Trellis – all about it(?) Conclusion

This little beast was more work to build than anticipated, however the page has been turned. I thought I’d share some pictures of the assembly and installation.

I started with the lower tier of the trellis, fitting the battens into the central rib, and then one pair of legs with associated stretcher:

All the battens employ bare-face tenons on both ends.

On goes the other pair of legs and stretcher:

That completes the assembly of the lower portion. On to the top tier, again beginning with attachment of the individual battens to the central rib:

The remainder of the upper frame is brought together, and then entire assembly is plopped onto the four posts:

Time for some wedging on those post tenons:

Then trimmed flush:

With tenons on the posts now firmly attached, I moved on to putting the stubby shachi sen into place to lock up the corner joints:

The second one driven in completes the outer rod tenon connection:

Then the pins are trimmed:

Another corner – here’s the site of the patch you saw in the first post in this thread:

A look at an inside corner, where there is also a rod tenon and a pair of ‘parallelogram-shaped tapered wedging pins’ (It is quicker to say shachi sen, yes?):

The connections drew up fairly well with the pins in place:

Another corner:


Considering I spent very little time fussing the fits of those joints, trying to complete the work as quickly as I could, they turned out decently I thought. I had planned on doing a separate beveled cap structure as well, but decided to pass on that plan. That might be a mistake, as I think the cap would have added considerable durability. That said, the cap can be added later if need be, so no worries. We’ll see how the Jatobá hangs in there over the seasons.

The lower tier of trellis has a central beam fixed to the stretchers using double-wedged tenons:

The corner connections of post and stretcher, employing haunched half tenons and wedging:

The upper tier’s central beam is attached using a 0.25″ Ipé peg, and I did simple mason’s miters at the meeting of upper arrises:

I used Jatobá for the wedges and shachi sen, and Ipé for the pins.

I gave the piece a quick an cursory swipe with the plane along the top and took the piece back to my house for the install. The two-tier trellis fit, as it was meant to, upon the Black Locust raised bed frame, now keeping ‘tower of power’ company:

Once screwed down with some stainless 2″ screws, I dabbed some paint on bits of exposed end grain. It’s done!

Here’s a few more shots of the piece:

It’s a climbing gym for plants:

It’s nice to complete the main portion of garden construction for this year. I’m planning to pull the front axle out of my truck and rebuild it in the next week, so I wanted to get the trucking of pieces like this all wrapped up before I lose use of the vehicle.

The garden architecture is a bit of a laboratory for me. I can compare how Spanish Cedar, Teak, Jatobá and Black Locust stand up to the conditions, which in Westerm Mass orbit between well below zero and 0% humidity, to quite hot and 100% humidity. I have various different forms of connections and forms of making the attachments that can also be compared over time. It should be educational. I am looking forward to seeing the plants draped all over the two trellises later in the summer and will post up a pic or two so you can see too. This is all a big experiment I guess, as growing cukes, melons and squash on a trellis is not commonly done I don’t think.  I’m a novice gardener so it’s all about learning from experience here.

Thanks for coming by the Carpentry Way.

16 thoughts on “Trellis – all about it(?) Conclusion

  1. Very nice. A little over designed perhaps, but really, really nice. I thought I had a problem adding complication to simple building tasks but now I see by comparison, I must be normal.

    Harlan Barnhart

  2. Harlan,

    If I had spent more time designing it, or the entire garden set up for that matter, I might have come to some simpler and more elegant solutions, no doubt. Glad you liked it despite your reservations.


  3. Those are some very fortunate plants. Apart from seeing how those woods weather, which of the three do you most prefer to work with? It would be fun to see your take on a garden gate some day, especially after that extensive series on Japanese gates! -Mike

  4. Mike,

    thanks for your comment. I think the teak was easily the most workable material, with a curious combination of being relatively soft and yet quite abrasive. Polishes nicely, however frequent trips to the sharpening pit are in order for really clean surfaces. The Jatoba is like working stone, and is a tough SOB. Seeing curly Jatoba for the first time was an unexpected surprise. The Spanish Cedar smell I find detestable. The Black Locust I worked green, which made it easy going, however not quite a fair comparison to the other materials in that respect.

    I may have that gate opportunity one day soon, we'll see – thanks for your encouragement.


  5. Chris

    Very entertaining to see the 'plant stand' come together. Could you explain the wedging techniques you used? I would have been tempted to double wedge everything, 'a foolish consistency' perhaps.

    Oh, you didn't mention working with Ipe. I found that pretty stone-like, but with a much better smell than Sp. cedar.


