Sunday, December 25, 2016

Project #16 Handle for Watanabe Mukimono

This project was commissioned at the same time as the project #15, but here I was given a 'free hand' on the handle materials and design. I could not resist to try something new - see below how did that work out :)

The knife in questions is a small mukimono made by Shinichi Watanabe. It had a 'standard' burned chestnut D shaped handle and so the handle removal was very simple, since these are not glued on, but just burned-on the tang.

Design & Materials

For a long time I have concentrated only on natural woods, but I was always attracted to dyed materials. I have recently bought a few blocks from different sources and this handle was a good opportunity to test one of them.

I have also wanted to try water buffalo horn (a 'standard' ferrule material on traditional Japanese handles) as well as some metal spacers. So the final selection was:
  • stabilized & dyed maple burl (from stabwoodlab)
  • water buffalo horn
  • black fiber paper spacers
  • 1 mm thick nickel silver spacers

Starting point - original handle and materials for the new one.

The handle block

Before I did the easy part - square the wood up, clean the gluing side of the horn, cut the pieces of fiber paper and nickel silver and glue them all together - I considered different options on how to deal with the fact, that there are metal spacers that need to be drill in some way. 

  1. One option would have been to drill all components beforehand and glue them directly with a dowel (that would not have a tang slot at that point) and then drill the tang slot afterwards. Doable, but because of the thinness of the blade the making the tang slot would have been a lengthy job.
  2. I have decided to go with second option - glue all the components together and do the drilling afterward. I will discuss the pluses and minuses later. I have used BladeBond Ultra epoxy for the first time here (will be relevant below).
  3. Third option that did not come to my mind would have been to drill the holes in the horn, nickel silver, fibre paper and a shallow hole in the wood, than make a shot (1 - 2 cm) dowel from soft wood that would only serve to align all the parts and would be later drilled-through with the same drill bit to allow to mount a slotted hardwood (beech) dowel.

About to glue it all together.

Gluing the handle block
Ready for the next step.

Glue excess ground off, ferrule shaped to align with the rest of the block.

Once the handle was glued I cleaned the block up (remember, it was already squared up before gluing) and drew centre lines on all sided, top and bottom. I also gently tapped the bottom to keep the track of where the centre was.

After all lines were drawn I realized that I forgot to square-up the bottom of the handle. The point of doing so was to make alignment in the vice for drilling easier.

Squaring up the bottom side of the handle.

Drilling the opening for the dowel

Note: Since I did several unfortunate choices in this part of the process, I have decided to describe it in more detail.

As mentioned above - I have decided to do the drilling after all the components were glued together. The main issue to deal with were of course the metal spacers. Even though not too thick, these still needed to be drilled with different drill than the one I normally use (drill bit for wood with the guiding 'tip').

I have first drilled a 'normal' hole with 10 mm diameter (first mistake, I should have started with a much thinner drill bit and use the hole in the horn as guide while drilling through the metal spacers) with the drill bit for wood and swapped to a drill bit for steel (second mistake - apparently not a really sharp one) once I reached the metal spacers.

Starting to drill - so far so good.

I have underestimated (third mistake) the amount of heat that will be produced while drilling through the nickel silver spacers. When I was half way through the first spacer I noticed that the handle block is loosing its mechanical stability and a few seconds later the horn cleanly delaminated from the rest of the block. Bummer :(

And the ferrule went clean-off when the drill bit
touched the first spacer.
Just drilling a little more to see how the metal spacer holds on the
underlying fiber spacer. Poorly.

Close inspection showed that the glue detached from the surface of the nickel silver (which I cleaned with acetone, but - fourth mistake - forgot to roughen with sanding paper and left it glossy polished).

Checking out the second nickel silver spacer which was sandwiched between to fibre paper spacers showed that this one was already delaminating, even though it had seen only a very short drilling (just the tip of the drill bit). I was able to detach it from both fibre spacers - again - there was no glue left on its surface.

