Modular Tansu – Veneered Doors

The doors for the series of modular tansu that I’ve been covering in this series of posts are styled after traditional tabi-dansu doors like the one on the cabinet on the left below:

The doors on these cabinets are made from edge glued quartersawn panels with a mitered frame. All the tabi-dansu that I’ve seen (like the one above) are made from solid pauwlonia, and the door panels have no provision for wood movement of any kind. Pauwlonia is soft and relatively stable, especially when quartersawn, and Japan’s relative humidity is much more consistent than many parts of the world, but cracks still develop. That said given the age of some of these cabinets, if cared for they do hold up surprizingly well. (The dark cabinet on the far left in the picture above is nearly 100 years old and the door is in really good shape.)

For the cabinets that I’m building here, I considered two main options for the door panels; the solid wood approach or a veneered approach. I’m really partial to the solid wood option and I explored a bunch of different possibilites. These cabinets will be headed to a much drier part of the world than Japan, and they are quite massive in size, so accounting for wood movement was one of the biggest issues. The best approach I could envision was to used dovetailed battens on the interior side of the doors, allowing a thin panel to move while also keeping it flat. Below are some of my early concepts drawn in Blender.

While I really like the solid wood approach, there’s a limit to how deep the battens on the interior side can be, and if they are too thin they become ineffective and the door could still warp causing all sorts of trouble. Weight was also a concern. A solid panel, plus battens, would certainly be heavier than a veneered panel, and these cabinets are being made from wood much heavier than the traditional paulownia (cherry, keyaki and chestnut).

So long story short, after giving it lots of thought I opted to go with veneered panels, but I planned to cut my own veneer giving me much more control over grain selection.

The first step was to mill up some billets to rip strips of veneer from. Below is a chunk of quartersawn keyaki from which I ripped 2.5 mm thick strips to make up the door panel for the keyaki tabi-dansu.

I used a Hitachi cb75F bandsaw and it’s stock 3″ wide rip blade for cutting the strips. Definitely overkill, but also definitely made easy work of the situation, and the results were really clean.

The process of ripping involved joining the billet, making one rip, rejoining that face, rinse and repeat… That way each strip ends up with one face joined and ready to be sent through the planer.

Here are the finished strips after ripping on the bandsaw. I kept the strips in order as I ripped them, so I could align the grain in a slip-match pattern for the panels later on. There are two sets, one for the front of the door and one for the back.

Below are the strips after running them through the planer bringing them down to 2mm in thickness (on the right is the stock for the door frame after rough milling).

Next I needed to edge join all of the strips, and to make that task easier and safer I attached a feather board to my jointer.

Here are the first set of keyaki slats sequencially slip-matched after edge joining. I love the look of slip-matched panels. The grain has a quiet consistency, yet you can still trace changes and movement in the grain as your eye moves across the panel.

And here is the other keyaki panel after edge gluing. I used liquid hide glue on the seams and masking tape to hold everything tight while the glue dried.

Now to address the actual method I chose for veneering. Again I thought about a lot of options for how to glue up these door panels, with a vacuum press being high on the list as one of the best options. But in all honestly this was one of the first times I’ve needed to veneer something on a large scale, and it’s not something I plan to do regularly, so investing in a vacuum press seemed unappealing. Not to mention it’s not a commonly used method here in Japan, and I was having no luck finding any sort of kit or components from woodworking suppliers.

Then, unrelated to this project, last fall I setup a basic vacuum chuck system for my lathe, and suddenly the prospect of setting up a vacuum press started to feel a lot more doable. After a bit of searching I was able to cobble together all the necessary supplies from our local home center. There’s plenty of crudity to this setup, but it actually ended up working really well, and since I’m not planning to use it regularly, a somewhat makeship vacuum press fit the bill.

One of my first test runs with the vaccum press was for a small test panel. I ended up doing a variety of test panels to work out kinks with the process, and make sure the vacuum press was effective.

When I was researching vacuum press systems I referred often to Joewoodworker.com which has tons of great info on setting up and using a vacuum press. One important bit of information I gleaned from the site was how much pressure to use. The site states that “you’ll need a minimum of 18″ of Hg. However, the recommended pressure for most adhesives is 21″ of Hg.” I couldn’t find any vacuum gauges here in Japan that measure inches of mercury (Hg), so I went with was was available and converted the measurements. My gauge measures megapascal, and with the bleed valve on my vacuum press shut, pressure maxes out at 0.085 Mpa, which is equal to 25″ of mercury. I’m sure I could tighten up all the little leaks in the system to get even more pressure but as is, I had more than enough.

Here’s the first test panel after going through the vacuum press. The veneer strips are quartersawn white oak, milled to 2mm thick, and glued onto both sides of a piece of 6mm mdf. The adhesive I used was epoxy.

