue sticks to.

To use glue is applied to the cherry pieces then the maple piece is set between them hanging over on the far and near faces of the cherry.  The two horizontal orange clamps pull the cherry tight to the backstop and the blue clamp pulls the three pieces tight together.  The notch in the backstop lets the maple piece extend beyond the face of the cherry without interfering with the cherry faces alignment. 

After letting the glue cure overnight, the clamps are removed and here are two views of what the piece looks like.  Note that the maple piece that was down during glue-up (side 3) is flush with the cherry while the other three sides of the maple piece are proud of the cherry.  You can also see numbers on the cherry faces.  Each of the four faces are numbers one through four and I will explain what they are for in just a bit.

Before moving on to the next step all the edges of the maple have to be brought down flush with the cherry.  That’s done with a no-set pull saw followed by some light sanding using a sanding block with 120 grit paper.

Making the cut for the second piece of maple is next.  Here is where the numbers come in.  The blank is put back in the jig butted up against the stop with the number two up.  The first piece was cut with the number one up, this time the cut is made with the number two up.  The same process will be used later for faces three and four. 

Once cut the glue-up process is the same as with face one.

After the glue cures the blank gets unclamped and the maple edges are trimmed flush with the cherry body using the no-set pull saw followed by some light sanding.  The top photo shows faces one and two with the meeting corners (red arrows) of the maple and cherry all lining up as expected.  However, in the middle photo one of the joints did not line up as expected.  The vertical maple piece on face 4 identified by the top red arrow should be over to the right so it meets the vertical piece identified by the bottom red arrow.  Now that’s a problem which when the cut was made for face three really showed up since the vertical maple pieces on the cut face (circled in red) should line up and they clearly don’t.  At first, I was at a loss for why since the 3D drawing did not show this as a problem.  After a thinking about it for a while I think I know what happened.  Because I wanted the individual lines of the knot to be thin the thickness of the maple piece is .085” thick but the saw kerf is listed on the blade as .125”.  This means that each time I make a cut and glue the maple piece in there is a net loss of .068”.  I know the difference between .125” and .085” is only .04” but due to the angled cuts the actual loss is .068”.  This matches up with the photo in that the offset is slightly less than the .085” thickness of the maple piece.  By the time I make fourth and last cut I think the difference will be a lot more and any possibility of getting a Celtic knot out this test piece is gone.  However, I am going to go ahead and finish it up then turn it round to see what the results look like.

Here the top two photos are after the third piece of maple has been glued in place and brought down flush with the cherry.  The bottom photo shows after the cut has been made so the fourth maple piece can be inserted.  As you can see the alignment of the maple pieces is getting more and more out of whack.  It will be interesting to see just how bad the pieces will be misaligned after the fourth piece is glued in place.

While the glue was curing, I made a cut on the table saw in a scrap piece using the same blade as I use when making the four angled cuts above.  Now the blade says it makes a .125” kerf but having been burned once when the maple inlay didn’t match the blade kerf, I wanted to make darn sure it matched this time.  The top photo below shows what thickness piece of maple fit in place with no slop and it’s not .125” but is a little thicker at .1305”.  The bottom photo shows the piece milled to fit in the saw kerf.  Now I know the difference is only about five thousandths of an inch but as there are four cuts the total difference is .020” and that’s about the thickness of my fingernail.  Nevertheless, because the finished Celtic knot is going to be less than a half inch in diameter, I am concerned that the misalignment would stand out if I used a .125” piece.

Next Up – Turning 1st & 2nd Test, Gluing Inlay Blank

Celtic Knot – #1 The Start, Design & Material Selection

Over the years I have borrowed designs from one medium to another like from Native American pottery and Oriental Bronzes into wood turning or Frank Lloyd Wright style textiles into stained glass.