There is not glue attached to the nickel silver. The scratchy spots are
from flattening with a hammer prior to gluing.

At this stage I should mention that I have used the BladeBond Ultra epoxy (a dedicated epoxy for knifemaking) which I did not test before with the materials in question (fifth mistake - untested glue). Afterward I have used the G/flex with the two spacers just to see how they will hold and looked fine, but I did not try to drill them.

So to save what still could be saved I sanded the remaining fibre spacer off the wood, sanded both wood and horn and glued them together.

Handle shaping

Once the glue cured I have proceeded in the standard way. I prepared a dowel from a 10 mm round beech, cut the tang slot with a hack saw, widened it with a belt sander and #120 belt until it would fit well on the tang and glued the dowel inside the handle block with the G/flex epoxy (wood glue would have worked as well). I have briefly inserted the tang inside and made sure that the blade is under right angle against the upper and lower side of the block. Indeed - a slight rotation could be corrected later, but it makes your life simpler if you do not have to correct this.

Note: Do not use too much glue for the dowel, as it will find its way inside the slot and you will have tough time to file it away so that you can fit the tang inside later.

Once the dowel fits the blade and also the hole in the block
(without too much resistance) it can be glued it. 
Checking handle alignment relative to the handle block after gluing.

Once the dowel was glued it was time to proceed with rough shaping. But before I would do that I would use a thin needle rasp (which I flattened from one side with a belt grinder to be able to sand tang slots as thinner than 3 mm) and make sure that the blade tang would fit in nicely. Then I would mark the handle with a pencil on the sides of the handle, grind off the excess of the material with a sharp coarse #40 belt, clean up on the disc grinder and do the same for the remaining two sides.

Note: If you are not sure about what size the handle should have when finished, then keep do this in a few steps. Ideally you would get the idea on the size from some other existing handle, as at this stage you can not really test how the handle feels in hand. 

Rough shaping the handle.

Rough shaping finished - cleaning up on a disc grinder.
Once the rough shaping was finished I used #240 sanding paper to flatten all sides (I can not really get perfectly flat grind on my disc sander because it is really small before I would go back to the disc sander and grind the handle to octagonal shape.

Flattening the handle before giving it octagonal shape.

Final shaping is one of the sensitive steps - I am doing it 'free hand' with the disc sander. I first try to get a reasonable angle, grind just a bit and move to the next 'corner' and work on all sides iteratively until I like the result. The reason for this is that it is rather easy to make a mistake and grind under wrong angle. Since this work does not take that much time, I really recommend not to hurry this step.

Final shaping finished - next step is handle finishing.

Once the octagonal shape of the handle is finished, I take the handle to #240 sand paper attached to sanding stone and make sure that all 8 facets are flat and I have removed all the scratches left by the disc grinder.

Note: Be careful here. #240 sandpaper is still fast enough and if you use too much pressure you can easily remove considerably more material than you have planned. Make sure that you apply the pressure evenly left-right and front-back as otherwise you will chance the angle of the facet you are sanding.

The wide facet has the longitudinal scratches from the hand sanding,
the narrow one has still the disc sander finish.

Once the handle is clean I move to sanding with a small hard wood block ( ca 20 x 60 x 100 mm ) with a double-sided Tesa scotch tape on the narrow long sides which I use to attach different sanding papers. I follow with 400, 1000 and 2000. At each step I round the edges on the ferrule and sand under constant angle the edges on the opposite end.

After #2000 grit sanding paper.

After #2000 grit sanding paper.

After #2000 grit sanding paper.
Normally I would stop between #600 and #1000 grit, but wit this handle material both the horn and stabilized wood seemed to need a finer finish. So I moved to micro mesh pads and went from #2400, #3200, #4000, #6000 and #8000. I was a bit surprised to see that the horn did not want to reach a deep polish. It turned out (in direct comparison with several Japanese handles I have) that horn has its own structure and the fine grooves were filled with very fine wood dust what made the impression of deep scratches even stronger. Cleaning the handle and wiping it with board butter made it look much nicer.

