Alaska Wood: Veneering - Where Angels Fear to Tred, I rushed in.

I have been working for the last several months, off and on, to perfect veneering a 42 inch square coffee table built to resemble what the decorator said was a Chinese Scroll Table. I have no idea how the Chinese did it, but here is my shot at it.

I decided to use the torsion-box concept since that would be lightweight and strong. As you can see in the picture, it is built up of a grid of half-lapped 1/2 inch plywood units. There are 84 joints, each of which is glued with urethane construction adhesive. A 3/8 wiggleboard skin was glued to each side, but required kerfing to make the tight radius.

Such a structure achieves its remarkable strength from the simple transfer of load. As a load is appled to the table, it caused the top skin to compress (squish together) and some of that pressure is transferred via the skeleton (the web) to the lower skin which is stretched. You see that now the load is no longer just downward on the parts, but sideways! As long as there is a good bond between the web and the skin, you have a very, very strong unit. I weigh 180 lbs. and could jump up and down on the table without any problem.

Version 1:

My first attempt at veneering this beast was to use the iron-on technique with PVA glue. I coated 10mil paper-backed mahogany and the wiggle-board substrate with a generous layer of Titebond2. After drying, I began ironing it on with a regular iron set on high.

Everything seemed to go OK, until I put the finish on. I could then see some bubbling. It didn't seem too bad, but when I delivered it, the veneer began to bubble, probably because of the high water content in PVA glue. The next month I brought the table home, pulled off the veneer, cleaned it up and decided to use contact cement. That ought to hold 'er down!

That was a nightmare. I decided to try using a 6mil plastic barrier between the veneer and table while I pressed from the center out pulling plastic as I went..

Well, the plastic got stuck in the contact cement while pressing down the large chuck of veneer and totally ruined the lamination. Arrrghh!

Ripped that off, and contact cement isn't very happy about releasing prisoners, so I ripped up little chunks of the substrate too. After filling and sanding, did it again, using a joint down the center, so I would have two relatively narrow pieces to work with...but it didn't look perfect, and I was out of time. Had to finish and deliver. I Told the customer I would replace it as soon as I could. It looked pretty good...but soon it developed bubbles and still had a few uneven spots I missed while cleaning up the old contact cement.

As I look back, there are two glaring issues. One, using a water-based PVA glue added too much moisture to the system. Even if the project laid flat initially, the moisture was going to try to get out and cause bubbles. PVA is just not appropriate for a large project anyway because of its tendency to creep. That's why the old timers used hide glue...no creep.
Number two - don't use contact cement...it has no advantages and lots of disadvantages.

My wife says she is really hating this table. But I can't just give up--no matter what it costs me. Besides, this can be a learning situation -- you know, making lemonade when life hands you lemons.

As an old friend in the computer systems administration business told me once. "Don't feel bad when it takes you a long time to figure something out...it just means that once you've got it, you're that much farther ahead of the rest who don't know how to do it--and you're valuable now."

So I went home and thought about it, did a little internet research, and decided to try vacuum bag pressing. At joewoodworker.com I found and bought the vacuum pump kit and enough 30mil urethane film to make a 4.5 ft x 9 ft. bag. The pump will pull 5.5 cfm, which should be good enough to handle the big bag I want to make.

The pump I built works flawlessly, and after testing, I pressed up some flat table tops. Using Unibond-800, I find I get a very flat, bubble free and so far, stable laminate. I have to give praise to Joe Gorleski for creating a supremely practical and functional pump kit.

Version 2:

So I rebuilt the table, making the shoulder radius a little larger (8 in) so I didn't have to kerf the wiggle board.

Another change, I realized I really didn't need to glue

each joint in the skeleton...nor did I have to bother with all those time-sucking lap joints. The wiggle board just needs support to withstand vacuum pressure and enough web to transfer load to lower skin. So I reduced the number of cross pieces and just brad-nailed them together, then used blue 1 1/2 in construction foam as a space filler and skin support. The foam also adds an echo-killing feature. The table won't sound like a drum when thumped.

I decided I would just pull the vacuum bag over the table and veneer like putting a sock on your foot. The vacuum should pull everything flat.

An assistant and I worked to get the Unibond applied, lay on the veneer, staple the veneer on one end, and start suckin' air. Pulling a 30 mil plastic bag over that table was not easy - definitely a two man job.

Alas, and alack. got a big wrinkle in the veneer on the end by the staples....I tried repositioning and rolling out the wrinkle area and removing the staples...but it was another disappointing and now expensive failure.

Version 3:

I figured I better do some drastic design change. I opted for rebuilding the skeleton in 5 sections. A 26.5 inch flat section for top, two curved sections 1/4 of 8 in radius, and two 8.5 in. leg sections. After pressing the flat panels and the outside of curve sections, I will glue and screw everything together. I will have to hand-veneer the inside of the curved section. Your prayers are welcome.

