Mike, who runs the estate shop here, recalls that in some System Three literature there is a suggestion that prior to using T-88 Adhesive a System Three penetrating epoxy should be used to prime the wood to be glued.
This not correct.
For all intents the failure of the glue actually has the look of a starved joint.
Even though we have a requisite glue line.
A Technical Data Sheet for T-88 states:
(no mention of wood moisture content)
"Glue line thickness is not critical and clamping is not necessary...
"Oak is a highly porous wood with a strong tendency to absorb resin and yield starved joints of substandard strength. Preferred practice is to apply a liberal coat to both surfaces and without mating allow the parts to stand open for 30-45 minutes. Dull spots indicate complete absorption and should be touched up... then assembled and clamped with minimal pressure - just enough to ensure contact. Or the oak can be presealed with a thin coat on both mating surfaces and cured separately, then sand lightly, apply fresh T-88 and join lightly."
Teak surfaces require vigourous wiping with lacquer thinner and while still moist dried with rag. Repeat if the surfaces don't have a whitish color "indicating the extraction of surface oil."
My ancient dry honduras mahogany could be considered porous, I suppose.
I did not do the 1/2 hour open time to see if absorption was happening. However, there arguably was plenty of glue applied to both surfaces and a mild pressure was continuously applied while hardening that might have closed nano gaps caused by liquid being absorped by the wood. The block when dry had an even glue line all round. This indicates no glueless areas caused by absorption. Wrong.
The problem starts to be noticed at the large endgrain area (the top of the post) and there are indications (from the cutoff wedge pieces used outside to prop the mast) that the wood pulls away evenly all along the glueline.
Smith & Co.'s Tropical Hardwood Epoxy (Jamestown) requires NO PRETREATMENT to stick any kind of wood together permanently. From Lignum Vitae to Teak to Mahogany.
You have to know something about the glue. End-grain gluing is ALWAYS a problem with any glue. This is not end grain gluing but matched flat to flat grain. But Smith's doesn't need a crutch.
Now, to be absolutely sure we would need a fair comparison in a second post glued up with Smith's Tropical.
SO NOW, LET'S OPEN UP THE JOINT:
I forced a chunk of the laminate apart on one of the cut-offs with a stiff-blade putty knife.. These are the large 'waste' pieces now disintegrating full time in the sun. The wood has 100% pulled away from the joint line. Yet the glue did not want to let go right away.
....Inside the mating surfaces show visually and by feel that the adhesive was evenly distributed on both surfaces. But one surface shows the layer of glue - the mating surface shows no sign of adhesion!
The only conclusion is that the glue was widely absorbed on one side leaving few areas of connection.
The glue line turns out is essentially a figment.
Even though there is a apparent visual line of glue
one side or the other side seems to have let go!
There is no solid connection of wood to glue to wood except one small area WHERE THE WOOD RIPPED OFF FROM THE MATING SIDE. Quite odd.
This could be a lack of experience with both materials on my part. We've had a hot summer here and the wood is very dry and has very low moisture content. I didn't put a meter on it.
Taking the 15X loupe (Lee Valley) to the glue lines shows some lines 'sticking' intermittently to one or other side. While other lines have separated a complete length from one side.
I remember when the T-88 was mixed it didn't feel like the thick body of Smith's epoxy I'm used to.
I don't believe the Smith's Tropical epoxy could have just disappeared like the T-88 did.
Why one identical mating surface accepting glue and not the other?
Why one surface appearing not absorbed (evident layer of glue) and the opposite surface starved? Insane. These are identical sequential pieces.
Because of the weirdness of T-88, my goof or not, I'm sticking to the other product I know and trust. I'm positive I could never force a Smith's joint apart with a PUTTY KNIFE.
Smith's CPES, a penetrating sealer, is not a recommended surface pre-treatment for Tropical Wood Epoxy.
To be absolutely fair, the pieces could be primed with thin epoxy and allowed to set. Then the block would be sanded and assembled and glued with structural epoxy. THis is probably the way I should have done them.
__________________________________________________ ___________________________
THE FIX IS IN.
Mike suggests that with a wide thin resaw blade on the Laguna bandsaw we could easily cut along the glueline - ERASE THE JUNK ENTIRELY and then stick the block back together with another glue that has balls.
Minimal dimensional change.
Some cross dowels will be needed to key the blocks from sliding when gluing. I'll drill those in before separating the blocks. Next time: RESORCINOL.
Plastic Resin (urea-formaldehyde) will not survive soaking or swell/shrink heat/cold cycles of the coaming post environment. Epoxy has an elevated temp inadequacies and similar problem with high moisture or wet that coaming posts are subject to. Resorcinol is the ONLY glue for the job. R. is the ONLY glue that survives BOILING tests.
__________________________________________________ ______________________________
Type into your search engine APPENDIX C SUPERIOR ADHESIVES FOR THE MILLENIUM. This is Larry Pardey's condemnation of epoxy and its major purveyor's. It's a great article, from a book whose title escapes me. A professional woodworker's perspective. Epoxy has its place in layup, but never as a wood glue. The tropics will destroy epoxy in a couple seasons. Really can't use it in any exterior wood glue up. Guess I got lucky!
Reply With Quote