Note: contrary to sanding paper which can be mounted on hard surface to have a well defined sanding surface, the micro mesh pads are soft. This inevitably means that there will be tendency to round over edges and the only control you have is how long you will be sanding with the pads.

After micro mesh treatment.

After micro mesh treatment. See the dust filled grooves on the ferule?

Once all the sanding was finished the handle was treated with a board butter.

Finished handle.
Finished handle.
Finished handle.
Finished handle.
Finished handle.
The only work left to do was the glue the handle onto the tang. I have used the G/flex epoxy which is fluid enough to (albeit slowly) flow inside the narrow tang slot. At the same time the glue as 45 minute pot time, so you do not need to hurry and once you get the blade in place you have still time to clean all the glue from the blade and the handle.

Getting ready for the final gluing.
Blade is in place, everything was cleaned and looks good :)

Lessons learned

  • Test the process first - see my problems using the BladeBond Ultra epoxy to glue the nickel silver spacers. I suppose this is a mistake on my side (see the test of this glue by Walter Sorrells on YouTube)
  • If you are using a new concept for the first time - think it through before you execute it. I did not and at the end I got lucky. I will proceed differently next time.
  • Be careful when hand sanding (cleaning up & flattening) the facets of the handle with coarse or medium grit sandpaper - you may remove more material than you expect.
  • Even if you have full contact of the flat area with the sand paper, the material will be removed there, where you apply pressure.
  • I have used here a new source of stabilized wood (from the xx642011 on eBay - also know as stabwoodlab on Instagram) . It turned out to be of good quality, but should that not be the case, I would 'only' have to make a new handle from a different material. I would have not used material of unknown quality if the handle design would have required to shape the handle after it was glued on the tang.

Finished handle

Thursday, December 22, 2016

Project #15 - Handle for a Kato 80 mm petty

This project together with the #16 is the first commission work which was agreed under special terms, as I am still learning basic skills and trying new things.

The wish for this particular knife was a nimble, thin handle, than would allow to use the knife as a parer.

I was also rather nervous, because it involved removing a previous custom made handle and it was a strange feeling to undo someone else's work. But hey, whatever makes the customer happy :)

The knife with the original handle and the ironwood block.


The final choice of design and materials was relatively simple - a 3mm thick brass ferrule with ironwood handle. This also implied a challenge - since the bolster was going to be press-fit onto the tang, the handle would be shaped already glued onto the tang. This meant that if anything should go south, I would have to grind the complete handle off the tang and start from the scratch. I am rather happy to say that that did not happen.

Handle removal

As already mentioned - I had to start with removing the original handle which was made from some black wood, spacer and epoxied onto the tang. This was done rather quickly with a coarse belt, but I had to be careful not to scratch the uncovered part of the tang, as otherwise it would have been necessary to re-finish the whole blade - a job I was trying to avoid.

To grind off the old handle took just a few minutes.

Once the handle was removed I have ground very tiny 'shoulders' onto the tang so I would be able to get a nice fit with the bolster. I used #120 belt as I was trying to remove as little material as possible. The job was finished with files.

Grinding 'shoulders' onto the tang.

Bolsters (yes, 2 of them)

Having the experience with the 2 funayuki knives I have rehandled before working on this one taught me that getting a bolster fit on a Japanese kitchen knife is not easy, because the tangs are not ground to shape and thus their cross section where the bolster should sit is irregular.

I have decided on 3 mm brass as 5 mm seemed too much for a handle that was requested to be rather thin & narrow. 

Blade, ironwood and brass - keeping it simple :)

The process was as usual  - cutting out a piece of brass, scribing the centre lines and the shape of the tang, drilling 3 holes with 2 mm drill bit and filing with needle files to shape, keep testing how the bolster fits the tang and hammer-in the last few mm. But I failed. Even though I tried to flatten the tang from sides before I started,  I did not manage to avoid gaps on the sides of the bolster.