The panel on left shows the temporary pine edges with blue masking which keeps the over-size veneer from breaking over the edge. Center shows how veneer will fit on the assembly, and the right is completed top section.

Also switched to Oakwood Veneer product called Bubble Free. It is a 20mil product that includes a moisture barrier. I will use Unibond-800 for completely water free non-creeping adhesive.
--------------------------------------------

Update - Looking good so far. I got the large center section veneered without a problem, thanks to my partner/wife who had a good suggestion about taping the veneer down till we could get into the vacuum bag. We had to mix up two batches of Unibond, but it seems to cover about 10 square feet per cup of resin. After five hours in the bag at about 70 degrees (electric blanket), lamination was perfect. I really like the 20 mil thickness because it doesn't telegraph the substrate irregularities like the 10mil.

Just popped the 'leg' sections in the bag and will let them cook for five hours while I go try to find an angle attachment for my drill. Screwing the curved shoulders to the flat sections will require some tight-quarters right-angle drilling to bring the surfaces together tight and true.

------------------------------------------

Update - The veneer attached very nicely to the curved 'shoulders' but the high pressure of the vacuum bag was too much for the wiggle board and the curve was not fair. As you can see from the picture above, which is built the same way, there were just 5 curved sections with 4 long pieces. That allowed for too much cross-grain flexibility. Perhaps I should adjust the vacuum pressure to just pull the veneer down and stop there. Currently set at 21 in. Hg = 1485 pounds/sq ft.

Pump pressures

in Hg lb/in lb/ft
10 4.91 707.27
11 5.40 778.00
12 5.89 848.72
13 6.39 919.45
14 6.88 990.18
15 7.37 1060.90
16 7.86 1131.63
17 8.35 1202.36
18 8.84 1273.08
19 9.33 1343.81
20 9.82 1414.54
21 10.31 1485.27

I have plenty of wiggle board and veneer, so I decided to rebuild the shoulders using 13 curved sections and a single row of spacers at the center. That should give the curve some integrity and indeed it looks very nice after gluing down the wiggle board.

Re: screwing the sections together, there's not enough space for right-angle drill attachment, so I'll have to use 1/4 x 11/2 lag screws with a socket wrench.
---------------------------------------

Well, it's about time! The veneering of the new shoulder worked perfectly... I had lowered the vacuum pressure to 15 inches Hg (1000 lb/sq ft) and the surface is quite fair. Victory at last...shucks, it's only been 11 months.

Here's a shot of the shoulder right out of the bag...

The sections are a little time consuming to attach using lag screws when I can only get 1/4 turn or so, but they will come together nicely...here is a dry-fit view.

The joint will be routed after gluing for an inlay of mahogany strip. Customer wants the inlay flush, so I'll probably do a lot of careful work with a scraper blade.

Still have to build the feet and edges, but we're getting close.

Hey, just bought a very cool tool....since the socket wrench was a little too big to use efficiently in such a narrow space, I went to Lowe's and looked for a ratcheting wrench. I bought the Flex GearWrench for $11.50. It is a dream to use, very nicely made and even has little marks on one side of the wrench to let you know which direction you're turning the target.

A well-made and useful tool is a beautiful thing. I wonder if God is pleased that way when one of His 'tools' is useful? I'd like to think so.

All the glue-up of structure is done, the substrate for inside of curve is glued, so now just have to veneer those... two little inside curves. This might be a case for the return of contact cement. Such a narrow piece should be easy to lay. It will be out of sight and out of harm's way -- so why not?

Indeed, the water-soluble contact adhesive worked very well for the inside curve, thanks to my beautiful assistant - the blonde cowgirl Ms. Tinklewater. Will be milling some mahogany for the inlay strips.

Note for future construction: I think it would be best to build a framework to hold the sections while screwing together. Perhaps just a temporary 'edge' piece to keep section aligned the long way.

Routed 1/4 in. groove on top to inlay strip to mask joint. the side joints are 1/2 in. The color difference between veneer and inlay should be minimized by the dye coat.

And added the feet, lovingly carved with the table saw, plane and sander. Edge pieces made of some Hondo I have in stock. Biscuit joined the three boards making up each edge assembly and glued the oversized assembly to the edges. A flush trim router bit will bring it all home and be ready for round-over. Routing those edges took about 2 hours and lots of dust and cramps in my hand, blood on the wood.

I was doing the last six inches and grabbed a little too much wood with the router and split off a big chunk. Glued it back on and finished, but isn't that the way things go? Ya think yer home free and wham.

Flush bit work done, and round-over was quick work.

And now she's ready to sand and finish! The sanding of the edges on ends was pretty slow-going hand work to get some of those router blips smoothed out. One just needs patience, but if it is going to be touched by the user it has to be smooth.