For the last four days, I've been trying to figure out how to make this part:
It looks simple enough on the plans, all innocent and perfectly drawn. Yet it's taken me four days so far, 17.5 hours of work, with more to come. Most of these brackets I've been making take maybe 4-8 hours to sort out. But this one's special.
The journey starts with: How to bend that hoop? Quarter-inch 4130 steel is no slouch in the strength department, though with a long enough lever it's easily bent. My first try was just to clamp the thing in a vise, perpendicular to a length of stout 1" tubing I had lying around (for the 1/2" radius bend), then bend it over the tubing by grabbing the long end and giving it a shove. This worked, but the resulting bends had a disagreeable curve to the "straight" parts that I didn't love. My math was also poorly understood, so it took a few tries before I got the right size, and I only had 17" of the stuff to work with. Finally I gave up and decided I could live with the poorly-bent hoops, just so I could make progress.
So, hoops done, check. I had the plates already, thanks to my various Ken Brock bracket kit purchases. I held the hoop up to the plate and tried to imagine how I'd actually hold any of this together. It's pretty fiddly, and clearly a quality jig was in order.
The first problem with the jig is that the space between the plates needs to be 2 1/8". I had 2" square tubing, which I'd used to make jigs for other brackets, but I had to find that extra 1/8" of thickness somewhere. I have some 1/16" thick sheet lying around from a previous project (making the drag strips, which I ultimately scrapped in favor of wires), so I decided to weld two layers of that onto my jig, which should work out to exactly 2 1/8" thick.
Of course, it did not. It was a little bit thin for some reason, and since these brackets tend to shrink a little after they're welded, I really didn't want to start off too small, or I'd never get the bracket onto the wing when the time came. So I did the only reasonable thing I could think of, and ran a weld bead down each of the 1/16" strips to make them a bit thicker.
Then it was a whole lot of filing to get them back down to the right size, after I managed to crash the head of my milling machine when the jig leapt up into the cutter and stopped the works very quickly indeed -- I still need to open that thing up and replace the busted gears. But I got it to the right size, and I'm going to have very odd musculature in my arms when this is all over.
However, my skill with a file is not expert-level, so the strips are only about right, they're not perfectly flat. I decided this was an acceptable compromise because it's all getting welded anyway, and striving for perfection there is an exercise in frustration thanks to the expansion and contraction of the steel as it's heated and cooled.
In any case, I took my imperfect jig to the drill press, and carefully drilled out the hole that would locate the side plates. Not carefully enough, though: the entry hole was perfect, but the exit hole was at least 1/16" off, probably because I'd been resting the jig on those imperfectly welded and filed strips. So I tried a trick I've never done before, and welded up the hole, having to weld from both sides. It worked, to my amazement.
None of this addresses the original problem, though: how to deal with that floppy hoop that has practically no mechanical connection to the plates until it's welded. Finally, I decided that I'd make a crossbar with slots in it that would hold the hoop, and the milling machine came out again (I'm telling this out of order; I made the crossbar before the head crashed). That piece at least was pretty easy to make.
The final problem to solve (so I naively thought) was how to make sure the side plates, which are held down by a single bolt, couldn't rotate. When installed on the plane, they interlock with a big round piece, but I didn't want to have to make that up for my jig. Instead, I ended up making a little shelf out of metal for the legs of the plate to rest on.
Finally, late in the evening, having fussed with the spacing on the imperfect hoop for several hours off and on, I turned on the torch, and applied heat to metal, tacking the hoop in place. As soon as I did, I realized that my design was lacking, and there was no way to get the tacked bracket off the jig. The crossbar was locking it in place, and I ended up (the next morning) having to saw the ends off the crossbar, and file rounded ends in it so it could get past the plates while still holding on to the arms of the hoop.
This was ok, because a discussion on the Biplane Forum had convinced me that I needed to try making stainless steel hoops. The idea was that there was no way to chip the paint off them, and they wouldn't rust. There followed a feverish study of stainless steel properties and welding techniques, and a quick order to the metal shop. I was glad to be able to get it on Friday afternoon, so I could have the project to work on over the weekend.
I had called my dad to discuss the stainless steel (he knows more about stainless than I do), and he reminded me that I had a better way to bend the rods than just clamping them next to a tube and pushing hard: I have a press. And, I realized, I built a really moosey tailspring rebending tool which might just work as an improvised press brake. The idea is that you use the press to push on the center of the thing you want to bend, while it's resting between two bars, or a V-block, or something like that. The idea struck me, so I spent most of Friday trying it out. Fortunately, I got three feet of the stainless steel rod, because I made bad bend after bad bend (not the press brake's fault, I just kept getting the math wrong). Eventually, though, I ended up with two perfect hoops, exactly the right size, and with lovely square corners. Shown here before the ends were trimmed, so their perfection isn't quite as obvious:
I decided, wisely, that before I would commit my hard-to-reproduce side plates to this project, I would try welding some of my incorrectly bent stainless rod to some mild steel, just to see if I could do this dissimilar-metals welding well enough to put on an airplane.
I welded the first hunk of rod to a failed experiment that was a sort of box of square tubing, which was certainly a decent stand-in for the side plate. Once it cooled down I hit it with a hammer a few different ways, finally sticking a piece of MDF (not the strongest material, but the right size, so why not?) through the center of the box and tried pressing the rod away from its plate. With one moderate hit of the hammer, it ripped right off.
Well, that's not good.
I tried again, on a different spot. Same thing. I made a smaller test piece, thinking maybe the big box was sucking up too much heat. Nope, same thing. For some reason, the welding rod I was using to build up the weld was porous and quite weak.
Just to make sure it wasn't my technique for welding a rod to a plate that was at fault, I tried the same thing with a piece of the 4130 that didn't get bent well. Making the same test, the MDF was destroyed, and it barely moved the rod. But it did move it enough that I could flip the piece and whale on it with the big hammer. It bent, then bent further, then bent almost double. The weld held, finally just starting to tear as the rod took its last few hits. That was much more like it. That's a weld I'd trust to hold down an airplane, or jack up an airplane.
So, the stainless steel experiment was a bit of a bust, but at least I found out before making a bit expensive mistake, or even worse, putting it on the plane and thinking it was doing well until the one thing that would doubtless be a catastrophic failure.
Now I just have to find more 4130 rod. But once I have it, I have the perfect bending technique all lined up and ready to use. Five days to produce two tie-down rings? Why not.