Yep, that is a gap. Let's try it again.

I came up with only one way how to assure that I will get the bolster fit without any gaps - and that was to grind shoulders on the sides of the tang as well. It was a bit tricky, because the tang was not too thick so I could only afford to remove very little material, before putting the strength of the whole blade in danger.

So I went back to file guide, sanded a fraction of millimetre of steel from each side and again finished with files. The tiny shoulders on left and right side gave me just enough safety margin so that after the new bolster was fit onto the tag I would get a much better fit.

Once the new bolster was fitting nicely, I would remove it and finish the front surface on a marble stone with sandpaper up to grit #2000.

Bolster with #2000 grit finish.
The front surface was covered with tape and I used the belt grinder to get it closer to its final size. I was trying to minimize the amount of grinding once the handle will be glued as the brass gets blistering hot very fast and I did not want to overheat the epoxy joint.

Mounting the handle

Before the wooden block could have been glued onto the tang 2 steps had to be accomplished - drilling and filing the tang slot and sanding the block until it would fit without gaps on the tang.

The drilling was done with a 4 mm drill bit. This was a bit larger than actually necessary, but on one hand this was not going to be visible, but more importantly I needed to be able to fit a needle rasp in the drilled hole to be able to make the tang slot.

Squaring-up the block by hand with #120 sanding paper on the stone block.

Handle block ready for gluing.

Once all parts were ready I could proceed with the gluing. I have first covered most of the handle with a tape and fixed it carefully in a vice. As usually I have used G/flex epoxy as it has relatively low viscosity (so it will albeit slowly flow inside the narrow tang opening) and long pot time so I did not have to hurry.

Important: During the gluing it is important to allow the air which is trapped inside the tang opening to escape. This can be done by slowly moving the tang in and out and going deeper just step by step. If too much air will remain trapped inside the handle it could push the blade up as the glue cures and you will be left with a gap between the wood and and bolster. This can in particular be an issue with a handle this small & light.

I have then came up with an idea to use 2 clamps to make sure that there will be no gaps - a usable idea in general, but one needs to be careful - if the force (of the clamps) on the two sides is not equal, it may actually induce a gap.

Gluing it all together - watch out for gaps!

Handle shaping

Once the glue cured the next step was to shape the handle. At this stage one should really not hurry. The handle is fixed with an epoxy to the knife - any larger mistake will mean that the whole handle - including bolster - will have to be removed and once can start all over.

Cleaning up the handle after gluing.

Tip: I was told that if one needs to removed a glued handle from a knife, the knife could be boiled (! yes - that is not a typo) and once the handle heats-through the epoxy should soften. It sounds terrible, but 100 degrees Celsius is not enough to negatively influence the blade hardness (since tempering temperatures during heat treat are around 200 degrees Celsius, give or take) and a short boil should not cause rust. But I have not tried this method yet, so use at your own risk :)

Before I would start with rough shaping of the handle I would first check whether the wood (which at this stage is nicely squared up and thus has a rectangular cross section) is in line with the blade, or whether there is some slight relative rotation. If there were any I would adjust the work rest on the grinder and first fix this.

Alignment-grinding. What you can not see is that the work rest is under
angle slightly larger than 90 degrees to the grinder surface.
Once I was sure that the handle is aligned with the blade I would draw centre lines on top and bottom of the handle as well as on the back side. Then I would draw lines that would reflect the taper if the handle width and use the belt grinder with #60 (#40 would be fine too) belt to do the rough material removal and follow with #80 grit disc grinder to clean up and flatten the sides.

About to grind the taper to the sides of the handle.

I would follow in a similar manner to grind the taper to the top and bottom side of the handle. I would of course have to be careful to tough the running blade with the choil of the knife.

Vertical and horizontal taper ground to the handle.

Once the vertical and horizontal taper was ground it was time to check in hand whether the size of the handle is about where it should be. Since the handle at this stage is just a rectangular block it is not easy to judge how it will feel once ground to octagonal shape. Sill - it should be tested in hand. Over time one develops the 'feel'. 

