Carving Grape Onlay 3

A carving breaks on the second side after the first side has already taken its support away. Wood held from two sides is now held from one. Your tool arrives, the fibres have nowhere left to move, and something lets go. A shallow channel, taken first, is the whole answer.

I came back to this panel straight from a week of teaching in Michigan, four and a half days of architectural carving, and the thing students asked me about most was not grapes at all. Breakage. How does a piece survive an hour of heavy work and then snap during a light pass?

The third workshop of the grape onlay has one sitting right there on the board. A leaf running uphill, already excavated on one side, with a big berry that has to be outlined on the other. Outline it the obvious way and the leaf comes off in your hand.

Key takeaways

• A fragile element is safe until you take material from its second side.
• Release the load first with a shallow channel, roughly four millimetres deep, cut along the outside of the fragile shape.
• The weak direction is short grain, and every twisting curl has some in it.
• Leave shapes fatter than the drawing.
• Work a junction from two sides rather than driving through from one.
• Small tools reach tight spots, though one that is too small leaves stripes behind.

What actually holds a carved leaf up?

Wood holds itself up along the grain, and almost not at all across it. Picture the fibres as bundles of drinking straws running the length of the board. Pull along a straw and it is strong. Bend across a bundle and it separates without much complaint.

A carved leaf keeps its strength while both of its sides are still buried in solid stock, because the fibres running through it continue into wood that is not going anywhere. Take the stock off one flank and the situation changes. Whatever is left is now a peninsula, and its only anchor is whatever grain direction happens to run through it.

What makes this dangerous rather than merely delicate is short grain. When a curl twists, the fibres inside it stop running the length of the shape and start running across it, so you end up with a piece that is only a few fibres long in the direction that matters. On a tendril that measures two millimetres across, short grain is the difference between a part that flexes and a part that is already in two pieces.

Most of the time you can see it coming. Follow the line of the carving with your eye. Wherever it turns hard, the grain underneath is crossing the shape rather than travelling with it, and that is the spot to treat carefully for the rest of the project.

What makes the second side the dangerous one?

The second side breaks it because your tool is now a lever acting on wood that can no longer push back. Everything you remove from flank one was, until that moment, holding flank two steady. It is not the size of the pass that matters here, and that is what confuses people.

Watch the sequence and you can see the trap. Morning: the leaf is buried, you excavate its lower side heavily, nothing moves, confidence high. Afternoon: one light outlining pass on the upper side, a pass so small you barely notice it, and the leaf snaps at the narrowest point in its grain.

The force had nowhere to go. A gouge is a wedge, whatever its sweep, and a wedge spreads the fibres sideways as it travels. When solid stock sits on both sides of your edge, that spreading is absorbed by wood you have not touched. When one side is air, the spreading has only one place to go, which is straight into the weakest section of whatever you are trying to keep.

My own panel carries the evidence. I had already lowered one section too far in an earlier sitting, and by the time I came back it was undercut without my planning it that way. There was a change of design, a lower area, a new movement. Not a disaster, though. It was a decision the wood made for me.

What is a relieving channel?

A relieving channel is a shallow groove taken on the far side of a fragile shape before you do any real carving near it. Mine on this panel went in with a number 11 at 4 mm, running along the outside of the leaf, and I took it down about 4 mm deep. A little over an eighth of an inch.

That is all. The groove removes the wood that would otherwise resist when the wedge starts spreading fibres from the other direction. Once it exists I can outline the big berry beside the leaf freely, cutting toward it, away from it, from above, without the leaf loading up at all.

Depth is the part to get right, and shallow is the answer. I am not excavating. I want the leaf to survive to the end of the project, which means keeping its future thickness intact, so the channel only has to be deep enough to break the load path. Go deeper and you have simply moved the problem to a thinner place.

Stage	Tool	What it buys you
Relieving channel	number 11, 4 mm wide, taken 4 mm down	Lets you carve the shape next to it without loading the fragile one
Outline afterwards	the half inch number 5	Free movement in every direction
Tight junction	an 8 mm number 2	Accuracy without stripes in the surface
The smallest holes	a Swiss made 5 mm	Reaches inside a hollow that a half inch tool cannot enter at all
Cleaning a corner	skew knife, point angled down	Opens the sharp inside angle where two shapes meet

None of this is exotic. A veiner and a little patience is the whole method, and it costs about ninety seconds per fragile element. Set against gluing a broken leaf back on, ninety seconds is nothing. If you have broken one already, the floral and molding course shows the repair, and thinned glue brushed into the fibres will strengthen an edge that is merely thin rather than loaded.

When should you leave the wood oversize?

Always, at the rough stage. A shape that comes out bigger than you intended is a shape you can still correct, and on this panel I deliberately left several berries fat on the first pass. Chopping more away later takes a minute. Putting wood back takes a glue joint and an afternoon.