I have always liked the Celtic designs that run from fairly simple to really complex and had thought about trying to incorporate one as an inlay into a piece.  The drawing on the left looked like a possibility and after seeing examples (right) where that design had been incorporated into a pen turning, I decided to explore how it was done.  After seeing a short live stream from Record Power out of New Zealand and a little research it looked like a project to tackle.  First to give me a better understanding of how the 13 different pieces interact to make the completed design I did some 3-D drawings varying some different parameters to see what I liked the best.  The bottom rendering is what I settled on or at least what I am shooting for.

Next is to pull out the possibilities for the various pieces to be used.  For the body cherry, mahogany and mesquite were considered.  I decided on using the mahogany as it’s a good turning wood whose grain will not compete with the Celtic knot.  The light banding wood will be maple and the dark center piece between the maple banding is walnut.  I want a really dark wood for the center and walnut may not be the best choice.  As I get into the build and some mockups get done, I should get a better idea of what looks the best.

First is to cut a piece off the mahogany board.  Here it’s set up on the chop saw ready to go.  In this photo you can see a red arrow that indicates the diagonal grain direction in the piece I want to use to make the pen blanks. 

To get straight grain blanks the piece gets cut on the bandsaw at an angle matching the grain direction.  I will get two blanks out of this piece with each of them long enough for two pens.

Ripping the blank using the bandsawn edge against the table saw’s rip fence for reference gives me a good straight generally parallel edge to the bandsawn edge.  That’s because the bandsawn edge is a rough and not perfectly straight.

To get straight parallel edges the newly cut edge is put up next to the rip fence and the bandsawn edge gets trimmed enough to remove the bandsaw blade’s rough surface.  This cut edge gets flipped and the first face gets a skim trim cut.  Usually this is enough to bring both faces parallel and flat but if not, the process can be repeated again.  Last is to rip the blank so its width is the same or a shade wider than the thickness of the blank.  The top photo shows that completed while the bottom two photos show the bandsawn edge before and after trimming on the table saw. 

While the width and thickness of the piece is close it’s not close enough.  To make the finished knot look right those two dimensions need to be very, very close which is done with the thickness sander.  The photos below show that after being run through the thickness sander the width and thickness only differ by two thousandths of an inch.

Working on the three-layer maple/walnut stack is next and it’s at this point a change is made.  Looking at the walnut piece I was worried that it may not be dark enough to provide the desired contrast between it and the maple.  So, it was back into the wood stock to see what other dark material was available.  In addition to the walnut (bottom sample in photo below) four other woods were potential candidates, Mesquite, Katalox, Wenge and Ebony.  Rather quickly I narrowed it down to the Katalox and the Ebony either of which will work better than the walnut.  I decided to go with Katalox because when finished it goes almost black with just a tinge of dark red which should work well with the mahogany body of the pen.  It is also already the width and is just a little thicker than I need.

Starting with the maple which it is just a little wide and thick.  Once it was ripped to the correct width it gets cut in half to make the two outer layers.  Because the piece is just a tiny bit longer than the piece of Katalox I have rather than use the chop saw to cut it in half a Japanese pull saw it used since its kerf is much thinner than the chop saw blade.  To get a square cut a reference block is clamped to a jig.

The two pieces start out at a little over 1/8” thick and are brought down to 3/64” which is closer to the final thickness.  Here the photo shows one of the initial pieces ready to go.  The carrier it’s on does a couple of things.  First, it provides for a solid larger supporting base as the piece gets thinner and thinner.  Second, the piece of cherry on the right end of the carrier gives a positive stop that keeps the maple piece from slipping while being sanded.  Because the pieces are so thin, once sanded to thickness I clamped them between two ¾” thick pieces of oak to keep them from curling and set them aside.

Next up is to start thinning the katalox using the thickness sander to bring it down near to its final thickness.  Just as I got ready to make its first pass through the sander it occurred to me that before I go through the process of getting the maple and katalox to their final thickness and gluing them up into a three-layer stack it would be a good idea to do a simple proof of concept to see if the process would work as planned or not.

Next Up – Proof of Concept & a Problem