If everything is right than the next step is grinding the octagonal shape. This is for me a tricky part as while one is grinding on the disc grinder - it is not easy to judge the angle precisely so I would keep checking constantly. I would also not try to grind one corner completely before moving onto the next one as I wanted to keep some material as an error margin should one corner be ground under angle that would deviate from the mirrored one (left-right symmetry).

At this stage I would say that I do not aim for a top-bottom symmetry. I tend to grind the bottom corners a little more and under different angle.

Final shaping finished. Next - finishing the surface.

Final shaping finished.


Now comes my favourite part. During the finishing the wood starts to reveal its beauty - and this is in particular the case for Ironwood.

Before I moved on with the handle I took a small piece of iron wood (and off-cut) and did a quick test on how it will respond to finishing. I gave it a quick polish with 240, 400 and 600 sandpaper and then applied a coat of True-oil. As it turned out this would have not been a good choice since ironwood is an oily wood by itself and thus the Tru-oil did not want to harden properly. I also found out that the wood reacted interestingly to steel wool. If rubbed with it for too long the grain structure would stand out as the filaments in the wood had different hardness. Still - this was an interesting test to do and, together with some information I collected online I have decided to just sand the handle and then coat it with board butter.


One important step that I did not photo-document properly was chamfering the edges on the front and rear side of the handle and the tools I used to finish the handle. To sand the handle I have used a small block of hard wood with dimensions about 20 x 60 x 100 mm. On the long narrow sides (20 x 100 mm) I have attaches a double-sided Tesa tape (the side that sticks stronger facing the wood) and then I would use it to attach the sanding paper.  The tape worked so well, that I had to change it only once during the whole process. Since the paper would clog fast with the wooden dust, I would use a small hard brush to give it a quick rub, so I did not have to change the sanding paper too often. This brushing worked of course better with coarser papers as with grit #1000 and finer the brush would not manage to get between the grains quite as efficiently. Still - it did help.

To chamfer the bolster I had to use a much smaller tool as the blade was getting in the way. I use a small piece of off-cut steel of about 3 x 20 x 40 mm (give or take) and used the double-sided tape as above to attach sanding paper.

I would sand the front and rear edges on the handle with every grit that I would use on the handle, but while I rounded the edges on the front side, I sanded the rear side under constant angle (about 45 deg) and did some rounding only with the finest paper. I simply like that look better.

Left: Handle after #240 grit,
Right: piece of sanded wood treated with Tru-oil.
So I have proceeded with sanding the handle. I really enjoyed how every finer grit revealed more and more structure of the wood.

After #240 grit.
After #400 grit.
After #600 grit
In principle I could have stopped after #600 grit, but since the sides of the bolster were also sanded at the same time, I went all the way up to #200 grit so that I would get (at least comparable) finish as I had on the front face of the bolster.

After #2000 grit.
After #2000 grit
After #2000 grit
After #2000 grit.

Warning: When sanding handle with metal parts you may notice that at the finer grits there will be tendency for the metal dust to 'stain' the wood in its vicinity. What happens is that the metal dust is so fine that it clogs the fine pores of the wood. I noticed it happening here at #1000 and #2000 grits. To minimise that I would only sand in the direction from the wood towards the brass and clean or change the paper after just a few strokes.

Once the handle was finished I applied a coat of board butter.

Lessons learned

  • Prepare the tang carefully for the press-fitting of the bolster. Do not allow the thickness change to be too abrupt towards the final position of the tang.
  • If the bolster does not want to fit with light hammering - take it off and work on it or the tang, otherwise it can bend and also the blade could get damaged
  • Be 110% sure that the blade is properly wrapped and will be safe from any damage (scratches, chips, broken tip, pressure marks, etc).
  • Becareful when sanding close to wood/metal boundary with fine grits as the metal dust could stain the wood.

Finished handle

The fine grain of the wood really pops unde the right light.