  6. Tom,

    good to hear from you. Not quite sure what you are specifically looking for in terms of 'wedging techniques', but basically the mortises are flared 1/10, or 3mm over 30mm, and I use a paring guide to chop that slope in the mortise (a bit over 5˚). Then I kerf the tenons with a rip saw, and prepare the wedges using my Kapex. I hammer the wedges in dry when the assembly is tight together. In some places I then trim them flush, or, in the case of the stretchers, I drove the wedges in until they were flush to the tenon ends and called it a day. That's about it.

    As for double-wedging the stretcher tenons, that was an option however since the tenons lap one another the chopping of the wedging slope in the mortise would cut partially into the opposing tenon, so I decided against that. A better way to do that is to use a through wedge above the tenon and make the tenons half-dovetails and drive them against one another.

    In hindsight, I put the stretchers in at the same height by habit, however in this piece there was no real need for them to be the same height. It would have been structurally sounder to stagger the stretcher height sufficiently that full height tenons could have been employed, and these could have been then double pinned. Would have been stronger and easier for demounting if required.

    Yah, Ipé is even harder than Jatobá, however I only made a couple of pins out of some 0.25″ stock, so I didn't feel a need to comment on its workability. I found it was plane-able in a previous project – with a 60˚ bedded plane.


  7. Chris

    That explains the single offset wedged tenons on the stretchers, thanks. On the tenons on the top, some wedges are offset and one at least is on an angle. Closing up gaps or…? Just trying to pry out some of your secrets!


  8. Naw, was just working fast and ripping the wedge kerfs without being fussy. One ended up on an angle. The tenons themselves are only 0.5″ square, and it is usually better to run the kerfs down the tenon so as to not provoke splitting beyond the end of the kerf. One solution is to drill a tiny hole at the end of the kerf, and another is to run the kerf so it does not run straight down the grain. Hence the offsets at the visible end – if the wedge was centered and the kerf run off at an angle for 3cm, then it would be awfully close to running out the side of the tenon, which would be a recipe for breaking off when the wedge is driven.

    Keep prying – I don't mind.


  9. Thanks, that makes a lot of sense. I've suffered 'splitting beyond the kerf' before. Its maddening!


  10. I find that there's a kind of joy in overdoing it, so to speak. My wife calls it 'getting all wood-nerdy'. The complications are the best part.

  11. Adam,

    nice to receive your comment.

    The thing is, what is one person's 'overdoing' is another's 'superficial attempt'. In this second trellis for the garden the joinery was, to me, very straightforward and in most cases choosing a simplified form. No compound angles. In example of the simplifications, the tusk tenons for the central ribs did not have mitered stubs on them to more cleanly transition the chamfering on the pieces – hence the recourse to mason's miters. I would never do that simplification on a piece of furniture or timber frame. The corner connections are little more involved to build, however it is always a certain difficulty in realizing 3-way connections, especially when adding in weatherproofness and durability to the equation. I also omitted a cap assembly to save time. I could have placed sloping wooden caps over the exposed tenons as you see in Japanese torii and bridges, so there is another shortcut.

    In many respects, I consider this trellis construction to be 'simple', a piece adequate to the task but not over-elaborated. Others may well scale it differently of course. It was a fun project in any case and I have no regrets about spending the time on it.


  12. Just to clarify Chris, my only “reservation” is that it's too nice to put out for the cucumbers to climb on. Nice work as always.

    Harlan B.

  13. hey Chris, I was curious about the subtlety that you didn't “fuss” with the (fantastic) joints much. would said fussing involve a GF paring chisel that the Jatoba would then respond to? Is Jatoba one of the most stubborn woods you've worked, worse than bubinga? I'm mainly impressed you could “fuss” those joints any better with some ungodly sharp tool.. and which width chisel would you prefer for those largest 2″ joints?

    what a project! strong as steel and cool test bed.. I'm also very curious as to the weathering properties!


  14. Will,

    by fussing I meant taking some additional time to have the joints close up without any gaps whatsoever. While perfection is an elusive goal, I would certainly have spent more time on them had the trellis been a piece of furniture.

    I find Jatobá fairly tough to work mostly due to its hardness. fortunately, it does not seem overly plagued with interlocking grain, unlike some other species. It is noticeably harder than bubinga. I find it can be chopped sawn and planer well enough with hand tools, though, like other woods in this range of hardness it is absolutely unforgiving of any miss-steps with the handsaw and will happily pull and bend teeth. If the chisel is sharp the wood will pare fairly well actually.

    I'll post up some pictures later in the summer of the various garden pieces to see how they are faring. so far they have been subject to several wash, rinse, and bake dry cycles. Hanging together so far…

    Thanks for your comment and questions.


Anything to add?