Episode 37 Diesel Engine Part 8 - Glue Fail

My epoxy glue failed. 

I had cut my new beds in four pieces from 48mm Merbau timber, and glued them in pairs to form two beds, both 96mm wide, using Norglass Epoxy resin. I noticed that something was wrong a few hours after gluing because the excess epoxy resin still felt liquid-soft. After a week, it had lost its tackiness, but was still soft enough for me to indent it with a fingernail. Worse, I could scrape off the excess epoxy resin with a paint scraper, and roll the excess into a ball with my fingers. Finally, I cut of the end and found that the joint was not difficult to break, leaving the timber in tact. Whatever the breaking strain of the glue in the joint, the epoxy was weaker than the timber, which was unacceptable.

I tested a sample of the remaining resin and hardener, and found that it would not set. My best guess is that during the pour for the floor plate, I exposed the bottles of the resin and hardener to the sun, which had done something bad to the chemistry. Thankfully, the resin for the floor plate in the boat had set as solid as a rock, and it had consumed nearly all of the 6 litre pack that I had bought. I was using what was left, thinking that it was OK.

This was an inconvenient set-back, but not as inconvenient as, say, the beds disintegrating at sea. Rather than risk my newly constructed beds, or try to fix them up with bolts or screws, I decided to throw them out. I had lost about $120 for the timber, about a day’s work and a week of elapsed time.

The new, new beds would be constructed from a single piece of timber, but the only convenient source of suitably sized and priced timber was the Demolition Yard, nearby. I selected a 120mm x 120mm square post, 1.5m long, paid $30, and got to work. I don’t exactly know what timber it was, only that it had been classified as “hardwood”. I do know, however, that it had already had a lifetime of seasoning, unlike much of the new timber in the yards.

The first job was to scrape off the old paint by hand. Although this was physically demanding and slow, I could check all the surfaces for any remaining nails or screws (I found the remains of one nail). I could also get a feel for where the splits, knots and soft timber lay. 

My templates and failed Merbau beds were put to use to find the best orientation for the new beds, and the setting-out of the long cuts. My circular saw cuts to a depth of 63mm, so the long cuts were made by cutting in from both sides to the half-way mark. About half the cuts met in the middle, and half did not. I used my reciprocating saw to clean out the connected cuts, and to connect the unconnected cuts. Although the reciprocating saw can cut through thick timber (the thickness depends on the length of the removable saw-blade) it has almost no directional control, which is why I needed the circular saw-cuts as a guide. In hindsight, it would have been better to take the timber to someone with a band-saw. 

For the short cuts, I used my newly acquired Bahco tenon saw, which, at about $50, was the most expensive on the rack at the hardware store. My previous tenon saw was cheaper, but had a tendency to wander off line and bind.

The next step was to clean up the timber with a plane. I had previously bought a hand plane at a closing-down sale, but knew little about how to sharpen the blade or set it up. The YouTube channel Third Coast Craftsman was an education, which was followed by a lengthy on-line search to find a real bricks-and-mortar shop nearby that would sell me a sharpening jig. It took considerable effort to persuade Google that I was not in search of an expensive power plane, and eventually found the desired item in the Carbatec shop in Wakerley. I went in with the intent of buying a $20 sharpening jig, and came out with the sharpening jig, sharpening stones and a strop, all for about $130.

The sharpening gear made a huge difference to my hand-planing. I could not quite get to the consistent waxy-smooth finish, or the uninterrupted whooshing sound of a perfectly formed shaving being parted from the parent timber that is featured on so many on-line videos. I did, at least, get away from the juddering, ripping and gouging of my previous attempts, much to the relief of my wife, who had hitherto tolerated my house-shaking planing with remarkable fortitude. 

The steps in the timber beds complicated the working of the lower steps, and accounted for some of the skin on my knuckles. However, the “free” surfaces on the bottom, sides and top, came up nicely. I know that these are only engine beds, destined to lie hidden under the diesel engine, but I found immense satisfaction seeing the beautiful patterns of the grain emerge in a way that cannot be done through sanding.

My saw-cuts and consequent trimming were not as accurate as I would have liked. There are still small differences in angle and depth, but they amount to less than 3mm from one bed to another, or from one end to another, which is probably OK. The only remedies I can think of are to clamp the two beds together and work over them with an accurate bandsaw, or a table saw, or a router in a frame, none of which are available to me.

Further rough dry-fitting confirms that the beds are good to go, and I am back at the stage where I need to drill the holes for the coach-bolts to bolt the new beds to the floor plate. For this I will need a pillar drill, or drill press, which means further calls on my mates.