For the most part things went well, but I had a bit of bleed through here and there, which was really surprizing. Vacuum pressure is amazing! I was not expecting to see the thick epoxy seep through the 2mm thick veneer.

In anycase it was a good test. I learned a lot about how much glue to apply, and from here on out started turning down the pressure during glue ups.

Initially I was really worried about the glue bleeding through, since 2 of the cabinet/door combos that I’d be making required stain. But I was surprized to see that after handplaning and removal of any glue on the surface, staining was no problem.

Here is the same oak test panel with a iron stain applied after handplaning. There are color variations, but they’re not the result of the bleed through.

From here I did a few other test panels, to see how other glues besides epoxy would fare. The two small keyaki panels below are for a test glue-up using liquid hide glue. I pre-sized one of the panels with liquid hide glue to see if I could “seal” the wood with glue prior to putting it in the vacuum press, and hopefully prevent any bleed-through. (at this point in the process, I was still concerned bleed-through would have a negative effect on finishing/staining.)

Here is the keyaki test panel after coming out of the press. Still using too much glue at this point, hence all the messy squeeze out around the edges. And I still got bleed-through, even on the side that I sized.

Here’s the same panel after a bit of handplaning.

… and after staining. Again bleed through had no effect on the final stain, which was a relief.

Eventually it became clear that bleed-through looks much worse than it actually is. When it comes out of the vaccum press the glue has been smooshed onto the surface of the wood making it look like a major mess. But after a couple of passes from the plane, all that surface glue is removed, leaving only a tiny bit in the pores of the wood here and there, which really didn’t seem to have any effect on the stains I was using.

One final test I did was a test panel using oak strips which I did not edge glue ahead of time. I had read about people people gluing thinner veneer without edge gluing, and so I thought I’d give it a shot…

… it did not turn out well. It only took a little bit stress testing to get the seams to open up. So that option was out.

So with all the veneer tests done and a bit more experience under my belt, I cautiously started work on the first actual cabinet door. I started with the smallest of the batch, which was the keyaki tabi-dansu door. I ended up settling on 6mm birch ply for the core material, and epoxy for an adhesive. Liquid hide glue actually worked out okay, but the epoxy test panels felt just a bit more stable after exposing them to some temp/humidity extremes.

It was mid-December when I was gluing up this door, and in an unheated shop 90min epoxy takes FOREVER to cure! So out came our kerosene heater to help warm up the shop and speed up the cure time.

Here’s my basic pressure regulator/gauge setup. This is exactly how I had it setup on my lathe, and it was easy to use the same arrangement for the vacuum press as well. The hose is pulling air out of the bag moving from right to left. The little down leg has a needle bleed valve to dial in the pressure. And the two cheap automotive gas line filters are there just to keep any stray dust out of the pump (which is especially crucial when using this setup on the lathe).

For the door glue ups I lowered the pressure to see if I could keep bleed-through to a minimum. (0.06 MPa = 17.7″ Hg)

All-in-all the glue up went really well, though still lots of bleed-through.

Here’s the panel after trimming to size. I was glad to see a nice clean lamination, with no gaps or voids of any kind. Vacuum pressure is pretty incredible.

A couple weeks later I went through the same process gluing up the panel for the chestnut cabinet, which you can see below. At this point the panel had been run through the vacuum press and trimmed roughly to size, in preparation for finish planing.

Quartersawn chestnut is a joy to plane.

Here’s the panel after planing. There’s nothing like a clean hand planed surface.

At long last it was time to tackle the biggest door of the bunch… the 1280mm wide (~50.5″) cherry door. I went through the same process for this door as the others, ripping/milling strips of 2mm slip-matched veneer, and edged glued them with liquid hide glue. The panels were so wide that I ended up gluing the inner and outer slats into 2 halves respectively. After each half was glued up, I then glued them together to form the inner and outer sheets of veneer.

And into the vacuum press. By this point I had a good feel for the process, and my glue calculation was pretty spot on. I used a toothed glue spreader to apply the glue to the birch ply, and I also started pre-heating the epoxy prior to mixing, which made mixing much easier, and sped up the cure time.

Still lots of bleed through, despite lowering the pressure, but again thanfully there were no issues in the final finish.

Here’s the cherry panel in the midst of hand planing. It’s such a wide panel that half of it was sticking off the side of my bench.

And a sneak peak at the inside face of the cherry door after gluing up the door frame.

The outside faces of the doors have the panels running flush with the surrounding mitered frame to keep the look similar to the old tabi-dasu style door pictured at the top of this post.

Compared to the veneering process, the frames were really straightforward to make, and I’ll share some pics and details on that next time.

Thanks for reading.