People get excited and go too far, and it happens fastest on small elements. The berry shrinks past the size that reads as fruit, the tendril stops being delicate and becomes invisible, and the leaf edge thins until the next pass takes it off entirely.

Oversize also protects you from the load problem in the first place. Extra material is extra support, so a fat berry beside a fragile leaf is a berry that is holding that leaf steady while you work elsewhere. I bring those down last, once everything around them is settled and there is nothing left to protect.

There is a limit in the other direction as well. Leave a shape too fat for too long and the whole area loses its movement, because you cannot judge depth against a lump. I take the mass down in stages and check the piece from the side after each one.

Where does the wood tell you to put the next grape?

The wood tells you through the cuts already in it. On this bunch I made several deep passes to shape the mass, and the hollows those passes left behind decided where the next berries went. Not the drawing. The panel.

That sounds vague until you have done it. A big excavation leaves a hollow with a rim, and a berry sitting on that rim reads as one that is pushing forward, while a berry placed inside the hollow reads as one tucked behind. The drawing could not have known either of those spots existed, because I made them five minutes ago.

So I work with a pencil close at hand and change things constantly. A berry gets promised to one spot, the excavation beside it opens up, and it moves. The bunch that comes out at the end of the session is related to the pattern rather than identical to it, which is why the finished drawing for a project like this can only be made after the carving.

Real fruit is not the target either. Look at grapes on a vine and you see gaps and voids underneath the bunch that no tool you own will ever reach. Carving in wood sits closer to interpretation than to copying, and the same freedom runs through every Baroque panel worth studying.

How do you spot a weak place before you carve it?

You read the grain, and then you read the shape against it. Those two lines agree for most of a panel, and where they stop agreeing is where the wood will let you down.

Start with the board. On a plain piece of basswood the fibres run the length of it, so anything I carve running that same way has long grain underneath and will take a surprising amount of abuse. Anything crossing it does not. A leaf tip pointing across the board is weaker than the same tip pointing along it, and no amount of skill changes that arithmetic.

Then look for the turns. Every curl, every scroll, every place a vine doubles back has a stretch where the grain crosses the shape, and those stretches are short grain whether you planned them or not. I mark the worst of them with a pencil before I start so I do not forget mid pass.

Defects deserve their own look. I found a small knot in this panel while I was working, and a knot sitting inside a fragile element is a break waiting for an excuse. Grain swirls around it, the density changes, and a tool that was slicing happily a second earlier suddenly grabs. The Wood Handbook explains the mechanics better than I can, though a carver mostly needs the practical version, which is to leave extra wood around a knot and deal with it last.

The final check is your hand. Press a finished element gently sideways with one finger. If it moves, it is thinner than it looks, and everything you do near it from that point on happens with small passes and a sharp edge. The second workshop covers the stage before this one, and the fourth is where the really fragile work starts.

Which tools reach a tight junction safely?

Narrow ones, and not as narrow as you would guess. A number 2 at 8 mm handles most of the tight work on this panel. Where a hollow is too small even for that, I move to a Swiss made 5 mm, which slips inside and lifts the waste without touching the walls.

Going smaller than the job needs is its own mistake. A tiny gouge leaves a surface covered in narrow parallel tracks, a striped result that you then have to blend away with something wider, so you have made two jobs out of one. Reach for the smallest tool that fits and no smaller.

The knife earns its place in the corners. A skew knife with the point angled down opens the sharp inside angle where shapes meet, which is the one place a chisel or gouge cannot follow. The difference between an opened corner and a rounded one is most of what people mean when they call a carving crisp.

Whatever the tool, come at a junction from both ends. Fibres reverse depending on which way you travel, so a pass that tears travelling left will slice cleanly travelling right, and two short approaches beat one long shove every time. Sharpness decides the rest, which is why sharpening sits underneath everything on this site.

Does any of this apply outside a grape panel?

All of it. The load rule is not about fruit, it is about wood, and I said so out loud while I was carving because it matters more than the project it appeared in.

Any time you are about to take material from the second side of something slender, stop and ask what is still holding it. A Greek acanthus tip, a pierced rosette petal, a run of grape leaves on a vase, the thin scrolls in a Gibbons drop, the tips of an acanthus ornament. Same wood, same straws, same answer.

Refining has no natural end, by the way. A shape keeps improving until the light changes and shows you something else, and I have carved for over forty years without once running out of things to adjust. Knowing when to stop is a separate skill, and the design course is where I try to teach it.

One last habit worth stealing. I do not do the final cleaning until the whole panel is finished, because my hands darken basswood as I work and every surface gets touched again anyway. Clean early and you will clean twice. If none of this is familiar yet, the beginner course starts further back, and the tools course will keep you from buying the wrong twenty gouges.

Shall we carve?