Failed epoxy resin glue joint - the glue should be stronger than the timber

Scraping old paint off the 120mm x 120mm recycled timber post

The hazards of recycled timber. In this case, the pointy end of a mild steel nail

New beds, rough cut. Clearly visible are the joins where the cuts from both sides had not met in the middle.

New beds after much planing. The mitre in the port side bed (the far side) is a cut-out for the water pump on the Diesel engine. The thick and thin ends of the wedges still need to be trimmed.

Preliminary dry fitting of port (larboard) bed. The thin piece of timber in the middle of the floor plate, stuck into one of the holes in the shaft flange, is a strut intended to keep some pressure on the drive shaft so that the seal does not leak as it did previously.

Preliminary dry-fitting of starboard bed

Episode 36 Diesel Engine Part 7 - The art of jigs

I am a confirmed bad dad dancer, and have the video footage to prove it. Unlike some, however, I prefer to keep the evidence buried 1000ft down in an underground bunker on a remote, uncharted island in one of the less-frequented, storm-ravaged polar oceans. 

Those are not the kinds of jigs I am describing here.

The kinds of jigs I describe here are the jigs, templates, gismos, whatever-you-call-them things that help you take measurements or line things up. The fact that they never make it into the final build usually relegates them to a zone of inferior quality and hasty make-do. That was how I regarded them when I started this project, which was a mistake.

I have made about a dozen jigs for this project. The first were crude attempts that anticipated a simple re-seating of the engine on fixed-up beds. As the problem of the beds emerged, the jigs became more sophisticated. I found I needed ways to measure the dimensions of the engine. This was much harder than expected, partially because moving the 65kg lump around on the work bench needed special care to avoid crushed fingers, or the unplanned descent of the whole assembly onto the concrete floor below, and partially because it was impossible to take a ruler to measure the dimensions from one point to another, say the distance between the bearings, because the intervening space was occupied by the engine. These measurements were necessary because of the dearth of useful dimensional information anywhere, on the internet or otherwise.

Then, there was the issue of the angle between the axis of the engine/prop shaft tube, and the slope of the curved hull. I thought this would become clearer after I had set the floor plate in place but, alas! The floor plate gave me a plane and some nicely defined edges to work from, but the original arrangement defied explanation. More on this below, but back to the jigs.

I say more sophisticated, rather than more complicated, purposefully. The jig, or template for the underside of the engine was a case in point. It now looks like a simple, single piece of dressed pine with some lengthy notches cut into the bottom. However, because of the care taken to get those notches in the right place, it has proved a brilliant tool for checking the expected clearances under the engine.

I also made up what I should have made much earlier - a kind of bottomless box that encloses the engine and mates with the flange plates of both engine and prop shaft. This box now travels between my workshop and the boat, and I confident that whatever fits in the box, will fit in the boat. I used it to measure, as accurately as I could, the angle between the old, upper step and the new floor plate, and its dimensions from the flange plate on the drive shaft. The latter was not an easy task, because of the angles - nothing was straight, or parallel, or flat. I then used the measuring-marks on the box to cut templates for the wedges that will form the new beds.

Big tip - whenever you are in Bunnings, check the free bin by the timber saw, for useful off-cuts or rejects. I found two 3mm ply boards, about 900 x 1200, and promptly claimed them for my template. I refute any claim to be a bin-diving free-loader on the basis that I had already spent plenty of money at retail prices at the store.

Before my saw got anywhere near the timber for the new beds, I measured, marked, cut, re-measured, re-marked, re-cut my templates. They look simple now, and that is the point, but they took about a week to develop. 

I confirmed that I would have to replace the step, rather than keep the fore and aft bearings on the same level. To keep the bearings on the same level, I would need to dig about 30mm into my newly made floor. The step complicated the measuring, but would not greatly complicate the final assembly. Further, I did not want to damage the structural integrity of the new floor. Also, if I had lowered the aft bearings, I would have less timber below them to receive and anchor the holding-down bolts, unless I drilled right through the hull.

I found that I had made an error in the height of the step. It should be 75mm, not 85mm. I had made the rookie mistake of not accounting for the 10mm thickness of the bracket, when I turned it upside-down. Still, my measurements came out with a 10 to 30mm difference, which is significant because the support bolts for the bearings have only about 20mm adjustment in them. 

I previously commented that the previous arrangement defied explanation, and this is why. Before I took the engine out, one of the four bearings had snapped, and all four were wound down so that the engine has been lowered down to the stops. Also, the lower steps were not parallel to the upper steps, being about 10 to 15 degrees out of whack. I referred back to the large piece of cardboard onto which I had traced the outline of the old beds and, try as I might, I could not find a nice fit between the old bearings, in their location and orientation, and my new template. The forward bearings sloped down - should I align the new beds to the lower or upper end of that slope? Furthermore, the old orientation of the engine was not aligned to the angle of the old, top step.

The orientation of the engine was a mystery; was it intentional, or was it was the result of poor workmanship or neglect? The engine ran and pushed the boat forward and backward, so the previous owners might be forgiven for thinking that there was nothing broken that needed fixing. It presented a dilemma. If it was intentional, I would need to restore the orientation of the old engine, although it would not be aligned to the angle of the old, top step. If it was not intentional, I would need to recreate the right orientation, which was probably, but not indisputably, aligned to the angle of the old, top step.

In the end I decided on a compromise and a guess, making the beds about 20mm lower than my jigs indicated. If needed, I could raise the bearings on packers, which could be easily made up from 19mm dressed hardwood, or something similar. It would be much easier to raise the engine on packers than to try to lower it below the level of the new beds. If the angles were all wrong, the packers could be cut into mini-wedges to compensate. Finally, if the beds were irredeemably misaligned, they could be unbolted from the floor plate and replaced.

Thus, the templates for the beds, cut from 3mm ply, took on a settled form. I made two, by simply cutting through two sheets instead of one. I took them to the timber yard, where they instantly showed which pieces of timber would be large enough for the beds. 

For the beds themselves, I found two merbau step-treads, dressed to a thickness of 49mm. My plan was to cut four wedges from the step-treads and glue them in pairs to make up a thickness of 98mm, which was wider than the width of the bearings, but less than the 130mm available on the floor plate. An important feature of this plan was my ability to cut thick timber, which was limited to the 63mm saw-cut depth of my circular saw. I had phoned around some milling yards to see if I could find a way to cut thicker timber, but with no positive results. 

The templates proved useful in setting out the saw-cuts. The cuts themselves were tricky, because I do not have a table-saw. A table-saw would make life easy for making parallel cuts, but I could not justify the investment on the basis of a one-off project like this. I had to make do with clamping a guide to the timber at the appropriate offset from the saw blade, and running the baseplate of the circular saw along the guide. The cuts were not as accurate as I would have liked, so I will have to finish off the surfaces with a smoothing plane. Fortunately, being wedges, any reduction in thickness from my planing can be made up by sliding the wedges further up the slope of the floor plate. The corresponding loss of timber at the bottom of the slope would be inconsequential, apart from some compromised aesthetics.

Finally, I used the templates to drill out the holes for the secret dowels for the glue-joints. All I needed to do was to drill through the template into one merbau step, then flip the template and do the same on its complementary other-half.

Merbau is a naturally oily timber, and these pieces had been subject to further oiling in the timber yard. The oiliness makes for a durable, rot resistant timber nicely suited to outdoor furniture and decking, but it also compromises gluing. I cleaned the surfaces with every solvent I could find in my workshop, finishing up by swabbing it several times with methylated spirit and leaving it to dry for a couple of hours. For glue, I mixed up some epoxy resin from the leftovers of the floor poor. The internet tells me that epoxy glue has a stronger bond than water-based glue, when gluing naturally oily timber, such as merbau.

The next steps will be to clean and square the upper and lower surfaces, then to drill holes for the holding bolts, then to bolt both beds to the floor plate. For the holding down bolts, I bought a dozen M12 100mm stainless steel coach bolts (the box for the M10 bolts was empty). If I can retain about 30mm between the bolt-head and bottom of bed, then I can embed the coach bolts into the 60mm of timber available in the floor before I hit the hull. I will need to saw off the tips of the coach bolts to reduce their overall length, which is OK, because I will screw them into pre-drilled holes. They will not miss the loss of their sharp points, which could otherwise puncture the hull.

The inside of the jig or bottomless box showing the port-side bearings. In this configuration, the aft bearing is down. The bracket needs to be turned upside-down for the step. 

The outside of the jig or bottomless box, showing its alignment to the engine flange, which will mate with the flange on the prop shaft in the boat. The cut-out for the flange is the point of reference for all other dimensions and the slop at the bottom of the box is aligned to the slope of the new floor plate in the boat.

Using the bed templates to mark up the cuts on the merbau step-treads.

The beds following the major long cuts

After the short cuts to form the steps

Using the templates to drill holes for the secret dowels

Gluing four into two. I have used clear packing tape around the sides to contain the excess epoxy resin, which will be removed in the next stage.